University of Southern California Dissertations and Theses

Using CRISPR-mediated deletion to study prostate cancer regulatory elements located at loop anchors identified by Hi-C

(USC Thesis Other)

Using CRISPR-mediated deletion to study prostate cancer regulatory elements located at loop anchors identified by Hi-C

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content

Using CRISPR-mediated Deletion to Study Prostate
Cancer Regulatory Elements Located at Loop Anchors
Identified by Hi-C

Jiani Shi
Mentor: Dr. Peggy J. Farnham
Department of Biochemistry and Molecular Medicine Master of Science
University of Southern California
August 2019

1
Acknowledgements
I want to express my appreciation to my mentor Dr. Peggy Farnham. To me, she is a great educator
who truly care about the students. She has provided me with numerous helps and advice. I admire
her attitude toward research and her great personality. I feel more than happy to have spent my last
two years in the Farnham lab.
I want to express my appreciation to my mentor Dr. Suhn Kyong Rhie. She has given me supports
on my thesis project and invested a great amount of time helping me dealing with any problems I
have encountered. I thank her for her encouragement and advice throughout the last two years.
I want to express my appreciation to my teachers and lab members: Shannon Schreiner who has
taught me lab techniques, Yu (Phoebe) Guo who has advised me a lot on my project, Jenevieve
Polin who has edited my thesis and helped with my project, Zhifei Luo, Andrew Perez and Weiya
(Stephanie) Ni who have helped me with my project and all my questions during my studying here.
I want to thank Wei Zhu and Karly Nisson for their supports and friendship during the years we
spend together, and Carol Munoz who always kindly deals with any problems we have.
I especially thank my tutor in the Farnham lab, Charlie Nicolet, who has not only trained me but
also encouraged and inspired me a great deal.
I want to express my appreciation to my committee members Dr. Michael Stallcup and Dr. Joseph
Hacia for their suggestions for improving my project and thesis.
I want to express my appreciation to my teachers and advisors Dr. Pragna Patel and Dr. Judd Rice.
They have given me supports on both academic studies and my individual development.
I want to express my appreciation to our program advisors Monica Pan and Joyce Perez who have
done a great job in improving the program for the students.

2
Table of Contents
CHAPTER 1 INTRODUCTION 10
1.1 Regulatory elements 10
1.2 Open chromatin 11
1.3 Histone modifications 11
1.4 Promoters and enhancers 12
1.5 Chromatin organization 14
1.6 Compartments, TADs and loops 15
1.7 Boundaries of compartments, TADs and loops 17
1.8 CTCF and cohesin 18
1.9 Using the CRISPR system to study gene regulation 23
1.10 A high-resolution 3D epigenomic map reveals prostate cancer-specific alterations in
chromatin architecture near the Androgen Receptor (AR) gene 25
CHAPTER 2 MATERIALS AND METHODS 27
2.1 Cell culture 27
2.2 Plasmids construction 27
2.3 CRISPR/Cas9-mediated deletion experiments 28
2.4 Single-cell colony isolation 30
2.5 RNA extraction and RT-qPCR 30
CHAPTER 3 Using CRISPR-mediated Deletion to characterize a putative enhancer that is
looped to the Androgen Receptor (AR) promoter 32
3.1 CRISPR/Cas9-mediated deletion of a putative AR enhancer (AR EN) in a 22Rv1 cell
population 34
3
3.2 Generation of single-cell derived clones carrying the AR EN deletion using 22Rv1 cells 37
CHAPTER 4 Using CRISPR-mediated deletion to characterize four CTCF anchor sites near
the AR gene in 22RV1 cells 39
4.1 CRISPR/Cas9-mediated deletion of the four CTCF anchor sites in a 22Rv1 cell population
41
4.2 Generation of single-cell-derived clones carrying deletions of the four CTCF anchor sites
46
4.3 CRISPR/Cas9-mediated deletion of both CTCF site 1 and site 2 in a CTCF site 2 deletion clone
52
CHAPTER 5 Control experiments: CRISPR/Cas9-mediated single gRNA cutting near loop
anchors and deletion of a region in the middle of the AR enhancer-promoter loop 56
CHAPTER 6 DISCUSSION 62
6.1 AR expression level decreased by less than 2-fold after CRISPR/Cas9-mediated deletions of a
putative enhancer and four loop anchors (individually) near the AR gene in 22Rv1 cells 62
6.2 CRISPR/Cas9-mediated DSBs at loop anchors affect nearby gene expression 63
6.3 Interpretation of small fold changes (< 2 by RT-qPCR) observed after CRISPR experiments
and clonal variation in single-cell-derived clones 65
6.4 Results from the depletion of CTCF/cohesin subunits studies 66
REFERENCES 71

4
List of Figures
Figure 1.1. Illustration representing a promoter, an enhancer, an insulator and their associated
epigenetic features.
Figure 1.2. An example of a Hi-C interaction map from the GM12878 human lymphoblastoid cell
line.
Figure 1.3. Illustration representing a loop constrained by a pair of CTCF (in convergent direction)
and a ring-like cohesin complex.
Figure 1.4. Illustration representing a “CTCF-CTCF loop”.
Figure 1.5. Illustration representing “enhancer hijacking”.
Figure 1.6. Illustration representing an “enhancer-promoter loop”.
Figure 1.7. Illustration showing CRISPR/Cas9-mediated deletion via the NHEJ pathway.
Figure 3.1. Hi-C chromatin interaction maps of a 3 Mb region containing the AR gene in normal
prostate (RWPE1) and prostate cancer (22Rv1) cells.
Figure 3.2. CRISPR/Cas9-mediated deletion of a putative AR enhancer (AR EN).
Figure 3.3. Expression of AR after CRISPR/Cas9-mediated deletion of the putative AR enhancer
(AR EN).
Figure 3.4. Expression of AR in clones carrying the AR EN deletion.
Figure 4.1. CRISPR/Cas9-mediated deletions of CTCF site 1, 2, 3 and 4.
Figure 4.2. Expression of AR after CRISPR/Cas9-mediated deletions.
Figure 4.3. Expression of EDA2R, OPHN and YIPH6 after CRISPR/Cas9-mediated deletions.
Figure 4.4. Expression of AR in clones carrying the CTCF site 1 deletion.
Figure 4.5. Expression of EDA2R, OPHN and YIPH6 in clones carrying the CTCF site 1 deletion.
Figure 4.6. Expression of AR in clones carrying the CTCF site 2 deletion.
Figure 4.7. Expression of EDA2R, OPHN and YIPH6 in clones carrying the CTCF site 2 deletion.
5
Figure 4.8. Expression of AR in clones carrying the CTCF site 3 deletion.
Figure 4.9. Expression of AR after CRISPR/Cas9-mediated deletions of both CTCF site 1 and site
2.
Figure 4.10. Expression of AR in clones carrying deletions of both CTCF site 1 and 2.
Figure 5.1. CRISPR/Cas9-mediated single-gRNA-cutting near AR EN, CTCF site 1 and CTCF
site 2.
Figure 5.2. Expression of AR after AR EN deletion and single-gRNA-cutting near AR EN.
Figure 5.3. Expression of AR after single-gRNA-cutting near CTCF site 1 and 2.
Figure 5.4. CRISPR/Cas9-mediated deletion of a region in the middle of the AR enhancer-promoter
loop (MID).
Figure 5.5. Expression of AR after deletion of the region in the middle of the AR promoter-enhancer
loop.
Figure 6.1. Location of gRNAs used in both single-gRNA-cutting and deletion experiments near
AR EN and CTCF site 1.

6
List of Tables
Table 2.1. gRNAs used in the CRISPR/Cas9-mediated deletion experiments.
Table 2.2. Genotyping primers used in the CRISPR/Cas9-mediated deletion experiments.
Table 2.3. RT-qPCR primers used in the CRISPR/Cas9-mediated deletion experiments.
Table 3.1. RT-qPCR results on AR, EDA2R, OPHN1 and YIPH6 after deletions in cell pools.
Table 3.2. RT-qPCR results on AR after deletions in cell pools (from other replicates of
CRISPR/Cas9-mediated deletion experiments).
Table 4.1. Meta data of clones and CRISPR/Cas9-mediated deletion experiments performed in
cell pools.

7
Abstract
Previously, members of the Farnham lab have generated Hi-C chromatin interaction maps in both
normal (RWPE1) and cancer (22Rv1) prostate cell lines. By combining the Hi-C data with
epigenetic features and transcriptome profiling, chromatin loops surrounding the Androgen
Receptor (AR) gene were identified, which are present in cancer cells (22Rv1) but absent in normal
cells (RWPE1), suggesting that these loops may contribute to the large upregulation of AR
expression in the 22 Rv1 cancer cells. My studies have focused on 5 chromatin loops near the AR
gene. Ones of these loops connects the AR promoter to a putative enhancer and the 3 other loops
are anchored by convergent CTCF sites. I performed CRISPR/Cas9-mediated deletion of the
putative AR enhancer and the four other CTCF loop anchor regions surrounding the AR gene in
22Rv1 cell pools and generated single-cell-derived clones harboring these deletions. After
performing these deletions in cell pools, I only observed modest changes on AR expression.
Similarly, little change on AR expression can be found in clones carrying these deletions. My
studies will help in understanding the relationship between chromatin looping and gene regulation.

8
List of Abbreviations
3C: Chromosome Conformation Capture
ADT: Androgen Deprivation Therapy
AR: Androgen Receptor
AR-FL: Androgen Receptor Full-length Variant
ATAC-seq: Assay for Transposase-Accessible Chromatin using Sequencing
Cas9: CRISPR-associated Protein-9 Nuclease
ChIA-PET: Chromatin Interaction Analysis by Paired-End Tag Sequencing
ChIP-seq: Chromatin Immunoprecipitation Sequencing
CNON: Cultured Neuronal Cells Derived from Olfactory Neuroepithelium
cRE: Candidate Regulatory Element
CRISPR: Clustered Regularly Interspaced Short Palindromic Repeats
CRPC: Castration-resistant Prostate Cancer
CTCF: CCCTC-binding Factor
DHS: DNase I Hypersensitive Sites
DSB: DNA double-stranded break
ENCODE: Encyclopedia of DNA Elements
FAIRE: Formaldehyde-assisted Isolation of Regulatory Elements
FISH: Fluorescence in situ Hybridization
gRNA: Guide RNA
GWAS: Genome Wide Association Studies
9
NGS: Next-generation Sequencing
NHEJ: Nonhomologous End Joining
PCa: Prostate Cancer
RNA-seq: RNA Sequencing
SNP: Single Nucleotide Polymorphism
TAD: Topologically Associating Domain
TF: Transcription Factor
TSS: Transcription Start Site

10
CHAPTER 1
INTRODUCTION
Regulatory elements
Less than 3% of the human genome consists of protein coding genes. The rest of the genome (i.e.
the non-coding regions) harbors a great number of functional elements which are critical for
transcription regulation. (Consortium, 2012) These non-coding functional elements have the
ability to regulate transcription of neighboring (or, in some cases, distal) genes. These regulatory
elements include promoters, enhancers and insulators. (Wittkopp e Kalay, 2011) With data
generated by next-generation sequencing (NGS) methods, our knowledge of regulatory elements
has become more enriched. Several criteria have been used to annotate regulatory elements in the
human genome, including transcription factor binding, chromatin accessibility, histone
modification, and chromosome conformation (Figure 1.1). The Encyclopedia of DNA Elements
(ENCODE) has established a registry of candidate Regulatory Elements (cREs) using data
produced by the ENCODE and Roadmap Epigenomics Consortia. These cREs are classified into
promoter-like, enhancer-like, and CTCF-bound insulator-like groups. In the following sections,
the defining epigenetic features of these three groups of cREs will be discussed in detail.
(Consortium, 2012) I will also discuss how chromatin organization can influence the functioning
of regulatory elements.

11
Open chromatin
Open chromatin indicates the presence of nucleosome-depleted regions. ENCODE has
characterized open chromatin by mapping the DNase I hypersensitive sites (DHSs) using Dnase-
seq and formaldehyde-assisted isolation of regulatory elements (FAIRE), and also now includes
open chromatin mapped using the more recent technique of Assay for Transposase-Accessible
Chromatin using sequencing (ATAC-seq). (Consortium, 2012) Open chromatin results from the
binding of transcription factors in place of nucleosomes. (Thurman et al., 2012) In fact, the
majority (98.5%) of the transcription factor (TF) binding sites mapped by ENCODE using
chromatin immunoprecipitation sequencing (ChIP-seq) fall within regions of open chromatin.
(Consortium, 2012) Open chromatin regions are associated with both active promoters and
enhancers, with about 3% of DHSs localizing to transcriptional start sites (TSSs) and ~95% being
located in distal non-coding regions. (Thurman et al., 2012)

Histone modifications
Most histone modifications are chemical residues added to amino acids near the N terminus of a
histone; these modifications are hypothesized to affect DNA wrapping around nucleosomes and
therefore affect DNA accessibility and gene regulation. Chromatin Immunoprecipitation
Sequencing (ChIP-Seq) is used to map regions of the genome that harbor modified histones.
Specific modifications have been showed to correlate with transcribed genes (e.g. H3K36me3),
constitutive heterochromatin (e.g. H3K9me3), polycomb repression (e.g. H3K27me3), and active
regulatory elements (e.g. H3K27ac). Also, H3K4me3 is highly enriched in promoter regions and
12
H3K4me1 is enriched in enhancer regions. The ratio of these two marks has been applied to
distinguish between promoters and enhancers. (Consortium, 2012; Holtzman e Gersbach, 2018)

Promoters and enhancers
Conventionally, a promoter is defined as a DNA region that is very near the 5’end of a gene and
governs its expression, while an enhancer can affect the expression of a target gene without strict
regard to direction or distance. (Yao et al., 2015) The accepted model for gene regulation suggests
that although both promoters and enhancers interact with transcription factors and co-factors, the
promoters recruit RNA polymerase to initiate transcription at a nearby TSS, while enhancers
enhance transcription by coming in contact with promoters through looping. Enhancers may also
recruit chromatin remodeling complexes, which are thought to enhance the open chromatin state.
(Ho et al., 2019) However, studies in the past decades have reported phenomena that challenge the
classic model: (1) rather than just enhancing transcription initiation at a TSS, an enhancer can
produce RNA (eRNA), and the level of eRNA has been suggested to correlate with target gene
expression, (2) some promoters can influence the regulation of other distal genes, thus functioning
like an enhancer, (3) the enhancer-gene relationship is promiscuous meaning that an enhancer can
regulate multiple genes, and a gene can be regulated by multiple enhancers; studies have suggested
that an average of four enhancers contact an active gene, (Merkenschlager e Nora, 2016) and (4)
looping between an enhancer and a promoter is found to be confined within megabase-sized
chromosome structures termed topologically associating domains (TADs) (discussed in a
following section); thus there may, in fact, be distance constraints on enhancer function. (Ho et
al., 2019) These findings showed that despite differences in location, enhancers and promoters
13
may share similar mechanisms regarding their regulatory functions. There is no clear universal
definition of a promoter or an enhancer; rather, different studies usually define these elements
based on the method in use or the purpose of the study. For example, the Registry of Candidate
Regulatory Elements from ENCODE annotates candidate promoters (candidate cis-Regulatory
Elements with promoter-like signatures, ccRE-PLS) as regions overlapped with sites of open
chromatin, proximal to a known TSS and enriched for H3K4me3 (although regions distal to a
known TSS with low a H3K27ac signal and a high H3K4me3 are also included). Similarly,
candidate enhancers are annotated as regions overlapped with sites of open chromatin, distal to a
known TSS and enriched for H3K27ac (although TSS-proximal regions with low H3K4me3
signals are also included). (Consortium, 2012)
To summarize, active regulatory elements are located within open chromatin and enhancers are
distinguished from promoters by the distance to a TSS as well as by the enrichment of different
histone modifications. Other information such as DNA methylation, TF binding patterns, and the
presence of eRNAs further helps identify enhancers that are active in a cell-type specific manner
(Figure 1.1).

14
Figure 1.1. Illustration representing a promoter, an enhancer, an insulator and their associated epigenetic
features.
An active enhancer is indicated by the presence of histone modifications (H2K27ac and H2K4me1), open
chromatin (highlighted by grey squares), eRNA, and TF binding. The inactive state of an enhancer is indicated
by the presence of DNA methylation. An active promoter is marked by the histone modification H3K4me3 and
is located in open chromatin. Insulators are shown as CTCF binding sites. This figure is adapted from Yao et al.
(Yao et al., 2015)

Chromatin organization
The importance of chromatin structure in the control of gene regulation has been posited for
decades; however, direct evidence in support of this hypothesis has been, until recently, hard to
gather. The recent development of high-throughput chromosome conformation capture methods
now allows the generation of genome-wide interphase chromatin interaction maps (Figure 1.2).
The Chromosome Conformation Capture (3C) method was first described by Dekker et al. (Dekker
et al., 2002) In this method, DNA molecules are cross-linked to surrounding proteins, then digested
into small fragments. The cross-linked fragments that are in close spatial proximity are ligated
together so they can be detected by PCR using primers specific to the hypothesized interacting
locations. More recently, a family of proximity-ligation-based methods (“C-methods”) have been
developed, such as Hi-C and Chromatin Interaction Analysis by Paired-End Tag Sequencing
(ChIA-PET). These methods include an additional biotinylation step, which allows ligated DNA
fragments to be gathered by a streptavidin-based enrichment assay and then subjected to high-
throughput sequencing. In this way, interactions across the genome, rather than just limited target
regions, can be captured. In the in situ Hi-C protocol, a ligation step is performed in intact nuclei
and provides a better resolution of the interaction regions. (Lieberman-Aiden et al., 2009) ChIA-
15
PET uses antibodies to further select for DNA fragments that bind to any protein of interest.
(Dowen et al., 2014; Ji et al., 2016) Using chromatin interaction datasets, each chromosome can
be segmented into smaller sections, primarily based on the intensity of interaction signals between
different chromosomal regions. As described in more detail below, promoters and enhancers are
either confined, separated, or connected by such structures, and it is thought that this organization
provides a mechanism for specificity of gene regulation by distal enhancers.

Compartments, TADs and loops
Using Hi-C, different layers of interphase chromosome structures have been revealed. In 2009,
Lieberman-Aiden et al. described A and B compartments. (Lieberman-Aiden et al., 2009) Using a
contact matrix at 1 megabase resolution, they found that each chromosome can be segregated into
two sets of loci, termed compartments A and B, and that contacts within each set are enriched
whereas contacts between sets are depleted. On a linear map, the genome can be shown to
alternately switch from A to B compartments (and vice versa). Compartment A includes more
accessible and transcriptionally active chromatin (it is enriched for H3K36me3, transcribed genes,
and DNaseI hypersensitive sites), whereas the chromatin in compartment B is more densely
packed, gene-poor and transcriptionally inactive. Also, compartments have been found to be
associated with replication timing.; compartment A is early replicating while compartment B is
late replicating. (Lieberman-Aiden et al., 2009; Pope et al., 2014)
More recent Hi-C studies with better resolution (i.e. more sequenced reads) have detected smaller
structures residing within the two large compartments. Independently, several groups have
16
generated contact matrices at less than 100 kilobase resolution, and identified megabase-sized local
chromatin interaction domains, later referred as topological associating domains (TADs). (Dixon
et al., 2012; Nora et al., 2012; Sexton et al., 2012) A TAD is defined as a chromosome region
having significantly more frequent interactions within itself than with regions outside of the TAD
boundaries. TADs are represented by squares along the diagonal in a contact map or as triangles
when the maps are turned sideways. Hi-C maps constructed using high resolution can identify
smaller sized domains, sometimes called “sub-TADs”. (Phillips-Cremins et al., 2013; Eagen,
2018) Rao et al. performed in situ Hi-C in human GM12878 cells and constructed a contact map
at 1 kilobase resolution, identifying TAD ranging in size from 40 kb to 3 Mb, with a median size
of 185 kb. (Rao et al., 2014) They also identified pairs of loci within a TAD that show significantly
closer proximity with one another than with the region lying between them, which suggests the
presence of intra-TAD chromatin loops. (Rao et al., 2014)

17
Figure 1.2. An example of a Hi-C interaction map from the GM12878 human lymphoblastoid cell line.
The Hi-C interaction map shows a region surrounding the TBX5 gene. CTCF motifs are labeled with blue
(forward) and red (reverse) lines (loops are created by CTCF binding to two sites in convergent orientation). The
tracks below the map show data from CTCF ChIP-seq, Rad21 ChIP-seq, RNA-seq and H3K27ac ChIP-seq. This
figure is adapted from Merkenschlager and Nora. (Merkenschlager e Nora, 2016)

Boundaries of compartments, TADs and loops
A compartment is composed of several adjacent TADs. Interestingly, although the TAD
boundaries have been found to be very similar when comparing different cell types, the
compartment that a TAD is located in can switch between active chromatin (an A compartment)
and silenced chromatin (a B compartment) in different cell types. (Dixon et al., 2015) For example,
during stem cell differentiation, 36% of the genome switched compartments while the TAD
boundaries remained stable. This suggests that TADs may be the units of dynamic activity in
chromosome compartments. (Dixon et al., 2015) Most loops are confined within a TAD, and long-
range interactions across TADs are very rare. Within TADs, changes in loop structures are
observed in different cell types, during cellular differentiation, and during disease states. (Dixon
et al., 2015; Yu e Ren, 2017) Compared with TADs, chromatin loops are relatively cell-type
specific and are considered to be correlated with nearby gene expression. (Ruiz-Velasco et al.,
2017) The chromatin architecture detected by the C-methods has, in some cases, been confirmed
by other studies. For example, fluorescence in situ hybridization (FISH) studies have showed the
average distance between two chosen loci is shorter if they are in the same compartment, TAD or
connected by a loop. (Lieberman-Aiden et al., 2009; Nora et al., 2012; Rao et al., 2014; Wang et
al., 2016) However, it should be noted that the true function of the loop structures detected by the
18
“C-methods” is not yet known. The chromosomes are not in a static state and the interaction
structure models are probabilistic. They emerge from accumulated observations within many cells,
rather than reflecting stable connection. (Merkenschlager e Nora, 2016)

CTCF and cohesin
Researchers have been looking for the key factors that specify the location of the boundaries of
TADs and the formation of chromatin loops. The involvement of CCCTC-binding factor (CTCF)
and the cohesin complex is currently supported by many studies. (Merkenschlager e Nora, 2016)
For example, in one study it was shown that most loop boundaries are bound by CTCF (86%) and
the cohesin subunits RAD21 (86%) and SMC3 (87%). (Rao et al., 2014) CTCF is a DNA-binding
protein with 11 zinc-finger domains. CTCF binding has been attributed with the insulator property,
because its presence can disrupt enhancer function in reporter plasmids. (Merkenschlager e Nora,
2016) CTCF directly interacts with cohesin subunits (SA1 and SA2 in human, Scc3 in yeast) in
vitro. The cohesin complex mediates the cohesion between sister chromatids from the time of
DNA replication in S phase until cell division. It can form a ring-like structure that is large enough
to topologically enclose two chromatin fibers, which may explain the high interaction frequency
and close proximity of distal regions of the chromosome (which would represent loop anchor
points) (Figure 1.3). (Merkenschlager e Nora, 2016) Rao etc. found that the vast majority of loops
are associated with pairs of CTCF motifs in a convergent orientation (>90% versus 25% expected
by chance). (Rao et al., 2014) CRISPR-mediated inversion of a CTCF binding site can either
abrogate or introduce new loop interactions, while the recruitment of CTCF to the binding site
remains unaffected. These studies suggest that the convergent orientation is crucial to loop
19
formation. (De Wit et al., 2015; Guo et al., 2015) The loop extrusion model suggests that
chromosome extrudes through a cohesin ring until it is stopped by a pair of CTCF protein in a
convergent direction. This model could explain some observations, such as: (1) the preferred
convergent orientation of the CTCF motif at loop boundaries and (2) after depletion of key proteins
involved in the cohesin loading or releasing process, TADs and loops show different changes in
strength and size (reviewed in detail by Eagen, 2018). (Eagen, 2018) This model is supported by
polymer simulation results, but it lacks direct evidence. (Sanborn et al., 2015; Fudenberg et al.,
2016; Merkenschlager e Nora, 2016) Although natural mutations or CRISPR-mediated deletion of
individual CTCF sites can disrupt specific loop formations, (Ruiz-Velasco et al., 2017) the
majority of CTCF sites (even if properly oriented) do not engage in looping. (Rao et al., 2014;
Merkenschlager e Nora, 2016) It is unknown how cells decide which sets of properly oriented
CTCF sites are utilized.

Figure 1.3. Illustration representing a loop constrained by a pair of CTCF (in convergent direction) and
a ring-like cohesin complex.

Although we are beginning to define how cells organize regulatory elements in a spatially and
temporally order, we do not yet understand how distal regulatory elements control transcription.
20
Specifically, it is difficult to predict which promoter might be regulated by a specific distal
enhancer. (Yao et al., 2015) For example, there are tens of thousands of potential enhancers
(according to analysis of epigenetic marks) in a given cell type and we do not know if all of the
enhancers are critically important for gene regulation. Annotation of regulatory elements with
epigenetic marks and chromosome structure information may help us predict functional
connection. For example, some studies have annotated distal enhancers using Genome-Wide
Association Studies (GWAS)-identified risk single nucleotide polymorphism (SNPs) that are
enriched in non-coding regions. These SNPs may cause gene dysregulation (and thus cause a
change in cellular phenotype) by disrupting enhancer function and/or chromosome structure. (Tak
e Farnham, 2015) Therefore, some investigators have prioritized enhancers to study by the
presence of a SNP associated with a disease. For example, in a previous study, Hazelett et al.
annotated 727 prostate cancer risk-associated SNPs, and 88% of them are located in putative
enhancer regions. (Hazelett et al., 2014) In another study Yao et al. identified 28 putative
enhancers harboring colorectal cancer (CRC) risk-associated SNPs. (Yao et al., 2014)
Two models have been developed to help understand enhancer-mediated gene regulation. Both
models involve CTCF. The importance of CTCF is confirmed by studies showing that its DNA
binding stability can be affected by GWAS SNPs that fall within or near its motif sequence. Also,
mutations or changes in DNA methylation at CTCF sites frequently occur in cancers.
(Merkenschlager e Nora, 2016) (Katainen et al., 2015) The first model, a “CTCF-CTCF loop”, is
based on the premise that regions within a CTCF-CTCF loop represent an “insulated
neighborhood” (Figure 1.4). (Hnisz, Day, et al., 2016; Ruiz-Velasco et al., 2017) Promoters and
enhancers enclosed by the same TAD/loop are close in 3-dimensional space, and thus are thought
to be more likely to interact with each other. A study has shown that genes within the same TAD
21
have a more similar regulation pattern than do genes in different TADs during embryonic stem
cell differentiation. (Nora et al., 2012; Merkenschlager e Nora, 2016) If a CTCF-mediated loop
anchor point is deleted, this could allow an enhancer from an adjacent TAD/loop to interact with
a promoter that was originally sequestered in an insulated neighborhood. When the 3D structure
is changed, the promoter could then be activated by “hijacking” enhancers in adjacent regions
(Figure 1.5).

Figure 1.4. Illustration representing a “CTCF-CTCF loop”.
A pair of CTCF proteins and a cohesin complex bind at loop anchors. A gene and an enhancer are constrained
in the same loop and they are thought to be more likely to interact with each other.

22
Figure 1.5. Illustration representing “enhancer hijacking”.
The expressed gene is activated by an enhancer located in the same loop, while another gene outside the loop is
sequestered from the enhancer and stays silent. After disruption of the loop, the previously silent gene “hijacks”
the active enhancer and become activated. This figure is adapted from Hnisz et al. (Hnisz, Day, et al., 2016)

The second model, here termed an “enhancer-promoter loop”, is based on the fact that a small, but
significant number of chromatin loops represent interactions between enhancers and promoters
(Figure 1.6). For example, among the loops identified in cultured neuronal cells derived from
olfactory neuroepithelium (CNON) that have at least one end mapping to an active promoter
(115,039 loops identified using a 5kb resolution Hi-C map), 11% have an enhancer at the other
end (12,970 of 115,039). (Rhie et al., 2018) When a promoter and an enhancer are brought in close
spatial proximity via looping, the gene transcription could be enhanced by the shared interactions
of transcription factors and cofactors bound at both elements.

Figure 1.6. Illustration representing an “enhancer-promoter loop”.
23
A gene and an enhancer are brought in close spatial proximity by a loop. Although this illustration shows CTCF
proteins and the cohesin complex binding at the loop anchors, they are not required for enhancer-promoter loops,
which are called based on the Hi-C interaction signals, independently of the presence of CTCF.
Although the two models provide insights into how enhancer-mediated gene regulation may work,
few studies have directly tested the function of a specific regulatory element in its natural
chromosomal location. The studies that have tested these models have used the clustered regularly
interspaced short palindromic repeats (CRISPR) system to delete, insert, mutate or epigenetically
modify specific CTCF binding sites or TAD/loop anchor points.

Using the CRISPR system to study gene regulation
The CRISPR/Cas9 system uses the Cas9 nuclease derived from Streptococcus pyogenes to
introduce DNA double-stranded breaks (DSBs) at a targeted location. A 20-nt guide sequence
(gRNA) directs Cas9 to its paired DNA target. The DNA target must immediately precede a 5′-
NGG PAM sequence. By bringing in two Cas9 proteins using two different guide RNAs in an
adjacent region, two DSBs can be created which will introduce a deletion via the nonhomologous
end joining (NHEJ) pathway, in which the ends of each DSB are directly ligated together (Figure
1.7). (Ran et al., 2013)

24

Figure 1.7. Illustration showing CRISPR/Cas9-mediated deletion via the NHEJ pathway.
The desired deletion region is flanked by a pair of gRNAs. Both gRNAs (along with Cas9) can cause DNA
double-stranded breaks (DSBs). The entire region in between of the gRNAs can be lost during the NHEJ repair
process. This figure is adapted from Ran et al. (Ran et al., 2013)

Using a CRISPR-based approach, many studies have demonstrated that deletion of TAD
boundaries or individual CTCF binding sites can cause alterations in local structure and gene
expression. (Hnisz, Day, et al., 2016; Ruiz-Velasco et al., 2017; Guo e Dean, 2018) Hnisz et al.
identified recurrent deletions in T cell acute lymphoblastic leukemia (T-ALL) genomes at TAD
boundaries that contain prominent T-ALL proto-oncogenes. After CRISPR/Cas9-mediated
deletions of two of such boundaries in nonmalignant cells, the proto-oncogenes (TAL1 and LMO2)
showed a ~2-fold increase in their expression levels. (Hnisz, Weintraub, et al., 2016) Guo et al.
deleted a CTCF loop anchor region using the CRISPR/cas9 system in 22Rv1 prostate cancer cells
and found that a gene in the loop anchored by the original CTCF site showed a hundred-fold
increase in its expression level after the deletion. (Guo et al., 2018) These results provide support
for the enhancer “hijacking” model described above.
25
A high-resolution 3D epigenomic map reveals prostate cancer-specific
alterations in chromatin architecture near the Androgen Receptor (AR) gene
Prostate cancer is the leading cancer type for the estimated new cancer cases in men in the USA.
(Siegel et al., 2018) A better understanding of the impact of chromosome structure on the prostate
cancer transcriptome may help improve diagnosis and treatment. Previous studies in the Farnham
lab (Rhie et al. manuscript in revision) have generated 3D epigenetic maps in both prostate cancer
cells and normal prostate cells, allowing the identification of prostate cancer-specific TAD
structures and enhancer-promoter loops that encompass promoters of genes that are up-regulated
in the cancer cells. These studies used 22Rv1, a diploid prostate carcinoma epithelial cell line
derived from a xenograft that was serially propagated in mice after castration-induced regression
and relapse. (Sramkoski et al., 1999; Knouf et al., 2009)
My project is focused on the characterization of a specific locus near the androgen receptor (AR)
gene in 22Rv1 prostate cancer cells. AR is expressed at low levels in normal prostate cells but at
high levels in primary prostate cancer and in metastases. The progression of prostate cancer
depends on AR signaling. Thus, prostate cancer patients are treated with androgen deprivation
therapy (ADT). However, a subset of patients become resistant to the treatment, and the cancer
develops into a lethal state of castration-resistant prostate cancer (CRPC). CRPC is found to have
greatly elevated AR expression level, and it is often associated with AR gene mutations,
amplification and abnormal expression of AR mRNA and AR protein variants. (Culig e Santer,
2014; Tan et al., 2015) AR is a type of nuclear receptor and has three major functional domains,
the N-terminal domain, the DNA binding domain, and the C-terminal ligand binding domain. After
being activated by androgenic hormones in the cytoplasm, it will translocate into the nucleus, and
26
act there as a DNA binding transcription factor. The AR-v7 variant, which is highly expressed in
CRPC, lacks the ligand binding domain and thus produces a constitutively active form of the AR
protein. (Tan et al., 2015) Notably, 22Rv1 cells express high levels of both the AR full-length
variant (AR-FL) and the AR-v7. The goal of my project was to test if the deletion of a specific
enhancer and/or CTCF binding sites located at regions of cancer-specific chromosomal structure
alterations in 22Rv1 prostate cancer cells is sufficient to cause the misregulation of AR and/or other
nearby genes.
27
CHAPTER 2
MATERIALS AND METHODS
Cell culture
The human prostate cancer cell line 22Rv1 (ATCC # CRL-2505) was obtained from ATCC
(https://www.atcc.org/). Cells were cultured in the ATCC-formulated RPMI-1640 medium,
supplemented with 10% fetal bovine serum (Gibco by Thermo Fisher Scientific, Waltham, MA,
USA) and 1% penicillin and streptomycin. Cells were cultured at 37 °C and 5% CO2. All cell
stocks were authenticated at the USC Norris Cancer Center cell culture facility by comparison to
the ATCC and/or published genomic criteria for that specific cell line; all cells were documented
as free of mycoplasma.

Plasmids construction
gRNAs were designed using website tool CRISPOR. (Haeussler et al., 2016) Double strand DNA
oligos of gRNAs were synthesized by Integrated DNA Technologies, Inc. (IDT). gRNAs used in
this project are listed in Table 2.1. The gRNA cloning vector used in this project is pSpCas9(BB)-
2A- Puro (PX459) V2.0 plasmid (from Feng Zhang, Addgene plasmid #62988;
http://n2t.net/addgene:62988; RRID: Addgene_62988). This plasmid contains Cas9 from S.
pyogenes plus a puromycin resistance gene (2A-Puro), and a cloning backbone for gRNA. This
plasmid was linearized using FastDigest BpiI (Thermo Fisher, Catalog number: FD1014). gRNAs
were cloned into linearized vector using NEBuilder HiFi DNA Assembly Master Mix (NEB,
28
Target Region Deletion Size (bp) gRNA gRNA Sequence （with PAM)
AR EN 7768
AR_EN_L TCTATATTAGGACGACGTGT GGG
AR_EN_R AAGACCTAAACAAACCACCG TGG
AR_EN_s ATGCTCCTCTAGTACATGAG TGG
CTCF site 1 2769
site_A_L2 CTCTGCTGCACGACGTGTAC GGG
site_A_R1 GGGTTGGGAGTCACGCTAAT AGG
site_A_s GATAAGTAACTCATGGACCC TGG
CTCF site 2 823
site_B_L CCCTTACTCCCCGAATTGTT TGG
site_B_R ATCCACCAGACTGATGGATT TGG
CTCF site 3 14010
site_D_G1 GAGATGACAGCTAGTTAGTA AGG
site_D_G6 TTGTTTGGTTCAAGCACAAT GGG
CTCF site 4
334 (left peak,using gRNA L2R2) site_E_L1 TACAGTCTTTACAGGTTACC TGG
20805 (both peaks, using gRNA L1L2)
site_E_L2 TGTCGCCTTGCGTTTAGAGT AGG
site_E_R2 CCTCAGTGGGATTGACCCCC AGG
Middle loop 3561
ARmidG2 CTCCTTCTAAGCCTAATCCA AGG
ARmidG3 ACCAGTTATCTCTCTTACAC AGG
Catalog number: E2621S) according to the manufacturer’s protocol. NEB 5-alpha Competent E.
coli was used for cloning (NEB, Catalog number: C2988J). To confirm successful cloning, the
resulting plasmids were subjected to sanger sequencing at GENEWIZ. E. coli stocks of
successfully cloned plasmids containing correct gRNA sequences were kept in -80 ℃.

Table 2.1. gRNAs used in the CRISPR/Cas9-mediated deletion experiments.

CRISPR/Cas9-mediated deletion experiments
22Rv1 cells were transfected with plasmids expressing Cas9 and guide RNAs using Lipofectamine
3000 Reagent (Thermo Fisher, #L3000015) according to the manufacturer’s protocol. In all
experiments, a total of 3.75 ng plasmid was used to transfect with 22Rv1 cells per well on 12 well
cell culture plates; thus, for deletions which require using two gRNAs, 1.875ng plasmid containing
29
each gRNAs were used. For single gRNA cutting and empty vector control experiments, 3.75ng
of one plasmid was used. 2 days after transfection, cells were treated with 2ug/mL puromycin. 2-
4 days after puromycin selection, cells were harvested for RNA extraction and genotyping. For
genotyping, cells were subjected to DNA extraction using the QuickExtract DNA Extraction
Solution (Epicentre #QE9050) according to the manufacturer’s protocol and genotyped by PCR
using primers listed in Table2.2.

Table 2.2. Genotyping primers used in the CRISPR/Cas9-mediated deletion experiments.
Target Region Genotyping Primer Sequence
AR EN
AR_EN_flanking_F GGGCTTGTTTAGGGCACTCA
AR_EN_flanking_R CACTGTCTTGCTGGTTTTGGTT
AR_EN_junction_R CCAGGCTCTGGGAAACAG
AR_EN_junction_F AGCTCCAAGTAGTTCCCCCT
AR_EN_inside_F TGGGGAAATACAGACATTAAACCG
AR_EN_inside_R CTGACCAAAGTAGCCCTCCAG
CTCF site 1
site_A_flanking_F1 CTGAATTCCAAAGGCTGCGG
site_A_flanking_R1 CCTATCAGCAAACTCTAAATGGCA
site_A_flanking_F4 CCATAGTTCCTGGTGACGCAG
site_A_flanking_R4 TGACTCCACCTCTATCAATGCC
site_A_inside_F TTGAGGCCGAGTCAGAGTTT
site_A_inside_R TGACCTTCACCACTGCTACA
CTCF site 2
AR-B_pL1 AGAGTTGTTTTCCTGCCTTCT
AR-B_pL2 ACATGGGCTCTGTTCTCTGC
AR-B_pR3 TTGGCTGTGTGTTGAAACCA
AR-B_pR4 TGGTGCCTAGTACAGAATTTGG
CTCF site 3
site_D_L1 ACCCAACACCCCTTTCTACC
site_D_R2 AGTCAGAACATACCCTTGTAACA
site_D_R3 TTTCTTGGACATCGCAGGTG
CTCF site 4
site_E_F TTGCTTCCTTGACACTCCTACT
site_E_inside_R GGAAAGAATACAAGGCAGGGC
site_E_flanking_R1 AGCTGTCAAACTCCCGGATG
site_E_flanking_R2 AGGAATTTTTGGGGGCTTGAA
Middle loop
ARmidintPL5 TAATCACCCAATGCCATCAG
ARmidPR8 GCATCGAAATACCATCCAGAC
ARmidPL2 TCTGCCTTTGTTCTCTGTCG
ARmidPR7 AGCAACTACACCACAGCTCAG
30
Single-cell colony isolation
22Rv1 cells were transfected and selected with puromycin as described above. The resulting
transfected cells were disassociated and sorted into 96 well plates with 1 cell/well using the BD
FACS Aria II cell sorter in the USC Flow Cytometry Facility. After sorting, the cells were kept in
RPMI-1640 Medium (as described above) supplemented with 20% fetal bovine serum and 1%
penicillin and streptomycin. After about 5 weeks, the sorted single cells had grown into colonies
visible to the naked eye. A portion of these colonies were subjected to DNA extraction (as
described above) and genotyped by PCR using primers listed in Table 2.2. Colonies carrying the
desired deletion were transferred into 6 well plates for further analysis.

RNA extraction and RT-qPCR
RNA from transfected cell pools or single-cell colonies was extracted using TRI reagent (Zymo
Research, Cat. No. R2051), according to the manufacturer’s protocol. The resulting RNAs were
converted into cDNA libraries using the iScript cDNA Synthesis Kit (BIO-RAD, #1708891).
Quantitative real-time PCR was performed using SYBR Green (Bio-Rad, #1725275) and a Bio-
Rad CFX96 machine (Bio-Rad, #1855196). RT-qPCR primers are listed in Table 2.3. Samples
tested were in technical triplicates using GAPDH as a reference mRNA.

31

Table 2.3. RT-qPCR primers used in the CRISPR/Cas9-mediated deletion experiments.
Target RT-qPCR Primer Sequence Product Size (bp)
AR ALL**
AR_fwd GTGTCAAAAGCGAAATGGGC
145
AR_rev GCTTCATCTCCACAGATCAGG
ARFL*
AR_uc004dwv.2_5_1_1f AGTGTGTCCGAATGAGGCA
124
AR_uc004dwv.2_5_1_1r TGATTTTTCAGCCCATCCAC
AR-V7
AR-v7_Phoebe_F AAAGAGCCGCTGAAGGGAAA
156
AR-v7_Phoebe_R TGCAATTGCCAACCCGGAAT
EDA2R*
EDA2R_uc004dwq.3_2_1_2f TACCGAAAGACACGCATTGG
124
EDA2R_uc004dwq.3_2_1_2r TGGGTGTATCTGCCTCCACT
OPHN1*
OPHN1_uc004dwx.3_2_1_2f TACTGGATCGGCACTTACACC
84
OPHN1_uc004dwx.3_2_1_2r GCCTCTCCTTGTCCACCTGT
YIPF6*
YIPF6_uc011mph.2_4_1_2f CTTCATGGTTCGGCTTTTTG
81
YIPF6_uc011mph.2_4_1_2r TGGCTATCAGCAAGGAAAGC
STARD8*
STARD8_uc004dxc.4_5_1_2f TACCGCGTTCATGGAGAAGT
91
STARD8_uc004dxc.4_5_1_2r ATCTGGGGTCTTGTTCCTCC
HEPH*
HEPH_uc011mpa.2_13_1_1f TACTGTCTGGCCACTGGCTG
73
HEPH_uc011mpa.2_13_1_1r GCCAGACCTCTCTGGGATG
VSIG4*
VSIG4_uc011moy.2_3_1_1f GGCAACCAAGTCGTGAGAGA
109
VSIG4_uc011moy.2_3_1_1r TCAAGGGGTATGTCATGGTTG
GAPDH
GAPDH_F AATCCCATCACCATCTTCCA
105
GAPDH_R CTCCATGGTGGTGAAGACG
RPLP0
RPLP0_F CAGGGAAGACAGGGCGAC
163
RPLP0_R GCCCATCAGCACCACAGC
* These primers are from the Human (hg19) Whole Transcriptome qPCR Primers track on UCSC Genome Browser
(https://genome.ucsc.edu/index.html)
** Primers for AR ALL are from Takeda et al.
32
CHAPTER 3
Using CRISPR-mediated Deletion to characterize a putative enhancer that is
looped to the Androgen Receptor (AR) promoter

My project is based on previous studies in the Farnham lab that have employed in situ Hi-C, ChIP-
seq for CTCF and H3K27ac (a histone modification mark for active regulatory elements), and
RNA-seq datasets from normal prostate (RWPE1) and prostate cancer (22Rv1) cells. (Guo et al.,
2018) (Rhie et al. under revision) In these studies, active promoters were defined as regions of
±2kb from the TSS of all expressed genes (identified using RNA-seq data for each cell line).
Enhancers were defined as regions corresponding to the top 25,000 H3K27ac peaks located greater
than 2 kb from a known TSS of all genes identified by GENCODE. The Hi-C datasets were derived
from ~1 billion reads per cell line. TADs were called using normalized contact matrices binned at
40kb resolution and loops (50kb-10Mb) were called using normalized contact matrices binned at
10kb resolution. To study enhancer-promoter loops, loops were selected that have an active
enhancer at one end and an active promoter at the other end. Next, enhancer-promoter loops were
selected that are (1) found in the cancer cells but not in the normal cells (“cancer-specific loops”),
(2) anchored at enhancers that are only present in cancer cells, and (3) anchored at the promoters
of genes that are statistically significantly higher expressed in cancer cells.
One of the chromosomal regions that showed the most dramatic structural change was near the AR
gene (Figure 3.1). As described above, the AR protein is a major regulator of prostate cancer. My
specific goals were to characterize the chromosomal region surrounding the AR gene. RNA-seq
33
data showed that the expression level of AR is elevated by about 300-fold in 22Rv1 prostate cancer
cells, as compared with the RWPE1 normal prostate cells.

Figure 3.1. Hi-C chromatin interaction maps of a 3 Mb region containing the AR gene in normal prostate
(RWPE1) and prostate cancer (22Rv1) cells.
The Hi-C interaction maps show loop structures that are present in 22Rv1 cancer cells but are absent in RWPE1
normal cells. Tracks show H3K27ac ChIP-seq, CTCF ChIP-seq, RNA-seq and UCSC gene. AR expression level
is significantly high in 22Rv1 cells as indicated by RNA-seq. Squares indicate locations of CTCF binding sites
(yellow, named CTCF site 1, 2, 3, 4) and H3K27ac peaks (green, named AR EN) that are overlapped with Hi-C
loop anchors. The blue square indicates the location of a region in the middle of the AR enhancer-promoter loop.
These sites will be individually removed via CRISPR/Cas9 editing. (Guo et al; Rhie et al. under revision)

CTCF
RNA-seq
H3K27ac
UCSC genes
CTCF
RNA-seq
H3K27ac
UCSC genes
CTCF Site 1 CTCF Site 2 CTCF Site 3 CTCF Site 4
1 2 3 4
Putative AR Enhancer
AR EN
Middle Loop Control
MID
CTCF Site 1 CTCF Site 2 CTCF Site 3 CTCF Site 4
1 2 3 4
Putative AR Enhancer
AR EN
Middle Loop Control
MID
34
CRISPR/Cas9-mediated deletion of a putative AR enhancer (AR EN) in a
22Rv1 cell population
Our studies have identified a 22Rv1 cancer-specific loop that connects the AR promoter to a
cancer-specific H3K27ac peak located ~770 kb away. This H3K27ac peak is absent in RWPE1
cells. Therefore, my first set of experiments was to test the hypothesis that the cancer-specific
enhancer-promoter loop shown in Figure 3.1 contributes to the upregulation of the AR gene in
22Rv1 cells.
I designed two gRNAs encompassing a 7.8 kilobase region containing the putative AR enhancer
(termed AR EN). This region also harbors a CTCF binding site. I decided to remove this CTCF
binding site together with the putative enhancer (i.e. the H3K27Ac mark). If this led to any effects
on AR expression, a more precise deletion could then be created to distinguish functions of the
CTCF site and the putative enhancer. I transfected 22Rv1 cells with gRNAs according to the
CRISPR/Cas9 deletion protocol described in the methods section. Upon transfection of a pool of
cells, only a portion of the cells will be carrying the desired deletion. However, if the target gene
expression level changes greatly after deletion, the effect can still be detected by RT-qPCR, as
demonstrated in a previous study. (Guo et al., 2018)
The success in detecting a change in expression in a pool of cells depends on the deletion
efficiency, which in turn depends on the efficiency of interaction of the gRNAs with the target site
(which is related to genomic context) and transfection efficiency. I used website tools such as
CRISPOR (http://crispor.tefor.net/) (Haeussler et al., 2016) to ensure that my gRNA designs were
as good as possible. To test if my gRNA pair can introduce correct deletion in the transfected cells,
I designed two sets of PCR primers, as indicated in the following figure (Figure 3.2): (1) a flanking
35
primer pair that will only produce a small sized product when the targeted region is deleted and
(2) a junction primer pair (one primer is inside the deletion region) that only produces products if
the targeted region is still present (not deleted). I used these two sets of primers to run a PCR with
DNA samples extracted from the CRISPR/Cas9-edited cells. As showed in Figure 3.2, when
amplified using the junction primer pair, DNA samples from both AR EN gRNAs transfected cells
(lane 2) and empty vector transfected cells (lane 7) produced correct products. This is because the
AR EN gRNAs transfected cells are a mixture of deleted and non-deleted cells. When amplified
using the flanking primer pair, only DNA samples from AR EN gRNAs transfected cells produced
correct products (lanes 3,4,5); the cells transfected with the empty vector did not produce a band
because it would have been 7.8 kb which does not PCR with high efficiency (lane 6). These results
indicate that my gRNAs can successfully delete the AR EN in a pool of cells.

Figure 3.2. CRISPR/Cas9-mediated deletion of a putative AR enhancer (AR EN).
Left: Schematic showing the location of AR EN deletion gRNAs (red lines) and genotyping primers (black
arrows). Tracks show H3K27ac ChIP-seq, CTCF ChIP-seq, and CTCF motif in 22Rv1 cells (arrow shows the
CTCF motif orientation). Grey rectangle indicates the targeted deletion region (7.8kb). (Guo et al; Rhie et al.
under revision). Right: Agarose gel analysis of PCR samples. Flanking primer products (884 bp) can only be
amplified from chromosomes carrying the deletion. They were present in AR EN gRNA transfected 22Rv1 cells
Sample AR EN gRNA Empty vector
WT 433 / / / / 433
Deletion / 884 884 884 / /
Expected Product Size
Primers Junction Flanking Junction
0.5 kb
1.0 kb
1 2 3 4 5 6 7
Flanking Primers
Junction Primers
Deletion
gRNA
Deletion
gRNA
Deletion 7.8 kb
22Rv1 H3K27Ac
22Rv1 CTCF
CTCF motif
36
while absent in empty vector transfected cells. The junction primer product (433 bp) can be amplified from both
the empty vector transfected cells and the cells transfected with the AR EN gRNAs.

Having determined that the AR EN deletion had been deleted in some of the transfected cells, I
tested AR mRNA levels in empty vector transfected cells (control) and AR EN gRNAs transfected
cells. As described above, AR is expressed as both the normal, full length allele (AR-FL) and as a
shortened, constitutively active variant (AR-v7) in 22Rv1 cells. I designed RT-qPCR primers that
would detect AR-v7 and both (AR all) transcripts after CRISPR/Cas9-mediated deletion in the
22Rv1 cell population. I observed that the expression level of AR-v7 decreased by ~35%, AR all
variant by ~20% (Figure 3.3). The figure shows results from one transfection experiment,
performed with triplicates for the control and experimental samples. I repeated this experiment
several times, obtaining similar results; results from other transfection experiments are listed in
Table 3.2.) These initial results suggested that deletion of the AR EN affected the transcript level
of AR, suggesting that perhaps this enhancer was a critical regulator of the AR promoter.

0
0.2
0.4
0.6
0.8
1
1.2
AR ALL AR-V7
Relative Normalized Expression
Cas9-CTRL Cas9-AR EN
*** ***
37
Figure 3.3. Expression of AR after CRISPR/Cas9-mediated deletion of the putative AR enhancer (AR EN).
All transfections were performed in triplicate in 22Rv1 cells then analyzed by RT-qPCR. DNA samples were
extracted from the same cell samples to confirm successful deletion by PCR. AR expression was normalized to
GAPDH. Bar plots show the average of the triplicate experiments, with error bars showing the standard error of
the mean (SEM). Significance was determined by a Student’s t test (AR EN deletion vs. empty vector control:
AR ALL ***p < 0.001, AR-V7 ***p < 0.001)

Generation of single-cell derived clones carrying the AR EN deletion using
22Rv1 cells
As discussed above, only a small portion of the transiently transfected cells carry the desired
deletion. To test if the AR expression level would show a more significant decrease in a cell
population in which all cells are carrying the AR EN deletion, I generated single-cell derived AR
EN deletion clones using 22Rv1 cells.
I transfected 22Rv1 cells with AR EN deletion gRNAs. As a control, I transfected 22Rv1 cells
with the empty plasmids. These cells were sorted into 96 well plates, one cell per well. After about
4 weeks, the single cell-derived colonies were screened by PCR using genotyping primers listed
in Table 2.2 for the AR EN deletion. It should be noted that one disadvantage of using clones is
that clonal variation cannot be avoided. When the effect itself is small, a statistically significant
interpretation may require a large sample size (i.e. a large number of clones). Therefore, I screened
58 single-cell clones and identified 11 clones that were fully deleted for the AR EN. I also
generated 4 control clones isolated from cells transfected with empty plasmids without gRNAs. I
observed a great clonal variation in AR expression level, even in the control clones (Figure 3.4).
38
With this limited sample size, I could not detect a significant difference between deletion clones
and control clones.

Figure 3.4. Expression of AR in clones carrying the AR EN deletion.
DNA samples were extracted from each clonal population from the enhancer deletion transfection to confirm
successful deletion by PCR; 11 AR EN deletion clones were identified. Control clones are derived from empty
vector plasmid transfected 22Rv1 cells; 5 control clones were tested. AR expression was normalized to GAPDH
and relative to control clone “C60”. Bar plot shows average with error bar as standard error of mean (SEM).
Significance was determined by Student’s t test: deletion clones were not significantly different from the empty
vector control clones.
AR ALL AR−V7
C60 C59 C61 C62 C63 EN3 EN5 EN6 EN15 EN16 EN18 EN26 EN30 EN37 EN50 EN54
0.0
0.5
1.0
1.5
2.0
0.0
0.5
1.0
1.5
2.0
Sample
Relative Normalized Expression
Group
CTRL
DEL
39
CHAPTER 4
Using CRISPR-mediated deletion to characterize four CTCF anchor sites
near the AR gene in 22RV1 cells

In the next set of experiments, I tested a second hypothesis: that the cancer-specific loop structures
contribute to the regulation of AR and/or other nearby genes. As showed in Figure 3.1, there are
four CTCF binding sites that overlap with Hi-C loop anchors near the AR gene. I named these
anchor CTCF sites 1, 2, 3 and 4. One CTCF-CTCF loop is anchored by CTCF site 1 and CTCF
site 2; this large loop contains the enhancer-promoter loop discussed above. The AR promoter is
also located in a CTCF-CTCF loop anchored at CTCF site 2 and CTCF site 3. Finally, a smaller
loop is anchored at CTCF site 3 and CTCF site 4. Apart from AR, there are six other genes in this
region. As showed in Figure 4.1, CTCF site 2 is located in the first intron of AR. CTCF site 3
contains two CTCF peaks and the deletion of both peaks will remove three exons of the OPHN1
gene. Similarly, CTCF site 4 also contains two peaks. And only the deletion of both peaks will
remove the first exon of the YIPF6 gene.

40

CTCF Site 1
CTCF Site 4
CTCF Site 3
CTCF Site 2
41
Figure 4.1. CRISPR/Cas9-mediated deletions of CTCF site 1, 2, 3 and 4.
Left: Schematic showing the location of deletion gRNAs (red lines) and genotyping primers (black arrows).
Tracks show H3K27ac ChIP-seq, CTCF ChIP-seq, and CTCF motifs in 22Rv1 cells (arrows show the CTCF
motif orientation). Grey rectangle indicates the targeted deletion region. (Guo et al; Rhie et al. under revision)
Right: Agarose gel analysis of PCR samples. Flanking primer products can only be amplified from chromosomes
carrying the deletion due to the large size of the product if there is no deletion. The expected deletion product
was observed when analyzing deletion gRNA transfected 22Rv1 cells but was absent in empty vector transfected
cells. The inside/junction primer product was observed to be amplified from both the empty vector transfected
cells and deletion gRNA transfected cells, indicating that the pool of cells contained both deleted and undeleted
alleles. For CTCF site 4, the figure shows a deletion of the left CTCF peak only and a deletion of both CTCF
peaks.

CRISPR/Cas9-mediated deletion of the four CTCF anchor sites in a 22Rv1 cell
population
My hypothesis is that disruptions of these loops via deletion of the individual CTCF binding sites
will cause expression level changes of nearby genes. Following the same procedures as in the AR
EN deletion experiments, I designed gRNAs to delete the four CTCF sites individually using the
CRISPR/Cas9 system. As showed in Figure 4.1, PCR experiments have demonstrated that the
gRNAs for CTCF site 1, site 2 and site 3 can successfully produce the correct deletion in
transfected 22Rv1 cells. However, for CTCF site 4, I was only successful in producing a deletion
of the left peak. However, I will describe below that, after using a different gRNA, I could
successfully delete both peaks of site 4.
I tested for expression level changes of AR and the six other nearby genes in cells harboring these
four deletions using RT-qPCR. However, gene VSIG4, HEPH and STARD8 are lowly expressed
42
in 22Rv1 cells and no significant increase of expression level was observed after any of these
deletion experiments. Thus Figure 4.2 and Figure 4.3 show only the RT-qPCR results for AR,
EDA2R, OPHN1 and YIPH6. Details of these results, including t-tests can be found in Table 3.1.
After deletion of CTCF site 1 and CTCF site 2, individually, the AR variant expression level
showed a significant decrease (AR ALL decreased by ~40% upon deletion of either CTCF site 1
or CTCF site 2), while deletions of CTCF site 3 and CTCF site 4 had little effects on AR variant
expression level. Expression level changes of EDA2R, OPHN1 and YIPH6 after deletion of each
of the CTCF sites were generally modest. However, after deletion of AR EN, CTCF site 1 and
CTCF site 4 (left peak only), OPHN1 and YIPH6 had a statistically significant increase of
expression level by~20%. For CTCF site 1 and site 4, this could be explained by the insulated
neighborhood model, in which the disruption of loop anchors may lead to alterations in
chromosome structure and thus it changes the expression level of the nearby genes.

43

AR ALL ARFL AR-V7
0
0.2
0.4
0.6
0.8
1
1.2
AR ALL
Relative Normalized Expression
Cas9-CTRL Cas9-AR EN
0
0.2
0.4
0.6
0.8
1
1.2
ARFL
Relative Normalized Expression
Cas9-CTRL Cas9-AR EN
0
0.2
0.4
0.6
0.8
1
1.2
AR-V7
Relative Normalized Expression
Cas9-CTRL Cas9-AR EN
0
0.2
0.4
0.6
0.8
1
1.2
AR ALL
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 1
0
0.2
0.4
0.6
0.8
1
1.2
ARFL
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 1
0
0.2
0.4
0.6
0.8
1
1.2
AR-V7
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 1
0
0.2
0.4
0.6
0.8
1
1.2
AR ALL
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 2
0
0.2
0.4
0.6
0.8
1
1.2
ARFL
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 2
0
0.2
0.4
0.6
0.8
1
1.2
AR-V7
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 2
0
0.2
0.4
0.6
0.8
1
1.2
AR ALL
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 3
0
0.2
0.4
0.6
0.8
1
1.2
ARFL
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 3
0
0.2
0.4
0.6
0.8
1
1.2
AR-V7
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 3
0
0.2
0.4
0.6
0.8
1
1.2
AR ALL
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 4
0
0.2
0.4
0.6
0.8
1
1.2
ARFL
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 4
0
0.2
0.4
0.6
0.8
1
1.2
AR-V7
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 4
RNA-seq
22Rv1
CTCF Site 1 CTCF Site 2 CTCF Site 3 CTCF Site 4
1 2 3 4
Putative AR Enhancer
AR EN
44
Figure 4.2. Expression of AR after CRISPR/Cas9-mediated deletions.
Results show CRISPR/Cas9-mediated deletions of the putative AR enhancer (AR EN) and CTCF binding sites
1, 2, 3 and 4 (left peak), individually, in 22Rv1 cells. All samples in RT-qPCR experiments were in triplicates.
DNA samples were extracted from the same cell samples to confirm successful deletion by PCR. AR expression
was normalized to GAPDH. Bar plot shows average with error bar as standard error of mean (SEM). Significance
was determined by Student’s t test, listed in Table 3.1.

45

EDA2R OPHN1 YIPF6
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
EDA2R
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
OPHN1
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
YIPF6
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 1
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
EDA2R
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
OPHN1
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
YIPF6
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 2
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
EDA2R
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
OPHN1
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
YIPF6
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 3
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
EDA2R
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 4
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
OPHN1
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 4
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
YIPF6
Relative Normalized Expression
Cas9-CTRL Cas9-CTCF site 4
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
EDA2R
Relative Normalized Expression
Cas9-CTRL Cas9-AR EN
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
OPHN1
Relative Normalized Expression
Cas9-CTRL Cas9-AR EN
0
0.2
0.4
0.6
0.8
1
1.2
1.4
1.6
1.8
YIPF6
Relative Normalized Expression
Cas9-CTRL Cas9-AR EN
RNA-seq
22Rv1
CTCF Site 1 CTCF Site 2 CTCF Site 3 CTCF Site 4
1 2 3 4
Putative AR Enhancer
AR EN
46
Figure 4.3. Expression of EDA2R, OPHN and YIPH6 after CRISPR/Cas9-mediated deletions.
Results show CRISPR/Cas9-mediated deletions of the putative AR enhancer (AR EN) and CTCF binding sites
1, 2, 3 and 4 (left peak), individually, in 22Rv1 cells. All samples in the RT-qPCR experiments were performed
in triplicates. DNA samples were extracted from the same cell populations to confirm successful deletion by
PCR. Expression of EDA2R, OPHN and YIPH6 were normalized to GAPDH. Bar plots show the average
expression with error bars showing the standard error of mean (SEM). Significance was determined by Student’s
t test, listed in Table #3.1.

Generation of single cell-derived clones carrying deletions of the four CTCF
anchor sites
Next, I generated single cell-derived clones carrying deletions of CTCF site 1, 2, and 3
individually. Similar to the deletion clones of the putative AR enhancer, in the deletion clones of
CTCF site 1, 2 and 3, I observed a great clonal variation in expression level of not only AR, but
also the other three genes, EDA2R, OPHN and YIPH6 (Figure 4.4, Figure 4.5, Figure 4.6, Figure
4.7, Figure 4.8). Unexpectedly, in most of the deletion clones, the expression of AR was higher
than in the control clones (Figure 4.4).

47

Figure 4.4. Expression of AR in clones carrying the CTCF site 1 deletion.
DNA samples were extracted from each clonal population from the CTCF site 1 deletion transfection to confirm
successful deletion by PCR; 10 CTCF site 1 deletion clones were identified (labeled numbers begin with “A”).
Control clones are derived from 22 Rv1 cells transfected with an empty vector plasmid; 4 control clones were
tested (labeled numbers begins with “C”). AR expression was normalized to GAPDH and relative to control
clone “C1”. Bar plots show an average, with error bar as standard error of mean (SEM). Significance was
determined by Student’s t test: AR ALL: p < 0.01, AR-V7: deletion clones were not significantly different from
the empty vector control clones.

AR ALL AR−V7
C1 C2 C4 C5 A4 A14 A15 A17 A25 A30 A32 A42 A53 A56
0.0
0.5
1.0
1.5
2.0
2.5
0.0
0.5
1.0
1.5
2.0
2.5
Sample
Relative Normalized Expression
Group
CTRL
DEL
48

Figure 4.5. Expression of EDA2R, OPHN and YIPH6 in clones carrying the CTCF site 1 deletion.
Expression of EDA2R, OPHN and YIPH6 was normalized to GAPDH and relative to control clone “C1”. Bar
plot shows average with error bar as standard error of mean (SEM). Significance was determined by Student’s t
test: deletion clones were not significantly different from the empty vector control clones.

EDA2R OPHN1 YIPF6
C1 C2 C4 C5 A4 A14 A15 A17 A25 A30 A32 A42 A53 A56
0.0
0.5
1.0
1.5
2.0
0.0
0.5
1.0
1.5
2.0
0.0
0.5
1.0
1.5
2.0
Sample
Relative Normalized Expression
Group
CTRL
DEL
49

Figure 4.6. Expression of AR in clones carrying the CTCF site 2 deletion.
DNA samples were extracted from each clonal population from the CTCF site 2 deletion transfection to confirm
successful deletion by PCR; 9 CTCF site 2 deletion clones were identified (labeled numbers begin with “B”).
Control clones are derived from empty vector plasmid transfected 22Rv1 cells; 4 control clones were tested (the
same 4 control clones as showed in Figure 4.4 and Figure 4.5, labeled numbers begin with “C”). AR expression
was normalized to GAPDH and relative to control clone “C1”. Bar plot shows average with error bar as standard
error of mean (SEM). Significance was determined by Student’s t test: deletion clones were not significantly
different from the empty vector control clones.

AR ALL AR−V7
C1 C2 C4 C5 B1 B3 B14 B19 B36 B55 B59 B66 B79
0.0
0.5
1.0
0.0
0.5
1.0
Sample
Relative Normalized Expression
Group
CTRL
DEL
50

Figure 4.7. Expression of EDA2R, OPHN and YIPH6 in clones carrying the CTCF site 2 deletion.
Expression of EDA2R, OPHN and YIPH6 was normalized to GAPDH and relative to control clone “C1”. Bar
plot shows average with error bar as standard error of mean (SEM). Significance was determined by Student’s t
test: deletion clones were not significantly different from the empty vector control clones.

EDA2R OPHN1 YIPF6
C1 C2 C4 C5 B1 B3 B14 B19 B36 B55 B59 B66 B79
0.0
0.5
1.0
1.5
2.0
0.0
0.5
1.0
1.5
2.0
0.0
0.5
1.0
1.5
2.0
Sample
Relative Normalized Expression
Group
CTRL
DEL
51

Figure 4.8. Expression of AR in clones carrying the CTCF site 3 deletion.
DNA samples were extracted from each clonal population from the CTCF site 3 deletion transfection to confirm
successful deletion by PCR; 7 CTCF site 3 deletion clones were identified (labeled numbers begin with “D”).
Control clones are derived from empty vector plasmid transfected 22Rv1 cells; 4 control clones were tested
(labeled numbers begin with “C”). AR expression was normalized to RPLP0 and relative to control clone “C6”.
Bar plot shows average with error bar as standard error of mean (SEM). Significance was determined by
Student’s t test: deletion clones were not significantly different from the empty vector control clones.
AR ALL AR−V7
C6 C7 C8 C9 D1 D2 D13 D15 D28 D30 D44
0.0
0.5
1.0
1.5
0.0
0.5
1.0
1.5
Sample
Relative Normalized Expression
Group
CTRL
DEL
52
CRISPR/Cas9-mediated deletion of both CTCF site 1 and site 2 in a CTCF site
2 deletion clone
To see if greater changes on gene expression level would occur when the CTCF sites at both loop
anchors were removed, I performed a transient transfection to delete CTCF site 1 in a single-cell-
derived clone carrying the CTCF site 2 deletion. In the same experiment, CTCF site 1 deletion was
performed in wild-type 22Rv1 cells as a control. I tested the expression level of AR variants,
EDA2R, OPHN1 and YIPF6 in these samples by RT-qPCR. When using single-cell derived clones
to test gene expression level changes, controls must be carefully designed to obtain a meaningful
interpretation of the results. In this experiment, there is an internal difference between the wild-
type 22Rv1 cells and the CTCF site 2 deletion clone (B1) and it would not be meaningful to
compare them directly. However, we can see that CTCF site 1 deletion in a single-cell-derived
clone carrying the CTCF site 2 deletion led to further decrease of AR expression (Figure 4.9). I
further repeated this experiment in a different site 2 clone (Figure 4.9). However, I found the AR
expression varies in different clones. Results of these experiments can be found in Table 3.1.

53

Figure 4.9. Expression of AR after CRISPR/Cas9-mediated deletions of both CTCF site 1 and site 2.
CTCF site 2 clone B#1 (up) CTCF site 2 clone B#14 (down) were used in these experiments. All transfections
were performed in triplicate in 22Rv1 cells then analyzed by RT-qPCR. DNA samples were extracted from the
same cell samples to confirm successful deletion by PCR. AR expression was normalized to RPLP0. Bar plots
show the average of the triplicate experiments, with error bars showing the standard error of the mean (SEM).
Significance was determined by a Student’s t test (***p < 0.001, **p < 0.01) (p-values are listed in Table 3.1)

0.00
0.20
0.40
0.60
0.80
1.00
1.20
AR ALL AR-V7
Normalized Relative Expression
WT + empty vector WT + site#1 deletion
site#2 clone + empty vector site#2 clone + site#1 deletion
***
**
***
**
0
0.5
1
1.5
2
2.5
3
AR ALL AR-V7
Normalized Relative Expression
WT + empty vector WT + site#1 deletion
site#2 clone + empty vector site#2 clone + site#1 deletion
***
***
**
***
54
I generated single cell-derived clones carrying deletions of both CTCF site 1 and 2 using the CTCF
site 2 deletion clone (B14). In these clones the expression of AR was lower compared to control
clones. These control clones derived from wild-type 22Rv1 cells were transfected with empty
plasmids without gRNAs. Because in the original CTCF site 2 deletion clone B14 the expression
of AR was lower than the average expression among CTCF site 2 deletion clones, the lower
expression of AR in these clones may be affected by this selection. However, the difference in AR
expression was still limited (Figure 4.10).

55

Figure 4.10. Expression of AR in clones carrying deletions of both CTCF site 1 and 2.
DNA samples were extracted from each clonal population from the transfection to confirm successful deletion
by PCR; 6 deletion clones were identified (labeled numbers begin with “AB”). Control clones are derived from
empty vector plasmid transfected 22Rv1 cells; 4 control clones were tested (labeled numbers begin with “C”).
AR expression was normalized to RPLP0 and relative to control clone “C6”. Bar plot shows average with error
bar as standard error of mean (SEM). Significance was determined by Student’s t test: AR ALL: p<0.05 AR-v7:
deletion clones were not significantly different from the empty vector control clones.
AR ALL AR−V7
C6 C7 C8 C9 AB4 AB8 AB18 AB26 AB28 AB57
0.0
0.5
1.0
1.5
0.0
0.5
1.0
1.5
Sample
Relative Normalized Expression
Group
CTRL
DEL
56
CHAPTER 5
Control experiments:
CRISPR/Cas9-mediated single-gRNA-cutting near loop anchors and deletion
of a region in the middle of the AR enhancer-promoter loop
To further investigate if the putative AR enhancer (AR EN) and the four CTCF sites located at loop
anchors are critical to AR expression regulation, I repeated the transfection experiments, but this
time comparing the results using the paired set of deletion gRNAs vs a single gRNA that is located
in adjacent areas. In the case of AR EN, I used a different gRNA located at ~4kb away from the
H3K27Ac peak. Similarly, for CTCF site 1, I used a different gRNA ~500 bp away from the CTCF
peak. For CTCF site 2, I used one of the gRNAs in the deletion experiment (the one on the right
side) (Figure 5.1). Unexpectedly, I observed a greater decrease of AR expression level when
targeting with the single gRNA near AR EN (Figure 5.2). For CTCF site 1 and site 2 (Figure 5.3),
I observed similar level of decrease of AR expression after the single-gRNA-cutting as compared
to it after the deletions of CTCF site 1 or 2 showed in Figure 4.2.

57

Figure 5.1. CRISPR/Cas9-mediated single-gRNA-cutting near AR EN, CTCF site 1 and CTCF site 2.
Schematic showing the location of gRNAs used in the deletion experiments (red lines) and gRNAs used in the
single-gRNA-cutting experiments (purple lines). Tracks show H3K27ac ChIP-seq, CTCF ChIP-seq, and CTCF
motif in 22Rv1 cells (arrow shows the CTCF motif orientation). Grey rectangles indicate the targeted deletion
regions. (Guo et al; Rhie et al. under revision).
CTCF motif
Deletion
gRNA
Deletion
gRNA
Deletion 7.8 kb
Single gRNA
cutting
22Rv1 H3K27Ac
22Rv1 CTCF
AR EN
58

Figure 5.2. Expression of AR after AR EN deletion and single-gRNA-cutting near AR EN.
All transfections were performed in triplicate in 22Rv1 cells and analyzed by RT-qPCR. For the deletion
experiments, DNA was extracted from cell samples to confirm successful deletion by PCR. AR expression was
normalized to GAPDH. Bar plot shows average with error bar as standard error of mean (SEM). Significance
was determined by Student’s t test: AR EN deletion vs. empty vector control. (AR ALL ***p < 0.001, AR-V7
***p < 0.001) Single gRNA cutting vs. empty vector control. (AR ALL ***p < 0.001, AR-V7 ***p < 0.001)

0
0.2
0.4
0.6
0.8
1
1.2
AR ALL AR-V7
Relative Normalized Expression
Cas9-CTRL Cas9-AR EN Cas9-Single gRNA
*** ***
*** ***
59

Figure 5.3. Expression of AR after single-gRNA-cutting near CTCF site 1 and 2.
All transfections were performed in triplicate in 22Rv1 cells then analyzed by RT-qPCR. DNA samples were
extracted from the same cell samples to confirm successful deletion by PCR. AR expression was normalized to
GAPDH. Bar plots show the average of the triplicate experiments, with error bars showing the standard error of
the mean (SEM). Significance was determined by a Student’s t test. CTCF site 1 single gRNA vs. empty vector
control: AR ALL **p < 0.01, AR-V7 ***p < 0.001. CTCF site 2 single gRNA vs. empty vector control: AR
ALL **p < 0.001, AR-V7 ***p < 0.001.

The CRISPR experiment using a single gRNA will not delete the targeted region (which I
confirmed by PCR) but will still cause a single double-strand DNA break (DSB). Thus, it was
possible that the AR expression changes observed after deletion may have been due to a transient
0.00
0.20
0.40
0.60
0.80
1.00
1.20
AR ALL AR-V7
Relative Noermalized Expression
Cas9-CTRL Cas9-CTCF site #1 Single gRNA
Cas9-CTCF site #2 Single gRNA
**
** **
*
***
60
activation of the DNA damage and repair pathway, caused as a result of CRISPR-mediated genome
cutting rather than due to the removal of a critical regulatory element. To test this hypothesis, I
deleted a region harboring no CTCF binding sites, H3K27Ac signal or genes. This region lies ~340
kb from the AR EN (toward the AR gene body) (Figure 3.1 and Figure 5.4). After this deletion, the
AR variants expression level stayed the same (Figure 5.5).

Figure 5.4. CRISPR/Cas9-mediated deletion of a region in the middle of the AR enhancer-promoter loop
(MID).
Left: Schematic showing the location of MID deletion gRNAs (red lines) and genotyping primers (black arrows).
The gRNA on the left side is used for single-gRNA-cutting. The ChIP-seq tracks show that there is no H3K27ac
or CTCF, at this location and the RNA-seq track shows that there is no gene expression in this region. In 22Rv1
cells. Grey rectangle indicates the targeted deletion region (3.6 kb). Right: Agarose gel analysis of PCR samples.
Flanking primer products (268 bp) can only be amplified from chromosomes carrying the deletion. They were
present in MID gRNA transfected 22Rv1 cells while absent in single gRNA and empty vector transfected cells.
The junction primer product (500 bp) can be amplified from all samples.

Flanking Primers
Junction Primers
gRNA gRNA Deletion 3.6 kb
22Rv1 H3K27Ac
22Rv1 CTCF
22Rv1RNA-seq
Single gRNA cutting
61

Figure 5.5. Expression of AR after deletion of the region in the middle of the AR promoter-enhancer loop.
All transfections were performed in triplicate in 22Rv1 cells and analyzed by RT-qPCR. For the deletion
experiments, DNA was extracted from cell samples to confirm successful deletion by PCR. AR EN single-
gRNA-cutting (AR EN SINGLE-CUT) was used as positive control. AR variants expression was normalized to
GAPDH. Bar plot shows average with error bar as standard error of mean (SEM). Significance was determined
by Student’s t test: AR EN single-gRNA-cutting vs. empty vector control. (AR ALL ***p < 0.001, AR-V7 ***p
< 0.001) Results from the middle loop region deletion (MID DELETION) and corresponding single-gRNA-
cutting (MID SINGLE-CUT) were not significantly different from the empty vector control.

These results suggest that CRISPR-mediated genome cutting alone is not sufficient to cause AR
expression level changes. The location targeted by gRNAs was relevant to AR expression. The AR
EN is proximal to a loop anchor, as showed in the Hi-C map in Figure 3.1. Although the single-
gRNA-cutting does not introduce a deletion, DSBs or small deletions at the targeted location may
still disrupt looping. In the case of AR EN, the DSBs or small deletions may affect their
interactions with unknown TFs or co-factors.
0
0.2
0.4
0.6
0.8
1
1.2
AR ALL AR-V7
Relative Normalized Expression
CTRL AREN SINGLE-CUT MID DELETION MID SINGLE-CUT
*** ***
62
CHAPTER 6
DISCUSSION

AR expression level decreased by less than 2-fold after CRISPR/Cas9-mediated
deletions of a putative enhancer and four loop anchors (individually) near the
AR gene in 22Rv1 cells
In this project, I have introduced five different deletions individually using the CRISPR/Cas9
system in 22Rv1 cells. These five deletions are: (1) a putative enhancer that is looped to the AR
promoter (which also harbors a CTCF binding site) and (2) four CTCF binding sites at four loop
anchors distributed near the AR gene. The deletion experiments were first performed in a pool of
cells, thus only a portion of the cells was carrying the desired deletion. After deletion of the
putative AR enhancer (AR EN), CTCF site 1 and CTCF site 2, expression of AR decreased by
20%-40%. As negative control experiments, I introduced single-gRNA-cutting near the putative
enhancer, CTCF site 1 and CTCF site 2 using the same protocol. Also, I deleted a region in the
middle of the AR enhancer-promoter loop. Surprisingly, all the single-gRNA-cutting experiments
led to a similar or even greater decrease of the AR expression level. While the deletion in the
middle of the loop did not cause any AR expression changes. These results suggest that the
CRISPR/Cas9-mediated cutting near loop anchor regions (with 4kb being the maximum distance
from the loop anchor for the gRNA near AR EN) can affect nearby gene expression, but that the
CTCF binding sites at the loop anchors may not be involved in gene regulation. The unchanged
AR expression level after deletion in the middle of the loop indicates that not all regions cause the
63
same effects when targeted by Cas9. It is possible that the single cuts that caused changes in gene
expression also caused changes in the 3D chromatin structure. However, Hi-C experiments are
needed to make sure if the chromosome structures were truly altered after either deletion of CTCF
binding sites or the single-gRNA-cutting. Furthermore, even if the loops were disrupted, such
changes in chromosome structure still had little effects on the nearby gene expression.

CRISPR/Cas9-mediated DSBs at loop anchors affect nearby gene expression
In the deletion experiments in cell pools, after transfecting cells with deletion gRNAs, the number
of cells actually carrying the deletion is relatively small. Compared to this, DSBs at gRNA targeted
locations would be happening at a higher frequency. Thus, the decrease on AR expression we saw
in cell pools after deletions of AR EN, CTCF site 1 and site2 may mostly result from Cas9-
introduced DSBs. This could also explain why we saw little effect on AR expression in clones
carrying the desired deletions. The reasons why the single-gRNA-cutting experiment near AR EN
yielded better results than the deletion experiment of AR EN could be: (1) the efficiency of the
gRNA used in the AR EN single-gRNA-cutting experiment was better than those used for deletion.
This may be due to both the propriety of gRNA sequence and the genomic location. (2) in single-
gRNA-cutting experiments, the amount of the same gRNA used in transfections was doubled
compared to it in a deletion experiment, which may cause a more efficient cutting at the targeted
location.

64

Figure 6.1. Location of gRNAs used in both single-gRNA-cutting and deletion experiments near AR EN
and CTCF site 1.
Schematic showing the location of gRNAs used in the deletion experiments (red lines) and gRNAs used in the
single-gRNA-cutting experiments (purple lines). Tracks show H3K27ac ChIP-seq, CTCF ChIP-seq and TADs
identified using Hi-C data in 22Rv1 cells. Grey rectangle indicates the targeted deletion regions.

Since we think that the CTCF binding sites at the loop anchors may not be involved in gene
regulation, it would be interesting to know what caused AR expression changes after the Cas9-
introduced DSBs. We cannot exclude the possibility that 3D chromatin structure is relevant to the
AR expression changes. As showed in Figure 5.1, the gRNAs targeted near AR EN and CTCF site
1 all located around a ~100kb region near the TAD boundary. The other boundary of this TAD is
located at the AR promoter. Whether this 100kb region is important to AR expression needs further
investigations (Figure 6.1).
65
Interpretation of small fold changes (< 2 by RT-qPCR) observed after CRISPR
experiments and clonal variation in single-cell-derived clones
Attention should be paid to the meaningful interpretation of any subtle fold changes (< 2 by RT-
qPCR) observed after CRISPR/Cas9-mediated deletion experiments no matter using a wild-type
cell population or single-cell-derived clones. These changes may be due to random variations or
artifacts brought by the CRISRP system. In such situations, it will be critical to: (1) repeat the
CRISPR/Cas9-mediated deletion experiments; (2) design negative and positive control
experiments; (3) for RT-qPCR experiments, test each primer pair and use at least two reference
genes. I have observed variations on the expression level of GAPDH in clones. In later
experiments, I used both GAPDH and RPLP0 as reference genes.
If the RT-qPCR results were to be consistent, RNA-seq could be performed in deletion clones to
further investigate the function of targeted functional elements. However, simply using RT-qPCR,
I have already observed variations of the expression level of AR and nearby genes. These clonal
variations may be due to several reasons: (1) The expression of genes in the original wild-type
cells were not on the same level but were distributed within a certain range. After the selection
process during colony expansion, such variations became exaggerated in single-cell-derived
clones. (2) The original wild-type cells carry heterozygous SNPs. In the isolated single cell,
differences in epigenetic features such as DNA methylation may cause one allele to be expressed
differently than the other allele, and the allele that dominates could be different in each clone. (3)
The single cell that was isolated may carry somatic mutations. AR has been observed to be
frequently mutated in PCa cells, including SNPs, gene amplifications and rearrangements. When
under pressure some beneficial mutations can be enriched in the surviving clones. For hypothesis
66
testing on gene expression level change using such clones, the number of clones (sample size) and
correct statistical tests are critical for data analysis.

Results from the depletion of CTCF/cohesin subunits studies
Although the correlation of CTCF and cohesion with loop structure is quite strong, the significance
of these interactions for gene regulation is not yet clear. Recently, different groups have performed
transient depletion of CTCF and cohesin subunits using the auxin-mediated acute protein
degradation system. The results of these studies differ in details but are consistent in general,
showing that after depletion of either CTCF or the cohesin core subunit RAD21: (1) loops and
TADs are abrogated on a genome scale, (2) the boundaries and epigenetic properties of
compartments remain mostly unaffected, and (3) strikingly, the changes in genome-scale gene
expression are quite subtle. This latter finding contradicts with studies in which deletions of
specific TAD boundaries or loop anchors cause significant fold changes in nearby genes
expression. (Nora et al., 2017; Rao et al., 2017; Schwarzer et al., 2017; Wutz et al., 2017; Eagen,
2018; Guo et al., 2018) Also, in a previous study in the Farnham lab, deletion of a CTCF binding
site caused a 100-fold increase in expression of a nearby gene. (Guo et al., 2018) However, these
individual observations do not indicate causal relationships between chromatin structure and gene
regulation. Biological and technical variations in Hi-C experiments should also be considered.
These contradictory observations with my work suggest that changes in chromosome structure and
changes on gene expression are not always correlated. We note that we observed a dramatic change
in 3D chromosome structure via the Hi-C interaction maps when comparing data obtained from
67
normal (RWPE1) and cancer prostate cells (22Rv1). In the case of AR, its ~300-fold up-regulation
in 22Rv1 cancer cells may be determined by other factors, which make it less sensitive to changes
in chromosome structures. Thus, the cancer-specific loops we identified may not be the most
influential factor on the regulation of nearby genes.
Taken together, these studies suggest that TAD and/or loop structures may not be sufficient to
drive gene expression changes, and its influences on gene expression vary upon genomic context.
The actual physical contacts between two locations connected by looping are likely to be weak
and dynamic. In some cases, gene expression can be influenced by TAD/loop structures, but they
may be determined by other mechanisms.

68
AR EN deletion AR ALL AR-FL AR-V7 AR EN deletion EDA2R OPHN1 YIPF6
Cas9-CTRL 1.000 1.000 1.000 Cas9-CTRL 1.000 1.000 1.000
Cas9-AR EN 0.802 0.890 0.717 Cas9-AR EN 0.982 1.269 1.545
SEM AR ALL ARFL AR-V7 SEM EDA2R OPHN1 YIPF6
Cas9-CTRL 0.021 0.020 0.040 Cas9-CTRL 0.035 0.029 0.040
Cas9-AR EN 0.014 0.019 0.023 Cas9-AR EN 0.023 0.037 0.044
AR ALL ARFL AR-V7 EDA2R OPHN1 YIPF6
T TEST 1.81E-06 0.001 3.62E-05 T TEST 0.672 4.34E-05 9.99E-08
*** ** *** . *** ***
CTCF site #1 deletion AR ALL AR-FL AR-V7 CTCF site #1 deletion EDA2R OPHN1 YIPF6
Cas9-CTRL 1.000 1.000 1.000 Cas9-CTRL 1.000 1.000 1.000
Cas9-CTCF site A 0.640 0.820 0.570 Cas9-CTCF site A 1.176 0.992 1.295
SEM AR ALL ARFL AR-V7 SEM EDA2R OPHN1 YIPF6
Cas9-CTRL 0.030 0.020 0.030 Cas9-CTRL 0.040 0.022 0.017
Cas9-CTCF site A 0.020 0.010 0.010 Cas9-CTCF site A 0.024 0.035 0.041
AR ALL ARFL AR-V7 EDA2R OPHN1 YIPF6
T TEST 3.10E-06 4.54E-05 1.90E-06 T TEST 0.002 0.845 4.41E-05
*** *** *** ** . ***
CTCF site #2 deletion AR ALL AR-FL AR-V7 CTCF site #2 deletion EDA2R OPHN1 YIPF6
Cas9-CTRL 1.000 1.000 1.000 Cas9-CTRL 1.000 1.000 1.000
Cas9-CTCF site B 0.671 0.766 0.560 Cas9-AR EN 0.975 0.971 0.906
SEM AR ALL ARFL AR-V7 SEM EDA2R OPHN1 YIPF6
Cas9-CTRL 0.042 0.045 0.083 Cas9-CTRL 0.024 0.021 0.026
Cas9-CTCF site B 0.031 0.027 0.021 Cas9-AR EN 0.025 0.048 0.030
AR ALL ARFL AR-V7 EDA2R OPHN1 YIPF6
T TEST 1.38E-05 6.84E-04 6.16E-04 T TEST 0.473 0.588 0.031
*** *** *** . . .
CTCF site #3 deletion AR ALL AR-FL AR-V7 CTCF site #3 deletion EDA2R OPHN1 YIPF6
Cas9-CTRL 1.000 1.000 1.000 Cas9-CTRL 1.000 1.000 1.000
Cas9-CTCF site D 0.900 0.870 0.740 Cas9-CTCF site D 1.106 0.924 0.938
SEM AR ALL ARFL AR-V7 SEM EDA2R OPHN1 YIPF6
Cas9-CTRL 0.057 0.041 0.060 Cas9-CTRL 0.025 0.048 0.038
Cas9-CTCF site D 0.022 0.039 0.033 Cas9-CTCF site D 0.025 0.022 0.037
AR ALL ARFL AR-V7 EDA2R OPHN1 YIPF6
T TEST 0.131 0.033 0.002 T TEST 0.009 0.180 0.265
. * ** ** . .
CTCF site #4 deletion AR ALL AR-FL AR-V7 CTCF site #4 deletion EDA2R OPHN1 YIPF6
Cas9-CTRL 1.000 1.000 1.000 Cas9-CTRL 1.000 1.000 1.000
Cas9-CTCF site E 0.819 0.951 0.853 Cas9-CTCF site E 0.869 1.244 1.196
SEM AR ALL ARFL AR-V7 SEM EDA2R OPHN1 YIPF6
Cas9-CTRL 0.049 0.085 0.054 Cas9-CTRL 0.027 0.042 0.046
Cas9-CTCF site E 0.033 0.090 0.041 Cas9-CTCF site E 0.045 0.048 0.066
AR ALL ARFL AR-V7 EDA2R OPHN1 YIPF6
T TEST 0.013 0.703 0.054 T TEST 0.028 0.002 0.029
* . * * ** *
AR EN single gRNA cutting AR ALL AR-V7 CTCF Site #1/ #2 single gRNA cutting AR ALL AR-V7 AR-FL
Cas9-CTRL 1.000 1.000 Cas9-CTRL 1.000 1.000 1.000
Cas9-AR EN 0.645 0.400 Cas9-A Single-cut 0.649 0.490 0.751
Cas9-Single-cut 0.312 0.214 Cas9-B Single-cut 0.689 0.578 1.020
SEM AR ALL AR-V7 SEM AR ALL AR-V7 AR FL
Cas9-CTRL 0.034 0.056 Cas9-CTRL 0.070 0.068 0.075
Cas9-AR EN 0.021 0.023 Cas9-A Single-cut 0.010 0.001 0.019
Cas9-Single-cut 0.006 0.009 Cas9-B Single-cut 0.037 0.001 0.119
T TEST AR ALL AR-V7 T TEST AR ALL AR-V7 AR FL
Cas9-AR EN 6.58E-07 1.06E-06 Cas9-A Single-cut 0.004 5.68E-04 0.751
*** *** ** *** .
Cas9-Single-cut 2.18E-08 4.10E-07 Cas9-B Single-cut 0.005 5.72E-04 0.888
*** *** ** *** .
Middle loop deletion AR ALL AR-V7
CTRL 1.000 1.000
AREN SINGLE-CUT 0.563 0.400
MID DEL 0.933 0.928
MID SINGLE-CUT 0.919 1.002
SEM AR ALL AR-V7
CTRL 0.042 0.019
AREN SINGLE-CUT 0.014 0.026
MID DEL 0.036 0.039
MID SINGLE-CUT 0.015 0.037
T TEST AR ALL AR-V7
AREN SINGLE-CUT 5.93E-05 1.35E-08
*** ***
MID DEL 0.258 0.138
. .
MID SINGLE-CUT 0.118 0.954
. .

69
Double deletion of CTCF site #1 & #2
Clone used in the
transient trasfection #B14
Clone used in the
transient trasfection #B1
Clone used in the
transient trasfection #A30
EXPRESSION AR ALL AR-V7 EXPRESSION AR ALL AR-V7 EXPRESSION AR ALL AR-V7
WT CTRL 1.000 1.000 WT CTRL 1.000 1.000 WT CTRL 1.000 1.000
WT ADEL 0.710 0.550 WT ADEL 0.335 0.193 WT BDEL 0.660 0.560
BDEL CTRL 1.050 2.690 BDEL CTRL 0.544 0.276 ADEL CTRL 0.990 0.550
BDEL ADEL 0.710 0.680 BDEL ADEL 0.306 0.098 ADEL BDEL 0.800 0.410
SD SD AR ALL AR-V7 SD
WT CTRL 0.030 0.060 WT CTRL 0.036 0.088 WT CTRL 0.100 0.120
WT ADEL 0.040 0.010 WT ADEL 0.018 0.020 WT BDEL 0.050 0.040
BDEL CTRL 0.030 0.040 BDEL CTRL 0.055 0.034 ADEL CTRL 0.060 0.070
BDEL ADEL 0.040 0.030 BDEL ADEL 0.019 0.010 ADEL BDEL 0.060 0.020
T TEST T TEST AR ALL AR-V7 T TEST
WT ADEL 3.26E-04 7.09E-04 WT ADEL 4.59E-07 1.77E-04 WT BDEL 0.017 0.014
*** *** *** *** * *
BDEL CTRL 0.251 4.87E-09 BDEL CTRL 8.77E-05 1.83E-04 ADEL CTRL 0.922 0.014
. *** *** *** . *
BDEL ADEL 7.05E-04 0.002 BDEL ADEL 3.30E-07 1.37E-04 ADEL BDEL 0.127 0.005
*** ** *** *** . **

Table 3.1. RT-qPCR results on AR, EDA2R, OPHN1 and YIPH6 after deletions in cell pools and clones.

clones carrying deletion vs. control clones
T TEST SITE 1 SITE 2 SITE3 SITE 1&2 AR EN
AR ALL 0.005 0.231 0.179 0.015 0.392
AR-V7 0.128 0.254 0.395 0.201 0.282
EDA2R 0.028 0.089 / / /
OPHN1 0.516 0.976 / / /
YIPF6 0.002 0.169 / / /
SITE 1 T TEST
MEAN EXPRESSION
RATIO
CTRL/DEL
CTRL DEL
AR ALL 0.005 0.80 1.42 1.78
EDA2R 0.028 0.99 0.66 0.67
YIPF6 0.002 1.00 0.61 0.61
SITE 1&2 T TEST
MEAN EXPRESSION
RATIO
CTRL/DEL
CTRL DEL
AR ALL 0.015 1.22 0.63 0.52
70

Table 3.2. RT-qPCR results on AR after deletions in cell pools (from other replicates of CRISPR/Cas9-
mediated deletion experiments).

Table 4.1. Meta data of clones and CRISPR/Cas9-mediated deletion experiments performed in cell
pools.
Deletion Plate Sample
AR all AR-FL AR-v7
Expression SEM Expression SEM Expression SEM
AR EN deletion
plate A B
Deletion 0.844 0.085 / / / /
Control 1.000 0.204 / / / /
plate D E
Deletion 0.605 0.026 / / / /
Control 1.000 0.133 / / / /
plate O
Deletion 0 .7 9 9 0 .0 5 5 0 .9 8 0 0 .0 6 5 0 .7 0 2 0 .0 8 0
Control 1 .0 0 0 0 .0 4 7 1 .0 0 0 0 .0 4 2 1 .0 0 0 0 .0 5 2
plate AP#1
Deletion / / 1.003 0.085 0.749 0.025
Control / / 1.000 0.116 1.000 0.091
plate AP#2
Deletion 0.861 0.022 0.976 0.034 0.707 0.018
Control 1.000 0.071 1.000 0.090 1.000 0.086
plate AP#4
Deletion 0.785 0.049 0.813 0.039 0.753 0.040
Control 1.000 0.028 1.000 0.036 1.000 0.044
CTCF site #1 deletion
plate AP#3
Deletion 0.645 0.015 0.824 0.017 0.570 0.014
Control 1.000 0.040 1.000 0.041 1.000 0.037
plate AP#4
Deletion 0.824 0.054 0.813 0.052 0.767 0.051
Control 1.000 0.049 1.000 0.046 1.000 0.061
CTCF site #2 deletion plate W
Deletion 0 .7 9 0 0 .0 3 7 0 .7 5 9 0 .0 3 7 0 .5 5 8 0 .0 2 4
Control 1 .0 0 0 0 .0 5 1 1 .0 0 0 0 .0 7 1 1 .0 0 0 0 .0 7 4
CTCF site #3 deletion plate X
Deletion 1 .0 1 8 0 .0 6 9 0 .9 6 0 0 .0 7 2 0 .8 6 3 0 .0 5 8
Control 1 .0 0 0 0 .0 4 7 1 .0 0 0 0 .0 4 8 1 .0 0 0 0 .0 7 4
Cell pools
gRNA qPCR samples Gene Tested
Test for AR
AR EN deletion plate P
AR, AR-V7, AR-FL, GAPDH
CTCF site #1 deletion plate AP#1
CTCF site #2 deletion plate W2
CTCF site #3 deletion plate X2
CTCF site #4 deletion plate Z
Test for other gene
AR EN deletion plate O
EDA2R, OPHN1, YIPF6, RPLP0, GAPDH
CTCF site #1 deletion plate AP#1
CTCF site #2 deletion plate W2
CTCF site #3 deletion plate X2
CTCF site #4 deletion plate Z
Single gRNA cutting
R EN_s plate AF AR, AR-V7, RPLP0, GAPDH
site_A_s plate Z AR, AR-V7, AR-FL, RPLP0, GAPDH
site_B_R plate Z AR, AR-V7, AR-FL, RPLP0, GAPDH
Middle loop deletion Middle loop deletion plate AH AR, AR-V7, RPLP0, GAPDH
CTCF site A&B deletion CTCF site #1 & #2 deletion plate AG AR, AR-V7, EDA2R, OPHN1, YIPF6, RPLP0, GAPDH
Clones
clones screened deletion clones tested CTRL clones Gene Tested
AR EN deletion 58 11 5
AR, AR-V7,
RPLP0, GAPDH
CTCF site #1 deletion 86 10
4
CTCF site #2 deletion 83 9
CTCF site #3 deletion 66 7
4
CTCF site #1 & #2 deletion 81 6
71
REFERENCES

RUIZ-VELASCO, M. ZAUGG, J. Structure meets function: How chromatin organisation
conveys functionality. Current Opinion in Systems Biology, 2017
CONSORTIUM, E. P. An integrated encyclopedia of DNA elements in the human genome.
Nature, v. 489, n. 7414, p. 57-74, Sep 2012. ISSN 1476-4687. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/22955616 >.

CULIG, Z.; SANTER, F. R. Androgen receptor signaling in prostate cancer. Cancer Metastasis
Rev, v. 33, n. 2-3, p. 413-27, Sep 2014. ISSN 1573-7233. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/24384911 >.

DE WIT, E. et al. CTCF Binding Polarity Determines Chromatin Looping. Mol Cell, v. 60, n. 4,
p. 676-84, Nov 2015. ISSN 1097-4164. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/26527277 >.

DEKKER, J. et al. Capturing chromosome conformation. Science, v. 295, n. 5558, p. 1306-11,
Feb 2002. ISSN 1095-9203. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/11847345
>.

DIXON, J. R. et al. Chromatin architecture reorganization during stem cell differentiation.
Nature, v. 518, n. 7539, p. 331-6, Feb 2015. ISSN 1476-4687. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/25693564 >.

______. Topological domains in mammalian genomes identified by analysis of chromatin
interactions. Nature, v. 485, n. 7398, p. 376-80, Apr 2012. ISSN 1476-4687. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/22495300 >.

DOWEN, J. M. et al. Control of cell identity genes occurs in insulated neighborhoods in
mammalian chromosomes. Cell, v. 159, n. 2, p. 374-387, Oct 2014. ISSN 1097-4172. Disponível
em: < https://www.ncbi.nlm.nih.gov/pubmed/25303531 >.

EAGEN, K. P. Principles of Chromosome Architecture Revealed by Hi-C. Trends Biochem Sci,
v. 43, n. 6, p. 469-478, 06 2018. ISSN 0968-0004. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/29685368 >.

FUDENBERG, G. et al. Formation of Chromosomal Domains by Loop Extrusion. Cell Rep, v.
15, n. 9, p. 2038-49, 05 2016. ISSN 2211-1247. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/27210764 >.

GUO, X.; DEAN, A. CRISPR/Cas9 offers a new tool for studying the role of chromatin
architecture in disease pathogenesis. Genome Biol, v. 19, n. 1, p. 185, 11 2018. ISSN 1474-
760X. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/30400943 >.

72
GUO, Y. et al. CRISPR-mediated deletion of prostate cancer risk-associated CTCF loop anchors
identifies repressive chromatin loops. Genome Biol, v. 19, n. 1, p. 160, 10 2018. ISSN 1474-
760X. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/30296942 >.

______. CRISPR Inversion of CTCF Sites Alters Genome Topology and Enhancer/Promoter
Function. Cell, v. 162, n. 4, p. 900-10, Aug 2015. ISSN 1097-4172. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/26276636 >.

HAEUSSLER, M. et al. Evaluation of off-target and on-target scoring algorithms and
integration into the guide RNA selection tool CRISPOR. Genome Biol, v. 17, n. 1, p. 148, 07
2016. ISSN 1474-760X. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/27380939 >.

HAZELETT, D. J. et al. Comprehensive functional annotation of 77 prostate cancer risk loci.
PLoS Genet, v. 10, n. 1, p. e1004102, Jan 2014. ISSN 1553-7404. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/24497837 >.

HNISZ, D.; DAY, D. S.; YOUNG, R. A. Insulated Neighborhoods: Structural and Functional
Units of Mammalian Gene Control. Cell, v. 167, n. 5, p. 1188-1200, 11 2016. ISSN 1097-4172.
Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/27863240 >.

HNISZ, D. et al. Activation of proto-oncogenes by disruption of chromosome neighborhoods.
Science, v. 351, n. 6280, p. 1454-1458, Mar 2016. ISSN 1095-9203. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/26940867 >.

HO, E. Y. et al. Shaping the nebulous enhancer in the era of high-throughput assays and genome
editing. Brief Bioinform, Mar 2019. ISSN 1477-4054. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/30895290 >.

HOLTZMAN, L.; GERSBACH, C. A. Editing the Epigenome: Reshaping the Genomic
Landscape. Annu Rev Genomics Hum Genet, v. 19, p. 43-71, 08 2018. ISSN 1545-293X.
Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/29852072 >.

JI, X. et al. 3D Chromosome Regulatory Landscape of Human Pluripotent Cells. Cell Stem
Cell, v. 18, n. 2, p. 262-75, Feb 2016. ISSN 1875-9777. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/26686465 >.

KATAINEN, R. et al. CTCF/cohesin-binding sites are frequently mutated in cancer. Nat Genet,
v. 47, n. 7, p. 818-21, Jul 2015. ISSN 1546-1718. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/26053496 >.

KNOUF, E. C. et al. Multiple integrated copies and high-level production of the human
retrovirus XMRV (xenotropic murine leukemia virus-related virus) from 22Rv1 prostate
carcinoma cells. J Virol, v. 83, n. 14, p. 7353-6, Jul 2009. ISSN 1098-5514. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/19403664 >.

73
LIEBERMAN-AIDEN, E. et al. Comprehensive mapping of long-range interactions reveals
folding principles of the human genome. Science, v. 326, n. 5950, p. 289-93, Oct 2009. ISSN
1095-9203. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/19815776 >.

MERKENSCHLAGER, M.; NORA, E. P. CTCF and Cohesin in Genome Folding and
Transcriptional Gene Regulation. Annu Rev Genomics Hum Genet, v. 17, p. 17-43, 08 2016.
ISSN 1545-293X. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/27089971 >.

NORA, E. P. et al. Targeted Degradation of CTCF Decouples Local Insulation of Chromosome
Domains from Genomic Compartmentalization. Cell, v. 169, n. 5, p. 930-944.e22, May 2017.
ISSN 1097-4172. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/28525758 >.

______. Spatial partitioning of the regulatory landscape of the X-inactivation centre. Nature, v.
485, n. 7398, p. 381-5, Apr 2012. ISSN 1476-4687. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/22495304 >.

PHILLIPS-CREMINS, J. E. et al. Architectural protein subclasses shape 3D organization of
genomes during lineage commitment. Cell, v. 153, n. 6, p. 1281-95, Jun 2013. ISSN 1097-4172.
Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/23706625 >.

POPE, B. D. et al. Topologically associating domains are stable units of replication-timing
regulation. Nature, v. 515, n. 7527, p. 402-5, Nov 2014. ISSN 1476-4687. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/25409831 >.

RAN, F. A. et al. Genome engineering using the CRISPR-Cas9 system. Nat Protoc, v. 8, n. 11,
p. 2281-2308, Nov 2013. ISSN 1750-2799. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/24157548 >.

RAO, S. S. et al. A 3D map of the human genome at kilobase resolution reveals principles of
chromatin looping. Cell, v. 159, n. 7, p. 1665-80, Dec 2014. ISSN 1097-4172. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/25497547 >.

RAO, S. S. P. et al. Cohesin Loss Eliminates All Loop Domains. Cell, v. 171, n. 2, p. 305-
320.e24, Oct 2017. ISSN 1097-4172. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/28985562 >.

RHIE, S. K. et al. Using 3D epigenomic maps of primary olfactory neuronal cells from living
individuals to understand gene regulation. Sci Adv, v. 4, n. 12, p. eaav8550, Dec 2018. ISSN
2375-2548. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/30555922 >.

RUIZ-VELASCO, M. et al. CTCF-Mediated Chromatin Loops between Promoter and Gene
Body Regulate Alternative Splicing across Individuals. Cell Syst, v. 5, n. 6, p. 628-637.e6, Dec
2017. ISSN 2405-4712. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/29199022 >.

SANBORN, A. L. et al. Chromatin extrusion explains key features of loop and domain
formation in wild-type and engineered genomes. Proc Natl Acad Sci U S A, v. 112, n. 47, p.
74
E6456-65, Nov 2015. ISSN 1091-6490. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/26499245 >.

SCHWARZER, W. et al. Two independent modes of chromatin organization revealed by
cohesin removal. Nature, v. 551, n. 7678, p. 51-56, 11 2017. ISSN 1476-4687. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/29094699 >.

SEXTON, T. et al. Three-dimensional folding and functional organization principles of the
Drosophila genome. Cell, v. 148, n. 3, p. 458-72, Feb 2012. ISSN 1097-4172. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/22265598 >.

SIEGEL, R. L.; MILLER, K. D.; JEMAL, A. Cancer statistics, 2018. CA Cancer J Clin, v. 68,
n. 1, p. 7-30, 01 2018. ISSN 1542-4863. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/29313949 >.

SRAMKOSKI, R. M. et al. A new human prostate carcinoma cell line, 22Rv1. In Vitro Cell
Dev Biol Anim, v. 35, n. 7, p. 403-9, 1999 Jul-Aug 1999. ISSN 1071-2690. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/10462204 >.

TAK, Y. G.; FARNHAM, P. J. Making sense of GWAS: using epigenomics and genome
engineering to understand the functional relevance of SNPs in non-coding regions of the human
genome. Epigenetics Chromatin, v. 8, p. 57, 2015. ISSN 1756-8935. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/26719772 >.

TAN, M. H. et al. Androgen receptor: structure, role in prostate cancer and drug discovery. Acta
Pharmacol Sin, v. 36, n. 1, p. 3-23, Jan 2015. ISSN 1745-7254. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/24909511 >.

THURMAN, R. E. et al. The accessible chromatin landscape of the human genome. Nature, v.
489, n. 7414, p. 75-82, Sep 2012. ISSN 1476-4687. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/22955617 >.

WANG, S. et al. Spatial organization of chromatin domains and compartments in single
chromosomes. Science, v. 353, n. 6299, p. 598-602, Aug 2016. ISSN 1095-9203. Disponível em:
< https://www.ncbi.nlm.nih.gov/pubmed/27445307 >.

WITTKOPP, P. J.; KALAY, G. Cis-regulatory elements: molecular mechanisms and
evolutionary processes underlying divergence. Nat Rev Genet, v. 13, n. 1, p. 59-69, Dec 2011.
ISSN 1471-0064. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/22143240 >.

WUTZ, G. et al. Topologically associating domains and chromatin loops depend on cohesin and
are regulated by CTCF, WAPL, and PDS5 proteins. EMBO J, v. 36, n. 24, p. 3573-3599, 12
2017. ISSN 1460-2075. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/29217591 >.

YAO, L.; BERMAN, B. P.; FARNHAM, P. J. Demystifying the secret mission of enhancers:
linking distal regulatory elements to target genes. Crit Rev Biochem Mol Biol, v. 50, n. 6, p.
75
550-73, 2015. ISSN 1549-7798. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/26446758 >.

YAO, L. et al. Functional annotation of colon cancer risk SNPs. Nat Commun, v. 5, p. 5114,
Sep 2014. ISSN 2041-1723. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/25268989
>.

YU, M.; REN, B. The Three-Dimensional Organization of Mammalian Genomes. Annu Rev
Cell Dev Biol, v. 33, p. 265-289, 10 2017. ISSN 1530-8995. Disponível em: <
https://www.ncbi.nlm.nih.gov/pubmed/28783961 >.

GUO, Y. et al. CRISPR-mediated deletion of prostate cancer risk-associated CTCF loop anchors
identifies repressive chromatin loops. Genome Biol, v. 19, n. 1, p. 160, 10 2018. ISSN 1474-
760X. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/30296942 >.

HAEUSSLER, M. et al. Evaluation of off-target and on-target scoring algorithms and integration
into the guide RNA selection tool CRISPOR. Genome Biol, v. 17, n. 1, p. 148, 07 2016. ISSN 1474-
760X. Disponível em: < https://www.ncbi.nlm.nih.gov/pubmed/27380939 >.

Abstract (if available)

Abstract Previously, members of the Farnham lab have generated Hi-C chromatin interaction maps in both normal (RWPE1) and cancer (22Rv1) prostate cell lines. By combining the Hi-C data with epigenetic features and transcriptome profiling, chromatin loops surrounding the Androgen Receptor (AR) gene were identified, which are present in cancer cells (22Rv1) but absent in normal cells (RWPE1), suggesting that these loops may contribute to the large upregulation of AR expression in the 22 Rv1 cancer cells. My studies have focused on 5 chromatin loops near the AR gene. Ones of these loops connects the AR promoter to a putative enhancer and the 3 other loops are anchored by convergent CTCF sites. I performed CRISPR/Cas9-mediated deletion of the putative AR enhancer and the four other CTCF loop anchor regions surrounding the AR gene in 22Rv1 cell pools and generated single-cell-derived clones harboring these deletions. After performing these deletions in cell pools, I only observed modest changes on AR expression. Similarly, little change on AR expression can be found in clones carrying these deletions. My studies will help in understanding the relationship between chromatin looping and gene regulation.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

Understanding prostate cancer genetic susceptibility and chromatin regulation

PDF

Using epigenetic toggle switches to repress tumor-promoting gene expression

PDF

Functional analysis of a prostate cancer risk enhancer at 7p15.2

PDF

Characterization of a new chromobox protein 8 (CBX8) antagonist in a model of human colon cancer

PDF

Functional characterization of colon cancer risk-associated enhancers: connecting risk loci to risk genes

PDF

Functional characterization of a prostate cancer risk region

PDF

Characterizing ZFX-mediated gene regulation to reveal possible candidates for clinical intervention

PDF

Homologous cell systems for the study of progression of androgen-dependent prostate cancer to castration-resistant prostate cancer

PDF

Do ZFX and ZNF711 regulate the same genes in HEK293T cells?

PDF

Modeling lung adenocarcinoma progression in vitro using immortalized human alveolar epithelial cells

PDF

Study of the role of bone morphogenetic proteins in prostate cancer progression

PDF

Identification of novel androgen receptor target genes in prostate cancer

PDF

Functional analyses of androgen receptor structure pertaining to prostate cancer

PDF

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

PDF

RNA methylation in cancer plasticity and drug resistance

PDF

Application of tracing enhancer networks using epigenetic traits (TENET) to identify epigenetic deregulation in cancer

PDF

The relationship between DNA methylation and transcription factor binding in colon cancer cells

PDF

Functional role of chromatin remodeler proteins in cancer biology

PDF

Dietary and supplementary folate intake and prostate cancer risk

PDF

G9a, GLP and HP1γ facilitate glucocorticoid-induced apoptosis in B-ALL cells

Asset Metadata

Creator Shi, Jiani (author)

Core Title Using CRISPR-mediated deletion to study prostate cancer regulatory elements located at loop anchors identified by Hi-C

School Keck School of Medicine

Degree Master of Science

Degree Program Biochemistry and Molecular Medicine

Publication Date 07/26/2019

Defense Date 06/07/2019

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

Tag CRISPR-mediated deletion,OAI-PMH Harvest,prostate cancer,regulatory elements

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Farnham, Peggy (committee member), Hacia, Joseph (committee member), Stallcup, Michael (committee member)

Creator Email jennynshi@gmail.com,Jianishi@usc.edu

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c89-193420

Unique identifier UC11663445

Identifier etd-ShiJiani-7636.pdf (filename),usctheses-c89-193420 (legacy record id)

Legacy Identifier etd-ShiJiani-7636.pdf

Dmrecord 193420

Document Type Thesis

Format application/pdf (imt)

Rights Shi, Jiani

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