University of Southern California Dissertations and Theses

Identification of gene expression regulated by 1,25(OH)₂D3 in human endometriosis cell lines with next-generation sequencing

(USC Thesis Other)

Identification of gene expression regulated by 1,25(OH)₂D3 in human endometriosis cell lines with next-generation sequencing

PDF

Download

Open document

Flip pages

Download a page range

Download transcript

Copy asset link

Request this asset

Transcript (if available)

Content 1

Identification of Gene Expression Regulated by 1,25(OH)
2
D3 in Human
Endometriosis Cell Lines with Next-generation Sequencing

Name: Liang Wu
Student Number: 4003946292
Supervisor: Dr. Sue Ingles
University of Southern California

Degree Conferral Month: December 2014
Degree Being Conferred: Master in Science
Field of Degree: Molecular Epidemiology

2

Table of Contents
Abstract…………………………………………………………………………………...…........4
Introduction…………………………………………………………………………………........5
Material and Methods…………………………………………………………………...............8
Cell Culture and DNA Extraction………………………………………………………………....8
Library Construction and Next-generation Sequencing………………………………………......9
Bioinformatic Analysis………………………………………………………………………......10
Results…………………………………………………………………………………………...11
Overall Results from the First Round of Experiments……………………………………….......11
Gene Ontology and Pathway Analysis in EEC Sample from the First Round of
Experiments………………………...............................................................................................15
Gene Ontology and Pathway Analysis in ESC Sample from the First Round of
Experiments………………………...............................................................................................22
Overall Results from the Second Round of Experiments……………………………………......25
Gene Ontolgy and Pathway Analysis in ESC Sample from the Second Round of
Experiments………………………...............................................................................................26
Discussion……………………………………………………………………………………….31
DEG Ontology of EEC Samples from the First Round of Experiments……………………........31
3

DEG Ontology of ESC Samples from the First Round of Experiments……………………........33
DEG Ontology of ESC Samples from the Second Round of Experiments………………….......34
Conclusion…………………………………………………………………………………….....37
Acknowledgements......................................................................................................................38
References………………………………………………………………………………….........38

4

Abstract
Endometriosis is a common female reproductive disease. This disease is caused by invasion of
endometrial cells into other organs. Pain at the site of ectopic endometrium and infertility are two
common symptoms of endometriosis, and around 30% of infertility cases are associated with
endometriosis. Vitamin D3 is a type of secosteroid, and there are plenty of studies reported that
vitamin D3 treatment helps to improve fertility in both animals and human patients, as well as
enhance the success rate of in vitro fertilization. We are interested in how vitamin D3 impacts
gene expression in endometriosis. 1,25(OH)
2
D3, the active metabolite of vitamin D3, was used
to treat endometriotic epithelial and stromal cell lines. Differential gene expression for the
treated group versus the control group was identified with RNA-Seq. We found 149 genes whose
expression levels were altered by at least two fold in epithelial cells, but only 25 genes in stromal
cells. Among the 149 differentially expressed genes of the epithelial cells, we managed to locate
a group of genes that are important for cell motility and invasion, with gene PLAUR and CTTN
being most likely to involve in cell movement directly. Nevertheless, more evidence is needed to
reveal the roles of these genes in endometriosis. Compared with epithelial cells, endometriotic
stromal cells seem to be much less sensitive to 1,25(OH)
2
D3, according to our experiment.
However, this could be a consequence of the pathology of endometriosis, we decided to follow-
up by include more samples and different treatments in the future.
Keywords: Endometriosis, vitamin D, 1,25(OH)
2
D3, next-generation sequencing, DEG

5

Introduction
Endometriosis is one of the most common female health disorders. It is characterised by
growth of endometrial tissue outside of the uterus. The major symptoms are pain and infertility.
Approximately 10-15% of all women of reproductive age are affected by endometriosis. It is also
related to about 30% of infertility cases (1).

The cause of endometriosis is not completely clear. Among the many hypotheses, there is
most support for the role of retrograde menstruation (2). In this hypothesis, the endometrial cells
are carried out of the uterus during menstruation, and attach themselves to the peritoneal surface
so that further invasion into tissues is possible (3). Other hypotheses include mullerianosis,
coelomic metaplasia, and vasculogenesis. Mullerianosis explains that cells having the potential
to differentiate into endometrium become dislocated from the uterus during embryonic
development, and later in life function like stem cells to allow endometriosis to occur (4, 5). The
coelomic metaplasia hypothesis suggests that peritoneal cells, which have the same origin as
endometrial cells, can go through metaplasia and transform into endometrial cells (6). According
to the vasculogenesis hypothesis, endothelial progenitor cells are turned into the microvascular
endothelium of endometriotic tissues through vasculogenesis (7). These hypotheses may not be
mutually exclusive. In fact, pathological theories point out those multiple factors may contribute
synergistically to the disease (3).

Genetic predisposition has been implicated as a factor in endometriosis (8). Germline
(inherited) variation in regions10q26 on chromosome 10 (9) and 7q15.2 on chromosome 7 (10)
have been associated with endometriosis. In the first study, genome-wide linkage analysis was
6

performed on 1,176 affected sister pair families, using DNA extracted from peripheral blood
lymphocytes (9). The second study is a genome-wide association study (GWAS) with 3194 cases
and 7060 controls. DNA sample was obtained from blood as well (10). However, these findings
did not show great increased risk for subjects carrying genetic alterations in these chromosomal
regions. Considering the nature of endometriosis, which involves tissue invasion and metastasis
similar to the progression of cancer, endometriosis is likely to be a complex disease requiring
multiple genetic events to occur.

Steroid hormones and their nuclear receptors are involved in development of endometriosis,
too, especially estrogen and estrogen receptor (ER) (11). During endometriosis, higher levels of
estrogen metabolizing enzymes are synthesized, which in turn lead to accumulation of estradiol
(E2). E2 then stimulates abnormal proliferation of endometrial cells, and results in ectopic
endometrium (12, 13).

Some researchers proposed that immunology should not be overlooked in endometriosis
occurring in women who have retrograde menstruation (13, 14). Endometriotic lesions are often
associated with inflammation, which was induced by increased production of cytokines, growth
factors and other pro-inflammatory molecules. Elevated concentration of cytokine IL-1 and IL-
16, epidermal growth factor (EGF), and vascular endothelial growth factor (VEGF) are observed
in peritoneal fluid in endometriosis patients, and usually accompanied with presence of less
active macrophages (14, 15).

7

Vitamin D3 and its derivatives seem to possess properties that can counter endometriosis. The
active form of vitamin D3, 1, 25-dihydroxyvitamin D3 (1,25(OH)
2
D3), is widely known for its
abilities to regulate cell proliferation and the immune system. 1,25(OH)
2
D3 exerts its effect by
binding to the intracellular vitamin D receptor (VDR), which then forms a heterodimer with
retinal X receptor (RXR) in the nucleus. This complex interacts with DNA sequence, usually
enchancer regions, to regulate target gene expression (16). 1,25(OH)
2
D3 has been demonstrated
to lower cytokine production (17). Target genes that can be regulated by 1,25(OH)
2
D3 include
p21, TGFβ-1, TGFβ receptor type II, c-FOS, DNA protein kinase, IL-1β, IL-2, IL-8, IL-12, etc
(18).

This secosteroid hormone has drawn much attention in the female reproductive health field,
since population studies have pointed out that there is an association between 1,25(OH)
2
D3
levels and reproductive outcomes. For example, it is found that vitamin D enhances reproductive
rates from IVF (in vitro fertilization), specifically by its actions on the endometrium (19, 20).
Higher levels of 1,25(OH)
2
D3 intake is effective in reducing pre-eclampsia risk (21).
Furthermore, in a recent prospective cohort study, predicted plasma 25(OH)D3 levels based on
diet were found to be inversely associated with endometriosis risk (22). 25(OH)D3 is the
prohormone of 1,25(OH)
2
D3, which is converted to 1,25(OH)
2
D3 by 1-alpha-hydroxylase (23).

Reproductive effects of 1,25(OH)
2
D3 have been evaluated using murine models. Female rats
had impaired fertility after being fed a vitamin D3 deficient diet (24), while VDR knock out
female mice become sterile (25). HOXA10, a gene critical to implantation and fertility, was
8

found to be regulated by vitamin D through direct interaction between VDR and regulatory
sequence of HOXA10 gene (26).

Based on the beneficial reproductive effects 1,25(OH)
2
D3 discussed above, we hypothesized
that 1,25(OH)
2
D3 could alter gene expression in endometriotic tissues. And we will focus our
attention on genes most representative in the pathology of endometriosis, such as those involve
in cell movement and inflammation. To test our hypothesis, we treated two types of
endometriotic cells: epithelial cells and stromal cells, separately, with 1,25(OH)
2
D3, and used
next-generation RNA sequencing to profile differentially expressed genes in the treated cells
versus controls. This is the first study known to investigate the effect of vitamin D on gene
expression in endometriosis cells directly, which gives more insight in the genetics of
endometriosis.

Materials and Methods

Cell Culture and RNA Extraction
Endometriosis epithelial cell line EEC 12Z, and stromal cell line ESC 22B used in this study
were courtesy from Doerthe Brueggmann, MD, assistant professor of the University of Southern
California. Starting passage number was 31 for the epithelial cell line (EEC p31), and 23 for the
stromal cell line (ESC p23).

Culture medium was comprised of Dulbeco’s Modified Eagle’s Medium (DMEM)
(BioWhittaker
®
, Lonza, Walkersville, MD, USA), 10% fetal bovine serum (FBS) (Gibco
®
,
9

Invitrogen Life Technology, Calsbad, CA, USA), and 5% L-glutamine (Cellgro
®
, Mediatech
Corning, Manassas, VA, USA) with antibiotics Penicillin-Streptomycin (Cellgro
®
, Mediatech
Corning, Manassas, VA, USA). Epithelial cells were cultured for 22 days, and stromal cells were
cultured for 25 days, in a humidified incubator with 95% of air and 5% of CO
2
at 37
o
C before
1,25(OH)
2
D3 treatment. Each cell line was passaged twice, divided into 6 tissue culture dishes,
and allowed to proliferate to ~80% confluency 24 hours prior to treatment. At this point, cells
were placed into the same medium described, except that no FBS was added, to eliminate
potential influence of 1,25(OH)
2
D3 present in FBS. Upon treatment, the 6 dishes were assigned
into 3 treatment samples and 3 controls. In the treatment group, 0.1 μM 1,25(OH)
2
D3 dissolved
in 95% ethanol was added into each dish to mimic the normal physiological level of active
vitamin D3 (27). Cells were moved back into the incubator to culture for another 4 hours before
RNA extraction.
A second round of experiments was conducted only using doublet of ESC samples (2 treated
ones and 2 controls), since in the first round, a fair amount of ESC cells were lifted after adding
1,25(OH)
2
D3 dissolved in ethanol, but EEC cells adapted well. During the second round, DMSO
was used as the solvent for 1,25(OH)
2
D3, and all other conditions were kept the same.
Commercially available RNeasy
®
Mini Kit (Qiagen, Valencia, CA, USA) was used for RNA
extraction. Extracted RNA were stored in RNase-free water at -80
o
C until they were sent our for
sequencing.

Library Construction and Next-generation Sequencing
cDNA library construction and next-generation sequencing were carried out at the USC
Epigenome Center. RNA libraries were generated using Illumina TruSeq
®
RNA-Sample
10

Preparation Kits (Illumina, La Jolla, CA, USA) according to Illumina’s protocols, with starting
material of 1 µ g of total RNA. RNA quality was determined using 2100 Bioanalyzer (Agilent,
Santa Clara, CA, USA). Libraries were quantified using KAPA Library Quantification Kits
(Kapa, Woburn, MA, USA). Sequencing was performed at the Data Production Facility of USC
Epigenome Center. Sequencing platform was Illumina Hi-seq
®
2000 (Illumina, La Jolla, CA,
USA), sequencing strategy was 50bp paired-end (50 PE). Libraries were applied to an Illumina
version 3 flow cell at a concentration of 16pM. The flow cell runs on Hi-seq
®
2000 for 100x100
cycles with a 7bp index. Sequencing depth was kept at no less than 50 million PF reads for each
sample.

Bioinformatic Analysis
The USC Epigenome Center was also responsible for data generation. Image analysis and base
calling were carried out with RTA v1.7.48. Final data formatting, demultiplexing and FASTQ
generation were done by CASAVA v1.7.0. Both RTA and CASAVA are softwares from
Illumina. Genome alignment, quality control and differential gene expression was analyzed with
Partek
®
Flow and Partek
®
Genomic Suite (Partek, St. Louis, MO, USA). Trimming of raw reads
(both ends) was based on minimum quality score of 20 and minimum read length of 25.
Genomes were aligned by incorporated TopHat2. Gene-specific analysis (GSA), a statistical
modeling approach used to test differential gene expression in Partek
®
Flow was used to choose
a specific model for each transcript. Models to choose from included: Log-normal, Poisson,
Negative Binomial, Zero-Inflated Poisson and Zero-Inflated Binomial. Further gene ontology
(GO) and pathway analysis was done by Ingenuity
®
Pathway Analysis (IPA) (Ingenuity,
Redwood City, CA, USA) with genome reference information from the RefSeq database.
11

Results

Overall result from the first round of experiments
The quality of extracted RNA from the 12 samples from the first round of experiment was
satisfactory. With the exception of one sample (#5), the virtual gel electrophoresis generated by
2100 Bioanalyzer showed no evidence of RNA degradation (Figure 1). Sample #5 is one of the
treated EEC samples, and had a 28S/18S ratio of 0.71, and a concentration of 125.33 ng/μL,
which is about only half the concentration of the rest of the samples.

Figure 1. Virtual gel report of RNA quality generated by Agilent 2100 Bioanalyzer. In this
graph, the left most is the ladder for reference, sample 1-3 are controls of EEC cell line, 4-6 are
treated EEC cell line, 7-9 are controls of ESC cells while 10-12 are treated ESC cells. From this
graph we can see that the 5th lane shows some degree of degradation, which can affect the
12

sequencing outcome. For the rest of the lanes, two strong bands can be readily observed. The
upper band at about 3800bp represents the 28S ribosomal RNA subunit, and the lower band at
about 1750bp is the 18S subunit. Ratio of 28S/18S at around 2 means excellence in RNA quality,
and this can be visualized from the size of the bands on this graph.

We used post-alignment QA/QC data given by Partek
®
Flow, to have a quantitative view of
base call quality and mapping quality. The Phred Quality Score is used here. Since the 5th

sample (Treated 2) has average quality and average mapping quality scores only slightly lower
compared with other samples, we decided to include it in our following analyses.

Table 1. Post-alignment QC chart for EEC samples
Sample
Name
Total
Reads
Aligned Unique Coverage
Avg.
Coverage
Depth
Avg.
Length
Avg.
Quality
Avg. Mapping
Quality
%GC
Control 1
28,445,
643
100.00
%
4.93% 5.98%
16.75
(SD
135.03)
49.63 37.42
42.41 (SD
17.70)
50.39
%
Control 2
32,504,
320
100.00
%
4.64% 6.53%
17.56
(SD
143.03)
49.62 37.39
42.38 (SD
17.73)
50.57
%
Control 3
31,215,
278
100.00
%
4.84% 6.35%
17.38
(SD
141.92)
49.61 37.39
42.22 (SD
17.87)
50.55
%
Treated 1
29,773,
698
100.00
%
4.68% 5.64%
18.57
(SD
137.33)
49.61 37.32
42.49 (SD
17.62)
50.66
%
Treated 2
27,301,
324
100.00
%
10.08
%
4.64%
20.39
(SD
438.40)
49.54 37.21
40.81 (SD
19.02)
52.50
%
Treated 3
30,144,
749
100.00
%
6.40% 4.83%
21.72
(SD
266.93)
49.61 37.33
42.15 (SD
17.88)
52.31
%
Phred+33 quality score format is used here.

Table 2. Post-alignment QC chart for ESC samples (1st

round)
Sample
Name
Total
Reads
Aligned Unique Coverage
Avg.
Coverage
Depth
Avg.
Length
Avg.
Quality
Avg. Mapping
Quality
%GC
13

Control 1
28,016,
557
100.00
%
5.17% 5.05%
19.38
(SD
188.82)
49.65 37.5
42.55 (SD
17.53)
51.4
7%
Control 2
29,814,
135
100.00
%
5.46% 5.76%
18.16
(SD
190.24)
49.63 37.48
42.33 (SD
17.75)
51.5
5%
Control 3
32,181,
635
100.00
%
5.17% 5.77%
19.55
(SD
180.97)
49.64 37.48
42.53 (SD
17.57)
51.8
1%
Treated 1
30,693,
275
100.00
%
5.20% 5.71%
18.87
(SD
185.39)
49.62 37.4
42.37 (SD
17.71)
51.7
2%
Treated 2
30,505,
293
100.00
%
5.09% 6.17%
17.41
(SD
167.68)
49.62 37.43
42.34 (SD
17.75)
51.4
7%
Treated 3
28,372,
591
100.00
%
5.09% 5.63%
17.68
(SD
163.00)
49.64 37.49
42.46 (SD
17.63)
51
.65%
Phred+33 quality score format is used here.

Principle component analysis (PCA), processed by Partek
®
Genomic Suite with the transcript
RPKM algorithm, was used to cluster samples with respect to level of gene expression
(displayed for all sample types in Figure 2). In the figure below, all stromal samples are found on
the PC2 axis (with a PC1 value of ~0). Epithelial samples are found to the right, with PC1
values >0). No statistically significant difference was found between ESC 1,25(OH)
2
D3-treated
and untreated cells. However among EEC cells, treated and untreated demonstrate conspicuous
divergence except for one treated sample, which correspond to the 5th sample in Figure 1.
14

Figure 2. Principle component analysis mapping generated through Partek
®
Genomic Suite
shows overall gene expression profile for all samples. In this graph, principle component 1
(PC#1) and principle component 2 (PC#2) are the X-axis and Y-axis of the coordinates,
respectively. The blue dots represent the EEC controls; green dots represent the EEC treated,
yellow dots represent the ESC controls and red dots represent the ESC treated. Difference
between treated ones and controls is not significant among ESC cells, but is obvious among EEC
cells, except that one treated sample falls within the cluster of controls. This aberrant sample
corresponds to the 5th sample shown in Figure 1, which has poor RNA quality, and is considered
as an outlier.

Individual genes were analyzed for differential expression in treated vs. untreated cells.
Differentially expressed genes (DEG) were defined as those having Log
2
(Treated/Control
Ratio)≥2 and p-value ≤ 0.01, which are conventional cutpoints. Multiple testing correction with
FDR adjustment was not applied, because FDR adjustment would be too stringent to obtain any
significant results, especially for ESC samples. In agreement with the PCA results, we found
that vitamin D treatment had a larger impact on gene expression in the epithelial than in the
stromal cell line. A total of 153 transcripts, and 149 unique genes were differentially expressed
15

in EEC samples. In ESC samples, these numbers are only 27 transcripts belonging to 25 genes.
Surprisingly, we found DEGs do not overlap between EEC and ESC samples (Figure 3).

Figure 3. Overlap in DEG between EEC and ESC samples. DEGs in EEC samples are shown as
the light green circle to the left, while ESC samples are the pink circle to the right. Number of
DEG met the criteria is 149 for EEC and 25 for ESC. This figure shows that there is no
overlapped DEG between the two types of samples.

Gene Ontology (GO) and pathway analysis in EEC samples from the first round of
experiments
GO and pathway results of the 149 genes were computed by IPA. The most up-regulated and
down-regulated 10 genes from EEC samples are shown in Tables 3 and 4, respectively.

16

Table 3 Most upregulated genes in EEC samples.
Gene Fold Upregulated p-value
HAX1 9.328 9.15E-05
DSN1 4.250 6.48E-03
PLK3 4.214 8.41E-03
APEX1 4.000 9.74E-03
GYG1 3.991 1.12E-03
CCDC71 3.950 5.37E-03
CABYR 3.914 2.20E-03
POLE2 3.588 9.60E-03
FLRT3 3.526 8.35E-03
FOS 3.444 9.89E-03
Most upregulated 10 genes in EEC samples with corresponding fold change and p-values. P-
value of significance is set at ≤ 0.01, since the overall significant level of DEGs is not high. FDR
adjustment is not applied, either.

Table 4. Most downregulated genes in EEC samples.
Gene Fold Downregulated p-value
INTS9 7.376 4.81E-05
RAD17 5.183 5.45E-04
BACH2 4.814 9.00E-03
PRNP 4.600 5.86E-03
CTTN 4.413 1.56E-03
GNG10 4.308 9.01E-03
LRP12 4.062 7.37E-03
PLAUR 3.857 1.01E-05
RUNX1T1 3.777 7.17E-04
DSTYK 3.647 9.13E-03
Most downregulated 10 genes in EEC samples with corresponding fold change and p-values. P-
value of significance is set at ≤ 0.01, since the overall significant level of DEGs is not high. FDR
adjustment is not applied, either.

Subsequent pathway analysis of the 149 DEGs in epithelial cells shows the most regulated
cellular functions. Figure 4 displays two types of information: the DEG regulated cellular
functions in heat map (4A), and subdivided cellular functions that have most statistical
association to our DEGs in bar chart (4B). In 4A, each big square represents an aspect of cell
function, and each small square within a big block is a smaller category of that function. The size
17

of a square indicates the magnitude of the association between a set of DEGs and a function,
which is determined by -log (p-value) using Fisher’s exact test, with p-value of <0.05 (-log p ≈
1.3) being statistically significant. The bigger the square, the less chance the association between
DEGs and that particular function is due to randomness. Significance of regulation is determined
by Z score transformation. A positive Z-score means upregulation, while a negative Z-score
means downregulation. Z-score ≥2 or ≤2 indicates statistically significant change in gene
expression. Color intensity of a square is a reflection of Z-score. Cold color means
downregulation and hot color means upregulation. When IPA is unable to make prediction for a
cellular function, that square is shown in grey.

Figure 4B is a bar chart showing the most DEG associated cellular function networks. There
are function networks in 4B not shown in 4A, because in fact, each bar represents one aspect of a
small function category (small squares in 4A). For example, the first 3 bars, cellular development,
cell growth and proliferation and connective tissue development and function, are all functions
related with proliferation of fibroblast cell lines, which is a small category under the big category
“Cell Development” (the first big square in 4A). These functions therefore appear in 4B, but may
not in 4A. The threshold (yellow horizontal line in 4B) of significant association between DEG
and cellular function is also set by p-value of <0.05 and -log p ≈ 1.3. Taller bar means higher
significance of association.

18

4A Heat map of regulated cellular functions for EEC samples.

4B DEG associated cellular functions for EEC samples.

Figure 4. The most regulated cellular functions in EEC samples are shown in heat map and bar
chart in 4A and 4B, respectively. 4A Cold color (purple) means downregulation and hot color
(orange) means upregulation, white color means no regulation, and grey color stands for no
prediction was made by IPA for this cellular function. Intensity of color is proportional to the
degree of regulation. Z-score is the determinant of color intensity, with Z-score ≥2 or ≤2 means
statistically significant change. Each big square is an aspect of cellular function, and each small
19

square within is a specific category under that function. Size of a square is determined by
association between a set of DEGs and a cellular function. Fisher’s exact test is used to identify
such association, and a -log (p-value) of greater than 1.3 (p<0.05) is considered statistically
significant. 4B A bar chart of DEG associated cell functions is shown. Y-axis is the -log (p-value)
of the association between DEG and a cellular function, X-axis shows the names of the regulated
function. A threshold of significant regulation is set in the graph, which is -log (0.05). Height of
a bar represents significance of association.

Most regulated cellular functions were down-regulated for 1,25(OH)
2
D3 treated cells, versus
untreated, while upregulated aspects are of small magnitude. Down-regulated functions included
those involved in cell/tissue development, proliferation, cell/tissue morphology, cell
death/survival and cell function maintenance are observed. Cancer development functions are
downregulated, too. We found that apoptosis and cell death functions are both downregulated.
For apoptosis, Z-score is -1.214, and –log (p-value) is 1.648 (p=0.022). For cell death, Z-score is
-0.776, and -log (p-value) is 1.643 (p=0.023). Interestingly, one of the most down-regulated
genes, RAD17, has an upregulated alternatively spliced transcript. The RefSeq ID of the
downregulated transcript is NM_133340, which encodes a 505aa protein, while ID of the
transcript that was transcribed 2.668 folds more is NM_13343, which encodes a protein contains
670aa (p=9.44E-03).

Figure 5 shows the most regulated pathways and supplemental information. Similar to cellular
function, significance of association between DEG and a pathway is determined by Fisher’s
exact test, p<0.05 is considered statistically significant. Transformed p value, -log (p-value) is
proportional to the location of each pathway; the more left pathway has more significant
association. The Y-axis is the percentage of significant DEGs involved in a pathway out of the
number of all genes participating in this pathway, which is the number above each bar. The
downregulated genes are shown in green and the upregulated genes are red. Expressions of
20

DEGs associated with these pathways are mainly downregulated, similar to what we observe in
cellular functions.

Figure 5. Most regulated pathways in EEC samples. Y-axis is the percentage of DEG among all
genes in the pathway. Number of all genes involved in a pathway is shown on top of each bar.
Downregulated genes are shown in green, and upregulated genes are shown in red in each bar.
X-axis is the names of regulated pathways. Pathways with higher significance of association are
located further left on the X-axis. Significance of association is measured by Fisher’s exact test,
with p<0.05 being statistically significant. -log (p-value) which indicates level of significance is
shown in small orange squares on each bar.

According to IPA, the organization of cytoskeleton function was downregulated (-log(p-
value)=2.471, Z-score=-0.771), along with function of microtubule dynamics function (-log(p-
value)=1.837, Z-score=-0.581), yet the Z-scores were not great. DEGs involved in cytoskeleton
are listed in Table 5, with fold change and expected impact on the function.

Table 5. DEGs related with cell motility functions in EEC samples.
Gene Function Genes in dataset Prediction (based on expression direction) Fold Change
Cytoskeleton Regulation
FOS Affected 3.444
21

HAUS8 Increased 3.286
DYNLL1 Increased 2.205
PANK4 Decreased 2.134
PIK3R1 Affected -2.014
RIT1 Increased -2.080
STXBP5 Affected -2.110
RAPGEF2 Increased -2.133
NFIB Decreased -2.211
DYNC2H1 Affected -2.300
ROR1 Decreased -2.309
VPS54 Affected -2.315
SLC3A2 Increased -2.432
TNIK Affected -2.459
SKIL Affected -2.556
CNTF Decreased -2.828
AGER Decreased -3.249
PLAUR Decreased -3.857
CTTN Decreased -4.413
PRNP Affected -4.600

Invasion of Cells
APEX1 Increased 4.000
SLC12A2 Decreased -2.000
PIK3R1 Decreased -2.014
HIPK2 Increased -2.031
PLA2R1 Decreased -2.097
ACSL4 Decreased -2.333
FOXO3 Decreased -2.341
AGER Affected -3.249
ETV1 Decreased -3.284
PLAUR Decreased -3.857
CTTN Decreased -4.413

Formation of Cellular Protrusions
FOS Affected 3.444
RIT1 Increased -2.080
STXBP5 Affected -2.110
RAPGEF2 Increased -2.133
NFIB Decreased -2.211
DYNC2H1 Affected -2.300
ROR1 Decreased -2.309
SKIL Affected -2.556
CNTF Decreased -2.828
AGER Affected -3.249
CTTN Decreased -4.413
PRNP Decreased -4.600
22

DEGs related with cell movement in EEC samples are listed. The column “Prediction” shows
impact on a cellular function by the change in expression of a particular gene.

Gene Ontology (GO) and pathway analysis in ESC samples from the first round of
experiments
The most up-regulated and down-regulated 10 genes in ESC samples are shown in Table 6 and
7, respectively.

Table 6. Most upregulated genes in ESC samples (1st round).
Gene Fold Upregulated p-value
RNF8 3.223 3.36E-03
FBXW11 2.756 4.32E-03
UMPS 2.578 9.05E-05
CCDC121 2.397 3.48E-03
BZW2 2.234 2.17E-04
MAPK3 2.192 5.99E-06
GUK1 2.130 3.51E-03
TMEM218 2.106 4.00E-03
TRNAU1AP 2.079 2.21E-03
ZMAT4 2.068 2.86E-03
Most upregulated 10 genes in ESC samples with corresponding folds of change and p-values. P-
value of significance is set at ≤ 0.01, since the overall significant level of DEGs is not high. FDR
adjustment is not applied.

Table 7. Most downregulated genes in ESC samples (1st round).
Gene Fold Downregulated p-value
UBE2L3 5.889 7.93E-03
CHMP3 4.917 6.47E-04
IFI6 4.655 8.16E-05
ST6GALNAC4 4.199 1.47E-03
CAMKK2 3.590 3.90E-05
CDH13 3.251 1.83E-03
SMARCAD1 3.246 1.64E-03
HMHA1 2.827 1.64E-03
TIMM23 2.748 1.81E-03
AGAP6 2.183 4.17E-03
23

Most downregulated 10 genes in ESC samples with corresponding folds of change and p-values.
P-value of significance is set at ≤ 0.01, since the overall significant level of DEGs is not high.
FDR adjustment is not applied.

The same GO and pathway analysis as the EEC samples was conducted for the ESC samples.
However, the only cellular function found to be significantly regulated is the infection aspect,
and the direction of regulation is upward. Therefore, Figure 6A only shows the upregulated
infection functions. The three squares stand for: infection by HIV-1, infection of cells and viral
infection, from left to right. HIV-1 infection and viral infection aspects are slightly upregulated,
with Z-score of 0.192 on both functions. Viral infection receives little regulation, with Z-score of
0.054. Figure 6B still shows the cellular functions that significantly associated with DEGs we
found. Many of the most significantly associated cellular functions are metabolism related.
Discrepancy in direction of regulation for different transcripts of the same gene is observed in
ESC samples, too. One of the most upregulated genes shown in the table, FBWX11, has a
downregulated transcript with 2.026 folds of change (p=7.29E-03). The same occurs for one of
the most downregulated genes CAMKK2, which has an upregulated transcript with 3.030 folds
of change (p=3.00E-03).

6A Upregulated infection functions shown in heat map for ESC samples (1st round).

24

6B Regulated cellular functions shown in bar chart for ESC samples (1st round).

Figure 6. Most regulated cellular functions in ESC samples are shown in heat map and bar chart
in 6A and 6B, respectively. 6A A heat map of infection aspect of cellular functions is shown.
Cold color (purple) means downregulation and hot color (orange) means upregulation, grey color
stands for no regulation. Intensity of color is proportional to the degree of regulation. However,
in this graph, only upregulation of infection function is shown. Z-score is the determinant of
color intensity, with Z-score ≥2 or ≤2 means statistically significant change. Each big square is
an aspect of cellular function, and each small square within is a specific category under that
function. Size of a square is determined by association between a set of DEGs and a cellular
function. Fisher’s exact test is used to identify such association, and a -log (p-value) of greater
than 1.3 (p<0.05) is considered statistically significant. 6B A bar chart of DEG associated cell
functions is shown. Y-axis is the -log (p-value) of the association between DEG and a cellular
function, X-axis shows the names of the regulated function. A threshold of significant regulation
is set in the graph, which is -log (0.05). Height of a bar represents significance of association.

Figure 7 shows the most regulated pathways in ESC samples, just like Figure 5 for the EEC
samples.

25

Figure 7. Most regulated pathways in ESC samples shown in bar chart (1st round). Y-axis is the
percentage of number of DEGs associated with a pathway vs. number of all genes in this
pathway. Number of all genes involved in a pathway is shown on top of each bar.
Downregulated genes are shown in green, and upregulated genes are shown in red in each bar.
X-axis is the names of regulated pathways. Pathways with higher significance of association are
located further left on the X-axis. Significance of association is measured by Fisher’s exact test,
with p<0.05 being statistically significant. -log (p-value) which indicates level of significance is
shown in small orange squares on each bar.

Overall result from the second round of experiments on ESC cells
As we can see on the post-alignment QC chart (Table 8), mapping qualities of these samples
was not as good as those from the first round of experiment. However, we observed significant
differences in gene expression between the treated and the control. DEG were defined as those
having Log
2
(Treated/Control Ratio)≥2 as before, but p-value was set to be ≤ 0.05, because of
reduced sample size.

Table 8. Post-alignment QC chart for ESC samples (2nd

round)
Sample
Name
Total
Reads
Aligned Unique Coverage
Avg.
Coverage
Depth
Avg.
Length
Avg.
Quality
Avg. Mapping
Quality
%GC
26

Control 1
13,771,
163
100.00
%
3.52% 4.42%
14.43
(SD
165.19)
49.45 36.86
29.47 (SD
24.23)
50.9
4%
Control 2
14,259,
672
100.00
%
3.50% 4.60%
14.45
(SD
171.59)
49.45 36.85
29.22 (SD
24.27)
51.0
3%
Treated 1
13,995,
438
100.00
%
3.57% 4.84%
13.28
(SD
163.24)
49.45 36.87
29.79 (SD
24.16)
50.7
3%
Treated 2
12,613,
207
100.00
%
3.66% 4.53%
12.82
(SD
144.92)
49.45 36.85
29.77 (SD
17.75)
50.5
5%
Phred+33 quality score format is used here.

Gene Ontology (GO) and pathway analysis in ESC samples from the second round of
experiments
A total of 451 differentially expressed genes were identified for the ESC samples in this round
of experiments, and 443 of them were included in the final analyses. The most up-regulated and
down-regulated 10 genes in ESC samples are shown in Table 6 and 7, respectively. Down-
regulation is especially prominent here, with several genes reached over 100 fold decreased
expression.

Table 9. Most upregulated genes in ESC samples (2nd round).
Gene Fold Upregulated p-value
CYP24A1 129.490 2.06E-02
G0S2 5.417 1.68E-02
FENDRR 4.864 2.12E-02
PLAU 4.649 2.24E-03
FOXF1 4.177 9.64E-03
FHOD3 4.167 4.72E-02
27

INHBA 4.146 7.39E-03
GABBR2 4.111 1.68E-02
COL16A1 4.011 6.74E-03
XIST 3.941 6.53E-03
Most upregulated 10 genes in ESC samples with corresponding fold change and p-values. P-
value of significance was set at ≤ 0.05. FDR adjustment is not applied.

Table 10. Most downregulated genes in ESC samples (2nd round).
Gene Fold Downregulated p-value
SLIT3 135.000 2.03E-02
DSC3 129.500 2.06E-02
KRT8 116.227 1.13E-03
ALHD1A2 110.000 2.21E-02
LRRN4 73.500 2.64E-02
KRT19 71.000 3.96E-03
FAM84B 55.000 3.03E-02
TFAP2A 53.000 3.08E-02
NPY 52.500 3.10E-02
L1CAM 47.000 9.70E-03
Most downregulated 10 genes in ESC samples with corresponding folds of change and p-values.
P-value of significance was set at 0.05. FDR adjustment is not applied.

As shown in the Figure 8A and 8B, the most regulated cellular functions involve cellular
movement, cancer, tissue/ organismal development and cell-to-cell signalling and interaction, etc.
Reproductive disease is also among the most regulated functions, too.

Within the cellular movement block, the “cell movement” sub-function is most closely
associated with the DEGs, but is not highly regulated (-log(p-value)= 31.104, z-score= -0.135).
The cell movement of tumor cell lines sub-function received a fair amount of down-regulation (-
log(p-value)= 22.134, z-score= -0.448). Functions involving movements of blood cells and
endothelial cells have z-scores of 0.846 and 1.200, respectively.

28

In the bar chart, we can see reproductive system disease being the 9th most relevant function
to the DEGs. In the heat map, however, most of its sub-functions are shown in grey cells,
including the one for endometriosis (-log(p-value)= 5.788). This means IPA is unable to
determine level of regulation based on current data. Only 3 sub-functions were up-regulated,
they are genital tumor (-log(p-value)= 18.456, z-score= 1.067), breast cancer (-log(p-value)=
16.795, z-score= 1.067) and mammary tumor (-log(p-value)= 17.251, z-score= 0.571).

8A Regulated cellular functions shown in heat map for ESC samples (2nd

round).

29

8B DEG associated cellular functions shown in bar chart for ESC samples (2nd round).

Figure 8. Interpretation of 8A and 8B is the same as that of 4A and 4B/ 6A and 6B. The most
regulated pathway is clearly cellular movement.

The most regulated canonical pathways are shown in Figure 9. The VDR/RXR activation is
the most intriguing one. This pathway in IPA consists of 88 genes and only 10 of them exist in
our DEG list, yet it has the strongest association to the DEGs. 5 genes had up-regulation and
equal number of genes had down-regulation. Another pathway of interest is the inhibition of
matrix metalloproteases, which ranked the 4th in terms of association to our DEGs. It only
contains 40 genes, and 4 were up-regulated while 3 were down-regulated. DEGs that have
impact on our two pathways of interest are collected in Table 11.
30

Figure 9. Most regulated pathways in ESC samples shown in bar chart (2nd round).
Interpretation is the same as for Figure 5 and 7.

Table 11. DEGs related with vitamin D signaling and cellular movement functions in ESC
samples (2nd round).
Gene Function Genes in dataset Fold Change
VDR/RXR Activation
CYP24A1 129.490
SPP1 3.679
IGFBP5 2.997
COL13A1 2.643
VDR 2.271
IGFBP6 -2.060
IGFBP3 -3.227
SEMA3B -11.750
WT1 -12.250
CD14 -15.750

Inhibition of Matrix Metalloprotease
MMP2 3.416
TFPI2 3.339
TIMP1 3.200
MMP14 2.130
THBS2 -3.293
MMP15 -6.973
MMP24 -7.250
DEGs related with VDR/RXR activation and inhibition of matrix metalloprotease pathways in
ESC samples are listed.

31

Discussion
In our first round of experiments, no significant change in gene expression was observed on
endometriotic stromal cell line, but epithelial cell line showed some extent of regulation by
1,25(OH)
2
D3. Epithelial cell development and proliferation were downregulated after
1,25(OH)
2
D3 treatment, which is consistent with known function of 1,25(OH)
2
D3 (28). While in
the second round of experiments, using only the stromal cell line, we found large differences in
gene expression in samples with/ without 1,25(OH)
2
D3 treatment. Such discrepancy between the
two sets of experiments can only be explained by the effect of solvents of 1,25(OH)
2
D3. Ethanol,
which was used in the first round to dissolve 1,25(OH)
2
D3, seemed to have caused the stromal
cells to lose their adherence and float, and this could have led to cell death before any significant
change in gene expression could happen. The epithelial cells, on the other hand, tolerated ethanol
much better with only a small number lifted after treatment, therefore we didn’t study them in
the second round of experiment. We turned to DMSO as the solvent in the second round, and this
time the stromal cells tolerated fairly well. DMSO should be used as a standard solvent for
vitamin D and its derivatives in our future experiments. Detailed discussions of the two rounds of
experiment are listed in the following paragraphs.

DEG ontology of EEC samples from the first round of experiments
Since endometriosis cells have the ability to invade tissue, cell movement can be important for
disease development, and cytoskeleton remodelling is critical to cell mobility. So first, we
looked into how gene expression related to cytoskeleton is changed (29). Although FOS and
PRNP are involved in cytoskeleton organization function and their changes are considerably high
(3.444 folds for FOS and -4.600 folds for PRNP), the predicted direction of influence on this
32

function has not been verified by IPA. Genes like HAUS8 (3.286 folds) and CTTN (-4.413 folds)
have highly altered expression and greater impact on cytoskeleton organization as well as
microtubule dynamics; however, direct association between their ontologies and vitamin D has
not been reported. Gene HAUS8’s protein product is one subunit of augmin protein complex that
helps to localize gamma-tubulin to spindle microtubule (30). However, this process is related to
mitosis rather than cell movement. CTTN gene, on the other hand, encodes protein cortactin,
which is an essential component in invadopodia (31), a type of protrusion that is rich in actin and
forms on cell membrane. Invadopodia acts like feet of a cell, its actin structures allows it to
extend into extracellular matrix (ECM), followed by secretion of matrix metalloproteinase
(MMP) to degrade ECM. Invadopodia is known for its presence in invasive and metastatic
cancer cells, so CTTN may be an important gene in endometriosis (32).

Cellular functions such as formation of plasma membrane projections (-log(p-value)=1.980, Z-
score= -0.599) and formation of cellular protrusions (-log(p-value)=1.664, Z-score=-0.906) were
down-regulated, a result consistent with what happened on cytoskeleton organization. Again,
these processes mostly occur in neurons, rather than in invasive cells.

Looking at cellular movement functions directly, we found reduced invasive activities in cells
(-log(p-value)=1.761, Z-score=-1.433). While decrease in CTTN expression remains a big
contributor, another important gene involved in both cytoskeleton and invasion is PLAUR (-
3.857 folds). PLAUR encodes urokinase plasminogen activator receptor (uPAR), it resides on
cell surface, and has proteolytic property so that it can cleave plasminogen to active plasmin, a
serine protease that is capable of degrading ECM, which generates a signal that can be
33

transduced to intracellular environment. Besides, uPAR itself interacts with protein vitronectin, a
component of ECM, and this gives uPAR a role in non-integrin cell adhesion. Because of its
function in signalling and interaction with ECM, uPAR are found to be closely linked with
cancer invasion, especially in glioma (33). From the information above, uPAR protein and
PLAUR gene are very likely to participate in endometriosis development. Huber et al. reported
that expression of PLAUR is upregulated in endometriotic stromal cells after treatment with
human chorionic gonadotropin (hCG) (34). They claimed that hCG have therapeutic effect
against endometriosis, and proteins related with ECM degradation reduce invasion. But their
study didn’t mention how hCG affect epithelial cells, and increased PLAUR expression should
be pro-invasion according to existing knowledge.

There are a number of canonical pathways associated with DEGs in epithelial cells , but the
genes directly functions in these pathways are only a few. The most influenced canonical
pathway is glutamine degradation, 1 of 5 genes in this pathway was regulated. This gene is GLS,
it encodes a K-type mitochondrial glutaminase. This pathway is frequently used by proliferative
cells as an approach to generate energy, particularly in cancer cells, but whether it is used by
endometriotic cells remains unknown (35).

DEG ontology of ESC samples from the first round of experiments
Change in gene expression responding to 1,25(OH)
2
D3 was very limited. Neither function
network nor canonical pathway analysis is very informative. The only thing worth noticing is
that the treatment changed alternative splicing on gene FBWX11 and CAMKK2. FBWX11
encodes a F-box/ WD40 repeat gene protein, which is a subunit of SCFs, a ubiquitin protein
34

ligase complex. This complex is known to be responsible for ubiquitination of phosphorylated
NFkappa1A, this allows release of free NFkappaB, which then participate in the WNT pathway
(36). There is one study mentioned that FBWX11 was downregulated (-1.21 folds) in androgen
receptor (AR) knockdown endometrial stromal cells, but they classified this gene as “not known
to be regulated in endometrium”. They found that AR knockdown enhances the motility and
proliferation of decidualizing cells, a type of cell originates from differentiation of endometrial
stromal cells, and is critical for embryo implantation (37). We cannot conclude that FBWX11
has an important role in fertility, but it may adjust endometrium cells in several ways via
NFkappaB. Another article reported that endometriotic epithelial cells can induce MMP
expression in normal endometrial stromal cells, making them behave like inflammatory tissue,
and this process is NFkappaB dependent (38). Implication of FBWX11 may exist in this case.
How different transcript of this gene impact on endometriosis still needs investigation. As for
gene CAMKK2, its alternative splicing is only known to be regulated by protein kinase A (PKA).
PKA and 1,25(OH)
2
D3 have mutual influence on each other on many occasions (39, 40, 41). If
the alternative splicing of CAMKK2 in our study is due to PKA is not clear.
Given that ESC cells in the first round of experiments were affected by ethanol treatment, and
since these results were not in agreement with the ones in the second round of experiment, they
should be taken with caution.

DEG ontology of ESC samples from the second round of experiments
In contrast to what was in the first round of experiment, drastic change in gene expression was
observed after 1,25(OH)
2
D3 treatment in ESC samples. ESC appeared to be less tolerant to
35

ethanol than EEC, so ethanol should be avoided in future experiments. The differences in gene
expression hint us that ESC may be more sensitive to 1,25(OH)
2
D3 than EEC.

Cellular movement function had the second strongest association with our DEGs and was
most regulated. On the down-regulation side, some genes reached >100 folds less expression.
KRT8 and KRT19 are two strongly down-regulated genes; they belong to the keratin family.
Keratins are responsible of forming intermediate filament of epithelial cells, and are usually
expressed in basal level in stromal cells (42). In case of certain types of carcinoma, expression of
cytokeratin can be elevated in stromal cancer cells (43). In epithelial cells, less keratin 8/18 was
found to associate with metastasis during epithelial-mesenchymal transition (44), which is
probably why down-regulation of KRT genes were predicted to increase cell movement in Table
11. However, in stromal cells, this prediction may not be accurate. SLIT3 gene is the most down-
regulated gene with -135 folds of change in expression; however, IPA couldn’t determine
whether this increases or decreases cell movement because of lack of support from literature.
SLIT3 gene encodes a secretory protein that binds to its receptor ROBO, together they can guide
developing axons, which is why axonal guidance signalling being the third most important
canonical pathway to our DEGs (Figure 9). SLIT3 has repulsive function and was found to
inhibit cell invasion in some cases (45, 46). It seems down-regulation of SLIT3 might boost cell
movement, and we couldn’t find how SLIT3 and 1,25(OH)
2
D3 are directly connected in
literatures.

VDR/RXR activation pathway is the canonical pathway most closely associated with our
DEGs, and should be the most relevant pathway to our experiment, theoretically. The gene
36

received most regulation within this pathway is CYP24A1, which had 129.490 fold of increased
expression. CYP24A1 is a mitochondrial gene that belongs to the cytochrome P450 family. It has
vitamin D3 24-hydroxylase activity, and its major role is to metabolize 1,25(OH)
2
D3 and
25(OH)D3 (47). Presence of 1,25(OH)
2
D3 is well-known to have great impact on expression of
CYP24A1 (48). In the study by Anderson et al., CYP24A1 mRNA level was found to be
significantly higher in certain tumors than in normal tissues; VDR mRNA level was also found
to be higher in these tumors, however far less than that of CYP24A1. The researchers believe
higher CYP24A1 level would result in faster metabolism of vitamin D3, thus eliminates its anti-
tumor effect, and little up-regulation of VDR is enough to induce drastic over-expression of
CYP24A1 (49). Their findings coincide with our results; the VDR gene expression in our
samples was up-regulated only by 2.271 folds, but it stimulated 129.490 folds of up-regulation in
CYP24A1 gene. This further substantiates the credibility of our experiment, which proves that
ESC may be more sensitive to 1,25(OH)
2
D3 than EEC.

Inhibition of matrix metalloproteases is the fourth canonical pathways most closely associated
with our DEGs. Within this pathway, two metalloprotease coding genes MMP24 and MMP15
were down-regulated by 7.250 folds and 6.973 folds, while the other two such genes MMP2 and
MMP 14 were up-regulated by 3.416 folds and 2.130 folds, respectively. As described in the
EEC section, MMPs are essential for cell invasion (31, 32), by digesting ECM when cells move.
Besides, PLAU, the gene that encodes the ligand for PLAUR (discussed in EEC section) was the
most upregulated gene in cell movement function. Upregulation of PLAUR can enhance
degradation of ECM during cell movement.

37

To sum up, it seems 1,25(OH)
2
D3 is capable to affect various aspects of cell movement/
invasion, but it is hard to conclude whether it inhibits or stimulates movement in our ESC
samples.

Conclusion
There are not many studies focusing on how vitamin D and its derivatives modulate
endometrial cells. The few studies that investigated this mainly concentrated on normal stromal
cells instead of endometriotic cells or epithelial cells. We report here on DEGs and pathways that
may be involved in endometriosis. From our study, we can conclude that endometriotic stromal
cells are more receptive to 1,25(OH)
2
D3 treatment than epithelial cells, since there’s prominent
change in VDR/RXR pathway in stromal cells, and they may be more worth studying. Important
DEGs we found in epithelial cells include CTTN and PLAUR, which are regulators in cell
movement, thus they may affect invasion in endometriotic cells. Although they serve important
functions, their changes in expression are not high. In stromal cells, keratin family genes, MMP
family genes and SLIT3 gene are strongly regulated, and they all play critical roles in cell
movement and invasion. But according to existing knowledge, it is hard to determine whether
1,25(OH)
2
D3 stimulates or quenches cell invasion by altering expression of these genes. More
basic research will be needed to confirm and establish any causal relationship. However, our
samples are diseased cells, which differ from normal cells and can be resistant to certain
treatments.

In fact, there is one study claiming that 25(OH)D3, the precursor of 1,25(OH)
2
D3, is
significantly higher in the serum of women with endometriosis compared to controls (50). This
38

study reported that endometriosis patients had higher levels of 1,25(OH)
2
D3 as well, though this
was not statistically significant. These findings could indicate that endometriosis patients have
high levels of reserved 25(OH)D3 and 1,25(OH)
2
D3 because they are less able to utilize vitamin
D in the body. Similarly, in our case, drastic up-regulation of CYP24A1 in endometriotic ESC
could mean that they are resistant to vitamin D3 treatment.

DEGs of endometriotic epithelial and stromal samples did not overlap much, and this adds
difficulty for data interpretation. Our next step will be increasing sample size for this experiment,
focusing on examining additional stromal cell lines, and treating cells with 1,25(OH)
2
D3 at
various concentration in order to investigate if there is a dose dependent change in gene
expression.

Acknowledgements
I hereby acknowledge Dr. Sue Ingles for her supervision; Doerthe Brueggmann, MD for
providing endometriosis cells and help in cell culture; USC Epigenome Center for next-
generation sequencing, and Dr. Yibu Chen for bioinformatics support.

References
1. Sayasneh, A., Tsivos, D., Crawford, R. (2011) Endometriosis and ovarian cancer: a
systematic review. ISRN Obstet Gynecol. 2011:140310.
2. van der Linden, P.J. (1996) Theories on the pathogenesis of endometriosis. Suppl 3:53-65.
3. Fauser, B.C., Diedrich, K., Bouchard, P., Domí nguez, F., Matzuk, M., Franks, S., Hamamah,
S., Simó n, C.,Devroey, P., Ezcurra, D., Howles, C.M. (2011) Contemporary genetic
technologies and female reproduction. Hum Reprod Update. 17:829-847.
4. Signorile, P.G., Baldi, F., Bussani, R., D'Armiento, M., De Falco, M., Baldi, A. (2009)
Ectopic endometrium in human foetuses is a common event and sustains the theory of
39

mü llerianosis in the pathogenesis of endometriosis, a disease that predisposes to cancer. J
Exp Clin Cancer Res. 9;28-49.
5. Batt, R.E., Smith, R.A., Buck Louis, G.M., Martin, D.C., Chapron, C., Koninckx, P.R., Yeh,
J. (2007) Mullerianosis. Histol Histopathol. 22:1161–1166.
6. Wellbery, C. (1999) Diagnosis and Treatment of Endometriosis. Am Fam Physician.
60(6):1753-1762.
7. Laschke, M.W., Giebels, C., Menger, M.D. (2011) Vasculogenesis: a new piece of the
endometriosis puzzle. Hum Reprod Update. 17:628-636.
8. Rahmioglu, N., Nyholt, D.R., Morris, A.P., Missmer, S.A., Montgomery, G.W., Zondervan,
K.T. (2014) Genetic variants underlying risk of endometriosis: insights from meta-analysis of
eight genome-wide association and replication datasets. Hum Reprod Update. Epub ahead of
print.
9. Treloar, S.A., Wicks, J., Nyholt, D.R., Montgomery, G.W., Bahlo, M., Smith, V., Dawson,
G., Mackay, I.J., Weeks, D.E., Bennett, S.T., Carey, A., Ewen-White, K.R., Duffy,
D.L., O'connor, D.T., Barlow, D.H., Martin, N.G., Kennedy, S.H. (2005) Genomewide
linkage study in 1,176 affected sister pair families identifies a significant susceptibility locus
for endometriosis on chromosome 10q26. Am J Hum Genet. 77:365-376.
10. Painter, J.N., Anderson, C.A., Nyholt, D.R., Macgregor, S., Lin, J., Lee, S.H., Lambert,
A., Zhao, Z.Z., Roseman, F., Guo, Q., Gordon, S.D., Wallace, L., Henders, A.K., Visscher,
P.M., Kraft, P., Martin, N.G., Morris, A.P., Treloar, S.A., Kennedy, S.H., Missmer,
S.A., Montgomery, G.W., Zondervan, K.T. (2011) Genome-wide association study identifies
a locus at 7p15.2 associated with endometriosis. Nat Genet. 43:51-54.
11. Deligdisch, L. (2000) Hormonal pathology of the endometrium. Mod Pathol. 13:285-294.
12. Dassen, H., Punyadeera, C., Kamps, R., Delvoux, B., Van Langendonckt, A., Donnez,
J., Husen, B., Thole, H., Dunselman, G., Groothuis, P. (2007) Estrogen metabolizing enzyme
in endometrium and endometriosis. Hum Reprod. 22:3148-3158.
13. Giudice, L.C., Kao, L.C. (2004) Endometriosis. Lancet. 364:1789-1799.
14. Christodoulakos, G., Augoulea, A., Lambrinoudaki, I., Sioulas, V., Creatsas, G. (2007)
Pathogenesis of endometriosis: the role of defective 'immunosurveillance'. Eur J Contracept
Reprod Health Care. 12:194-202.
15. Lebovic, D.I., Mueller, M.D., Taylor, R.N. (2001) Immunobiology of endometriosis. Fertil
Steril. 75:1-10.
16. Pike, J.W. (2011) Genome-wide principles of gene regulation by the vitamin D receptor and
its activating ligand. Mol Cell Endocrinol. 347:3-10.
17. Zhang, Y., Leung, D.Y., Richers, B.N., Liu, Y., Remigio, L.K., Riches, D.W., Goleva,
E. (2012) Vitamin D Inhibits Monocyte/Macrophage Proinflammatory Cytokine Production
by Targeting MAPK Phosphatase-1. J Immunol. 188:2127-2135.
18. Segaert, S., Bouillon, R. (1998) Vitamin D and regulation of gene expression. Curr Opin Clin
Nutr Metab Care. 1:347-354.
19. Rudick, B., Ingles, S., Chung, K., Stanczyk, F., Paulson, R., Bendikson, K. (2012)
Characterizing the influence of vitamin D levels on IVF outcomes. Hum Reprod. 27:3321-
3327.
40

20. Rudick, B., Ingles, S., Chung, K., Stanczyk, F., Paulson, R., Bendikson, K. (2014) The
Influence of Vitamin D Levels on IVF Outcomes in Donor-Recipient Cycles. Fertil
Steril. 101: 447-452.
21. Grundmann, M., von Versen-Hö ynck, F. (2011) Vitamin D - roles in women's reproductive
health? Reprod Biol Endocrinol. 9:146.
22. Harris, H.R., Chavarro, J.E., Malspeis, S., Willett, W.C., Missmer, S.A. (2013) Dairy-
food, calcium, magnesium, and vitamin D intake and endometriosis: a prospective cohort
study. Am J Epidemiol. 177:420-430.
23. Barreto, A.M., Schwartz, G.G., Woodruff, R., Cramer, S.D. (2000) 25-Hydroxyvitamin D3,
the prohormone of 1,25-dihydroxyvitamin D3, inhibits the proliferation of primary prostatic
epithelial cells. Cancer Epidemiol Biomarkers Prev. 9:265-270.
24. Halloran, B.P., DeLuca, H.F. (1980) Effect of vitamin D deficiency on fertility and
reproductive capacity in the female rate. J Nutr. 110:1573-1580.
25. Yoshizawa, T., Handa, Y., Uematsu, Y., Takeda, S., Sekine, K., Yoshihara, Y., Kawakami,
T., Arioka, K., Sato, H., Uchiyama, Y., Masushige, S., Fukamizu, A., Matsumoto, T.,Kato, S.
(1997) Mice lacking the vitamin D receptor exhibit impaired bone formation, uterine
hypoplasia and growth retardation after weaning. Nat Genet. 16:391-396.
26. Du, H., Daftary, G.S., Lalwani, S.I., Taylor, H.S. (2005) Direct regulation of HOXA10 by
1,25-(OH)2D3 in human myelomonocytic cells and human endometrial stromal cells. Mol
Endocrinol. 19:2222-2233.
27. Ramagopalan, S.V., Heger, A., Berlanga, A.J., Maugeri, N.J., Lincoln, M.R., Burrell,
A., Handunnetthi, L., Handel, A.E., Disanto, G., Orton, S.M., Watson, C.T., Morahan,
J.M.,Giovannoni, G., Ponting, C.P., Ebers, G.C., Knight, J.C. (2010) A ChIP-seq defined
genome-wide map of vitamin D receptor binding: associations with disease and evolution.
Genome Res. 20:1352-1360.
28. Bouillon, R., Bischoff-Ferrari, H., Willett, W. (2008) Vitamin D and health: perspectives
from mice and man. J Bone Miner Res. 23:974-979.
29. Flamini, M.I., Sanchez, A.M., Goglia, L., Tosi, V., Genazzani, A.R., Simoncini, T. (2009)
Differential actions of estrogen and SERMs in regulation of the actin cytoskeleton of
endometrial cells. Mol Hum Reprod. 15:675-685.
30. Goshima, G., Mayer, M., Zhang, N., Stuurman, N., Vale, R.D. (2008) Augmin: a protein
complex required for centrosome-independent microtubule generation within the spindle. J
Cell Biol. 181:421-429.
31. Clark, E.S., Whigham, A.S., Yarbrough, W.G., Weaver, A.M. (2007) Cortactin is an
essential regulator of matrix metalloproteinase secretion and extracellular matrix degradation
in invadopodia. Cancer Res. 67:4227-4235.
32. Weaver, A.M. (2006) Invadopodia: specialized cell structures for cancer invasion. Clin Exp
Metastasis. 23:97-105.
33. Ferraris, G.M., Sidenius, N. (2013) Urokinase plasminogen activator receptor: a functional
integrator of extracellular proteolysis, cell adhesion, and signal transduction. Semin Thromb
Hemost. 39:347-355.
41

34. Huber, A., Hudelist, G., Knö fler, M., Saleh, L., Huber, J.C., Singer, C.F. (2007) Effect of
highly purified human chorionic gonadotropin preparations on the gene expression signature
of stromal cells derived from endometriotic lesions: potential mechanisms for the therapeutic
effect of human chorionic gonadotropin in vivo. Fertil Steril. 88:1232-1239.
35. Moreadith, R.W., Lehninger, A.L. (1984) The pathways of glutamate and glutamine
oxidation by tumor cell mitochondria. Role of mitochondrial NAD(P)+-dependent malic
enzyme. J Biol Chem. 259:6215-6221.
36. Winston, J.T., Strack, P., Beer-Romero, P., Chu, C.Y., Elledge, S.J., Harper, J.W. (1999) The
SCFbeta-TRCP-ubiquitin ligase complex associates specifically with phosphorylated
destruction motifs in IkappaBalpha and beta-catenin and stimulates IkappaBalpha
ubiquitination in vitro. Genes Dev. 13:270-283.
37. Cloke, B., Huhtinen, K., Fusi, L., Kajihara, T., Yliheikkilä , M., Ho, K.K., Teklenburg,
G., Lavery, S., Jones, M.C., Trew, G., Kim, J.J., Lam, E.W., Cartwright, J.E., Poutanen,
M., Brosens, J.J. (2008) The androgen and progesterone receptors regulate distinct gene
networks and cellular functions in decidualizing endometrium. Endocrinology. 149:4462-
4474.
38. Zhang, H., Li, M., Wang, F., Liu, S., Li, J., Wen, Z., Zhao, X. (2010) Endometriotic
epithelial cells induce MMPs expression in endometrial stromal cells via an NFkappaB-
dependent pathway. Gynecol Endocrinol. 26:456-467.
39. Young, M.R., Lozano, Y. (1997) Inhibition of tumor invasiveness by 1alpha,25-
dihydroxyvitamin D3 coupled to a decline in protein kinase A activity and an increase in
cytoskeletal organization. Clin Exp Metastasis. 15:102-110.
40. Santillá n, G.E., Boland, R.L. (1998) Studies suggesting the participation of protein kinase A
in 1, 25(OH)2-vitamin D3-dependent proteinphosphorylation in cardiac muscle. J Mol Cell
Cardiol. 30:225-233.
41. Hsieh, J.C., Dang, H.T., Galligan, M.A., Whitfield, G.K., Haussler, C.A., Jurutka,
P.W., Haussler, M.R. (2004) Phosphorylation of human vitamin D receptor serine-182
by PKA suppresses 1,25(OH)2D3-dependent transactivation. Biochem Biophys Res
Commun. 324:801-809.
42. Moll, R., Divo, M., Langbein, L. (2008) The human keratins: biology and pathology.
Histochem Cell Biol. 129:705-733.
43. Chia, Y., Thike, A.A., Cheok, P.Y., Yong-Zheng C. L., Man-Kit, T.G., Tan, P.H. (2012)
Stromal keratin expression in phyllodes tumours of the breast: a comparison with other
spindle cell breast lesions. J Clin Pathol. 65:339-347.
44. Fortier, A.M., Asselin, E., Cadrin, M. (2013) Keratin 8 and 18 loss in epithelial cancer cells
increases collective cell migration and cisplatin sensitivity through claudin1 up-regulation. J
Biol Chem. 288:11555-11571.
45. Denk, A.E., Braig, S., Schubert, T., Bosserhoff, A.K. (2011) Slit3 inhibits activator protein 1-
mediated migration of malignant melanoma cells. Int J Mol Med. 28:721-726.
46. Schubert, T., Denk, A.E., Ruedel, A., Kaufmann, S., Hustert, E., Bastone, P., Bosserhoff,
A.K. (2012) Fragments of SLIT3 inhibit cellular migration. Int J Mol Med. 30:1133-1137.
42

47. Sakaki, T., Kagawa, N., Yamamoto, K., Inouye, K. (2005) Metabolism of vitamin D3 by
cytochromes P450. Front Biosci. 10:119-134.
48. Ohyama, Y., Okuda, K. (1991) Isolation and characterization of a cytochrome P-450 from rat
kidney mitochondria thay catalyzes the 24-hydroxylation of 25-hydroxyvitamin D3. J Biol
Chem. 266:8690-8695.
49. Anderson, M.G., Nakane, M., Ruan, X., Kroeger, P.E., Wu-Wong, J.R. (2006) Expression of
VDR and CYP24A1 mRNA in human tumors. Cancer Chemother Pharmacol. 57:234-240.
50. Somigliana, E., Panina-Bordignon, P., Murone, S., Di Lucia, P., Vercellini, P., Vigano, P.
(2007) Vitamin D reserve is higher in women with endometriosis. Hum Reprod. 22:2273-
2278.

Abstract (if available)

Abstract Endometriosis is a common female reproductive disease. This disease is caused by invasion of endometrial cells into other organs. Pain at the site of ectopic endometrium and infertility are two common symptoms of endometriosis, and around 30% of infertility cases are associated with endometriosis. Vitamin D3 is a type of secosteroid, and there are plenty of studies reported that vitamin D3 treatment helps to improve fertility in both animals and human patients, as well as enhance the success rate of in vitro fertilization. We are interested in how vitamin D3 impacts gene expression in endometriosis. 1,25(OH)₂D3, the active metabolite of vitamin D3, was used to treat endometriotic epithelial and stromal cell lines. Differential gene expression for the treated group versus the control group was identified with RNA-Seq. We found 149 genes whose expression levels were altered by at least two fold in epithelial cells, but only 25 genes in stromal cells. Among the 149 differentially expressed genes of the epithelial cells, we managed to locate a group of genes that are important for cell motility and invasion, with gene PLAUR and CTTN being most likely to involve in cell movement directly. Nevertheless, more evidence is needed to reveal the roles of these genes in endometriosis. Compared with epithelial cells, endometriotic stromal cells seem to be much less sensitive to 1,25(OH)₂D3, according to our experiment. However, this could be a consequence of the pathology of endometriosis, we decided to follow-up by include more samples and different treatments in the future.

Linked assets

University of Southern California Dissertations and Theses

Conceptually similar

PDF

SLIT3 gene expression by 1,25(OH)₂D₃ in an endometriosis stromal cell line

PDF

Genes and environment in prostate cancer risk and prognosis

PDF

Gene expression and angiogenesis pathway across DNA methylation subtypes in colon adenocarcinoma

PDF

Dietary carcinogens and genetic variation in their metabolism: epidemiological studies on the risk of selected cancers

PDF

Dietary and supplementary folate intake and prostate cancer risk

PDF

Genomic risk factors associated with Ewing Sarcoma susceptibility

PDF

Risk factors of pelvic floor disorders in the multiethnic cohort study

PDF

Prostate cancer disparities among Californian Latinos by country of origin: clinical characteristics, incidence, treatment received and survival

PDF

The effect of inhibiting the NFkB pathway on myeloid derived suppressor cell cytokine production

PDF

Genetic epidemiological approaches in the study of risk factors for hematologic malignancies

PDF

Long-term effects on vertebral body height growth of dose sculpting intensity modulated radiation therapy for children with neuroblastoma

PDF

Understanding DNA methylation and nucleosome organization in cancer cells using single molecule sequencing

PDF

Carcinogenic exposures in racial/ethnic groups

PDF

Identification of molecular mechanism for cell-fate decision in liver; &, SARS-CoV replicon inhibitor high throughput drug screening

PDF

Chronic eye disease epidemiology in the multiethnic ophthalmology cohorts of California study

Asset Metadata

Creator Wu, Liang (author)

Core Title Identification of gene expression regulated by 1,25(OH)₂D3 in human endometriosis cell lines with next-generation sequencing

School Keck School of Medicine

Degree Master of Science

Degree Program Molecular Epidemiology

Publication Date 09/10/2014

Defense Date 07/18/2014

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

Tag 1,25(OH)₂D3,DEG,endometriosis,next-generation sequencing,OAI-PMH Harvest,vitamin D

Format application/pdf (imt)

Language English

Contributor Electronically uploaded by the author (provenance)

Advisor Ingles, Sue Ann (committee chair), Siegmund, Kimberly D. (committee member), Stern, Mariana C. (committee member)

Creator Email wlynchris@gmail.com,ynchris@hotmail.com

Permanent Link (DOI) https://doi.org/10.25549/usctheses-c3-473419

Unique identifier UC11287822

Identifier etd-WuLiang-2907.pdf (filename),usctheses-c3-473419 (legacy record id)

Legacy Identifier etd-WuLiang-2907.pdf

Dmrecord 473419

Document Type Thesis

Format application/pdf (imt)

Rights Wu, Liang

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