ENGRAILED-2 (EN2) EXPRESSION
IN OVARIAN CANCER
by
Sophie McGrath
Thesis submitted in accordance with the requirements of the
University of Surrey for the degree of Doctor of Philosophy
Faculty of Health & Medical Sciences
University of Surrey
May 2015
DECLARATION OF ORIGINALITY
This thesis and the work to which is refers are the results of my own efforts. Any ideas, data,
images or text resulting from the work of others (whether published or unpublished) are fully
identified as such within the work and attributed to their originator in the text, bibliography
or in footnotes. This thesis has not been submitted in whole or in part for any other academic
degree or professional qualification. I agree that the University has the right to submit my
work to the plagiarism detection service TurnitinUK for originality checks. Whether or not
drafts have been so-assessed, the University reserves the right to require an electronic version
of the final document (as submitted) for assessment as above.
Sophie McGrath
May 2015
I
ABSTRACT
Epithelial ovarian cancer (EOC) is a common malignancy, usually diagnosed at an advanced
stage, resulting in over 4,000 deaths each year in the UK. There is currently no National
Screening Programme and no clinically approved diagnostic biomarker. Engrailed-2 (EN2)
is a homeodomain-containing transcription factor, essential during embryological neural
development, which is dysregulated in several cancer types. We have studied the biology
and function of EN2, and have investigated the potential role as a biomarker in EOC tissue
and bodily fluids.
Elevated En2 mRNA expression levels and protein staining were observed in EOC tissue, but
levels were much lower in benign tumours, and negligible in normal ovary. Therefore EN2
could be utilised as a diagnostic biomarker, particularly in high-grade serous tumours
(HGSOC), however this would require an invasive biopsy specimen. Urine samples are more
easily obtainable and EN2 protein was positive in 86% of newly diagnosed HGSOC,
compared with female healthy controls, making it a potential non-invasive diagnostic
biomarker and possible screening tool.
In HGSOC patients who received neoadjuvant chemotherapy prior to surgery, elevated En2
mRNA levels and positive protein expression predicted a shorter progression-free survival
and resistance to platinum chemotherapy. Such information could be used as a prognostic
biomarker to guide post-operative treatment decisions.
In normal adult Purkinje neurons EN2 is located in the nucleus at 33kDa, however EN2 was
cytoplasmic and visualised at 43kDa or 50kDa in EOC, suggesting post-translational
modification from the native state.
En2 over-expression studies, including microarray
analysis and cisplatin-challenge, suggested a role in the development of platinum-resistance.
EN2 was also implicated in cell invasion and metastasis, via the process of epithelialmesenchymal transition (EMT).
EN2 appears to be actively involved in disease progression and treatment resistance in EOC,
and further research determining its role in EMT, and the TGFβ receptor and Wnt signalling
pathways is warranted. It also shows potential as a clinical biomarker, especially in urine
where the non-invasive nature of the test and the potential for detection prior to any clinical
signs or symptoms, make it a worthy candidate for ongoing research.
II
ACKNOWLEDGEMENTS
I would like to acknowledge and express my gratitude to all those who have helped me to
complete this thesis. Firstly, I would like to thank my supervisors, Dr Agnieszka Michael
and Dr Nicola Annels, who have tirelessly taught, advised and encouraged me throughout
this work. They have always remained enthusiastic and supportive despite set-backs in
certain projects, and they have inspired me to seek a future career in Academic Medicine.
I am very grateful to the Gynae-oncology Multidisciplinary Team at The Royal Surrey
County Hospital who helped with identification of patients, obtaining consent and sample
acquisition. In addition, I would also like to thank the members of the Oncology department,
within the Faculty of Health and Medical Sciences, for their scientific advice and expertise,
support and friendship.
Finally I must thank my family, in particular my husband, my two daughters and my parents.
They have been a constant source of love and support, especially when I have needed to run
experiments late into the evenings or on weekends. Pieter - thank you for always listening
and for sharing my enthusiasm over the past 4 years. I am “So Damn Lucky”. Amélie and
Olivia - you are loved more than one can describe and I hope that in the future, the hard work
and dedication that I have instilled in this work will inspire you to thrive, work passionately,
and fulfil your dreams.
III
TABLE OF CONTENTS
I
II
III
IV
IX
XV11
XX
Declaration of originality
Abstract
Acknowledgements
Table of contents
Abbreviations
Figures
Tables
CHAPTER 1
1
1.
INTRODUCTION
2
1.1.
Ovarian cancer
2
1.1.1. Demographics
2
1.1.2. Risk factors
3
1.1.3. Pathogenesis
5
1.1.4. Diagnosis
5
1.1.5. Histological classification
6
1.1.6. Clinical staging
9
1.2.
1.1.7. Treatment guidelines
10
1.1.8. Platinum resistance
11
Cancer biomarkers
14
1.2.1. Current biomarkers in ovarian cancer
17
1.2.1.1. Gene mutation and amplification
17
1.2.1.2. DNA methylation
18
1.2.1.3. Protein biomarkers in blood
18
1.2.1.4. Protein biomarkers in urine
22
1.2.1.5. Protein biomarkers in ascites
22
1.2.1.6. Autoantibodies
23
1.3.
The homeobox gene superfamily
24
1.4.
The biology of Engrailed
26
1.5.
Engrailed in cancer
29
1.6.
Engrailed as a potential biomarker
32
1.7.
Summary
34
IV
CHAPTER 2
35
2.
MATERIALS AND METHODS
36
2.1.
Patients and controls
36
2.2.
Cell culture media
38
2.3.
Human cell lines
39
2.4.
RNA extraction from cell lines
40
2.5.
RNA extraction from human tumour tissues
40
2.6.
Complementary DNA (cDNA) synthesis
41
2.7.
Genomic DNA (gDNA) extraction from cell lines
41
2.8.
Genomic DNA (gDNA) extraction from human tumour tissues
41
2.9.
McrBC endonuclease cleavage of methylcytosine-containing DNA (methylated DNA)
42
2.10. Semi-quantitative reverse transcriptase polymerase chain reaction (rt-PCR)
43
2.10.1. mRNA expression
44
2.10.2. En2 methylation status
45
2.11. Enzymatic immunofluorescent staining on cell lines
46
2.12. Enzymatic immunohistochemistry on patient slides and tissue arrays
47
2.13. Enzyme-linked immunosorbent assay (ELISA) for protein identification in cell line lysates 49
2.14. Enzyme-linked immunosorbent assay (ELISA) for autoantibody identification in
51
patient plasma or serum
2.15. Western blotting analysis for protein identification in cell line lysates and supernatants
51
2.16. Over-expression of En2 in the PEA1 cell line
53
2.17. Enzyme deglycosylation of proteins
55
2.18. siRNA-mediated En2 silencing in cell lines
56
2.18.1. KDalert™ GAPDH Assay to detect silencing of GAPDH (glyceraldehyde-3-
56
phosphate dehydrogenase)
2.18.2. siRNA-mediated En2 silencing in the PEA2 cell line
57
2.19. 50% tumour inhibitory concentration (IC50) estimation of cisplatin in cell lines
58
2.20. Microarray-based gene expression analysis of En2 over-expressed cell lines
58
2.21. Statistical analysis
61
2.21.1. Analysis of En2 mRNA expression differences in cell lines and human tissue
61
2.21.2. Analysis of EN2 protein expression in human tumours and tissue arrays
61
2.12.3. Analysis of EN2 protein expression in patient urine
61
2.12.4. Analysis of EN2 and NY-ESO-1 antibody levels in patient plasma
62
V
2.12.5. Analysis of En2 promoter methylation in cell lines and human tumours
62
2.12.6. Analysis of microarray data
62
CHAPTER 3
63
3.
EN2 EXPRESSION IN EPITHELIAL OVARIAN CANCER CELL LINES
64
3.1.
Introduction
64
Study objectives and hypothesis
66
Results
67
3.2.1. En2 gene expression in epithelial ovarian cancer (EOC) cell lines
67
3.2.2. EN2 protein expression in epithelial ovarian cancer (EOC) cell lines
69
3.2.
3.2.2.1. Immunohistochemical staining of EOC cell lines for the expression of
69
EN2 protein
3.2.2.2. Enzyme-linked immunosorbent assay (ELISA) analysis of EOC cell lines
74
for the expression of EN2 protein
3.2.2.3. Western blotting analysis of EOC cell lines for the expression of EN2
75
protein
3.3.
Discussion
77
3.4.
Conclusion
84
CHAPTER 4
85
4.
EN2 EXPRESSION IN HUMAN EPITHELIAL OVARIAN CANCER TISSUE
AND ASSOCIATED BODY FLUIDS
86
4.1.
Introduction
86
Study objectives and hypothesis
91
Results
92
4.2.1. EN2 expression in human ovarian tissue
92
4.2.
4.2.1.1. En2 gene expression in human ovarian tissue and correlation with
92
clinico-pathological characteristics
4.2.1.2. EN2 protein expression in human ovarian tissue and correlation with
clinico-pathological characteristics
VI
104
4.2.1.3. EN2 protein expression in human ovarian tissue arrays
115
4.2.2. EN2 expression in urine from ovarian cancer patients
125
4.2.3. EN2 expression in ascites from ovarian cancer patients
127
4.2.4. EN2 autoantibody levels in the plasma from ovarian cancer patients
129
4.3.
Discussion
131
4.4.
Conclusion
138
CHAPTER 5
139
5.
DNA METHYLATION OF THE EN2 PROMOTER REGION
140
5.1.
Introduction
140
Study objectives and hypothesis
144
Results
145
5.2.1. DNA methylation of the En2 promoter region in cell lines
145
5.2.2. DNA methylation of the En2 promoter region in ovarian tumours
148
5.3.
Discussion
153
5.4.
Conclusion
156
5.2.
CHAPTER 6
157
6.
THE ROLE AND FUNCTION OF EN2 IN ONCOGENESIS
158
6.1.
Introduction
158
Study objectives and hypothesis
163
Results
164
6.2.1. Secretion of EN2 protein from EOC cell lines
164
6.2.2. En2 over-expression in the PEA1 EOC cell line
166
6.2.3. Post-translational modification of EN2
170
6.2.4. siRNA-mediated En2 silencing in the PEA2 EOC cell line
172
6.2.5. The effect of En2 over-expression on platinum-resistance in EOC cell lines
175
6.2.6. Microarray analysis of the effect of En2 over-expression on oncogenic pathways
177
6.3.
Discussion
188
6.4.
Conclusion
196
6.2.
VII
CHAPTER 7
197
7.
DISCUSSION
198
7.1.
EN2 as a diagnostic biomarker in EOC
199
7.2.
EN2 as a prognostic biomarker in EOC
200
7.3.
EN2 as a treatment response biomarker in EOC
200
7.4.
The biology and function of EN2
201
7.5.
Future work
202
REFERENCES
205
PUBLICATIONS
227
VIII
ABBREVIATIONS
α-EN2
Anti-EN2 antibody
Ab
Antibody
AAb
Autoantibody
ABL
Abelson murine Leukemia viral oncogene homolog
ACC
Adenoid Cystic Carcinoma
ActRIIA/B
Activin receptor type II A/B gene
ADP
Adenosine Diphosphate
AFP
Alpha-fetoprotein
Ag
Antigen
AKT
v-akt murine Thymoma viral oncogene homolog
ALK
Anaplastic Lymphoma Kinase gene
AMV
Avian Myeloblastosis Virus
ANOVA
Analysis of Variance
APC
Adenomatous Polyposis Coli protein
APS1
A monoclonal mouse anti-EN2 antibody
ARID1A
AT-rich Interactive Domain-containing protein 1A gene
AUC
Area under the Curve
AVP
Arginine Vasopressin protein
βHCG
Human Chorionic Gonadotropin beta
BCA
Bicinchoninic Acid
BCR
Breakpoint Cluster Region gene
bp
Base pair
BRAF
V-Raf murine Sarcoma viral oncogene homolog B
BRCA1/2
Breast Cancer susceptibility gene1/2
BRN-3A(l)
Brain-specific homeobox protein 3A(l)
BSA
Bovine Serum Albumin
CA125
Carcinoma Antigen 125
IX
CA15-3
Carcinoma Antigen 15-3
CA 19-9
Carcinoma Antigen 19-9
Cas9
Clustered Regularly Interspersed Palindromic Repeats (CRISPR)
Associated protein 9
CD
Cluster of Differentiation
CDKN2A
Cyclin-dependent kinase inhibitor 2A gene
cDNA
Complementary Deoxyribonucleic Acid
CEA
Carcinoembryonic Antigen
CF
Cystic Fibrosis
CFTR
Cystic Fibrosis Transmembrane Conductance Regulator protein
CI
Confidence Interval
CIMP
CpG Island Methylator Phenotype
CK7
Cytokeratin 7 protein
CKIα
Cyclin-dependent Kinase Inhibitor protein alpha
CO2
Carbon Dioxide
CpG
Cytosine-phosphate-Guanine
CRHR1
Corticotropin-releasing Hormone Receptor 1
cRNA
Complementary Ribonucleic Acid
CSF
Cerebrospinal Fluid
CSH1
Chorionic Somatomammotropin Hormone 1 gene
CT
Cycle threshold
CTP
Cytidine Triphosphate
Cy3
Cyanine 3
CYP3A4/5
Cytochrome P450, family 3, subfamily A, 4/5
DAB
3,3'-diaminobenzidine
DAP
Death-associated protein
DEPC
Diethylpyrocarbonate
dg
Deglycosylated
DNA
Deoxyribonucleic Acid
dNTP
Deoxynucleotide Triphosphate
X
dT
Deoxy-thymine nucleotides
DTT
Dithiothreitol
EDTA
Ethylenediaminetetraacetic Acid
EGFR
Epidermal Growth Factor Receptor
EH
Engrailed Homology region
eIF4E
Eukaryotic Translation Initiation Factor 4E
ELISA
Enzyme-Linked Immunosorbent Assay
EMT
Epithelial-Mesenchymal Transition
Emx/EMX
Empty Spiracles Homeobox gene
En1/2
Engrailed 1/2 gene
EN1/2
Engrailed 1/2 protein
[EN2]
Engrailed 2 protein concentration
EOC
Epithelial Ovarian Cancer
Ep-CAM
Epithelial Cell Adhesion Molecule
ER
Oestrogen receptor
ERCC1
Excision Repair Cross-Complementing-1 gene
ERBB2
v-erb-b2 avian Erythroblastic Leukemia viral oncogene homolog 2
FANCF
Fanconi Anemia Group F gene
FCS
Fetal Calf Serum
FeLC
Friend murine erythroleukemia cells
FIGO
International Federation of Gynecology and Obstetrics
G418
Geneticin disulfate salt antibiotic
GADD45β
Growth Arrest and DNA-damage-inducible, beta gene
GAPDH
Glyceraldehyde 3-phosphate Dehydrogenase gene
GCT
Germ Cell Tumour
GDF
Growth differentiation factor
gDNA
Genomic Deoxyribonucleic Acid
GSK3β
Glycogen Synthase Kinase 3 beta protein
GSTP1
Glutathione S-transferase P1 gene
XI
GTP
Guanosine-5'-triphosphate
H2SO4
Sulphuric Acid
H&E
Haematoxylin and Eosin
HCC
Hepatocellular Carcinoma
HDAC4
Histone Deacetylase 4 protein
HE4
Human Epididymis protein 4
HER2/neu
Human Epidermal Growth Factor Receptor 2
HGSOC
High Grade Serous Ovarian Cancer
HNPCC
Hereditary Non-polyposis Colorectal Cancer
HOX
Homeobox gene
hPrEC
Human Prostate Epithelial cells
HRT
Hormone Replacement Therapy
HSP-90
Heat Shock Protein 90
IC50
50% Tumour Inhibitory Concentration
IFN-γ
Interferon-gamma
IgG
Immunoglobulin G
IgM
Immunoglobulin M
IHC
Immunohistochemistry
IL1A
Interleukin-1 A gene
IL1RAP
Interleukin-1 Receptor Accessory Protein gene
KLK
Kallikrein
KRAS
Kirsten rat Sarcoma viral oncogene homolog
LDH
Lactate Dehydrogenase
LRP
Low Density Lipoprotein Receptor-related Protein
MAPK
Mitogen-activated Protein Kinase
McrBC
Methylation-specific endonuclease
M-CSF
Macrophage Colony Stimulating Factor
mib-1
Mindbomb E3 ubiquitin protein ligase 1
MIF
Macrophage Migration Inhibitory Factor
XII
MMMT
Malignant Mixed Mesodermal/Müllerian Tumours
MMP
Matrix Metalloproteinase
MMR
Mismatch Repair
MMS
Multimodal Screening
MOPS
3-(N-morpholino)propansulfonic acid
mRNA
Messenger Ribonucleic Acid
MSI-H
Microsatellite Instability-High
MSR
Methylation Sensitive Enzyme Restriction
MSX
MSH homeobox gene
MTS
(3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4sulfophenyl)-2H-tetrazolium
MUC
Mucin
Na2CO3
Sodium Carbonate
NaHCO3
Sodium Bicarbonate
NER
Nucleotide-excision Repair
Ni-NTA
Nickel-nitrilotriacetic Acid
NR4A1/ NUR77
Nuclear Receptor Subfamily 4 group A member 1 gene
NS
Non-significant
NY-ESO-1
New York Esophageal Squamous Cell Carcinoma 1
OCP
Oral Contraceptive Pill
OD
Optical Density
OS
Overall Survival
OSE
Ovarian Surface Epithelium
Otx
Orthodenticle homeobox gene
OVX1
Ovarian Cancer Antigen X1
p16
Tumour Protein 16
p53
Tumour Protein 53
PAI1
Plasminogen Activator Inhibitor 1 gene
PALB2
Partner and Localizer of BRCA2 gene
PARP
Poly (Adenosine-diphosphate Ribose) Polymerase
XIII
PAX
Paired gene
PBS
Phosphate Buffered Saline
PBST
Phosphate Buffered Saline/0.05% Tween-20
PBX
Pre-B-cell Leukemia Transcription Factor
PCR
Polymerase Chain Reaction
PFS
Progression-free Survival
PIK3
Phosphatidylinositide 3-kinase gene
PLCB1
Phospholipase-C beta 1 gene
PNGase F
Peptide N-glycosidase F
pNPP
Para-nitrophenylphosphate
PR
Progesterone receptor
PRKAR2β
cAMP-dependent Protein Kinase A Regulated type II beta gene
PSA
Prostate Specific Antigen
pT
Primary Tumour
PTEN
Phosphatase and Tensin gene homolog
r2
Coefficient of determination
RARβ2
Retinoic Acid Receptor-β2 gene
RASSF1A
Ras Association Domain-containing Protein 1 gene
REC
Research Ethics Committee
rEN2
Recombinant Engrailed 2 protein
RIPA
Radioimmunoprecipitation Assay
RMI
Risk of Malignancy Index
RNA
Ribonucleic Acid
RNAi
Ribonucleic Acid Interference
ROC
Receiver Operator Characteristic
ROX
Reference dye
RR
Relative Risk
RSCH
The Royal Surrey County Hospital
rt-PCR
Reverse Transcriptase Polymerase Chain Reaction
XIV
SCTR
Secretin Receptor gene
SD
Standard Deviation
SDS
Sodium Dodecyl Sulphate
SDS-PAGE
Sodium Dodecyl Sulphate-polyacrylamide Gel Electrophoresis
SE
Standard Error
SEER
Surveillance, Epidemiology and End Results
SEPT9
Septin 9 gene
SERPINE1
Serine Protease Inhibitor 1 gene
SHOX2
Short Stature Homeobox 2 gene
siRNA
Small/short Interfering Ribonucleic Acid
Ski
v-Ski avian Sarcoma viral oncogene homolog
SMAD3/4
Mothers Against Decapentaplegic homolog 3/4 gene
SNAIL
Snail Family Zinc Finger genes
SNPs
Single-nucleotide Polymorphisms
STAT1
Signal Transducers and Activators of Transcription 1 protein
Tcf-4
Transcription Factor 4 gene
TGFβ
Transforming Growth Factor beta
TMB
Tetramethylbenzidine
TNC
Tenascin C gene
TNFAIP8
Tumour Necrosis Factor α–induced Protein 8
TP53
Tumour Protein 53 gene
TWIST2
Twist-related Protein 2 gene
UGT1A1
UDP Glucuronosyltransferase 1 family, polypeptide A1
USS
Ultrasound Scan
VEGF-A
Vascular Endothelial Growth Factor A
VIM
Vimentin gene
VUC
Voided Urine Cytology
WB
Western Blot
WGA
Wheat Germ Agglutinin
XV
Wnt
Wingless-type MMTV Integration Site family
w-t
Wild-type
WT1
Wilms Tumour 1 protein
x2
Chi-squared distribution
xg
Relative centrifugal force
YOPD
Young-onset Parkinson’s Disease
ZEB2
Zinc Finger E-box Binding Homeobox 2 gene
5-mC
5-Methylcytosine
∆CT
Change in cycle threshold
XVI
FIGURES
Figure 1.1.
Functional domains within the EN protein.
26
Figure 2.1.
Ovarian carcinoma tissue array panel display used for
immunohistochemical analysis of EN2 expression (OV2082-Biomax, USA).
37
Figure 2.2.
Female reproductive system tissue array panel display used for immunohistochemical analysis of EN2 expression (FRS801-Biomax, USA).
38
Figure 2.3.
Workflow for McrBC enzyme cleavage of methylated gDNA and
evaluation of En2 promoter methylation status.
43
Figure 2.4.
Workflow for RNA sample preparation and microarray processing.
59
Figure 3.1.
En2 mRNA expression in EOC cell lines and controls.
68
Figure 3.2.
Immunofluorescent staining of serous epithelial ovarian cancer cell lines
for the expression of EN2.
70
Figure 3.3.
Immunofluorescent staining of the platinum-sensitive/-resistant paired
serous epithelial ovarian cancer cell lines for the expression of EN2.
71
Figure 3.4.
Immunofluorescent staining of endometrioid and clear cell epithelial
ovarian cancer cell lines, and negative & positive control cell lines for the
expression of EN2.
72
Figure 3.5.
EN2 is not clearly demonstrated in the cell nucleus.
73
Figure 3.6.
EN2 may be present in the EOC cell membrane.
73
Figure 3.7.
EN2 protein expression in EOC cell lines and controls, as determined by
the Direct ELISA and Western Blotting.
75
Figure 3.8.
A timeline demonstrating the acquisition of paired patient tumour samples
used to develop platinum-sensitive and –resistant cell lines.
79
Figure 4.1.
En2 mRNA expression in human ovarian tumours, classified by
histological sub-type.
96
Figure 4.2.
A high degree of variability in En2 mRNA expression exists within the
histological sub-types of EOC.
97
Figure 4.3.
En2 mRNA expression in human high grade serous tumours.
98
Figure 4.4.
En2 mRNA expression is lower in human high grade serous tumours from
Stage 4 versus Stage 3 patients.
100
Figure 4.5.
En2 mRNA expression is lower in human high grade serous tumours
exposed to neoadjuvant chemotherapy.
100
Figure 4.6.
En2 mRNA expression is higher in human high grade serous tumours that
become platinum-resistant rather than platinum-sensitive.
101
XVII
Figure 4.7.
Overall survival analysis of the human ovarian tumours, classified by
histological sub-type.
101
Figure 4.8.
Overall survival analysis of the high grade serous human ovarian tumours,
comparing En2 mRNA expression.
102
Figure 4.9.
Progression-free survival analysis of the high grade serous human ovarian
tumours, comparing En2 mRNA expression.
102
Figure 4.10. Examples of EN2 protein expression in human ovarian tissue from Cohort
1, using enzymatic immunohistochemistry.
105
Figure 4.11. Examples of EN2 protein expression in the different histological sub-types
of human ovarian tissue from Cohort 1, using enzymatic immunohistochemistry.
106
Figure 4.12. Examples of EN2 protein expression in human ovarian tissue from Cohort
2, using enzymatic immunohistochemistry.
108
Figure 4.13. Progression-free survival analysis of the combined high-grade serous
human ovarian tumours, comparing EN2 protein expression.
114
Figure 4.14. Overall survival analysis of the combined high-grade serous human
ovarian tumours, comparing EN2 protein expression.
114
Figure 4.15. Examples of EN2 protein expression from the OV2082 ovarian tissue
array, using enzymatic immunohistochemistry.
115
Figure 4.16. Examples of EN2 protein expression from the CJ2 ovarian tissue array,
using enzymatic immunohistochemistry.
120
Figure 4.17. Overall survival analysis of the serous tumours from the CJ2 ovarian tissue
array.
123
Figure 4.18. Examples of EN2 protein expression from the FRS801 female
reproductive tissue array, using enzymatic immunohistochemistry.
124
Figure 4.19. EN2 protein concentration in urine samples from patients with EOC
compared with healthy controls, as determined by the Direct ELISA.
126
Figure 4.20. EN2 urinary protein concentration is significantly elevated in EOC
compared with healthy controls, as determined by the Direct ELISA.
126
Figure 4.21. EN2 protein concentration in ascites samples from patients with EOC
compared with breast cancer ascites and benign disease controls, as
determined by the Direct ELISA.
128
Figure 4.22. There is no significant EN2 IgG response in EOC patients compared with
healthy controls, as determined by ELISA.
130
Figure 4.23. The EN2 IgG response in Breast and Prostate cancer patients compared
with healthy controls, as determined by ELISA.
130
Figure 5.1.
146
En2 promoter methylation status and mRNA expression in EOC cell lines
and controls.
XVII
Figure 5.2.
En2 promoter methylation status does not correlate with En2 mRNA
expression in EOC cell lines.
147
Figure 5.3.
En2 promoter methylation status in individual human EOC tumours, and
normal tissue.
149
Figure 5.4.
Mean percentage En2 promoter methylation in human EOC tumours,
classified by histological sub-type.
150
Figure 5.5.
En2 promoter methylation status does not correlate with En2 mRNA
expression in EOC human tumours.
151
Figure 5.6.
Mean percentage En2 promoter methylation in human EOC tumours,
classified by platinum sensitivity status.
152
Figure 5.7.
Survival analyses of the human epithelial ovarian tumours, comparing
percentage En2 promoter methylation.
152
Figure 6.1.
The Wnt signalling pathway in the absence and presence of ligand
162
Figure 6.2.
Secreted EN2 protein expression in EOC cell lines and controls, as
determined by the Direct ELISA and Western Blotting.
165
Figure 6.3.
Images of the PEA1 EOC cell lines after transient forced over-expression
of En2.
166
Figure 6.4.
En2 mRNA expression in the PEA1 EOC cell line after transient forced
over-expression of En2.
167
Figure 6.5.
EN2 protein expression in the PEA1 EOC cell line after transient forced
over-expression of En2, as determined by Western Blotting.
167
Figure 6.6.
En2 mRNA expression in PEA1 EOC cell line clones after stable forced
over-expression of En2.
168
Figure 6.7.
EN2 protein expression in PEA1 EOC cell line clones after stable forced
over-expression of En2, as determined by Western Blotting.
169
Figure 6.8.
Enzymatic deglycosylation of EN2 protein in PEA1 EOC cell lines after
transient forced over-expression of En2, as determined by Western
Blotting.
171
Figure 6.9.
Images of the PEA2 EOC cell lines after siRNA-mediated En2 silencing.
172
Figure 6.10. En2 mRNA expression in the PEA2 EOC cell line after siRNA-mediated
En2 silencing.
173
Figure 6.11. EN2 protein expression in the PEA2 EOC cell line after siRNA-mediated
En2 silencing, as determined by Western Blotting.
173
Figure 6.12. A high level of En2 forced over-expression does promote resistance to
cisplatin in PEA1 EOC cancer cells.
176
Figure 6.13. Heatmap visualization of significantly regulated probe sets.
177
XVIII
Figure 6.14. Quantitative rt-PCR validation of microarray-identified up-regulated genes
in PEA1 En2 over-expressing clones.
185
Figure 6.15. Quantitative rt-PCR validation of microarray-identified down-regulated
genes in PEA1 En2 over-expressing clones.
186
Figure 6.16. Quantitative rt-PCR validation of genes involved in TGFβ and Activin A
signalling, from PEA1 En2 over-expressing clones.
187
Figure 6.17. The TGFβ receptor signalling pathway.
191
Figure 6.18. The Activin A signalling transduction pathway.
193
XIX
TABLES
Table 1.1.
The 5-year relative survival by stage for ovarian cancer.
2
Table 1.2.
Risk factors for ovarian cancer.
3
Table 1.3.
Risk of Malignancy Index (RMI).
6
Table 1.4.
Histological features of EOC sub-types.
8
Table 1.5.
FIGO staging for epithelial ovarian cancer.
10
Table 1.6.
Potential uses for cancer biomarkers.
15
Table 1.7.
Tumour markers in current use in the UK.
16
Table 1.8.
A summary of the evidence linking the En genes and proteins with cancer.
31
Table 2.1.
Cell culture media used.
38
Table 2.2.
Cell lines used, their growth media, tissue of origin and relevant drug
resistance.
39
Table 2.3.
Reaction volumes and concentrations for methylated DNA digestion using
McrBC endonuclease.
42
Table 2.4.
The forward and reverse primer sequences for the genes evaluated by rtPCR.
44
Table 2.5.
The forward and reverse primer sequences for the En2 promoter evaluated
using rt-PCR.
45
Table 2.6.
The assigned score (0-4) for the percentage of EN2 positive cells.
48
Table 3.1.
En2 gene expression in EOC cell lines and controls.
67
Table 4.1.
Full demographic data from the human ovarian tumours in RNA later.
93
Table 4.2.
Summary data from the human ovarian tumours stored in RNA later with
mean relative En2 mRNA expression.
96
Table 4.3.
EN2 protein expression in human ovarian tissue from Cohort 1.
107
Table 4.4.
Full demographic data from the human ovarian tumours in Cohort 2.
109
Table 4.5.
EN2 protein expression in human ovarian tissue from Cohort 2.
111
Table 4.6.
EN2 protein expression in the combined human ovarian tissue cohorts.
113
Table 4.7.
Full demographic data from the OV2082 ovarian tissue array, with
detailed EN2 scoring.
116
Table 4.8.
EN2 protein expression in the OV2082 ovarian tissue array.
119
XX
Table 4.9.
Full demographic data from the CJ2 ovarian tissue array, with detailed
EN2 scoring for the malignant and benign tumours of epithelial cell origin.
121
Table 4.10.
EN2 protein expression in the CJ2 ovarian tissue array.
122
Table 5.1.
Full demographic data from selected human ovarian tumours, normal
tissue and commercial ovary gDNA, with the resultant percentage of
methylated En2.
148
Table 6.1.
Microarray probes up-regulated in two PEA1 En2 over-expression clones
compared with wild-type PEA1 cells.
179
Table 6.2.
Pathway maps obtained from enrichment analysis of the up-regulated
genes in two PEA1 En2 over-expression clones compared with wild-type
PEA1 cells.
181
Table 6.3.
Process networks obtained from enrichment analysis of the up-regulated
genes in two PEA1 En2 over-expression clones compared with wild-type
PEA1 cells.
181
Table 6.4.
Microarray probes down-regulated in two PEA1 En2 over-expression
clones compared with wild-type PEA1 cells.
183
Table 6.5.
Pathway maps obtained from enrichment analysis of the down-regulated
genes in two PEA1 En2 over-expression clones compared with wild-type
PEA1 cells.
184
Table 6.6.
Process networks obtained from enrichment analysis of the downregulated genes in two PEA1 En2 over-expression clones compared with
wild-type PEA1 cells.
184
XXI
CHAPTER 1
INTRODUCTION
1
1.
INTRODUCTION
1.1.
OVARIAN CANCER
1.1.1. Demographics
Ovarian cancer is the fifth most common female cancer in the United Kingdom (UK), after
breast, bowel, lung and endometrial cancer. Around 6,500 women are diagnosed each year,
and over 4,000 die each year, making it the 4th most common cause of female cancer death
[1].
There is currently no UK National Screening Programme for early detection of ovarian
cancer, so most women are only diagnosed once they develop symptoms. However these
symptoms can be vague and non-specific, such as bloating, change in appetite, weight
change, altered bowel habit, intermittent pain or increased urinary frequency.
Such
symptoms have a broad differential diagnosis, and patients often undergo several
investigations over a period of time, before ovarian cancer is confirmed. Inevitably, this
frequently results in late presentation and advanced-stage, high-grade disease at diagnosis.
The primary tumour may also infiltrate the adjacent tissues and organs and metastasis to the
peritoneal cavity is a characteristic of advanced ovarian cancer.
60% of women have
advanced disease at diagnosis, and for these women, the 5 year survival rate varies from 18.6
to 46.7% [2, 3]. However women diagnosed with early disease, when the disease is confined
to the ovary, have an overall 5 year survival of over 90% [1, 4], as demonstrated in Table
1.1. Unfortunately the improvement in mortality over the last 30 years has been marginal.
Disease Stage
at Diagnosis
5yr Relative
Survival (%)
Stage I
Stage II
Stage III
Stage IV
All stages
90%
43%
19%
3%
39%
Table 1.1. The 5-year relative survival by stage for Ovarian Cancer. Statistics obtained
in women aged 15-99 years, within the Former Anglia Cancer Network, 2002-06 (adapted
from [1]).
2
Approximately 90% of ovarian cancer cases are epithelial ovarian cancers (EOC) and occur
in women over the age of 50, although they can also occur in younger, pre-menopausal
women. Traditionally such cancers were thought to derive from the surface epithelial cells of
the ovary. Both this and more recent theories are explained in detail below. Germ cell
cancers, originating from the ovum-producing germ cells of the ovary, account for around 5%
of ovarian cancer diagnoses, and are typically seen in younger women [1]. Stromal ovarian
cancers are rarer still, and develop from connective, stromal tissue. The treatments and
prognoses are very different for these histological sub-types, however the preponderance of
EOC in the clinic results in the overriding poor prognosis for advanced stage disease.
1.1.2. Risk factors
There are a number of factors that may influence the risk of developing ovarian cancer.
Nulliparity and infertility, hereditary genetic mutation, and hormone manipulation are the
most common. These factors are summarised in Table 1.2.
Relative
risk
Lifetime probability
percent[5]
Reference
1.0
1.4
[5]
BRCA1 gene mutation
35 to 46
[6]
BRCA2 gene mutation
13 to 23
[6]
Lynch syndrome (hereditary nonpolyposis colon cancer)
3 to 14
[7, 8]
General population
Infertility
2.67
[9]
Polycystic ovarian syndrome
2.52
[10]
Endometriosis (increased risk of clear cell, endometrioid,
or low grade serous carcinomas)
2.04 to 3.05
[11]
Cigarette smoking (increased risk of mucinous carcinoma)
2.1
[12]
Intrauterine device
1.76
[13]
Past use of oral contraceptives
0.73
[14]
Past breast feeding (for >12 months)
0.72
[15]
Tubal ligation
0.69
[16]
Previous pregnancy
0.6
[17]
Table 1.2. Risk factors for ovarian cancer (adapted from [18]).
3
In addition to infertility, early menarche, nulliparity, and late menopause are associated with
an increased risk of developing ovarian cancer [9, 19]. In addition, the use of Hormone
Replacement Therapy (HRT) for five or more years is associated with a 45% higher risk of
developing ovarian cancer compared with those who have never used HRT. However when
HRT is stopped, the risk is believed to return to normal over the course of a few years [20].
The Combined Oral Contraceptive Pill (OCP) has a protective effect, along with early age at
first pregnancy, prolonged lactation, multiple pregnancies, and older age of final pregnancy
[14, 15, 17]. In a meta-analysis of 45 case-control and prospective studies analysing the use
of oral contraceptives, it was shown that the relative risk reduction in developing ovarian
cancer increased further with the duration of OCP use, with those taking the medication for
10 years having a 60% reduction in ovarian cancer risk (RR 0.42) [14]. The risk is also
reduced in women who have undergone tubal ligation or hysterectomy, where oestrogen and
progesterone secretion are suppressed, and there is reduced exposure to exogenous irritants
(reviewed in [21]).
Mutations of the Breast Cancer Susceptibility Gene 1 (BRCA1) and Breast Cancer
Susceptibility Gene 2 (BRCA2) are associated with a lifetime risk of ovarian cancer of
between 13 and 46 %, and account for 5-13% of all ovarian cancers [6]. The typical age of
presentation is often lower than average, namely pre-menopausal. Such patients may also
have a personal history of breast cancer or a family history of breast and/or ovarian cancer.
Lynch Syndrome or Hereditary Non-Polyposis Colorectal Cancer (HNPCC) is another
hereditary genetic condition that results in an increased risk of developing ovarian cancer, as
well as uterine, colon and other gastrointestinal cancers [7, 8].
Some women opt for
prophylactic bilateral salpingo-oophorectomy if such genetic abnormalities are detected.
This has been shown to reduce the chance of developing ovarian, as well as breast cancer
[22].
Endometriosis, where endometrial tissue grows in sites other than the uterine lining, is
associated with an increased risk of developing ovarian cancer, particularly the clear cell,
endometrioid, or low grade serous carcinoma sub-types of EOC [23]. Lifestyle factors such
as smoking, obesity and a diet high in animal fat and low in fresh fruit and vegetables, have
also been linked to an increased risk of developing ovarian cancer [12, 24]. Other factors that
have been investigated include talc use, asbestos exposure, and other occupational chemicals,
but these have not been linked convincingly to ovarian or fallopian tube cancer [25-27].
4
1.1.3. Pathogenesis
Several theories have been proposed to explain ovarian cancer development involving the
multistep acquisition of epigenetic and genomic abnormalities. These may be caused by
previously mentioned environmental factors, such as talc or asbestos exposure, or prior
endometriosis, pelvic inflammatory disease or mumps infection [28]. Another theory, the
“incessant ovulation hypothesis”, suggests that women who are nulliparous undergo repeated
cycles of ovulation-induced trauma, inflammation, and repair of the ovarian surface
epithelium at the site of ovulation. This continual apoptosis and cellular repair may induce
genetic instability, predisposing the epithelial cells to carcinogenesis [29-31].
These mechanisms may play a role in the development of certain histological sub-types of
EOC, but the more recent consensus regarding high grade serous carcinoma is that its origin
lies in the fimbria of the fallopian tube. The occult intraepithelial carcinoma may develop in
the fimbria and directly spread to the ovary [32], or normal fimbrial epithelial cells may
implant on the ovarian surface post-ovulation, resulting in envelopment and subsequent
development of cortical inclusion cysts [reviewed in [33]]. Historically, only one or two
cross-sections of the fallopian tube would be sectioned during histological preparation of an
ovarian mass and, therefore, may not have included the fimbriated ends. This may explain
why this hypothesis for the pathogenesis of serous EOC has only recently come to light.
Nevertheless, it still does not explain the pathogenesis of all EOC histological sub-types.
1.1.4. Diagnosis
Along with a full clinical examination, the most common investigations used in the diagnosis
of ovarian cancer include pelvic and/or abdominal ultrasound, subsequent tissue biopsy or
fluid aspiration, and serological CA125 testing. CA125 is a protein that can be detected in
the serum by monoclonal antibody, and can aid diagnosis of ovarian cancer, although it is
only approved for monitoring treatment response and disease progression. This biomarker is
explained in more detail in Section 1.2.1.3.
Further imaging with Computed Tomograph
scanning of the chest, abdomen and pelvis will be required prior to any surgical or drugbased intervention.
The Risk of Malignancy Index (RMI) may be used as a tool to help the physician determine
the most appropriate management pathway for the patient [34]. The RMI is a product of the
5
score from three pre-surgical features: serum CA125 level (IU/ml), menopausal status and
ultrasound score (Table 1.3). The latter score is based on the presence of multilocular cysts,
solid areas, metastases, ascites and bilateral lesions. All women with an RMI score of 250 or
greater should be referred to a specialist multidisciplinary team.
Table 1.3. Risk of Malignancy Index (RMI) [18].
Since histological analysis is the mainstay of diagnosis, the patient may undergo an imageguided biopsy or laparoscopy to obtain tissue.
In certain circumstances a cytological
diagnosis may be satisfactory, especially if the patient is proceeding to drug-based treatment
prior to surgery (neoadjuvant treatment).
1.1.5. Histological classification
Epithelial ovarian cancer is sub-divided into four histological sub-types with different
microscopic appearances. They are thought to derive from varying pre-cursor cells, and
response to standard treatment strategies can vary. The serous sub-type is the most common,
resembling fallopian tube epithelium, and the cells frequently have round or oval nuclei with
irregular membranes, hyperchromasia, large nucleoli, and many mitotic figures [35]. Most
6
serous tumours are poorly differentiated. The other ovarian tumour histological sub-types
resemble
endometrial
tissue
(endometrioid),
mucin-secreting
(mucinous), and glycogen-filled vaginal rests (clear cell) [36].
endocervical
glands
The major histological
features of these sub-types, along with their common immunophenotype and genotype, are
summarised in Table 1.4, with images depicting the typical appearance.
It should be noted that some EOCs are thought to develop from, or at least co-exist with
borderline ovarian tumours.
These demonstrate many of the histological features of a
malignant ovarian tumour, but they do not invade the surrounding stromal tissue and are
therefore said to have low malignant potential. They grow slowly and are usually diagnosed
at an early stage, although they can form non-invasive implants on more distant sites of the
body [38].
The grade of the tumour will also be characterized on histological examination, as this may
influence future treatment decision and disease prognosis. Such borderline tumours are
usually characterized as Grade 0 i.e. non-invasive tumours. Grade 1/well-differentiated/lowgrade cells resemble normal cells, whereas Grade 3/poorly-differentiated/high-grade cells
bear very little resemblance to their tissue of origin.
More recently, a novel classification has been developed which combines the anatomical and
clinical features of ovarian cancer patients with their molecular abnormalities.
Type 1
tumours constitute 25% of epithelial ovarian carcinomas and comprise of low-grade serous,
low-grade endometrioid, mucinous and clear-cell carcinomas [35, 39].
They are slow-
growing tumours evolving from adenofibromas or borderline tumours in a slow step-wise
fashion involving gradual chromosomal instability [40], and often present as large cystic
unilateral masses. Approximately two thirds of these tumours have a mutually exclusive
mutation in one of the KRAS, BRAF or ERBB2 genes, in contrast to TP53 mutations which
are almost always associated with high grade serous carcinomas, known as Type 2 tumours
[41-43]. KRAS mutations are more common in mucinous tumours (50-68%) whereas BRAF
mutations are mostly reported in serous tumours (30-35%) [39]. These mutations are also
found in 60% of ovarian borderline tumours.
7
Histological
Sub-type
% of
EOC
Histological Features
Immunopheno- and
Geno-type
Image
64 Round/oval nuclei,
WT1 +
hyperchromasia, large nucleoli,
many mitotic figures
High Grade:
p53 overIrregular luminal contours, focal expression/mutation
slit-like spaces,
p16 expression
papillary & micropapillary
Loss of BRCA1 expression
architecture
Serous
Low Grade:
Low Grade: Serous Borderline BRAF/KRAS mutation
Tumour associated
Endometrioid
17 Endometrial-like, metaplasias,
secretory change, expansile
invasion
ER/PR +
Nuclear β-catenin +
WT1 β-catenin/PTEN/PIK3
mutation
Endometriosis/Endometrioid
Borderline Tumour/
Endometrioid Uterine
Carcinoma associated
Mucinous
Microsatellite instabilityhigh (MSI-H)
13 Intracytoplasmic mucin,
expansile invasion
CK7 >20
Retained SMAD4
expression
Intestinal Mucinous Borderline Nuclear β-catenin Tumour associated
ER p16 Mesothelin Fascin Racemase KRAS mutation
Clear Cell
6
ER/PR –
WT1 –
p53 –
mib-1 -
Papillary, tubulocystic, solid,
hobnail, frequently clear
cytoplasm; Glycogen-filled
vaginal rests
MSI-H/PTEN mutation
Endometriosis/Clear Cell
Borderline Tumour associated
Table 1.4. Histological features of EOC sub-types (adapted from [37]).
8
In addition, ß-catenin, PTEN and PIK3 mutations, and high microsatellite instability (MSI-H)
profiles have been identified in Type 1 tumours. ß-catenin mutations are present in up to
60%, and PTEN mutations in 20-31% of low-grade endometrioid carcinomas [44, 45]. PIK3
mutations are also mutually exclusive from KRAS and BRAF mutations, and occur in 33% of
clear-cell and in endometrioid ovarian carcinomas. MSI-H accounts for 12% of epithelial
ovarian cancer, particularly Type 1 non-serous tumours [39].
Type 2 tumours (75% of epithelial ovarian carcinomas) include high-grade serous, highgrade endometrioid, undifferentiated carcinomas, and carcinosarcomas, also known as
Malignant Mixed Mesodermal Tumours (MMMTs). These all display aggressive growth
rates and typically present at an advanced stage. TP53, and BRCA1 or BRCA2 mutations are
frequently present in these sporadic cancers [39, 46], with TP53 mutations occurring in 6080% of epithelial ovarian cancers [47]. BRCA1 or BRCA2 mutations or epigenetic losses
occur in 5-15% of ovarian cancers, predominantly of serous histology [39, 46]. ERBB2
amplification and over-expression are also present in Type 2 ovarian tumours but are
uncommon (occurring in 7% and 11%, respectively) [46].
In certain cases, it is difficult to macro- or microscopically determine the definitive origin of
the tumour specimen, as the histological features of serous tubal, ovarian and ‘primary
peritoneal’ epithelial carcinomas may be identical.
In fact there are few documented
differences between these tumours in terms of patient demographics, histology, and survival
[48, 49], so they are increasingly believed to have a common origin and pathogenesis, and are
treated in the same manner. They are often referred to as Pelvic Serous Carcinoma [33].
1.1.6. Clinical staging
There are currently two internationally recognised disease staging systems in routine clinical
use, although the International Federation of Gynecology and Obstetrics (FIGO) Staging
System is most commonly used in the UK. The features of this are detailed in Table 1.5. The
disease staging helps to direct further management, including the amount of surgical input
required, the need for chemotherapeutic drugs and the particular drug regime. Disease
prognosis also depends on accurate clinical staging, as the 5 year survival may be greater
than 90% in Stage I cases, but fall to below 20% in Stage IV [4].
9
In the case of a patient with an abdominal mass and pleural effusion, disease must be
confirmed cytologically in the pleural fluid, before it can be classed as a Stage IV disease.
Table 1.5. FIGO staging for epithelial ovarian cancer [50].
1.1.7. Treatment guidelines
The standard treatment for EOC involves surgery, the extent of which depends upon the
cancer stage, with or without the need for chemotherapy. The mainstay of surgery includes
hysterectomy, bilateral salpingo-oophorectomy, and omentectomy, along with visualisation
and biopsies from the diaphragm surface within the abdominal cavity. Random pelvic and
abdominal peritoneal biopsies are usually performed, as well as pelvic and para-aortic lymph
node sampling and peritoneal washings. In Stage Ia or Ib disease where the tumour is wellor moderately differentiated, surgery alone may be adequate treatment [51]. In patients with
a high-grade, or densely adherent, or stage Ic tumour, their risk of disease relapse may be as
high as 30%, so adjuvant chemotherapy with Carboplatin, is usually recommended [52-54].
Stage II patients will usually be advised to have adjuvant combination chemotherapy i.e
Carboplatin and Paclitaxel.
Likewise for Stage III and IV patients, the recommended treatment includes surgery and
chemotherapy although some chemotherapy may be given prior to the surgical debulking
10
procedure i.e. neoadjuvant chemotherapy. The surgeon will aim to debulk as much of the
gross tumour as possible as well as removing nodular disease from the serosal surfaces and
removing pelvic and para-aortic nodal groups. Traditionally the aim has been to debulk the
disease to <1cm of visible disease, although most surgeons will now try to achieve <0.5cm of
residual disease, as cytoreduction is an independent prognostic variable for survival [55, 56].
An intravenous, platinum-based chemotherapy combination is advised for a total of 6 cycles
(18 weeks) which may be administered after the surgery or split before and afterwards. The
response rate from these drugs regularly exceeds 60%, with a median time-to-progression
greater than 1 year, even in the sub-optimally debulked group [57].
Although original combination agent trials took place with Cisplatin, the better tolerated
drug, Carboplatin, has gained favour more recently, and several studies have proven noninferiority for Carboplatin-based versus Cisplatin-based regimens [58-61]. In recent years
there has also been the addition of Bevacizumab (Avastin) targeted therapy for those with
sub-optimally cytoreduced surgery or stage IV disease i.e. those at high risk of disease
progression.
Bevacizumab is a recombinant monoclonal antibody which inhibits
angiogenesis via its action on Vascular Endothelial Growth Factor A (VEGF-A). In initial
Phase III trials, when Bevacizumab was added to initial chemotherapy and continued as a
maintenance treatment for a total of a year, there was only a modest improvement in
progression-free survival (PFS) in all-comers [62, 63]. However when analysing data from
the high risk disease sub-groups, an improvement in PFS of 3.6 months and an Overall
Survival (OS) benefit of 7.8 months was achieved [63].
Traditionally all EOC histological sub-types have been managed using the same treatment
guidelines, although chemotherapeutic drugs are typically more effective against high-grade
tumours, with a high proliferative index, rather than low-grade, more slowly growing,
tumours.
1.1.8. Platinum resistance
Despite the modest improvements in median survival time seen with platinum-based
combination chemotherapy and cytoreductive surgery, with or without the addition of
Bevacizumab, the majority of patients still develop recurrence and eventually die of the
11
disease. Response to treatment in chemo-sensitive cancer is around 80%, but the prognosis is
poor in those developing resistance to chemotherapy at an early stage [64].
The mechanism of action of the platinum drugs involves covalent binding to purine bases on
the DNA, which can interfere with DNA repair mechanisms resulting in DNA damage and
eventual cellular apoptosis, unless the cell can repair the DNA [65-67]. The development of
resistance to these drugs may result from decreased membrane transport of the drug,
increased cytoplasmic drug detoxification or inactivation, increased DNA repair, or increased
tolerance to DNA damage. It is not yet known whether the chemotherapy resistant cells
derive from cancer stem cells within the tumour which are inherently chemotherapy resistant,
or whether resistance emerges as the cancer cells interact with the tumour microenvironment
[67]. Some EOC patients are deemed to have intrinsic resistance to platinum i.e. “Platinum
Refractory” disease, and so develop disease progression before completing their
chemotherapy. However the majority of patients acquire resistance during cycles of therapy
with Cisplatin and so their disease progression becomes evident within 6 months of
chemotherapy completion i.e. “Platinum Resistant” [65]. The latter patients may have
tumours with a heterogeneous population of intrinsically platinum-resistant, as well as
sensitive cells.
There are five major DNA-repair pathways, but nucleotide-excision repair (NER) is believed
to be the main pathway involved in repair of DNA due to Cisplatin damage. The excision
repair cross-complementing-1 protein (ERCC1) demonstrates higher expression in Cisplatinresistant ovarian cancer cells. Increased ERCC1 mRNA levels have been identified in EOC
patients with clinical resistance to platinum-based chemotherapy [68, 69]. More recently,
ERCC1 positive circulating tumour cells have been identified in the blood of EOC patients so
could act as an independent predictor of platinum resistance [70]. Loss of function of the
mismatch repair (MMR) pathway may also influence platinum-induced DNA damage,
particularly influencing development of acquired drug resistance [71].
Understanding the mechanisms behind the development of such drug resistance, may enable
the discovery of methods to overcome resistance within the tumour.
Along with the
development of new, improved platinum drugs, and improvement of drug delivery, it is
hoped that co-administration of platinum drugs with either pharmacological modulators of
resistance mechanisms or new molecularly targeted drugs will help to overcome drug
resistance.
Bevacizumab, the monoclonal antibody to VEGF-A, has shown promising
12
responses in platinum-resistant disease when combined with chemotherapy in the AURELIA
study, where there was an increase of progression free survival (PFS) from 3.4 to 6.7 months
[72]. There have also been encouraging results when combined with oral cyclophosphamide
or weekly paclitaxel [73-75]. Also in platinum-resistant disease, the TRINOVA 1 study
combining paclitaxel chemotherapy with an angiopoietin 1/2 inhibitor, trebananib, improved
PFS by 1.8 months, and in PRECEDENT (EC145, a folate targeted vinca alkaloid added to
liposomal doxorubicin) a 2.3 month PFS improvement was seen [76, 77].
Glutathione S-transferase P1 (GSTP1), which metabolises platinum drugs, has been
identified to varying degrees in ovarian tumours, and it is believed that such inter-tumour
differences in GSTP1 expression may influence response to platinum-based chemotherapy
[78]. TER286/TLK286 is a pro-drug that is preferentially activated by GSTP1 to release a
nitrogen mustard alkylating agent, so could exploit the increased levels of GSTP1 in
platinum-resistant cancers.
We await the results of several Phase III trials evaluating
TLK286 in combination with chemotherapeutic agents.
Patients harbouring germ-line mutations of BRCA1 or BRCA2 have a more favourable
clinical course including an increased sensitivity to platinum-based chemotherapy [79, 80].
Similar findings have been noted in tumours sharing phenotypic characteristics with tumours
associated with germ-line BRCA mutations i.e. the “BRCAness” phenotype, which may be
related to defective homologous recombination. In fact a BRCAness gene expression profile
based on samples of BRCA1/2-mutated tumours was found to predict responsiveness to
platinum and poly(ADP-ribose) polymerase inhibitors [81]. Other biomarkers have been
identified that could be used as independent predictors of platinum sensitivity or resistance,
or help to identify resistance at an earlier stage in disease. There is the case of ERCC1positive circulating tumour cells in the blood as previously mentioned, but also markers
within tumour tissue such as tumour necrosis factor α–induced protein 8 (TNFAIP8) overexpression, which is associated with platinum resistance [82].
At present we still do not know how best to select platinum-resistant or platinum-refractory
patients for treatment, or which combinations of drugs to administer.
treatment response rates and overall survival rates remain low.
13
As a result, the
1.2.
CANCER BIOMARKERS
Early detection of cancer and use of biomarkers of disease in clinical practice remains a
challenge. It is widely accepted that when cancer is discovered early, at a lower stage or
grade, the patient is likely to live longer, potentially require less extensive treatment, and
tolerate the treatment better [83]. This is strongly influenced by the visibility and ease of
detection of the tumour, either by clinical examination or imaging. Data collected by the
Surveillance Epidemiology and End Results (SEER) Program of the National Cancer Institute
indicates a 5 year survival rate approaching 95% for skin cancer, 75-90% for breast cancer,
45% for ovarian cancer, but only 6% for pancreatic cancer [84]. By the time patients present
with symptoms of their disease, the cancer is often quite advanced and treatment options
limited. In order to detect cancer at an early, asymptomatic stage, screening programmes
have been introduced, and others are currently in varying stages of evaluation. Many of these
utilise diagnostic biomarkers and, as a result, there is an on-going need for biomarker
discovery and improved detection techniques.
The definition of a biomarker according to the National Cancer Institute is “a biological
molecule found in blood, other body fluids, or tissues that is a sign of a normal or abnormal
process, or of a condition or disease” [85]. Table 1.6 shows potential uses for cancer
biomarkers, and gives examples currently in clinical use.
In the case of diagnostic
biomarkers of cancer, the particular biomarker should either be absent in corresponding
normal tissue or fluid, or should be expressed at a significantly different level, be that higher
or lower than in normal cases.
The term ‘tumour marker’ is frequently used by physicians to describe cancer biomarkers,
but does not delineate between the different uses in clinical practice. A panel of such
markers is frequently requested by physicians when they suspect a diagnosis of cancer,
although many of those are not specific for cancer and can be present in benign conditions.
Some are associated with only one cancer type, whereas others may be associated with two or
more types. Most markers are also made by normal as well as by cancer cells, but are usually
produced at much higher levels in the latter. Although more than 20 different tumour
markers have been characterised and are in clinical use, markers have not been identified for
every tumour type or site. The tumour markers that are in current use in the UK are listed in
Table 1.7. Early detection of such biomarkers is often very difficult as the high level is
14
frequently associated with advanced disease whereas low levels can be seen in benign
conditions.
Use
Estimate risk of developing cancer
Screening
Differential diagnosis
Disease prognosis
Predict response to therapy
Monitor for disease recurrence
Monitor for response/progression in
metastatic disease
Predict drug toxicity
Example
Reference
BRCA1 germline mutation
(breast and ovarian cancer)
Prostate specific antigen -PSA
(prostate cancer)
Immunohistochemistry to determine tissue of origin
e.g. cytokeratin stains
21 gene recurrence score
(breast cancer)
KRAS mutation and anti-EGFR antibody
(colorectal cancer)
HER2 expression and anti-HER2 therapy
(breast and gastric cancer)
Oestrogen receptor expression
(breast cancer)
CEA
(colorectal cancer)
CA 125
(ovarian cancer)
AFP, LDH, βHCG
(germ cell tumour)
CA15-3 and CEA
(breast cancer)
CA 125
(ovarian cancer)
CYP3A4, CYP3A5, UGT1A1
(acute lymphoblastic leukaemia)
[86, 87]
Table 1.6. Potential uses for cancer biomarkers (adapted from [99]).
15
[86]
[87]
[88]
[89-91]
[92]
[93]
[94]
[95]
[96]
[94]
[97, 98]
Tumour marker
Cancer types
Tissue analysed
Non-small cell lung cancer
Anaplastic large cell lymphoma
Hepatocellular carcinoma (HCC)
Germ cell tumours (GCT)
Tumour
Beta-human chorionic
gonadotropin (β-HCG)
Multiple myeloma
Chronic lymphocytic leukemia
Some lymphomas
Testicular cancer
Choriocarcinoma
Blood
Urine
CSF
Urine
Blood
BCR-ABL fusion gene
Chronic myeloid leukemia
BRAF mutation V600E
Cutaneous melanoma
Colorectal cancer
Breast cancer
Blood
Bone marrow
Tumour
ALK gene rearrangements
Alpha-fetoprotein
(AFP)
Beta-2-microglobulin
CA15-3
Blood
Blood
Use
Treatment choice
Prognosis
Diagnosis (HCC)
Staging (GCT)
Prognosis (GCT)
Treatment response
Prognosis
Treatment response
Staging
Prognosis
Treatment response
Diagnosis
Disease recurrence
To determine treatment
Treatment response
Disease recurrence
Treatment response
Blood
CA-125
Pancreatic cancer
Gallbladder/Bile duct cancer
Gastric cancer
Ovarian cancer
Calcitonin
Medullary thyroid cancer
Blood
Carcinoembryonic antigen
(CEA)
Colorectal cancer
Breast cancer
Blood
CD20
Chromogranin A
Non-Hodgkin lymphoma
Neuroendocrine tumours
Blood
Blood
c-KIT
Gastrointestinal stromal tumour
Mucosal melanoma
Non-small cell lung cancer
Tumour
Oestrogen (ER) &
Progesterone receptor (PR)
Breast cancer
Tumour
Treatment response
Disease recurrence
Diagnosis
Treatment response
Disease recurrence
Diagnosis
Treatment response
Disease recurrence
To determine treatment
Diagnosis
Treatment response
Disease recurrence
Diagnosis
To determine treatment
To determine treatment
Prognosis
To determine treatment
HER2/neu
Breast cancer
Gastric/Oesophageal cancer
Multiple myeloma
Waldenström macroglobulinaemia
Tumour
To determine treatment
Blood
Urine
Granulosa cell ovarian tumours
Mucinous epithelial ovarian cancer
Blood
Diagnosis
Treatment response
Disease recurrence
Diagnosis (Granulosa cell)
Treatment response
Disease recurrence
To determine treatment
CA19-9
EGFR mutation analysis
Immunoglobulins
Inhibin A & B
Blood
Tumour
Colorectal cancer
Non-small cell lung cancer
Lactate dehydrogenase (LDH) Germ cell tumours
Tumour
Prostate-specific antigen (PSA) Prostate cancer
Blood
KRAS mutation analysis
Thyroglobulin
Blood
Thyroid cancer
Tumour
Staging
Prognosis
Treatment response
Diagnosis
Treatment response
Disease recurrence
Treatment response
Disease recurrence
Table 1.7. Tumour markers in current use in the UK. (adapted from [100]).
16
1.2.1. Current biomarkers in ovarian cancer
Early detection of ovarian cancer is usually associated with earlier stage disease, hence
improved outcome, although potential screening tests for ovarian cancer have not yet been
shown to reduce mortality. There is currently no standardized test to detect ovarian cancer at
an early stage. A clinically useful ovarian cancer biomarker needs to be highly specific and
highly sensitive, and may have diagnostic, prognostic or predictive value. The main focus of
research to date has been on tumour tissue and serum, however urine and ascites samples are
also being investigated.
1.2.1.1. Gene mutation and amplification
As previously discussed in Section 1.1.5, around two thirds of Type 1 epithelial ovarian
carcinomas harbour a KRAS, BRAF or ERBB2 gene mutation, with KRAS mutations
particularly noted in mucinous tumours and BRAF mutations in serous tumours. In addition,
ß-catenin is mutated in up to 60% of grade I endometrioid carcinomas and is believed to be
associated with a good prognosis [44, 101]. PIK3CA mutations occur in 33% of clear-cell
and in endometrioid ovarian carcinomas, correlating with a young age at diagnosis and
resistance to chemotherapy [102]. MSI-H accounts for 12% of epithelial ovarian cancer,
particularly Type 1 non-serous tumours, but there is no clear consensus on the prognostic
value [39, 103].
TP53, BRCA1 or BRCA2 mutations, and ERBB2 amplifications are frequently present in
Type 2 epithelial ovarian cancers [39, 46]. TP53 mutation and over-expression correlates
with poor outcome [47]. Mutated TP53 has been shown to trigger autoantibody formation
and these serum antibodies (AAbs) have been detected in lung cancer prior to clinical
diagnosis [104, 105]. This prompted Lu et al to investigate the potential role of serum p53
AAbs as biomarkers in the early detection of ovarian cancer [106]. They demonstrated a
significantly different plasma level for p53 AAbs between type 2 tumours and healthy
controls, resulting in a sensitivity of 85.7% with a specificity of 100%, for p53 AAbs in
combination with CA-125.
BRCA1 or BRCA2 mutations or epigenetic losses occur in 5-15% of ovarian cancers,
predominantly of serous histology [39, 46], and convey a good prognosis [107, 108]. These
tumours have been shown to have an increased sensitivity to platinum-based chemotherapy
17
and to PARP inhibitor therapy [108, 109]. ERBB2/HER2 amplification and over-expression
are present in Type 2 ovarian tumours but are uncommon (occurring in 7% and 11%,
respectively), with no clear correlation with prognosis or treatment response [46].
1.2.1.2. DNA methylation
Aberrant DNA methylation has been observed in several tumour types, and there is
increasing interest in its use as a biomarker. DNA methylation usually occurs at CpG
dinucleotides which can cluster to span hundreds or thousands of base pairs, known as CpG
islands [110, 111]. These islands are often associated with the promoter region of genes and
are usually unmethylated, permitting transcription initiation. Hence a methylated promoter
CpG island represses promoter activity [112]. It is thought that DNA methylation changes
are early events in the development of cancer. They may be present in precancerous lesions
and aid early cancer detection [110, 113].
Both global hypomethylation and localised
increases in hypermethylation have been observed in cancer. Methylation patterns in primary
tumour specimens have been shown to mimic those in other biological fluids, such as plasma
and sputum, which supports the potential use as a diagnostic biomarker [114-124].
Up to 21% of epithelial ovarian carcinomas contain BRCA, FANCF or PALB2 promoter
methylation, which is associated with poor prognosis [125, 126]. Fiegl et al studied 71 genes
within ovarian cancers, and also found that methylation of HOXA10 and HOXA11 differed
compared with non-cancerous tissue [127].
HOXA11 methylation showed a particular
association with poor outcome, raising the possibility for use as a prognostic marker.
1.2.1.3. Protein biomarkers in blood
CA125, also known as mucin 16 (MUC16), is a high molecular weight mucinous transmembrane glycoprotein that is approved for monitoring ovarian cancer progression and
treatment response. It is produced by coelomic epithelium which includes mesothelial cells
and Mullerian tissues, is present on the cell surface of ovarian tumours and can be detected in
the serum by monoclonal antibodies such as OC125. It is often used as a screening method,
as it can be easily sampled from patient serum, although it has many limitations. Although
elevated in 80% of epithelial ovarian cancer, it has a very low sensitivity of 50-62% for early
18
stage disease [128-130], although this rises to 90% in advanced stage ovarian cancer. The
specificity for all stages of disease ranges from 94 to 98.5%. It can be detected in nonmalignant conditions such as endometriosis, benign ovarian tumours, fibroids, liver cirrhosis,
pelvic inflammatory disease, and other cancers especially if they produce a pleural or
peritoneal fluid. The specificity is particularly crucial in ovarian cancer because the majority
of women with a positive result will proceed to invasive surgery [131]. CA125 does not
provide any prognostic information such as the aggressive nature of the cancer, or how it may
respond to platinum-based chemotherapy. This is especially relevant given the heterogeneity
of ovarian tumour histology, with a marked variation in the number of Ki-67 positive tumour
cells reflecting cellular proliferation, hence a Ki-67 index ranging from 1 to >90% [35]. CA125 is used to assess treatment response as a 50% decrease in serum level has been shown to
correlate with disease response and a doubling from baseline is consistent with treatment
failure or progressive disease [94].
A large randomised controlled trial is currently being conducted in the UK called “The UK
Collaborative Trial of Ovarian Cancer (UKCTOCS)”. This opened to recruitment in 2001
with the aim of recruiting 200,000 post-menopausal women, in order to assess the cost,
acceptability and mortality benefit of population screening. Women were randomly assigned
to 3 groups:1) no treatment (control)
2) annual multimodal screening (MMS) where the CA125 is checked, with subsequent
trans-vaginal ultrasound scan (USS) if the CA125 is positive
3) annual trans-vaginal USS
The preliminary results were published in 2009 and demonstrated a sensitivity of 89.4% for
the MMS groups compared with 84.9% for the ultrasound only group [132].
The
specificities were 99.8% and 98.2% respectively. The positive predictive value was only
5.3% for the ultrasound only group, compared with 43.3% for the MMS group.
It is
interesting to note that 5.8% of women in the ultrasound only group were deemed to be at
elevated risk of ovarian cancer with an abnormal scan, compared with only 0.5% in the
MMS group. This resulted in 3.9% of women from the ultrasound only group being referred
for clinical evaluation, compared with 0.3% from the MMS group. Ultimately 87% of
women in the ultrasound only group underwent surgery for benign disease. The initial
19
results are encouraging for a combined approach of annual CA125 and trans-vaginal USS
screening for ovarian cancer, as large-scale population screening appears to be feasible and
does detect ovarian cancer in symptomless women. Nevertheless, the resultant impact on
mortality is yet to be determined.
There is also an on-going screening study for women with a family history of ovarian cancer,
the UK Familial Ovarian Cancer Screening Study (UKFOCSS). Familial ovarian cancer
represents 10-15% of the overall diagnoses. In those with a BRCA1, BRCA2 mutation or
Lynch syndrome (heritable non-polyposis colorectal cancer mismatch repair gene mutation),
there is a 35-46%, 13-23% and 3-14% lifetime probability of developing ovarian cancer
respectively. In the UKFOCSS study, the women undergo annual screening with CA125 and
trans-vaginal USS as well as 4-monthly blood testing for retrospective analysis of existing
and novel tumour markers.
CA125 has also been investigated as a prognostic biomarker. Budiu et al evaluated soluble
MUC16 (CA125) and MUC1 (CA15-3) levels as well as serum MUC16- and MUC1-specific
antibodies as potential prognostic biomarkers for platinum-resistant ovarian cancer [133].
Serum MUC1 (CA15-3) was raised in all of the platinum-resistant cases receiving
intraperitoneal interleukin-2 therapy, and significantly correlated with increased risk of death
(p=0.003). Anti-MUC1 IgG was higher in patients with progressive disease compared with
treatment responders, and anti-MUC1 IgM antibodies inversely correlated with overall
survival at both early (p=0.052) and late (p=0.009) time points. Similar results were not seen
with MUC16.
Other mucin-related glycoproteins that have been evaluated include OVX1, HE4 and
Mesothelin. Elevated OVX1 levels have been identified in 67% of patients with ovarian
cancer who were CA-125 negative [134]. When OVX1 was combined with CA-125 and MCSF, 84% of ovarian cancers were detected compared with 69% when using CA-125 alone,
although the improved sensitivity resulted in a reduction in specificity from 99% to 84% with
combined markers [135]. Mesothelin combined with CA-125 also gives a higher sensitivity
than the individual markers when comparing cancer patients to healthy controls, however this
sensitivity falls to 44.1% when distinguishing between malignant and benign ovarian tumours
[136]. Human epididymis protein 4 (HE4) is a protease inhibitor which has been shown to be
over-expressed by epithelial ovarian cancer tumours, and can be identified in both the tissue
and serum of patients [137]. HE4 was also elevated in over half of the cases where CA125
20
was absent [138]. A combination of HE4 with CA-125 was able to discriminate ovarian
tumours from healthy controls with a sensitivity of 92.9% and specificity of 95%, compared
with a sensitivity of 78.6% for CA-125 and HE4 alone [139]. Evaluation of this combination
for detecting stage I disease provided 46.1% sensitivity at 90% specificity, compared with
sensitivities of 23.1% and 46.2% respectively for CA-125 and HE4 alone [140]. These
authors along with others have demonstrated very limited expression of HE4 in the rare
mucinous ovarian tumour cell type [138, 140, 141]. However mucinous tumour patients do
demonstrate elevated serum levels of total inhibin and CEA [142-145].
Various cytokines and growth factors have been evaluated alone or in combination with CA125. Visintin et al combined CA-125 with five other serum biomarkers (leptin, prolactin,
osteopontin, insulin-like growth factor II and macrophage migration inhibitory factor) to
identify patients with ovarian cancer [146]. Whilst maintaining a specificity of 99.4%, they
demonstrated a sensitivity of 95.3% for the combination of markers compared with 72% for
CA-125 alone. Notably, this test could not be used for general population screening as the
positive predictive value based on the prevalence of ovarian cancer in the general population,
was only 6.5%. Vascular endothelial growth factor (VEGF) has been investigated as a
diagnostic marker but it appears to have more prognostic relevance. The presence of a serum
level greater than 380pg/ml and advanced stage disease, have been shown to be significant
risk factors for mortality in ovarian cancer [147].
There are 15 known human kallikrein (KLK) proteins which act as serine proteases and they
have been extensively studied in cancer, especially as they can be identified in biological
fluids such as serum, seminal fluid, and ascites [148]. Elevated KLK-6 and KLK-10 levels
have been found in ovarian cancer tissue and serum samples, with the latter present in 35% of
patients who were negative for CA125 [141, 149, 150]. When KLK-10 was combined with
CA125 in Stage I and II ovarian cancer patients, sensitivity was increased by 21% (with 90%
specificity) compared to CA125 alone [150]. Elevated levels of KLK-4, KLK-5, KLK-6,
KLK-10 and KLK-15 have also been associated with a poor prognosis in ovarian cancer
[149, 151-155]. KLK-4 has been shown to be associated with paclitaxel resistance and those
patients who showed higher levels of serum KLK6 protein were also non-responsive to
chemotherapy [149, 156]. Conversely elevated KLK-8, KLK-9 and KLK-11 have been
associated with a favourable prognosis [157-161]. The production of human kallikreins or
other proteases might contribute to the presence of cleaved protein products in the plasma of
ovarian cancer patients, providing the potential for non-invasive biomarkers [148].
21
1.2.1.4. Protein biomarkers in urine
There is growing interest in the use of urinary biomarkers given the simplicity and ease of
collection as well as the direct deposition of cancer-associated material into urine. Voided
urine cytology (VUC) is routinely used to aid bladder cancer diagnosis, and has a high
sensitivity (78-100%), but is a time-consuming, subjective method of analysis with the
potential for considerable inter-observer variability, and lacks robust sensitivity (12.2-84.6%)
[162-164]. As a result, the focus of research has moved to the detection of protein in urine.
Despite a number of studies investigating ovarian cancer biomarkers in urine, progress has
been slow. It is possible that low molecular weight proteins could be filtered through the
kidney glomerulus and detected in urine at an early stage of disease. In a particular cohort of
ovarian cancer patients with CA125 levels below 35 U/ml, compared with healthy controls,
the combination of urinary matrix metalloproteinase (MMP)-2 and -9 gave an AUC value of
0.773 (95% CI=0.702-0.843; p<0.0001) on multivariate logistic regression analysis [165].
When combined with age, the AUC increased to 0.881 (95% CI=0.831-0.930; p<0.0001),
yielding a sensitivity of 82% and specificity of 75%.
However, the samples were all
advanced i.e. stage III and IV disease, and no inflammatory ovary or benign tumour tissue
was included in the analysis. Badgwell et al investigated mesothelin levels in the serum and
urine of ovarian cancer patients, showing 42% elevated urine levels and 12% serum levels in
early stage disease, rising to 75% urine and 48% serum levels in advanced disease [166].
1.2.1.5. Protein biomarkers in ascites
Many patients with peritoneal carcinomatosis secondary to advanced epithelial ovarian
cancer can develop a protein-rich ascitic fluid collection within the abdominal cavity. As
ascites is often reflective of advanced disease, and obtaining ascites samples is an invasive
technique, it is unlikely that it would have much utility as a diagnostic biomarker.
Notwithstanding this, researchers have examined proteins within this exudative fluid as
potential prognostic biomarkers. A shorter progression-free and overall survival has been
noted in patients with elevated interferon-γ (IFN-γ) levels in their ascites [167], whereas an
elevated level of the serine protease human kallikrein 8 (KLK8), is associated with an
improved progression-free survival [159].
Within chemo-resistant serous EOC ascites
specimens, Huang et al identified significantly up-regulated levels of ceruloplasmin, so
22
analysis of ascites may also be able to predict treatment response or help to target treatment
[168].
1.2.1.6. Autoantibodies
Anti-tumour antibodies provide a novel field in diagnostic biomarker research as they allow
recognition of tumour antigen at an earlier stage in tumour development because the immune
system responds to tumour emergence very early. This enables antigen detection at very low
levels, perhaps lower than can be reliably detected by direct protein assay. As well as a
potentially useful early diagnostic tool, they may also provide anti-tumour defence.
Serum autoantibodies directed against tumour associated antigens have been frequently
detected in the sera of patients with different types of cancer [169]. Antibodies to epithelial
cell adhesion molecule (Ep-CAM) [170], HSP-90 [171, 172], HOXA7 [173], HER2 and
MUC1 [174, 175] have been identified in ovarian cancer. It is difficult to see how some of
these autoantibodies will prove useful in early diagnosis of cancer as they are often identified
in more than one tumour type, although combinations of antibodies may hold more promise.
The search for diagnostic biomarkers, prognostic factors, as well as an explanation as to why
patients develop resistance to drug treatment continue to be a major challenge in ovarian
cancer. Intensive research has identified several poor prognostic factors but the causes of
chemotherapy resistance, the ways of overcoming this problem and factors related to cancer
progression remain a major challenge.
23
1.3.
THE HOMEOBOX GENE SUPERFAMILY
The homeobox genes are a superfamily of regulatory genes that encode homeodomaincontaining transcription factors. They play a key role in early embryonic development but
have also been linked to disease, including cancer. The homeobox refers to a 183-bp DNA
sequence, originally identified in Drosophila [176], which encodes a 61-amino-acid
homeodomain within HOX proteins. This specific domain acts as a binding site for other
proteins to enable activation or repression of many downstream effector target genes [177,
178].
Over 100 homeobox genes have been identified, all of which share sequence similarity within
the homeodomains, and have been separated into individual groups such as the HOX gene,
the paired (PAX) gene, and the Engrailed (En) gene families. It is currently thought that
homeobox genes represent 0.1-0.2% of the entire vertebrate genome [179]. In 1994, Manak
and Scott reported the role of homeobox-containing genes in vertebrate embryonic
development [180]. Specifically, they have been shown to be involved in the control of cell
identity [181], cell growth and differentiation [182], as well as cell-cell and cell-extracellular
matrix interactions [183]. In mammals this extends to central nervous system, skeletal and
limb development, as well as organogenesis. Many genetic disorders have been associated
with deregulation of the homeobox genes, affecting multiple organ systems, and this can
involve an up- or down-regulation of expression [184]. Analysis of the homeobox gene
abnormalities has demonstrated the presence of deletion, insertion and substitution mutations.
Thirty nine HOX genes have been identified in humans and are implicated in apoptosis,
receptor signalling, differentiation, motility and angiogenesis [185]. They are fundamental to
normal limb and organ development along the anterior-posterior axis [186-189], as well as
blood vessel formation and prostate gland development [190, 191]. In the normal adult
human, specific patterns of HOX gene expression are involved in stem cell function as well
as haematopoietic lineage differentiation, particularly the HOXA10, HOXB3 and HOXC4
genes [182, 192, 193]. In addition, expression of HOXA10 and HOXA11 varies during the
course of the human menstrual cycle, with the most drastic increase occurring at the time of
implantation.
This continual process mirrors the HOX gene expression seen in the
developing reproductive tract [194-196].
24
Mutations in human HOX genes have been linked with developmental limb malformations
such as synpolydactyly (HOXD13) [197-199], radio-ulnar synostosis (HOXA11) [200] and
the hand-foot-genital syndrome (HOXA13) [201, 202], as well as malignancy. The HOXA9
gene is associated with a form of acute myeloid leukaemia [203, 204], and certain
lymphoblastic leukaemias [205], and confers a worse prognosis and treatment failure [206].
HOXC4-6 are expressed in non-Hodgkin’s lymphomas [207]. The HOX gene family have
also been implicated in the development of kidney tumours (HOXA9, -C9, -D10) [208-210],
colon (HOXA9, -B6-8, -C8-9, -D11) [211, 212], lung (HOXA1,-A7, -A9, -C5) [213-215],
breast (HOXA5, -A10, -B7, -B13) [216-225] prostate (HOXB13, -C8) [226-229], melanoma
(HOXB7) [230-234], endometrial (HOXD10) [235] and ovarian cancers (HOXA7, -A9-11, B13) [236-239].
Certain gene expression profiles are associated with therapy-resistant
cancers, and an increased risk of distant metastasis at recurrence [206, 219, 220, 223-225,
240].
Genetic control of the vertebrate anterior head and rostral brain development is influenced by
Emx and Otx genes, with heterozygous mutations in the human EMX2 gene contributing to
schizencephaly, a rare congenital brain malformation accompanied by craniofacial defects
[241-243]. The MSX gene family are closely linked with craniofacial and tooth development
[244] and MSX2 has been implicated in melanoma cell invasion and survival, with
cytoplasmic expression indicating improved prognosis [245].
PAX genes are mostly involved in the control of embryonic tissue development and cellular
differentiation [246, 247], including renal morphogenesis (PAX2,-8) [248, 249], Blymphocyte differentiation (PAX5) [250], and central nervous system development [251].
They show a strong correlation with human diseases, in particular renal defects and Wilms’
tumours (PAX2,-8) [248], microcephaly (PAX6) [252], glioblastoma multiforme (PAX5)
[253], rhabdomyosarcoma (PAX3) [254], congenital cataracts and other eye defects (PAX6)
[252, 255, 256], prostate cancer (PAX2) [257] and thyroid cancers (PAX8) [258].
25
1.4.
THE BIOLOGY OF ENGRAILED
Engrailed (En) is another member of the homeobox gene family, which was first
characterised in Drosophila.
The encoded homeodomain-containing transcription factor
again plays an important role in development, and has been identified in annelids [259],
molluscs [260], insects [261], echinoderms [262], chordates [263], and vertebrates [264].
Although these genes share a limited degree of sequence conservation at the protein level,
there are five En homology regions (EHs) which represent particular regions of similarity
[265]. Figure 1.1 depicts these functional domains within the EN protein. Tolkunova et al
demonstrated that EH1 mediates transcriptional repression by recruiting the co-repressor
groucho, and EH5 has a similar role [266]. EH4 is the homeodomain, which demonstrates
the highest level of conservation [265], whilst EH2 and EH3 bind PBX, another
homeodomain-containing transcription factor. The latter is able to modify the DNA binding
affinity and specificity of En [267, 268]. In addition to transcription, En protein may have a
regulatory role in translation, as it is able to bind directly to the eukaryotic translation
initiation factor 4E (eIF4E) with high affinity and specificity [269].
Figure 1.1. Functional domains within the EN protein [270].
26
The homeodomain sequences also facilitate the association of cytoplasmic En protein with
vesicles, enabling secretion of the protein from the cell [271-273]. En can also undergo
internalization by the cell, a mechanism also dependent upon sequences within the
homeodomain [274]. This process has been demonstrated both at 4 and 37°C and does not
appear to require a specific receptor, although both the secretion and internalization
mechanisms have yet to be fully elucidated [275, 276].
Following its identification in Drosophila, where the mutated En resulted in a malformation
of the border between the posterior and anterior wing compartments [277], further evaluation
of its role in embryological neural and axonal development has taken place. In vertebrates,
two En genes were discovered which differ slightly in their specific functions; En1 which is
located on chromosome 2 (2q14.2), and En2 which is located on chromosome 7 (7q36.3)
[278, 279]. High levels of both En1 and En2 are expressed in the alar (dorsal) cells of the
midbrain/hindbrain border region during brain development, and influence the survival of
mesencephalic dopaminergic neurons [23, 280-282]. Wilson et al recently demonstrated that
En1 and En2 are expressed in many cell types in the cerebellum, with expression still evident
at postnatal day 21 [283]. Between this point and the embryonic stages, there is a distinct
change in their cellular and spatial distributions, with the expression domains becoming
distinct. En plays a pivotal role in axonal guidance, as demonstrated in the developing chick
optic tectum [284], where En2 helps to establish rostro-caudal polarity. On-going work has
shown that EN2 protein continues to exert its actions even after internalization by the axons
[285].
Germline mutations in En are therefore likely to have profound effects on
embryological development.
Expression of En genes in the human foetus has been demonstrated in all of the neuronal
groups of the mid-gestational medulla and cerebellum, and is thought to be crucially involved
in the development and anatomic organisation of these structures [286]. A murine model of
homozygous mutation of the En1 and En2 genes results in disrupted formation of the
mesencephalon and metencephalon, with cerebellar hypoplasia, which mimics the anatomical
findings in two unrelated infants who were born at term but died in early infancy from lack of
central respiratory drive. It is believed that these cases resulted from a mutation or deletion
in the En2 gene [287].
27
The arcuate nucleus at the ventral surface of the medulla oblongata which aids in the control
of central chemoreception, cardioventilatory activity and blood pressure [288], also expresses
the En2 gene. Lavezzi et al demonstrated high EN2 protein expression in the arcuate nucleus
of the 17th to 22nd gestational week human foetus, with decreased expression up to the first
days after birth [289]. They also studied 13 cases of sudden infant death, where 61% had
hypoplasia of the arcuate nucleus with predominantly negative EN2 protein expression.
Their findings support the role of En2 in normal human neuronal development and anatomic
organisation, particularly of the arcuate nucleus.
The only known sites of normal adult EN2 expression are in the nervous system, particularly
the Purkinje neurones, and in the tubular epithelial cells of the kidney. This 333 amino acid
protein is identified in the nucleus of the Purkinje neurones [290] however, cytoplasmic
staining is evident in human kidney. Jankowski et al showed that gene over-expression
resulted in retardation in the maturation of the Purkinje cells, particularly in the timed
development of their dendritic tree [291].
Owing to its expression in mesencephalic
dopaminergic neurons, Rissling et al investigated single-nucleotide polymorphisms (SNPs) in
the promoter region or transcribed part of En2, in individuals with young-onset Parkinson’s
disease (YOPD) [292]. This suggested an association between SNP rs1345514 and the
development of YOPD, although this has not yet been verified in an independent sample.
En2 has been linked to the neuro-developmental disorder, autism, as it shares the
chromosomal region 7q36.3 which is thought to be an autism susceptibility locus. There
appears to be a genetic linkage in families with autistic members, but also broad similarities
between the neuropathology seen in En2 knockout mice and autistic individuals, including
cerebellar hypoplasia, decreased Purkinje cells and an anterior shift in the position of the
amygdala in the cerebral cortex [293-297]. Further cohort studies, mainly in Chinese and
Indian populations, support this [298-300].
28
1.5.
ENGRAILED IN CANCER
Over-expression of EN2 protein may be linked to tumour development in adult humans,
particularly in breast, prostate, melanoma and ovarian cancers.
En2 has been identified as a potential oncogene in a breast cancer model. Martin et al
identified ectopic expression of En2 in a number of breast cancer cell lines but only 7.3% of
human breast cancers [301]. There was no evidence of rearrangement or amplification of the
gene, so the ectopic expression may result from epigenetic modification. Non-tumorigenic
murine mammary cell lines were forced to express En2, and subsequently exhibited
malignant characteristics, including a reduction in cell cycling time, a loss of cell to cell
contact and a failure to differentiate in response to lactogenic hormones. The cell line
induced mammary tumours when transplanted into the cleared mammary glands of syngeneic
hosts. The same group showed that in a human breast cancer cell line, suppression of En2
resulted in a significant decrease in their proliferation rate.
They did not demonstrate
expression of En2 in normal breast epithelium, and En1 was not demonstrated in cell lines or
human breast tissue.
In the prostate, Bose et al initially demonstrated En2 over-expression in human prostate
cancer cells as compared to normal prostate epithelial cells [302]. SiRNA-mediated downregulation of En2 in the cell lines resulted in a decrease in PAX-2 expression, and vice versa,
and caused a dramatic decrease in prostate cancer cell proliferation, consistent with the
previous findings in breast cancer cells [301]. Morgan et al further identified En2 in prostatic
adenocarcinoma and demonstrated that it is not expressed in normal prostate tissue, normal
tissue adjacent to the cancer, benign hypertrophy, or high grade prostatic intraepithelial
neoplasia, a pre-malignant lesion [303]. Immunohistochemical staining of patient tumour
biopsies showed that EN2 protein expression was most intense in the ductal structures of
tumours, along with presence in the cytoplasm and basal membrane, however there was no
nuclear expression. EN2 containing blebs were also identified in prostatic acini and ducts
suggesting secretion into ductal lumen, confirming the known secretory properties of En2.
This group were able to identify EN2 in the urine of 66% of biopsy-proven prostate cancer
patients, some of whom had undetectable levels of serum prostate specific antigen (PSA).
This contrasted with <15% positivity in the low PSA control groups, prompting further
research into its use as a potential biomarker, as discussed later.
29
Subsequent work on human bladder cancer cell lines and patient tumour specimens supports
this data [304], and preliminary work on En2 expression in ovarian cancer has also
demonstrated high expression in some ovarian cancer cell lines, as well as in epithelial
ovarian cancer tissue [305]. Flow-cytometric analysis of ovarian cancer cell lines showed
positive EN2 cell surface expression. Immunohistochemical analysis of ovarian cancer tissue
arrays demonstrated that EN2 was present in approximately 80% of ovarian cancer tissues,
compared with low (<10%) or absent expression in normal ovarian tissue.
Hypermethylation of the engrailed genes has been identified in several cancers although its
specific role is yet to be characterised. Rauch et al found that all four HOX gene clusters
were preferential targets for DNA methylation in lung cancer cell lines, with additional
hypermethylation of both En1 and En2 [213]. The exact significance of this is unknown, but
the authors hypothesised that such DNA methylation markers could be useful in early
diagnosis of disease. Karpinski et al determined the methylation status of three CpG islands
at the 2q14.2 chromosomal band, including En1, in 148 sporadic colorectal cancers [306].
Generally 18% to 25% of sporadic colorectal cancers show CpG island methylator phenotype
(CIMP), and the average number of methylated sites was significantly higher in these
tumours, namely 70% En1 methylation was seen in CIMP positive tumours compared with
22% in CIMP negative tumours. The authors postulate that this hypermethylation, along with
the other CpG islands, may contribute to the specific characteristics of CIMP positive
tumours and their clinicopathologic features.
Of note, these studies did not investigate
subsequent En gene or protein expression. Hypermethylation of En1 has subsequently been
seen by Mayor et al in 90% of colorectal tumours [307], but also in astrocytoma [308], and
prostate cancers [309]. En2 hypermethylation, along with other homeobox genes, has been
identified in the follicular lymphoma cell line, RL, and in ten primary follicular lymphomas
[310].
However transcriptional down-regulation was not observed, indicating aberrant
epigenetic regulation in follicular lymphoma. Although En hypermethylation is present in a
number of malignancies its functional role is still unknown, as it does not appear to
contribute to gene silencing. In fact Bell et al demonstrated EN1 protein over-expression in
cases of adenoid cystic carcinoma (ACC), where significant hypermethylation of the En1
gene was observed [311]. To date, this is the only published evidence of EN1 protein overexpression in malignancy. ACC represents a rare and progressive malignancy of the salivary
glands in which the authors had previously observed significant hypermethylation at the
transcriptional start site of the En1 gene, prompting them to further evaluate EN1 protein
30
expression [312]. In a cohort of 200 patients, they demonstrated positive EN1 expression in
58.8% of tubular, 56.5% of cribriform, and 85.7% of the solid histological sub-types.
Nuclear staining of the inner ductal cells was observed with only faint expression in the
cytoplasm of salivary ductal cells. The solid pattern was predominantly poorly differentiated
and devoid of myoepithelial cells, suggesting that EN1 expression may correlate with a more
aggressive tumour with poor prognosis. The group mentioned that they had also performed
immunoreactivity on a triple-negative breast carcinoma microarray, which demonstrated
similar results for high-grade basaloid breast carcinomas.
Table 1.8 summarises the current evidence linking the Engrailed genes and their protein
products with cancer.
Evidence for involvement in
cancer
EN1 over-expression
(human tumours)
En2 promotes malignant
characteristics
(murine mammary cell lines)
En2 expression required for
cancer cell proliferation
(human cell lines)
En2 over-expression
(human cell lines & tumours)
Cancer type
Technique used
Salivary Gland
(Adenoid Cystic)
Breast
Immunohistochemistry
[313]
Forced over-expression of En2
[301]
Breast
RNAi down-regulation of En2
[301]
Prostate
RNAi down-regulation of En2
[302]
Breast
[301]
Prostate
Semiquantitative RT-PCR
Immunohistochemistry
Semiquantitative RT-PCR
Immunohistochemistry
Semiquantitative RT-PCR
Immunohistochemistry
Semiquantitative RT-PCR
Immunohistochemistry
ELISA
Bladder
ELISA
Lung
Colorectal
Astrocytoma
Prostate
Salivary Gland
(Adenoid Cystic)
Lung
Methylated-CpG island recovery
Methylation-specific PCR
Methylated-CpG island recovery
Methylated-CpG island amplification
and microarray
Methylated CpG island amplification
and microarray
Methylated-CpG island recovery
Follicular Lymphoma
Methylation-specific PCR
Prostate
Ovary
Bladder
EN2 secretion in urine
(human tumours)
En1 hypermethylation
(human cell lines & tumours)
En2 hypermethylation
(human cell lines & tumours)
Reference
[302, 303]
[305]
[304]
[303, 314]
[304]
[213]
[306, 307]
[308]
[309]
[312]
[213]
[310]
Table 1.8. A summary of the evidence linking the En genes and proteins with cancer
(adapted from [315]).
31
1.6.
ENGRAILED AS A POTENTIAL BIOMARKER
Several groups have reported the use of homeobox genes as diagnostic and prognostic
biomarkers, as certain gene expression profiles can be linked to tissue specificity, association
with early stages of carcinogenesis, and even therapy-resistant disease resulting in a worse
prognosis. The majority of this work has been with the HOX gene family [206, 316-318],
however, as more is known about En, its potential as a biomarker is growing.
There is currently a great deal of interest in the potential of EN2 as a biomarker in prostate
cancer. Recent published evidence of its use as a tumour specific urinary biomarker for the
early diagnosis of prostate cancer showed that the presence of EN2 in urine was highly
predictive of prostate cancer, with a sensitivity of 66% [303]. These were samples from
biopsy-proven prostate cancer patients, some of whom had undetectable levels of serum PSA.
This was in contrast to <15% positivity in control groups, giving a specificity for the test of
88.2%. Men with prostate cancer had a 10 fold higher level of EN2 in their urine versus noncancer controls. A receiver operator characteristic (ROC) analysis for this data (comparing
biopsy proven prostate cancer versus negative biopsies) gave an area under the curve (AUC)
of 0.8021 (p < 0.001), indicating a high diagnostic potential for EN2. This contrasted with an
AUC of 0.55 for PSA in the same study group. These findings were corroborated
independently in an equivalent population of patients and controls in another centre. The
Prostate Cancer Prevention Trial recorded a sensitivity and specificity of 24% and 93%
respectively for PSA, the current standard detection test for prostate cancer [319]. Hence it is
clear that the use of PSA is limited, although it could be used in combination with urinary
EN2 to reduce the need for prostatic biopsy.
Subsequently, Pandha et al evaluated the relationship between levels of pre-treatment urinary
EN2 and cancer volume in a Danish patient cohort who had undergone radical prostatectomy
for prostate cancer [314]. Seventy percent of patients were positive for EN2 in urine and
there was a strong relationship between urinary EN2 and prostate cancer volume by linear
regression (p=0.006). Higher EN2 levels also correlated with advancing tumour stage i.e.
pT2b versus pT3a (p=0.027). High EN2 levels were not simply reflective of prostatic
volume as no relationship between total prostate volume and EN2 was evident. Notably, no
such relationships were observed with serum PSA, although this may reflect the small cohort.
These observations were reinforced in a second prospective study of 57 men where pre32
surgical urinary EN2 levels were again correlated with tumour stage and volume
measurements from radical prostatectomy specimens [320].
Eighty five percent of
prostatectomy patients had detectable urinary EN2, and this correlated with increasing
tumour stage (pT2 versus pT3a p=0.0063), positive margins (p=0.0078), and with prostate
cancer volume in a linear regression analysis (p<0.0001). These two studies highlighted the
potential utility of EN2 to not only diagnose prostate cancer but also to assess the disease in
terms of tumour volume and, as a result, to distinguish those patients most suitable for active
surveillance. This has prompted the design of a prospective study of 6 monthly urinary EN2
testing along with other 'active surveillance' criteria.
In a cohort of urothelial bladder cancer patients, the mean urinary EN2 concentration was
significantly elevated compared to that for control subjects, with an overall sensitivity of 82%
and specificity of 75% [304].
High-grade tumours demonstrated higher mean EN2
concentration and higher sensitivity, namely 87%. Urinary EN2 testing could therefore be
used alone or in conjunction with other tests to direct further clinical investigations in those
with a high suspicion of bladder or prostate cancer. Multi-centre tests are on-going.
Hypermethylation of En1 and En2 has been identified in several tumour sites, as previously
discussed, and may have potential use as an early diagnostic marker. Mayor et al were able
to detect methylated En1 in stool DNA from patients with colorectal carcinoma with 44%
sensitivity and 97% specificity in patients with corresponding tumour methylation [307].
However the sensitivity fell to 27% when including the 27% of patients whose tumours did
not show En1 methylation. In serum, the sensitivity was only 11%. The authors concluded
that the presence of a methylated En1 CpG island in stool DNA could be developed as a
diagnostic biomarker. En1 was frequently methylated (65%) in primary prostate tumours,
significantly differentiating cancer from normal tissue when combined with SCTR
hypermethylation [309]. Combined methylation of these two genes may provide potential
novel biomarkers for prostate cancer detection. Bell et al initially suggested a potential role
for En1 methylation status as a biomarker in human salivary adenoid cystic carcinoma, given
the correlation with histological tumour grade, tumour location and patient outcome [312].
Subsequently they demonstrated a significant correlation between increased EN1 protein
expression and a lower survival rate (p=0.014) as well as a higher incidence of lymph node
metastasis, suggesting its role as a prognostic biomarker [311].
33
1.7.
SUMMARY
Based on the paucity of diagnostic and prognostic biomarkers in epithelial ovarian cancer,
and the growing body of evidence for the role of Engrailed 2 in other epithelial cancers such
as prostate, bladder and breast, the objective of this research was to evaluate En2 gene
expression as well as EN2 protein expression in a broad selection of EOC cell lines, human
tumour tissue, urine and ascitic fluid. Throughout this analysis, comparisons would be made
between the different histological sub-types, as well as tumour grades and stages. Potential
differences between platinum-sensitive and –resistant tumours would also be recorded.
Localisation of EN2 protein within the cell, cellular responses as a result of gene silencing
and over-expression, and analysis of closely related genes may help to further define the role
of En2 and its protein product in cancer cells.
Ultimately, if En2 mRNA and EN2 protein were found to be over-expressed in human EOC
tumour and fluid specimens, this may prove useful as a diagnostic, prognostic or treatment
response biomarker, helping to improve overall survival rates.
34
CHAPTER 2
MATERIALS AND METHODS
35
2.
MATERIALS AND METHODS
2.1.
PATIENTS AND CONTROLS
The collection of patient tissue, urine, ascites and blood was approved by the local research
ethics committee (REC no: 09/H1103/50). Suitable patients were identified at The Royal
Surrey County Hospital Gynae-oncology Multidisciplinary Team meeting, either by myself,
the Oncology Consultants, or by members of the Gynae-oncology Surgical Team. Patients
were contacted in advance, provided with an information leaflet about the research, and asked
to sign a written consent form, which was filed in the notes.
Fresh tissue was acquired at the time of primary or interval debulking surgery over a period
of 5 years, transferred to the laboratory by myself or a member of the aforementioned teams,
and then stored in RNAlaterTM preservative (Sigma-Aldrich, UK) at -20°C. These tumours
represented the different histological sub-types of EOC, MMMT, borderline and benign
tumours as well as normal ovary and fallopian tube. In total 111 tumours and 5 normal tissue
specimens were collected. The full demographic data from the human ovarian tumours in
RNAlaterTM is shown in Chapter 4, Table 4.1. When enzymatic immunohistochemistry was
performed, these specimens were referred to as Cohort 1.
Midstream urine samples were either collected from patients undergoing debulking surgery
for suspected or confirmed ovarian cancer, or from known ovarian cancer patients attending
the Oncology Outpatients’ clinic. Twenty-two urine samples were obtained from patients,
and 18 samples were obtained from female healthy volunteers.
Nineteen ascites samples were either collected at the time of debulking surgery, or from
patients undergoing ascitic drainage procedures on the wards. Dr Sandra Diebold at King’s
College, London kindly donated 3 control ascites samples from patients with benign
gynaecological pathology.
Ninety-eight plasma samples were collected from patients attending the Oncology
Outpatients’ clinic at The Royal Surrey County Hospital, whilst 123 samples were obtained
from female healthy volunteers. For control comparison of unrelated cancers, samples of
sera from 2 cohorts of breast cancer patients and 1 cohort of prostate cancer patients were
also analysed, along with healthy control samples (SUN study; REC no: 08/H1306/115).
36
Two samples of normal human ovary total RNA, two samples of normal human fallopian
tube total RNA, and two samples of normal human ovary genomic DNA were purchased
from OriGene Technologies Inc, USA, and stored at -80°C.
A separate cohort of 90 pre-cut slides from formalin-fixed, paraffin embedded ovarian
tumours was available for immunohistochemistry, again from The Royal Surrey County
Hospital (Cohort 2). The full demographic data from these specimens is shown in Chapter
4, Table 4.4. Slides from 4 normal ovaries were also donated by the Pathology Department,
although no demographic details were available for these patients.
Three 5µm-thick formalin fixed, paraffin embedded tissue arrays were also purchased: 1) OV2082 (US Biomax, Rockville, MD, USA) – An ovarian cancer tissue array
containing 104 cases of normal (5 cases), adjacent normal (5 cases), and malignant
(94 cases) tissues of the ovary in duplicate (Figure 2.1)
2) CJ2 (Super Bio Chips, Seoul, Korea) - An ovarian cancer tissue array containing 59
cases of benign (14 cases), borderline (9 cases) and malignant (36 cases) tissues of the
ovary, including survival data
3) FRS801 (US Biomax, Rockville, MD, USA) - A female reproductive tissue array
containing 80 cases of normal, inflamed, diseased and malignant tissues of the breast,
ovary, fallopian tube, cervix, endometrium, uterine wall and vulva. There were 12
ovarian cases in total (3 malignant, 3 inflammatory, 3 benign disease, 3 normal)
(Figure 2.2)
Figure 2.1. Ovarian carcinoma tissue array panel display
immunohistochemical analysis of EN2 expression (OV2082-Biomax, USA).
37
used
for
Figure 2.2. Female reproductive system tissue array panel display used for
immunohistochemical analysis of EN2 expression (FRS801-Biomax, USA).
The full demographic data from the OV2082 and CJ2 ovarian tissue arrays is shown in
Chapter 4, Tables 4.7 and 4.9. These show histological sub-type, grade and stage, with the
CJ2 array also including survival data.
2.2.
CELL CULTURE MEDIA
All the cell culture media used are shown below (Table 2.1) with their source and required
supplements.
Medium
RPMI 1640
DMEM
McCoy’s
MCDB
Medium 199
F-12K
Supplements
10% FCS, 50U/ml penicillin, 50µg streptomycin, 2mM
glutamax-I
10% FCS, 50U/ml penicillin, 50µg streptomycin, 2mM
glutamax-I
10% FCS, 50U/ml penicillin, 50µg streptomycin, 2mM
glutamax-I
15% FCS, 50U/ml penicillin, 50µg streptomycin, 2mM
glutamax-I, 1.5g/L NaHCO3
15% FCS, 50U/ml penicillin, 50µg streptomycin, 2mM
glutamax-I, 2.2g/L NaHCO3
10% FCS, 50U/ml penicillin, 50µg streptomycin, 2mM
glutamax-I
Table 2.1. Cell culture media used.
38
Source
Sigma-Aldrich, UK
Sigma-Aldrich, UK
GIBCO, UK
Sigma-Aldrich, UK
Sigma-Aldrich, UK
ATCC, USA
2.3.
HUMAN CELL LINES
All cell lines were obtained from the American Type Culture Collection (ATCC, USA) with
the exception of the ovarian cell lines PEO1, PEO14 and PEO23 which were obtained from
the Health Protection Agency (HPA, UK), PEO4, PEA1 and PEA2 which were obtained
from Professor Hani Gabra at Imperial College, University of London, and Fibroblasts which
were obtained from the University of Birmingham. All of the human cell lines listed below
(Table 2.2) were adherent lines and were maintained at 37°C in a humidified, 5% CO2
incubator.
Cell line
Medium & supplements
Tissue type
Histology
Drug resistance
CaOV3
DMEM/10% FCS
OVCAR3
RPMI 1640/10% FCS
COV318
DMEM/10% FCS
PEO1
RPMI 1640/10% FCS
PEO4
RPMI 1640/10% FCS
PEO14
RPMI 1640/10% FCS
PEO23
RPMI 1640/10% FCS
PEA1
RPMI 1640/10% FCS
PEA2
RPMI 1640/10% FCS
TOV112D
MCDB/15% FCS +
Medium199/15% FCS (1:1)
McCoy’s/10% FCS
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ovary)
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ascites)
Ovarian adenocarcinoma
(from ovary)
Ovarian adenocarcinoma
(from ovary)
Ovarian adenocarcinoma
(from ovary)
Prostate adenocarcinoma
Melanoma
Skin
Prostate stroma/fibroblasts
Mouse fibroblasts
High-grade
Serous
High-grade
Serous
High-grade
Serous
High-grade
Serous
High-grade
Serous
High-grade
Serous
High-grade
Serous
Low-grade
Serous
Low-grade
Serous
High-grade
Serous
High-grade
Serous
Endometrioid
-
SKOV3
MCDB/15% FCS +
Medium199/15% FCS (1:1)
McCoy’s/10% FCS
OV90
ES2
TOV21G
PC-3
A375M
Fibroblasts
WPMY1
NIH3T3
MCDB/15% FCS +
Medium199/15% FCS (1:1)
F-12K/10% FCS
DMEM/10% FCS
DMEM/10% FCS
DMEM/5% FCS
DMEM/10% FCS
Clear cell
Adriamycin, Cisplatin
Adriamycin, Cisplatin
Cisplatin
Cisplatin
Cisplatin
Cisplatin
Clear cell
Cisplatin, Doxorubicin,
Etoposide
-
-
-
Table 2.2. Cell lines used, their growth media, tissue of origin and relevant drug
resistance.
39
In the matched pairs, the first set of cell lines (PEO1, PEO14, PEA1) were derived early in
the patients’ diagnosis, whereby the second set (PEO4, PEO23, PEA2) were derived
following the onset of acquired clinical platinum resistance.
CpG Methylated Jurkat genomic DNA (New England BioLabs, USA), which is derived from
human acute T-Cell leukemia cells, was acquired as a positive control for DNA methylation
work, and was stored at -20°C.
2.4.
RNA EXTRACTION FROM CELL LINES
Total RNA was extracted from cell lines using the RNeasy® Plus Mini Kit (Qiagen, UK).
The cells were trypsinised and washed in Phosphate buffered saline (PBS)(Fisher Scientific,
UK). Cells were then treated according to the manufacturer’s instruction. Extracted RNA
was
eluted
in
RNase-free
water
and
quantified
using
a
Nanodrop
ND-1000
Spectrophotometer. RNA samples were stored at -80°C.
2.5.
RNA EXTRACTION FROM HUMAN TUMOUR TISSUES
Fresh human tumour samples obtained from patients at The Royal Surrey County Hospital
over a period of 5 years, were stored in RNAlater (Sigma-Aldrich, UK) at -20°C. Tumours
obtained prior to June 2011 had been stored at -80°C, therefore the tissue was thawed prior to
RNA extraction. For tumours obtained from June 2011 onwards, RNA extraction was
performed on the day of, or day after surgery, before storing the sample at -80°C.
After thawing, the tissues were initially homogenised using the gentleMACS Dissociator
(Miltenyi Biotec) according to the manufacturer’s instructions. The gentleMACS Program
RNA-01 was used.
After termination of the program, the M tube was detached and
centrifuged at 1500 rpm for 3 minutes. Total RNA was extracted from cells using the
RNeasy® Plus Mini Kit (Qiagen, UK) according to the manufacturer’s instructions.
Extracted RNA was eluted in RNase-free water and quantified using a Nanodrop ND-1000
Spectrophotometer. RNA samples were stored at -80°C.
40
2.6.
COMPLEMENTARY DNA (cDNA) SYNTHESIS
cDNA was synthesised from total RNA using the Cloned AMV First Strand cDNA Synthesis
Kit (Invitrogen, UK) following the manufacturer’s protocol. In summary, 1 µg of RNA was
incubated in a volume of 20µl at 50°C for 30mins, with final concentrations of 0.1M DTT,
10mM dNTP mix, 50µM Oligo(dT)20 primer, 15 U/µl of Cloned AMV reverse transcriptase,
and 40 U/µl of RNAseOUT. DEPC-treated water and cDNA Synthesis Buffer were added to
give a total volume of 20 µl. The cDNA synthesis reaction was terminated by incubation at
85°C for 5 minutes. cDNA samples were stored in a total volume of 100µl (10ng/µl) at 20°C.
2.7.
GENOMIC DNA (gDNA) EXTRACTION FROM CELL LINES
Genomic DNA was extracted from cell lines using the Gene JET Genomic DNA Purification
Kit (Thermo Scientific, USA). The cells were trypsinised, washed, and counted to acquire
20x105 cells. These cells were pelleted by centrifugation for 5 minutes at 250xg. The cells
were then treated according to the manufacturer’s instruction, involving incubation with
Lysis, Proteinase K and RNaseA solutions prior to the use of gDNA purification columns
with wash buffers. Extracted gDNA was eluted in 50µl of Elution buffer with the elution
step repeated for maximum DNA yield. This was then quantified using a Nanodrop ND1000 Spectrophotometer. The purified gDNA samples were stored at -20°C.
2.8.
GENOMIC DNA (gDNA) EXTRACTION FROM HUMAN TUMOUR
TISSUES
Genomic DNA was extracted from human tumour tissues and normal ovary and fallopian
tube using the Gene JET Genomic DNA Purification Kit (Thermo Scientific, USA).
Approximately 10mg of tissue was cut into small pieces, or disrupted using the TissueRuptor
homogenizer (Qiagen, UK), if necessary. The material was collected into a 1.5ml Eppendorf
tube and then treated according to the manufacturer’s instruction, involving incubation with
Digestion and Proteinase K solutions overnight (minimum 15 hours) at 56°C until completely
lysed. Subsequently, RNaseA, Lysis solution and 50% ethanol were added, prior to the use
of gDNA purification columns with wash buffers. Extracted gDNA was eluted in 50µl of
41
Elution buffer with the elution step repeated for maximum DNA yield. This was then
quantified using a Nanodrop ND-1000 Spectrophotometer. The purified gDNA samples
were stored at -20°C.
2.9.
McrBC
ENDONUCLEASE
CLEAVAGE
OF
METHYLCYTOSINE-
CONTAINING DNA (METHYLATED DNA)
McrBC endonuclease (10,000 U/ml) and reaction buffers (New England BioLabs, USA) were
incubated with 400ng of purified cell line or human tumour tissue gDNA for 16 hours at
37°C. The gDNA volume was made up to 15µl with nuclease-free water, to give a final
reaction volume of 40µl (Table 2.3). McrBC is a methylation-specific endonuclease, which
cleaves DNA containing 5-mC on one or both strands, but does not act on unmethylated
DNA.
Strand breaks induced by cleavage of the methylated DNA abrogate PCR
amplification, whereas unmethylated cytosines remain intact and can be detected by
quantitative PCR product recovery.
Stock Solutions
Sample gDNA
Nuclease-free water
McrBC endonuclease
BSA (100X)
GTP (100X)
NEBuffer 2 (10X)
Volume for one sample (µl)
15.0
19.2
1.0
0.4
0.4
4.0
Total = 40 µl
Final concentration
400ng
10U
100µg/ml
1mM
1X
Table 2.3. Reaction volumes and concentrations for methylated DNA digestion using
McrBC endonuclease.
Undigested gDNA served as the internal control i.e. the sample was exposed to the same
reaction conditions without the addition of the enzyme. The positive control was CpG
Methylated Jurkat gDNA (New England BioLabs, USA), which is derived from human acute
T-Cell leukemia cells, and is reported to be 100% methylated. The workflow for sample
preparation and evaluation of En2 promoter methylation status, is shown in Figure 2.3.
This protocol was kindly provided by Dr S.J. James, Department of Paediatrics, University of
Arkansas for Medical Sciences, USA (direct communication).
42
Figure 2.3. Workflow for McrBC enzyme cleavage of methylated gDNA and evaluation
of En2 promoter methylation status.
2.10. SEMI-QUANTITATIVE
REVERSE
TRANSCRIPTASE
POLYMERASE
CHAIN REACTION (rt-PCR)
Semi-quantitative rt-PCR was performed using the Stratagene MX3005P Real Time PCR
machine (Agilent Technologies, UK), measuring PCR product accumulation during the
exponential phase of the reaction by SYBR green fluorescence (SYBR® Green JumpStart™
Taq ReadyMix™ Kit, Sigma-Aldrich, UK). This technique was used to evaluate En2 mRNA
expression levels in cell lines and human tumour tissue, as well as additional genes identified
during a microarray analysis of En2 over-expressed cell lines. The methylation status of the
En2 promoter region in cell lines and human tumour tissue was also analysed using this
method with some modifications, as described below.
43
2.10.1. mRNA expression
The PCR Master Mix, was prepared using 1µl of cDNA, representing 10ng of total RNA,
6.25µl RNase Free Water, 12.5µl 2x SYBRGreen JumpStart Taq ReadyMix, and 0.25µl
Reference Dye (ROX). 5µl of En2 primer was added to the relevant wells of the PCR plate
and 20µl of the prepared Master Mix was added. ß-actin was used as a control gene.
Reaction conditions were 1 cycle of 94°C for 10 minutes, followed by 40 cycles of 30
seconds at 94°C, 1 minute at 60°C and 30 seconds at 72°C. The forward and reverse primers
for ß-actin, En2 and all of the microarray validation genes, are listed in Table 2.4.
Gene Name
Forward Primer
Reverse Primer
ß-actin
ATGTACCCTGGCATTGCCGACA
GACTCGTCATACTCCTGCTTGT
En2
GAACCCGAACAAAGAGGACA
CGCTTGTTCTGGAACCAAAT
PAI1/SERPINE1
CCGCCTCTTCCACAAATCAG
AATGTTGGTGAGGGCAGAGA
Activin (Inhibin βA
subunit)
TGTACCCAACTCTCAGCCAG
TGCCCTCCTTCCAATGTCAT
CSH1
CCTCGGACAGCGATGACTAT
CCGTAGTTCTTGAGCAGTGC
GADD45β
ATTGCAACATGACGCTGGAA
AAGGACTGGATGAGCGTGAA
TWIST2
TACAGCAAGAAGTCGAGCGA
CTTGCTCAGCTTGTCAGAGG
Ski
TACAAGAAGGAGAGCGCCAA
GAGTTGAGAATCTGCGGCAG
IL1RAP
ACTACAGCACAGCCCATTCA
TGCCAGTGTCATTGAGGAGA
IL1A
TCATTGGCGTTTGAGTCAGC
ACCACCATGCTCTCCTTGAA
MMP1
GATGAAGCAGCCCAGATGTG
GCTTGACCCTCAGAGACCTT
Tenascin C (TNC)
GACAATGAGATGCGGGTCAC
CGCTGACAGGAATGCTCTTC
PLCB1
GTCTCAGCCCCTTTCTCACT
TGGGTGATGACAGGTTCCTC
NR4A1/NUR77
AGAAGATCCCTGGCTTTGCT
CAGGGACATCGACAAGCAAG
PRKAR2β
GATCCGGAGCAGATGTCTCA
AGCTGCTCTGGGTGTATTGT
SMAD4
TCCAGCCTCCCATTTCCAAT
ACCTTGCTCTCTCAATGGCT
SMAD3
AGGAGAAATGGTGCGAGAA
CCACAGGCGGCAGTAGAT
TGF-β1
CTTTCCTGCTTCTCATGGCC
TCCAGGCTCCAAATGTAGGG
ActRIIA
GAAGTCACACAGCCCACTTC
GTCCTGGGTCTTGAGTTGGA
ActRIIB
TGACTTTGGCTTGGCTGTTC
CCTCAAAGGGCAGCATGTAC
Table 2.4. The forward and reverse primer sequences for the genes evaluated by rtPCR.
44
Fluorescence increased in accordance with increasing levels of PCR product and relative
expression was calculated using the ∆CT comparative method (2-∆Ct).
CT is the cycle
threshold, defined as the number of cycles required for the fluorescent signal to cross the
threshold (i.e. exceeding the background level).
In the results, the expression is shown relative to ß-actin (x100,000). The fibroblast cell line,
commercially available total RNA from human normal ovarian and fallopian tube tissue, and
normal ovarian tissue from theatre were used as negative controls for the PCR procedure.
2.10.2. En2 methylation status
The PCR Master Mix was prepared using 1µl of En2 promoter region forward primer and 1µl
of reverse primer, 7.25µl RNase Free Water, 12.5µl 2x SYBRGreen JumpStart Taq
ReadyMix, and 0.25µl Reference Dye (ROX).
3µl of digested and undigested gDNA,
representing 30ng of gDNA, was added to the wells of the PCR plate in triplicate, with 22µl
of the prepared Master Mix added. Reaction conditions were 1 cycle of 94°C for 10 minutes,
followed by 45 cycles of 30 seconds at 94°C, 1 minute at 60°C and 30 seconds at 72°C. The
forward and reverse primer sequences for the En2 promoter region are listed in Table 2.5.
Gene Name
Forward Primer
Reverse Primer
En2 promoter
GTGTCTCCGGGGCGCTTCAC
CCCCGCATCCCCAGAAAGCG
Table 2.5. The forward and reverse primer sequences for the En2 promoter evaluated
using rt-PCR.
Fluorescence increased in accordance with increasing levels of PCR product and the
percentage methylation of the En2 promoter was calculated using the CT for each cell line
or tumour sample, compared with CT for the CpG Jurkat DNA, where CT represents
CTdigested DNA - CTundigested DNA. An average of 3 biological repeats was taken for each cell line.
The human skin fibroblast cell line, the mouse fibroblast cell line NIH3T3, and the prostate
stromal cell line WPMY-1 were used as negative controls for the PCR procedure. Normal
ovarian and fallopian tube tissue from theatre, and two commercial samples of normal human
ovary genomic DNA were also used as negative controls. CpG Methylated Jurkat genomic
DNA was used as the positive control, representing 100% methylated DNA.
45
2.11. ENZYMATIC IMMUNOFLUORESCENT STAINING ON CELL LINES
Ovarian cancer and control cell lines were incubated for monolayer growth in 8-chambered
polystyrene culture treated glass slides (BD Biosciences, UK) with relevant media for each
cell line. Each cell line was grown in quadruplicate, so that the staining could be compared
in the absence of primary antibody, and in non-permeabilised and permeabilised cells. After
removal of the media, 500µl of 5µg/ml Wheat-Germ Agglutinin cell membrane stain
(Invitrogen, UK) was added to the chamber and incubated for 10 minutes at 37°C. After
removal of the stain solution, the cells were fixed using 500µl of warm 4%
paraformaldehyde, and incubated for 10 minutes at 37°C. The cells were then washed with
PBS three times. Cells in one of the duplicate chambers from each cell line were then
permeabilised with 0.2% Triton X-100 (Sigma-Aldrich, UK) for 10 minutes and washed with
PBS. All of the cells were subsequently blocked with 4% horse serum for 10 minutes. A
polyclonal goat anti-EN2 antibody (Abcam, UK) diluted 1:100 in 1% Bovine serum albumin
(1% BSA)(Sigma-Aldrich, UK) in PBS, was applied overnight. 1% BSA in PBS alone was
added to the negative control wells.
This primary antibody has been used within the
department for immunofluorescence work in multiple cells lines derived from varying tumour
types, with consistent results. The slides were washed and incubated with the secondary
antibody, Alexa Fluor 488 donkey anti-goat IgG (Invitrogen, UK) diluted 1:200 in 1%
BSA/PBS, along with the nuclear stain TO-PRO-3 (Life Technologies, UK) diluted 1:400, at
room temperature for 1 hour. Subsequently, the slides were washed and, after removal of the
chambers, mounted with Vectashield mounting medium (Vector Laboratories, USA). The
human fibroblast cell line was used as a negative control, and the A375M melanoma cell line
was used as a positive control. Cells were visualised using a Zeiss LSM 510 confocal laser
scanning microscope, and a Plan-Apochromat 40x oil immersion objective. Images were
recorded and analysed using ZEN 2009 capture software.
46
2.12. ENZYMATIC IMMUNOHISTOCHEMISTRY ON PATIENT SLIDES AND
TISSUE ARRAYS
Expression of EN2 in ovarian and fallopian tumours along with normal tissue was
investigated using: 1) Cohort 1 – 116 fresh tissues preserved in RNAlaterTM (Sigma-Aldrich, UK), which were cut, formalin-fixed, paraffin embedded, sectioned and mounted onto slides
2) Cohort 2 – 90 pre-cut slides from formalin-fixed, paraffin embedded tissues
3) OV2082 (US Biomax, Rockville, MD, USA) – A 5µm-thick formalin fixed, paraffin
embedded ovarian cancer tissue array containing 104 cases in duplicate
4) CJ2 (Super Bio Chips, Seoul, Korea) - A 5µm-thick formalin fixed, paraffin
embedded ovarian cancer tissue array containing 59 cases
5) FRS801 (US Biomax, Rockville, MD, USA) - A 5µm-thick formalin fixed, paraffin
embedded female reproductive tissue array containing 80 cases of normal, inflamed,
diseased and malignant tissues of the breast, ovary, fallopian tube, cervix,
endometrium, uterine wall and vulva.
The slides were dewaxed (if necessary), deparaffinised in a series of xylene and then
rehydrated in a series of 100%, 70% and 50% ethanol. The slides were then subjected to heat
mediated antigen retrieval in a microwave using Citrate buffer (10mM, pH 6.0). Slides were
blocked for 15 minutes with 4% horse serum followed by a streptavidin-biotin blocking step
(Vector Laboratories, USA). A polyclonal rabbit anti-EN2 antibody (LifeSpan Biosciences,
Seattle, USA, LS-B3477) was diluted in 1% Bovine Serum Albumin (BSA) in PBS and
applied overnight at a 1:20 dilution. 1% BSA in PBS alone was added to control slides. This
primary antibody was chosen as it has been validated for use in immunohistochemistry
against formalin-fixed paraffin-embedded tissues by LifeSpan Biosciences, in more than
twenty tissue types. Biotinylated universal secondary antibody from R.T.U. VECTASTAIN
ABC Kit (Vector Laboratories, USA) was added to all wells for 30 minutes, followed by
streptavidin-horseradish peroxidase. EN2 staining was revealed using a DAB Peroxidase
substrate kit (Vector Laboratories, USA).
Slides were counterstained with Haematoxylin
(Vector Laboratories, USA) for 45 seconds and dehydrated in a series of alcohols, before
being mounted with VectaMount mounting medium (Vector Laboratories, USA).
47
Normal ovarian tissue was used as a negative control for each staining procedure, whilst
normal kidney was used as a positive control as suggested by Abcam, UK who also produce
anti-EN2 antibodies for immunohistochemistry.
The Royal Surrey County Hospital Pathology Department kindly performed automated
Haematoxylin & Eosin (H&E) staining on adjacent sections from Cohort 1.
Slides were scored by two independent investigators for the immunointensity of EN2
staining: 0, negative; 1, weak; 2, moderate; 3, strong. The percentage of immuno-positive
cells was also recorded and scored from 1 to 4 (Table 2.6). The product of immunointensity
and immunopositivity was then calculated to produce overall values ranging from 0 to 12,
with 0-4 representing EN2 negative staining, and 5-12 representing EN2 positive staining
[321].
Percentage of Positively Stained Cells (%)
1-10
11-40
41-70
>70
Assigned Score
1
2
3
4
Table 2.6. The assigned score (0-4) for the percentage of EN2 positive cells (taken from
[321]).
Any discrepancy in scoring was reviewed and a collective decision was made. All tissue
array specimens were clearly labelled as normal, benign or malignant tissue, however an
independent gynaecological pathologist also reviewed these arrays along with the patient
tumour slides to confirm the diagnosis, and to ensure that EN2 positive staining corresponded
with malignant tissue.
48
2.13. ENZYME-LINKED IMMUNOSORBENT ASSAY (ELISA) FOR PROTEIN
IDENTIFICATION IN CELL LINE LYSATES AND SUPERNATANTS,
PATIENT URINE AND ASCITES
The EOC and control cell lines were grown in 175cm3 tissue culture flasks (Helena
Bioscience, UK) with their specific media until 70-80% confluent.
For preparation of cell lysates: The cells were washed with cold PBS twice, and lysed by adding RIPA buffer and inhibitor
cocktail mix to the flask. The mix consisted of 1000μl RIPA buffer (Thermo Scientific,
USA), 20μl phosphatase inhibitor, 20μl protease inhibitor and 20μl EDTA (Thermo
Scientific, USA) per flask.
The flask was then placed on ice for 10 minutes, mixing
occasionally. The cells were collated with the aid of a cell scraper if necessary, transferred to
a 1.5ml Eppendorf tube and centrifuged at 13,000rpm for 5 minutes. The supernatant was
then transferred to a clean tube and stored at -80C prior to analysis.
For preparation of cell line supernatants: The cells were washed with cold PBS twice, prior to incubation in 10mls of serum-free
specific media at 37°C. After 24 hours, the supernatant was transferred to a 10ml Universal
container and centrifuged at 1500rpm for 3 minutes to remove cellular debris.
The
supernatant was then transferred to an Eppendorf tube and stored at -80C prior to analysis.
A monoclonal mouse anti-EN2 antibody, APS1, was raised using the synthetically produced
C-terminal 100 amino acids of EN2 (Biosynthesis Inc., USA) as an antigen (Antibody
Production Services Ltd., UK). APS1 was conjugated to alkaline phosphatase using the
Lightning Link alkaline phosphatase conjugation kit (Innova Biosciences, UK) and stored at
4°C. EN2 dilutions were prepared in RIPA buffer (cell lysate analysis), serum-free media
(cell supernatant analysis) or PBS/0.1% Tween-20 (urine and ascites analysis) using the EN2
C-terminal fragment (10,000ng/ml).
The following dilutions were prepared with 200µl
added to the wells of a standard immunosorbent plate (Nunc, Germany) in triplicate:
49
250ng/ml, 125ng/ml, 62.5ng/ml, 31.25ng/ml, 15.625ng/ml and 0ng/ml. Subsequently, 100µl
of each cell line or clinical sample was added in triplicate and was incubated for 2 hours at
room temperature. The plate was then washed 8 times in PBS/0.1% Tween-20 before adding
100µl of the detection antibody, APS1-alkaline phosphatase, diluted 1:250 in the relevant
serum-free media/diluent. After 15 minutes at room temperature, the plate was again washed
4 times in PBS/0.1% Tween-20, followed by a single wash in 200µl of Tris-buffered saline
(Sigma-Aldrich, UK) to remove phosphate.
100µl of pNPP (Sigma-Aldrich, UK), a
colorimetric agent, was added and the absorption of light/optical density (OD) at 405nm was
measured after 30 minutes, using a Varioskan Flash plate reader (Thermo Scientific, USA).
The dilution series was used to generate a standard curve by which the concentration of EN2
in each sample was measured.
Individual patient urine readings were scored as EN2 positive if their concentration was equal
to or greater than two standard deviations from the mean of the control subjects.
A recombinant EN2 (rEN2) protein developed in E. coli was used as a positive control. This
was produced by Dr. Guy Simpson within the Oncology department, as follows: En2 cDNA with optimized codon usage was synthesized and cloned by GenScript, USA.
This En2 cDNA was recloned into the pQE31 plasmid (Qiagen, UK), placing it under the
control of a T7 promoter as well as adding an aminoterminal histidine tag. This plasmid was
transformed into E. coli host strain (M15; Qiagen), which contains multiple copies of the
pREP4 plasmid which constitutively expresses the lac repressor protein, resulting in
regulated efficient and controlled expression of E. coli proteins. The histidine tags allowed
purification of recombinant proteins by affinity chromatography on Ni-NTA resin (Qiagen)
under denatured conditions [322].
Supernatants from the mouse fibroblast cell line NIH3T3, the prostate stromal cell line
WPMY-1, and the melanoma cell line A375M were also included for comparison.
50
2.14. ENZYME-LINKED
IMMUNOSORBENT
ASSAY
(ELISA)
FOR
AUTOANTIBODY IDENTIFICATION IN PATIENT PLASMA OR SERUM
Maxisorp 96-well plates (Nunc, Germany) were coated with 4g/well of recombinant EN2 in
0.1M carbonate buffer (33.5mM Na2CO3, 0.1M NaHCO3, pH 9.6) and incubated overnight at
4C with gentle rocking. The plates were then blocked with 3% BSA in PBS containing
0.05% Tween-20 (3% BSA/PBST) for 2 hours at room temperature on a rapid shaker.
Subsequently, each patient or control serum/plasma sample was diluted 1:100 in 3%
BSA/PBST, with 100ul added to the wells in triplicate. The plate was incubated for 1 hour at
room temperature on a rapid shaker then washed 4 times in PBS/0.1% Tween-20. The goat
anti-human IgG antibody conjugated to horseradish peroxidase (Jackson, USA) was then
diluted 1:10,000 in 3% BSA/PBST with 100ul added to each well, and incubated for 1 hour
at room temperature on a rapid shaker. The plates were developed with tetramethylbenzidine
(TMB) (Sigma-Aldrich, UK) for 5 minutes at room temperature.
The reaction was
terminated by addition of 25µl of 2M H2SO4. The optical density at 450nm was measured
using a Varioskan Flash plate reader (Thermo Scientific, USA).
Individual patient sera/plasma readings were scored as EN2 positive if their OD value was
equal to or greater than three standard deviations from the mean of the control subjects.
For control comparison of unrelated cancers, samples of sera from 2 cohorts of breast cancer
patients and 1 cohort of prostate cancer patients were also analysed, along with healthy
control samples. The same protocol was also used to measure the spontaneous IgG immune
response against the comparative tumour antigen, NY-ESO-1.
2.15. WESTERN BLOTTING ANALYSIS FOR PROTEIN IDENTIFICATION IN
CELL LINE LYSATES AND SUPERNATANTS
The cell line lysates or supernatants were acquired as described in Section 2.13. The protein
concentration of each sample was determined compared to a protein standard, using the
Pierce BCA Protein Assay kit (Thermo Scientific, USA) according to the manufacturer’s
instruction. After incubation with the colorimetric agent for 30 minutes at 37°C, the optical
density at 562nm was measured, using a Varioskan Flash plate reader (Thermo Scientific,
USA). To enable uniform protein loading for each Western Blot, the samples were diluted to
the required concentration in their specific diluent, ensuring that a minimum of 5μg protein
51
was analysed for each experiment. Thirteen micro-litres of each sample was added to 2μl of
NuPAGE® Sample Reducing Agent (Life Technologies, UK) and 5μl of NuPAGE® LDS
Sample Buffer and placed on a heat block at 70°C for 10 minutes. The apparatus was
assembled for Sodium dodecyl sulphate-polyacrylamide gel electrophoresis (SDS-PAGE),
utilising NuPAGE® 4-12% Bis-Tris gels with MOPS SDS running buffer (Life
Technologies, UK). The Novex® Sharp Pre-stained marker (Life Technologies, UK) and
20μl of diluted cell line lysate/supernatant were loaded onto appropriate lanes of the gel.
Electrophoresis was carried out at 200V for 45 minutes. Proteins from the gel were then
transferred to a nitrocellulose membrane using the iBlot® Gel Transfer Device and Transfer
Stacks (Invitrogen, UK), electroblotting at 20V for 7 minutes, according to the
manufacturer’s instruction. The membranes were blocked, either overnight or for a minimum
of 2 hours, at 4°C in 5% milk in PBS/0.1% Tween-20, with gentle shaking. They were then
probed with the primary polyclonal goat anti-EN2 antibody (Abcam, UK) diluted 1:1666 in
blocking buffer, for 1 hour at room temperature with gentle shaking. The membranes were
washed for 4x10 minutes with PBS/0.1% Tween-20, and then probed with the secondary
donkey anti-goat IgG antibody conjugated to horseradish peroxidase (Jackson, USA) diluted
1:10,000 in blocking buffer, for 1 hour at room temperature with gentle shaking. This was
followed by 3x10 minute washes with PBS/0.1% Tween-20. The membranes were exposed
to SuperSignal® West Pico chemiluminescent substrate (Thermo Scientific, USA) for 1
minute, before a 1-10 minute exposure using the Chemi Doc-IT2 imager (UVP, UK).
Detected protein bands could then be visualised alongside the Novex® Sharp Pre-stained
marker, allowing confirmation of the molecular weight.
The recombinant EN2 (rEN2) protein was used as a positive control, with the later addition
of Rat Whole brain lysate (Abcam, UK) as an additional positive control, as suggested by
Abcam, UK who produced the primary antibody. Cell lysate and supernatant from the
normal human skin fibroblast cell line was used as a negative control. Supernatants from the
mouse fibroblast cell line NIH3T3, the prostate stromal cell line WPMY-1, and the
melanoma cell line A375M were also included for comparison.
52
Stripping blots for antibody re-probing
On certain occasions the nitrocellulose membrane was stripped in order to re-probe with a
second primary antibody, most commonly β-actin, to ensure equal loading of protein in all
lanes.
After initial exposure to the chemiluminescent substrate and subsequent imaging of the
protein bands, the membrane was placed in PBS/0.1% Tween-20 wash buffer for 5 minutes.
RestoreTM PLUS Stripping Buffer (15mls; Thermo Scientific, USA) was decanted into a tray
containing the membrane and rotated very slowly for 15 minutes. The membrane was then
rinsed in running water for 1 minute before 3 x 5 minute washes in Wash buffer, and block in
5% milk in PBS/0.1% Tween-20 at 4°C for 1 hour. The membrane was then probed with the
primary monoclonal mouse anti-β-actin antibody (Sigma-Aldrich, UK) diluted 1:2000 in
blocking buffer, for 1 hour at room temperature with gentle shaking. The membranes were
washed for 4x10 minutes with PBS/0.1% Tween-20, and then probed with the secondary
rabbit anti-mouse IgG antibody conjugated to horseradish peroxidase (Life Technologies,
UK) diluted 1:2000 in blocking buffer, for 1 hour at room temperature with gentle shaking.
This was followed by 3x10 minute washes with PBS/0.1% Tween-20. The membranes were
developed as previously described.
2.16. OVER-EXPRESSION OF EN2 IN THE PEA1 CELL LINE
PEA1 cells were seeded in 6-well tissue culture plates in their standard media, at 1.5x105
cells/well (1ml per well) made up to a total of 3ml with media, and incubated overnight at
37°C and 5% CO2. Before proceeding with the transfection, cells were observed under a
light microscope to ensure a 60-80% confluence.
The En2 DNA plasmid (TrueClone OriGene SC303017, En2 untagged, neomycin resistant;
Origene, USA) was reconstituted in water to 100ng/µl. En2 plasmid (0.2µg/well), OptiMEM
(Gibco by Life Technologies, UK) and PLUSTM reagent (Life Technologies, UK) were then
mixed together and incubated at room temperature for 5 minutes. A transfection only (no
DNA) control was also prepared by adding the appropriate volumes of Opti-MEM and PLUS
reagent.
53
Lipofectamine LTX (Life Technologies, UK) was subsequently diluted in Opti-MEM to give
a concentration of 2.5µl per well. This was vortexed and incubated at room temperature for 5
minutes, before combining with the DNA or transfection-only samples in a 1:1 volume ratio.
The solutions were incubated for 30 minutes at room temperature. During this incubation
time, the cell culture media was aspirated from the wells of the 6-well plate and replaced with
Opti-MEM, before returning the plate to the 37°C, 5% CO2 incubator. The wells of the plate
were aspirated and replaced with 200µl of the relevant transfection mix and 800µl OptiMEM. Opti-MEM alone was also added to appropriate wells for use as a non-transfection
control. The plate was placed on a plate shaker for 3 minutes to aid mixing before being
returned to the 37°C, 5% CO2 incubator for 5.5 hours. After this time, the transfection agents
were aspirated from the wells of the plate and replaced with 2ml of normal cell culture
medium before further incubation at 37°C, 5% CO2 for 24 hours.
The next day, the visual appearance of the cells was viewed by light microscope to check for
any signs of toxicity, prior to RNA extraction (Section 2.4.), cDNA synthesis (Section 2.6.)
and determination of En2 mRNA expression by semi-quantitative rt-PCR (Section 2.10.).
EN2 protein expression was also evaluated in cell lysates (protocol described in Section
2.13.), via Western Blotting (Section 2.15.).
Once over-expression of En2 was confirmed, the protocol was repeated in order to set-up
stably transfected cell lines harbouring En2 mRNA over-expression. On Day 3 i.e. 24 hours
after removal of the transfection agents, the cells were trypsinised and seeded in round tissue
culture dishes, in standard media. After a further 24 hours, the media was replaced with
selection media i.e. serum-free standard media with 200µg/ml of Geneticin disulfate salt
(G418) antibiotic (Sigma-Aldrich, UK). The plates were reviewed daily, observing an initial
period of predominant cell death caused by the antibiotic treatment, with subsequent
emergence of circular cell clones. These cell clones were assumed to contain the transfected
En2 gene, hence the neomycin resistance gene which prevented them from being killed by
G418. The clones were transferred to 96-well plates under sterile conditions, using a pipette
tip, and allowed to grow to 80% confluence, before serially increasing the culture dish
volume. Once confluent in 175cm3 flasks, the cells were trypsinised, prior to RNA extraction
(Section 2.4.), cDNA synthesis (Section 2.6.) and rt-PCR analysis (Section 2.10.). Cell
lysates were also acquired from a number of clones (protocol described in Section 2.13.), and
analysed for EN2 protein expression via Western Blotting (Section 2.15.).
54
2.17. ENZYME DEGLYCOSYLATION OF PROTEINS
This protocol utilised the PNGase F enzyme, O-Glycosidase, Neuraminidase, β1-4
Galactosidase and β-N-Acetylglucosaminidase to remove almost all N-linked and simple Olinked glycans from glycoproteins (Protein Deglycosylation Mix, New England BioLabs,
USA).
Using the cell lysates from PEA1 En2 DNA plasmid transfected cells (protocol described in
Section 2.13.), 100μg of protein was dissolved in 18μl of water and 2μl of 10X Glycoprotein
denaturing buffer was added, according to the manufacturer’s protocol. The glycoprotein
was heated at 100°C for 10 minutes, and then chilled on ice and centrifuged for 10 seconds.
A mixture of reaction buffers, water and deglycosylation enzyme cocktail was then added
with gentle mixing, and the reaction was incubated at 37°C for 4 hours.
Prior to analysis via Western Blotting, 5μl of NuPAGE® LDS Sample Buffer (Life
Technologies, UK) was added to the final reaction volume. Equal protein concentrations
were loaded on to the NuPAGE® 4-12% Bis-Tris gel, with the same concentration of
untreated protein added as a control. The SDS-PAGE, transfer steps and antibody probing
were carried out as documented in Section 2.15., however a concentration of 1:1250 primary
polyclonal goat anti-EN2 antibody was used, given the resultant small concentrations of
deglycosylated protein loaded on the gels.
Bovine fetuin, a glycoprotein containing sialylated N-linked and O-linked glycans, was used
as a positive control for endoglycosidase enzymes. The enzyme deglycosylation protocol
was carried out as above, with subsequent SDS-PAGE, however instead of proceeding with
protein transfer onto a membrane, the gel was placed in 10ml of Coomassie Blue dye (SigmaAldrich, UK) and left overnight at room temperature, with gentle shaking. The following
morning, the gel was washed with Coomassie wash solution (Sigma-Aldrich, UK) until bands
were clearly visible, and then imaged using the Chemi Doc-IT2 imager (UVP, UK) with the
trans-illuminator platform.
55
2.18. siRNA-MEDIATED EN2 SILENCING IN CELL LINES
The efficiency of siRNA transfection is strongly influenced by the concentration of
transfection reagent and the number of cells plated, therefore the KDalert™ GAPDH Assay
kit (Life Technologies, UK) was initially used to optimise these two parameters for
subsequent siRNA knockdown of En2.
2.18.1. KDalert™ GAPDH Assay to detect silencing of GAPDH (glyceraldehyde-3phosphate dehydrogenase)
The GAPDH siRNA and negative control #1 siRNA were diluted to a concentration of 2µM
in nuclease water, and mixed thoroughly. The diluted siRNAs were stored at -20°C for up to
3 months. For each set of transfection conditions, 3 replicate transfections were included.
The transfection conditions included:

GAPDH siRNA

Negative control #1 siRNA

Non-transfected control (cells that are mock-transfected with Opti-MEM
medium, but no siRNA).
Cells were trypsinised and re-suspended in cell culture medium at 1.5x105 cells/ml, but then
diluted to enable the later addition of 4, 8 and 12x103 cells per well (80μl/well) to a 96-well
plate. The re-suspended cells were stored at 37°C and 5% CO2 until required. The siPORT
NeoFX transfection agent (Life Technologies, UK) and Opti-MEM were warmed to room
temperature before use. The siPORT Neo FX was diluted in Opti-MEM medium to provide
concentrations of 0.2, 0.5 and 0.8µl per well, prior to incubation at room temperature for 10
minutes. The GAPDH and negative control #1 siRNAs (2μM stock) were also diluted in
Opti-MEM to provide 1.5μl per well, and incubated at room temperature for 10 minutes. The
diluted siPORT Neo FX transfection agent was then mixed with each diluted siRNA in a 1:1
volume ratio, and incubated at room temperature for another 10 minutes.
The
siRNA/siPORT Neo FX complexes were added to a 96-well tissue culture plate (20μl/well)
with Opti-MEM added as a non-transfection control. The incubated cells were then added to
the appropriate wells of the 96-well plate (80μl/well), which was gently mixed for a few
seconds on a plate shaker before incubation at 37°C and 5% CO2. After 24 hours, the
56
contents of the wells were aspirated and replaced with fresh culture medium before
incubation for a further 24 hours.
The visual appearance of the cells was viewed by light microscope to check for any signs of
toxicity, prior to evaluation of GAPDH knockdown using the KDalert™ GAPDH Assay kit.
The KDalertTM master mix was prepared according to the manufacturer’s instructions and
stored on ice until required. The media was aspirated from the wells of the 96-well plate and
replaced with 100µl of Lysis buffer, prior to incubation at 4°C for 20 minutes. The lysate
was then pipetted up and down several times to homogenise the lysate. Each cell lysate
(10µl) was transferred to a new 96-well tissue culture plate, with 10µl of water added to
control wells. The KDalertTM master mix (90µl) was added to each well and incubated at
room temperature for 15 minutes before reading the absorption of light/optical density (OD)
at 620nm using a Varioskan Flash plate reader (Thermo Scientific, USA).
The transfection conditions producing the greatest percentage of GAPDH knockdown with
the least cell toxicity in the given cell line, were then used in subsequent gene silencing
experiments. In the case of the PEA2 cell lines, this was 8000 cells/well, 1.0μl/well of
siPORT Neo FX and 1.5μl/well of siRNA.
2.18.2. siRNA-mediated En2 silencing in the PEA2 cell line
The KDalert™ GAPDH Assay protocol was followed using 8x103 cells per well (80μl/well),
1.0μl/well of siPORT Neo FX and 1.5μl/well of the En2 siRNAs #S4674, #4675 and #4676,
and a negative control siRNA (2μM stock; Life Technologies, UK). The siRNAs were all
diluted in standard serum-free media.
After addition of the siRNA/siPORT Neo FX
complexes to the 96-well tissue culture plate (20μl/well), serum-free standard media was also
added as a non-transfection control. As with the KDalert™ GAPDH Assay protocol, the
medium in all wells was aspirated after 24 hours and replaced with fresh culture medium,
before incubation for a further 24 hours.
Subsequently, the visual appearance of the cells was viewed by light microscope to check for
any signs of toxicity, prior to RNA extraction (Section 2.4.), cDNA synthesis (Section 2.6.)
and determination of En2 mRNA expression by semi-quantitative rt-PCR (Section 2.10.).
EN2 protein expression was also evaluated in cell lysates (protocol described in Section
2.13.), via Western Blotting (Section 2.15.).
57
2.19. 50% TUMOUR INHIBITORY CONCENTRATION (IC50) ESTIMATION OF
CISPLATIN IN CELL LINES
The PEA1 and PEA2 cell lines, along with selected PEA1 En2 DNA plasmid transfected
cells were grown to 70% confluency in a 96-well tissue-culture plate. Cisplatin (3.3mM
stock) was kindly donated by The Royal Surrey County Hospital and serially diluted in 2%
standard media to give concentrations of 50, 25, 12.5, 6.25 and 3.125µM. Each cell line was
treated with these cisplatin concentrations for 72 hours (100µl) at 37°C, as well as 2%
standard media alone as a control. This media was also added to empty wells to serve as a
‘Background’ control.
After the 72 hour incubation, 100µl of MTS reagent ((3-(4,5-
dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium;
Promega, UK) mixed with phenazine methosulfate, was added to selected wells and left for
45 minutes. The cell viability was determined by measuring the absorption of light/optical
density (OD) at 490nm using a Varioskan Flash plate reader (Thermo Scientific, USA) and
plotting a dose-response curve indicating the percentage cell survival relative to increasing
concentrations of cisplatin. The IC50 could then be obtained for each cell line with values
used to compare the levels of cytotoxicity between the different cell line groups.
2.20. MICROARRAY-BASED GENE EXPRESSION ANALYSIS OF EN2 OVEREXPRESSED CELL LINES
The Agilent One-Color Microarray-based Gene Expression Analysis was carried out in order
to investigate the difference in gene expression between two En2 over-expressed PEA1 cells
(clones E12 and H3) and wild-type PEA1 cells, as well as wild-type PEA2 cells. This
technique uses cyanine 3-labeled targets to measure gene expression. All of the kits and
equipment used for the gene expression analysis were obtained from Agilent Technologies,
USA, unless otherwise specified. The workflow for sample preparation and array processing
is shown in Figure 2.4.
58
Figure 2.4. Workflow for RNA sample preparation and microarray processing (taken
from Agilent One Color Microarray-Based Gene Expression Analysis Protocol;
http://www.chem.agilent.com/Library/usermanuals/Public/G4140-90040_GeneExpression_OneColor_6.7.pdf).
Total RNA from wild-type PEA1 and PEA2 cells along with two selected En2 overexpressed PEA1 clones (E12 and H3; Section 2.16), was prepared as previously described
(Section 2.4.). The RNA Spike-In Kit, One Color, and the Low Input Quick Amp Labeling
Kit, One Color were used to generate fluorescent complimentary RNA (cRNA) from 100ng
of this total RNA, according to the manufacturer’s instructions. The method uses T7 RNA
Polymerase Blend, which simultaneously amplifies and incorporates Cyanine 3-CTP into the
target material. The RNeasy® Mini Kit (Qiagen, UK) was used to purify the amplified
cRNA samples, prior to elution in 30μl of RNase-free water and quantification using a
Nanodrop ND-1000 Spectrophotometer. The Cyanine 3 dye concentration (pmol/μl) and
cRNA concentration (ng/μl) values were recorded for each sample, in order to determine the
59
yield of cRNA (μg) and specific activity (pmol Cy3 per μg cRNA) of each reaction using the
following calculations: (Concentration of cRNA) x 30μl (elution volume)
1000
Concentration of Cy3 x 1000
Concentration of cRNA
= μg of cRNA
= pmol Cy3 per μg cRNA
All of the samples achieved the threshold cRNA yield and specific activity to enable us to
proceed to the hybridization stage, using the Gene Expression Hybridization Kit and a whole
human genome microarray (“4x44K”) containing 44,000 probes, according to the
manufacturer’s instruction. The Cy3-labeled cRNA was initially fragmented before loading
onto the microarray slide, which was then encased within the hybridization chamber. This
was placed in rotisserie in a hybridization oven set to 65°C for 17 hours. The microarray was
subsequently washed in Gene Expression Wash Buffers, the hybridization chamber
disassembled, and the slide scanned in Cy3 channel using the Agilent DNA Microarray
Scanner (model G2505B).
The data was analysed by Dr Carla Moller-Levet, a Bioinformatics Experimental Officer at
the University of Surrey, using the publicly available programming environment R (R Core
Team, Austria)[323] and the commercial software MetaCoreTM (Thomson Reuters, USA).
Log2 values were quantile-normalized using the R Bioconductor package, Limma [324].
Probes with more than one flagged sample (Feature flags from Agilent Feature Extraction
Software; Version 9.5.1) were filtered out from the analysis. Probes with an absolute fold
change larger than three (>3fold) in both of the En2 over-expressed PEA1 clones compared
with wild-type PEA1 were classed as differentially expressed, with generation of a
hypermetric test p-value of intersection for both up- and down-regulated probes. The GeneGo
software from MetaCoreTM was then used to investigate the biological functions of these
preferentially up- or down-regulated genes, enabling the construction of gene pathway maps
and process networks.
60
2.21. STATISTICAL ANALYSIS
The GraphPad Prism software package was used to draw graphs and for all statistical
calculations, unless otherwise stated. The significance level for all tests was set at 5%, hence
a p-value of <0.05 could be deemed significant. The following symbols were used in
diagrams/figures to denote levels of significance: *=p<0.05; **=p<0.01; ***=p<0.001; ****=p<0.0001.
2.21.1. Analysis of En2 mRNA expression differences in cell lines and human tissue
The one-way ANOVA with Bonferroni correction was performed to evaluate the statistical
significance of differences in En2 mRNA expression of the cell lines, using the mean CT
value relative to ß-actin [239, 325, 326]. When comparing the expression levels between the
different histological sub-types of human tumour tissue, the non-parametric Kruskal Wallis
test with Dunn’s correction was used, as many of the groups were not normally distributed
and some were small in number. For comparison of two groups, the non-parametric MannWhitney U-test was used. Survival analysis was performed using the Log-rank (Mantel Cox)
test.
2.21.2. Analysis of EN2 protein expression in human tumours and tissue arrays
Comparative analysis of EN2 positive and negative tumour sections and tissue cores was
conducted using the Chi-squared test [327]. Survival analysis was performed for the human
tumours and CJ2 tissue array, using the Log-rank (Mantel Cox) test.
2.21.3. Analysis of EN2 protein expression in patient urine
Urine samples were deemed to have positive EN2 expression if the EN2 concentration lay
above the positive cut-off value, calculated as the mean plus 2 standard deviations of the
control group. The difference in mean EN2 concentration between patients and controls was
compared using the unpaired t-test with Welch’s correction.
61
2.21.4. Analysis of EN2 and NY-ESO-1 antibody levels in patient plasma
The IgG antibody responses to the recombinant proteins EN2 and NY-ESO-1 were deemed
positive if the value lay above the positive cut-off value, calculated as the mean plus 3
standard deviations of the control cohorts. The proportions of IgG responders to each antigen
were compared between patients and controls using the Chi-squared test.
2.21.5. Analysis of En2 promoter methylation in cell lines and human tumours
The Mann-Whitney U-test was used to compare the mean percentage En2 methylation for
invasive EOC tumours compared with borderline tumours, benign tumours and normal
ovary/fallopian tube specimens, respectively. This test was also used to compare the mean
percentage En2 methylation status of platinum resistant and platinum-sensitive tumours. A
linear regression analysis was performed to evaluate correlation between En2 promoter
methylation status and En2 mRNA expression in cell lines and human tumours. Survival
analysis of high versus low percentage En2 methylation in human tumours was performed
using the Log-rank (Mantel Cox) test.
2.21.6. Analysis of microarray data
As previously described in Section 2.20., the microarray data was analysed by Dr Carla
Moller-Levet, a Bioinformatics Experimental Officer at the University of Surrey, using the
publicly available programming environment R and the commercial software MetaCoreTM.
Log2 values were quantile-normalized using the R Bioconductor package, Limma. Probes
with a >3fold absolute change in both of the En2 over-expressed PEA1 clones compared with
wild-type PEA1 were classed as differentially expressed, with generation of a hypermetric
test p-value of intersection for both up- and down-regulated probes.
62
CHAPTER 3
EN2 EXPRESSION IN EPITHELIAL
OVARIAN CANCER CELL LINES
63
3.
EN2 EXPRESSION IN EPITHELIAL OVARIAN CANCER CELL LINES
3.1.
INTRODUCTION
Engrailed 2 (EN2) is a homeodomain-containing transcription factor which is known to play
a pivotal role in embryonic neural development and it can be identified in the nuclei of the
Purkinje neurones within the normal adult brain, although it has no defined role in the adult
human [290].
Its developmental functions may involve transcriptional and translational
regulation, as well as secretion and internalization [270].
Elevated expression of Engrailed 2 at the mRNA level has been demonstrated in several solid
tumours, including breast, prostate and bladder specimens. The resultant protein product has
also been identified and localised within the epithelial cells, although it is frequently seen in
the cytoplasm rather than within the nucleus. Much of this early work has been carried out
using cell lines, which represent the growing human tumour cells within a controlled, sterile,
in vitro environment. This method allows the researcher to study the presence of a gene, as
well as investigating its function under varying environmental and chemical conditions.
Martin et al demonstrated ectopic En2 expression in five breast cancer cell lines derived from
breast adenocarcinoma and ductal carcinoma, however only 7% of human specimens of these
same histological sub-types had elevated mRNA expression [301].
Non-tumourigenic
murine mammary cell lines did not express En2, neither did normal breast epithelium.
Bose et al also used cell lines in their initial evaluation of En2 expression in prostate cancer
prior to investigating the functional role of the gene [302]. Along with three epithelial
prostate cancer cell lines, they cultured human prostate epithelial (hPrEC) cells, and
demonstrated significantly higher En2 mRNA levels in the malignant cell lines compared
with the normal cells. These findings were confirmed at the cell line protein level. Certainly
in the case of prostate cancer, cell line work appears to directly mirror what is seen in human
prostate tumour biopsies and normal prostate tissue [303], and the majority of prostate
tumours stain for EN2 protein in the cytoplasm and basal membrane of the ductal structures,
with the notable appearance of EN2 containing blebs in prostatic acini and ducts suggesting
secretion into ductal lumen.
64
Although there is no published work regarding En2 expression in ovarian cancer, altered
expression of members of the HOX gene family, also homeodomain-containing transcription
factors, has been observed [236-238]. This family of 39 genes are implicated in apoptosis,
receptor signalling, differentiation, motility and angiogenesis in foetal development.
Overexpression of HOXA9 has been identified in the serous histological sub-type of EOC,
whereas HOXA10 is present in the endometrioid sub-type, and HOXA11 in the mucinous subtype [236]. Loss of expression of HOXA5 and HOXA9 has been associated with tumour
survival in EOC, as a result of resistance to apoptosis [240], however more recently, HOXA9
expression has been shown to increase in advanced disease, suggesting that its function may
change as the disease progresses [328]. Expression of the HOX genes often differs between
EOC cell lines, even if they represent the same histological sub-type, for example HOXA13,
HOXB5, HOXB9 and HOXD9 are all up-regulated in the SKOV3 serous cell line, but not in
the OV90 serous cell line [239].
Currently there is very little data regarding EN2 expression in ovarian cancer, although early
unpublished experiments looked at En2 mRNA levels in EOC cell lines and EN2 protein
levels in tumour specimens, suggesting an elevated expression compared with normal tissue
(personal communication with Professor R. Morgan, Oncology Department, University of
Surrey). However, there has not yet been sufficient work carried out to enable histological
sub-type analysis, or to accurately localise the EN2 protein within the cell.
Owing to the pelvic location of ovarian tumours and the preponderance for direct peritoneal
spread, the resultant malignant cells within ascites are often used to confirm diagnosis and
plan treatment, even if a direct tumour biopsy is not feasible. Therefore, ovarian cancer cell
lines derived from such ascitic fluid malignant cells are frequently used in the laboratory, to
represent EOC. Many commercial EOC cell lines are available representing the major
histological sub-types, with some harbouring platinum-resistance. Most of these are derived
from the ascitic fluid of patients with advanced ovarian cancer. Hence it appears logical to
evaluate the presence of En2 in such cell lines and to confirm the location of EN2 protein
within the cells, prior to studying its expression in human tumour specimens.
65
Study objectives and hypothesis
The objectives of this chapter were to evaluate the En2 gene expression as well as EN2
protein expression in a broad selection of EOC cell lines.
We hypothesised that En2 would be over-expressed in a number of EOC cell lines, in
comparison to a control human fibroblast cell line, and that the resultant protein product
would also be detectable within the cell. This level of expression may vary across the
different histological sub-types, and when comparing platinum-sensitive with platinumresistant cell lines. We looked at the expression of En2 at the mRNA level and at the protein
level in the same cell lines, with additional focus on localisation of the protein within the cell.
66
3.2.
RESULTS
3.2.1. En2 gene expression in epithelial ovarian cancer (EOC) cell lines
RNA was extracted from cultures of fourteen EOC cell lines representing a variety of
histological sub-types. The platinum-sensitivity status was also recorded for these cell lines.
A human normal epithelial ovarian cell line was not available therefore a human fibroblast
cell line was used as a negative control, and the A375M cell line, derived from melanoma,
was used as a positive control. The A375M cell line had previously been evaluated for En2
gene expression by other members of the research team. The cDNA template from all
samples was used to quantify En2 expression via quantitative PCR. The relative expression
value was calculated as a ratio with the housekeeping gene ß-actin (Table 3.1).
Histology
Serous
Endometrioid
Clear Cell
Cell Line
Platinum Status
En2 relative expression
OV90
Sensitive
41.14
SKOV3
Resistant
2.95
CaOV3
Sensitive
7.18
OVCAR3
Resistant
5.66
COV318
Sensitive
0.33
PEO1
Sensitive
43.24
PEO4
Resistant
289.55
PEO14
Sensitive
53.86
PEO23
Resistant
95.19
PEA1
Sensitive
8.37
PEA2
Resistant
177.25
TOV112D
Resistant
17.40
ES2
Resistant
5.71
TOV21G
Sensitive
99.14
Fibroblasts
-
0.35
A375M
-
413.74
Controls
Normal skin
Melanoma
Table 3.1. En2 gene expression in EOC cell lines and controls. The relative expression of the
En2 gene was determined by quantitative rt-PCR, with results expressed as a ratio with the housekeeping gene
-actin (x100,000), and the mean of three independent experiments. The histological sub-type and platinumsensitivity status for each EOC cell line is listed. PEO1 and PEO4 represent platinum-sensitive and –resistant
paired cell lines, as do PEO14 and PEO23, along with PEA1 and PEA2.
67
PEO1 represented a cell line derived from a platinum-sensitive, serous ovarian carcinoma.
The PEO4 cell line was derived from the same patient, once they had developed platinumresistant disease. Similarly, PEO14 and PEO23, and PEA1 and PEA2, were paired cell lines
derived from the same patient before and after developing platinum-resistant disease. For
each cell line, the mean Cycle Threshold (CT) value relative to ß-actin was calculated for
three independent rt-PCR experiments, representing En2 relative expression.
Evaluation of the EOC cell lines revealed En2 overexpression in the majority of cell lines,
but to a varying degree (Figure 3.1). The serous EOC lines PEO4 and PEA2 demonstrated
significantly higher En2 expression than fibroblasts (p<0.0001 and p<0.01 respectively). The
platinum-sensitive/resistant paired serous cell lines all showed elevated En2 expression,
however there was a significant increase in the platinum-resistant lines, compared with their
platinum-sensitive pairings. This was most notable with the PEO1/PEO4 (p<0.0001) and
PEA1/PEA2 (p<0.05) paired cell lines. En2 expression in the A375M positive control was
significantly higher than that in the fibroblast line (p<0.0001).
****
****
**
500
r e la t iv e t o B - a c t in
M e a n C T v a lu e
C o n tro ls
400
****
S e ro u s
300
*
E n d o m e trio id
200
C le a r C e ll
100
P la tin u m -re s is ta n t
50
b
ro
ib
7
3
F
A
la
s
ts
5
O M
S V9
K
0
O
C V
3
a
O O
V V
C 3
C A
O R
V 3
3
P 18
E
P O
E 1
P O
E 4
P O
E 1
O 4
P 23
E
T PEA1
O
V A2
T 11
O 2
V D
2
1
G
E
S
2
0
Figure 3.1. En2 mRNA expression in EOC cell lines and controls. Fourteen ovarian cancer
cell lines were analysed by quantitative rt-PCR. Fibroblasts were used as a negative control and the A375M cell
line as a positive control. The En2 mRNA expression is shown relative to the housekeeping gene β-actin
(x100,000). Error bars represent the SD (n=3) and the one-way ANOVA with Bonferroni correction was used
for analysis (*=p<0.05; **=p<0.01; ****=p<0.0001).
68
3.2.2. EN2 protein expression in epithelial ovarian cancer (EOC) cell lines
3.2.2.1. Immunohistochemical staining of EOC cell lines for the expression of EN2
protein
In order to assess the expression and location of EN2 at the protein level in ovarian cancer
cell lines, enzymatic immunofluorescent staining was performed on fourteen EOC cell lines,
along with normal skin fibroblasts and the melanoma cell line A375M. The individual
histological sub-types are as listed in Table 3.1.
This immunofluorescent staining and confocal analysis demonstrated that EN2 was expressed
at the protein level in all of the EOC cell lines tested, as well as in A375M melanoma cells,
however it was absent in the normal skin fibroblast cell line (Figures 3.2, 3.3, and 3.4). The
nuclear marker, TO-PRO-3, and the cell membrane marker, wheat germ agglutinin (WGA),
were used to help localize the EN2 protein within the cells. The cells were either nonpermeabilised to enable improved visualisation of membrane staining, or permeabilised with
Triton X to enable improved cytoplasmic and nuclear EN2 visualisation. EN2 appeared to be
predominantly expressed within the cytoplasm of the cancer cell lines. When the cells were
permeabilised, a stronger cytoplasmic staining was evident in most of the cell lines. If
nuclear EN2 was present, the co-localisation of nuclear TO-PRO-3 and EN2 protein staining
would result in an aquamarine colour pigment. This was not clearly evident in any of the cell
lines stained, and the intensity profile supports this (Figure 3.5). Co-localisation of cell
membrane WGA and EN2 protein staining resulted in a yellow pigment and this was evident
in certain cells (Figure 3.3, blue arrows). The intensity profile from the PEO4 cells supports
this (Figure 3.6). This was likely to represent EN2 protein in close proximity to the cell
membrane although we could not confidently say whether it was located on the outer or inner
surface.
Although the images were all taken keeping the confocal laser power and master gain
readings constant, it was difficult to accurately compare the staining intensity between the
cell lines, as they were not necessarily stained at the same time, or imaged at the same time
interval. Despite that, there appeared to be a correlation between the intensity of protein
staining and En2 gene expression, particularly with the low-expressers SKOV3, CaOV3 and
COV318, and the high expressers PEO4 and PEA2.
69
Figure 3.2. Immunofluorescent staining of serous epithelial ovarian cancer cell lines for
the expression of EN2. EN2 fluorescent staining (green) was present in the cytoplasm of the serous cell
lines, OV90, SKOV3, CaOV3, OVCAR3, and COV318. This was more clearly observed in those that had been
permeabilised with Triton X rather than the non-permeabilised cells. There was no EN2 staining evident in the
permeabilised negative control of each cell line, where only secondary antibody was added (inset). TOPRO
staining (blue) identified the cell nucleus and WGA staining (red) identified the cell membrane.
70
Figure 3.3. Immunofluorescent staining of the platinum-sensitive/-resistant paired serous epithelial ovarian cancer cell lines for the expression
of EN2. EN2 fluorescent staining (green) was present in the cytoplasm of the serous cell lines, PEO1, PEO4, PEO14, PEO23, PEA1 and PEA2. This was more clearly observed in
those that had been permeabilised with Triton X rather than the non-permeabilised cells. There was no EN2 staining evident in the permeabilised negative control of each cell line,
where only secondary antibody was added (inset). TOPRO staining (blue) identified the cell nucleus and WGA staining (red) identified the cell membrane. Co-localisation of cell
membrane WGA and EN2 protein staining resulting in a yellow pigment, can be seen in some of the non-permeabilised images (blue arrows).
71
Figure 3.4. Immunofluorescent staining of endometrioid and clear cell epithelial ovarian cancer cell lines, and negative & positive control cell
lines for the expression of EN2. EN2 fluorescent staining (green) was present in the cytoplasm of the clear cell cell lines TOV21G and ES2, and in the endometrioid cell line
TOV112D. This was more clearly observed in those that had been permeabilised with Triton X rather than the non-permeabilised cells. There was no EN2 staining evident in the
permeabilised negative control of each cell line, where only secondary antibody was added (inset). In the fibroblast negative control cell line, there was no EN2 staining evident,
however strong cytoplasmic staining was seen in the melanoma A375M positive control cell line. TOPRO staining (blue) identified the cell nucleus and WGA staining (red)
identified the cell membrane.
72
Figure 3.5. EN2 is not clearly demonstrated in the cell nucleus. The intensity profile measured
along the length of the white arrow demonstrates the three different fluorescent labels in a PEA2 cell. The red
peaks indicate the cell membrane, whilst the blue peak demonstrates the nucleus. The green intensity peak
shows the presence of EN2 in the cytoplasm, between the cell membrane and nucleus, but there is very low
intensity in the nucleus.
Figure 3.6. EN2 may be present in the EOC cell membrane. The intensity profile measured
along the length of the white arrow demonstrates the three different fluorescent labels in a PEO4 cell. The red
peaks indicate the cell membrane, whilst the blue peak demonstrates the nucleus. The green intensity peak
shows the presence of EN2, suggesting co-localisation with the cell membrane as the red and green peaks
directly overlap. In the cell, this co-localisation is shown as a yellow pigment.
73
3.2.2.2. Enzyme-linked immunosorbent assay (ELISA) analysis of EOC cell lines for the
expression of EN2 protein
In order to quantify the relative expression of EN2 at the protein level in the ovarian cancer
cell lines, a Direct ELISA was performed including whole cell lysates from the fourteen EOC
cell lines, along with normal skin fibroblasts.
The recombinant EN2 (rEN2) protein
developed in E. coli was used as a positive control. The EN2 protein concentration within
each cell lysate is illustrated in Figure 3.7(A), with the histological sub-type and platinum
status detailed.
EN2 expression in the rEN2 positive control was predictably high, whilst protein expression
in the fibroblast cells was negligible. Evaluation of the EOC cell line lysates revealed
demonstrable EN2 protein in the majority of cell lines, but to a varying degree. The serous
EOC lines PEO1 and -4, and PEA2 demonstrated particularly high concentrations of EN2
protein. Protein at a lower concentration was detectable in the serous cell lines OV90,
PEO14 and -23, and PEA1, along with the endometrioid cell line TOV112D and the clear cell
cell line TOV21G. However, there was no detectable protein in the serous SKOV3, CaOV3,
OVCAR3 and COV318 cell lines.
There was no clear correlation between the protein
concentration and platinum sensitivity within the cell lines.
These results for EN2 protein concentration in the cell lines did concur with the En2 mRNA
expression levels shown in Figure 3.1 (Section 3.2.1.).
74
3.2.2.3. Western Blotting analysis of EOC cell lines for the expression of EN2 protein
In order to further compare the expression of EN2 at the protein level between the ovarian
cancer cell lines and controls, gel electrophoresis and protein immunoblotting were
performed on the fourteen EOC cell lines, along with normal skin fibroblasts and the rEN2
protein (Figure 3.7(B)).
A
B
Figure 3.7. EN2 protein expression in EOC cell lines and controls, as determined by the
Direct ELISA and Western Blotting. Whole cell lysates from fourteen ovarian cancer cell lines were
analysed by Direct ELISA (A) and Western Blotting (B) techniques. Fibroblast cell line lysates were used as a
negative control and rEN2 as a positive control. The histological sub-type and platinum-sensitivity status for
each EOC cell line is depicted. The EN2 protein concentration (ng/ml) for each cell line was calculated relative
to a RIPA buffer standard curve. Uniform total protein loading was performed for the Western Blot, according
to the BCA Protein Assay, therefore the blot area and intensity reflect the relative EN2 protein concentration
between the cell lines.
75
The rEN2 positive control demonstrated a strong band at approximately 45kDa, whilst no
band was seen in the fibroblast cells. Evaluation of the EOC cell line lysates revealed
demonstrable protein bands at 50kDa in the majority of cell lines, but to a varying degree.
The serous EOC cell lines OV90, PEO1 and -4, PEO14 and -23, and PEA1 and -2
demonstrated particularly strong bands for EN2 protein. Weaker bands were detectable in the
serous cell lines OVCAR3 and COV318, along with the endometrioid cell line TOV112D
and the clear cell cell line TOV21G. However, there was no detectable protein in the serous
SKOV3 and CaOV3, and clear cell ES2 cell lines. There was no clear correlation between
the protein concentration and platinum sensitivity within the cell lines.
Aside from the OVCAR3 and COV318 samples, these results for EN2 protein concentration
in the cell lines did concur with the protein concentrations determined by the Direct ELISA
(Figure 3.7(A)), and the En2 mRNA expression levels shown in Figure 3.1 (Section 3.2.1.).
76
3.3.
DISCUSSION
The aims of this chapter were to evaluate the En2 gene expression and resultant protein
product in a number of cell lines representative of the major histological sub-types of
epithelial ovarian cancer. The majority of cell lines were derived from patient ascites and
represented poorly differentiated/high grade tumours, although the PEO14 and PEO23 cell
lines were derived from a well-differentiated/low grade serous adenocarcinoma [329]. The
serous histological sub-type is the most common EOC type in clinical practice, representing
over 60% of EOCs and two-thirds of all ovarian cancer deaths [330, 331]. This is reflected in
the greater availability of serous cell lines. We were able to obtain one cell line to represent
endometrioid carcinoma, TOV112D, and two cell lines representing clear cell carcinoma,
TOV21G and ES2.
EN2 appeared to be over-expressed in a number of EOC cell lines, both at the mRNA and
protein level. Eleven of the cell lines were derived from advanced serous ovarian cancers and
all but one of these demonstrated detectable En2 relative expression >1.0, although only
PEO4 and PEA2 were significantly elevated compared with the fibroblast control cell line.
The endometrioid and clear cell lines did express En2 but there were no significant
differences in expression compared with the fibroblast cell line. Neither was there a clear
difference in expression between these histological sub-types.
It does not come as a great surprise that some of the serous cell lines gave contrasting results,
as epithelial ovarian tumours are known to be heterogeneous and often demonstrate variable
genetic mutations [332, 333]. Cancer cell lines are ubiquitously used as in vitro tumour
models, however Domcke et al recently demonstrated that there are significant differences in
molecular profiles between many of the frequently used EOC cell lines, including high-grade
serous examples [334]. They analysed and compared 47 different EOC cell lines particularly
looking for the following major genomic features of HGSOC, frequent copy number
alteration, TP53 mutation, and low frequency of somatic mutation in protein-coding regions.
SKOV3, which is the most published EOC cell line, does not in fact possess a TP53
mutation, but does demonstrate PIK3 and ARID1A mutations, which are more commonly
associated with low-grade endometrioid or clear cell carcinomas [335, 336]. OVCAR3 and
OV90 cell lines contain the TP53 mutation, which supports their HGSOC phenotype,
however the frequency of copy number alteration was relatively low, hence the authors
77
concluding that they only possibly represent HGSOC. However, COV318 proved to be their
6th highest ranking cell line in terms of HGSOC likelihood, as it contained frequent copy
number alteration, and a mutated TP53, however this cell line had not been cited in any
PubMed articles at the time. Their analysis of the two clear cell cell lines used in our
research, TOV21G and ES2, demonstrated that the former is hypermutated with PIK3, PTEN
and ARIDA1 mutations but absence of TP53 mutation, supporting the clear cell phenotype.
Interestingly KRAS mutations were present which are present in the majority of mucinous
carcinomas [337]. Conversely, ES2 did not contain any of these mutations but did possess
mutated TP53 and BRAF, perhaps suggesting the presence of both low- and high-grade
serous features [41, 337]. Seemingly, the histological sub-type assigned to cell lines initially,
does not always match their molecular profiling, and this may stem from the fact that the cell
lines are often derived from ascitic fluid or metastatic deposits in the peritoneum, rather than
the primary ovarian tumour. It is also interesting to note that the twelve cell lines that most
resemble high-grade serous carcinoma in Domcke’s analysis, only account for 1% of
PubMed citations. Obviously this publication did not include all of the cell lines used for the
analysis of En2 in EOC, notably omitting the paired platinum sensitive/resistant cell lines, but
when considering their revised histological sub-typing, there still did not appear to be any
correlation with En2 mRNA expression levels.
Several of the EOC cell lines are known to be resistant to platinum chemotherapy. The cell
line pairings of PEO1 and PEO4, PEO14 and PEO23, and PEA1 and PEA2, are particularly
interesting as they are derived from tumours from the same patient when they are initially
platinum-sensitive and then when they clinically develop resistance.
The timeline for
derivation of these cell lines is shown in Figure 3.8 (the PEO6 cell line was not available to
us for this project). These cell line pairings have been well published since their derivation in
the early 1980s [329, 338] [339, 340]. Genomic analysis of these paired cell lines has
demonstrated that they do share some genomic features, for instance widespread pathogenic
TP53 mutations, however mutually exclusive genomic characteristics are also seen in
sensitive and resistant cells derived from the same patient [339]. These findings lend more
weight to the suggestion that genetically heterogeneous, platinum-resistant clones exist at low
level in the original patient tumour, even when this tumour is deemed to be platinumsensitive, rather than the direct linear development of resistant cells in response to
chemotherapy treatment. Hence the disease relapse results from eventual growth of the
resistant cell clones, which were not killed by the cytotoxic therapy. If such clones of
78
platinum-resistant cells are already present in the original tumour, albeit in small volume,
they may harbour particular gene signatures which can be detected prior to, or in the early
stages of treatment, and help to direct therapy.
Figure 3.8. A timeline demonstrating the acquisition of paired patient tumour samples
used to develop platinum-sensitive and –resistant cell lines (adapted from [339]). Cell lines
from 3 individuals were established at different time-points through disease progression. All patients received
platinum-based chemotherapy. The grey boxes indicated platinum-resistant cell lines.
Our quantitative real-time PCR analysis suggested a trend towards a higher level of En2
expression in the platinum-sensitive PEO1 and PEO14 cell lines compared to the other
sensitive serous cell lines, CaOV3 and COV318. It is known that PEO1 and PEO14 are
derived from patient tumours that subsequently develop resistance to platinum chemotherapy,
so this elevated En2 level may reflect expression within the resistant cell clones which are
already present within the tumour at an early stage, as suggested by Cooke et al [339].
Perhaps such levels of En2 mRNA expression are an early indicator of the likelihood of
platinum-resistance developing in these patient tumours. However this finding was not seen
in the PEA1 cell line.
As well as using the platinum-sensitive cells of the paired EOC cell line series to look for
predictive biomarkers of platinum resistance, comparative analysis of the sensitive and
resistant pairs can be conducted to identify alterations in genes and their protein products,
once resistance develops. In the case of En2 mRNA, higher levels of expression were seen in
each of the platinum-resistant cell lines compared to their sensitive pairing. In particular,
statistically significant En2 overexpression was seen in PEO4 compared with PEO1, and in
PEA2 compared with PEA1. The lack of a statistically significant increase in expression
between PEO14 and PEO23 could be explained by the fact that these cell lines were derived
79
from a low-grade serous tumour rather than high-grade serous tumours like the PEO1/4 and
PEA1/2 pairings. Although a clear increase in En2 mRNA expression occurred in these
platinum-resistant paired cell lines, at this stage we cannot say whether this overexpression
directly influences the development of platinum resistance, or whether it occurs as a result of
upstream genetic influence. In addition, it is noted that although there was a clear increase in
the level of En2 expression in the platinum-resistant pairings of these cell lines, this level of
over-expression was not seen in the other serous platinum-resistant cell lines SKOV3 &
OVCAR3, or in the endometrioid TOV112D cell line. Again, this may be explained by the
findings of Domcke et al, where SKOV3 was shown to bare very little resemblance to the
standard gene mutation signature of HGSOC [334]. Likewise, they suggested that OVCAR3
is only possibly representative of the HGSOC phenotype, but they did not analyse TOV112D.
As well as evaluation of the genetic expression of En2 within the EOC cell lines, the EN2
protein was qualitatively and quantitatively investigated. By studying differences in location
and concentration between the histological sub-types and platinum sensitive/resistant cells,
the functional roles of EN2 within the cell may be better understood. Early work involving
chick EN2 protein transfected into COS-7 fibroblast-like cells, demonstrated EN2
predominantly in the nucleus but with presence in the cytoplasm and membrane owing to
vesicular transport [271-273].
In the normal adult human, EN2 has been documented to be
present in the nucleus of the Purkinje neurones [290] and in the cytoplasm and some basal
membranes of the tubular epithelial cells of the kidney [341]. Secreted blebs staining for
EN2 were also noted in renal tubules and the lumen of the collecting tubules. In human
prostate tumours, EN2 has been identified in the cytoplasm and basal membrane of tumour
cells, with absence of nuclear expression [303], however in transitional cell carcinoma,
squamous cell carcinoma and adenocarcinoma sub-types of bladder cancer, cytoplasmic
staining with some nuclear positivity was detected [304]. In Clear cell renal cell carcinoma,
the level of EN2 expression was down-regulated compared with normal renal cells and
tubules, but any detectable staining was still within the cytoplasm with absence of any
nuclear staining [341].
The EN2 protein was detected in the cytoplasm of all of the EOC cell lines, along with the
A375M positive control, however it was absent in the fibroblast negative control cell line.
The intensity of expression appeared stronger in the cell line pairings of PEO1 and PEO4,
PEO14 and PEO23, and PEA1 and PEA2, along with the serous cell line OV90, when the
confocal microscope settings were maintained at equal levels. This was in keeping with the
80
demonstrated overexpression at the mRNA level, however it was difficult to draw too many
conclusions from this, as the cell lines were not all stained at the same time, and some cells
were imaged within 24 hours whereas others were imaged within 48 hours. The latter time
interval difference may have affected the intensity of the staining somewhat. The nonpermeabilised and permeabilised images from each cell line could be directly compared
however, as these were always prepared, stained and analysed at the same time and under the
same conditions.
Theoretically, the non-permeabilised process should enable better
visualisation of EN2 on the cell membrane, where the co-localisation of cell membrane
staining and EN2 staining results in yellow pigmentation. This is evident in limited areas in
the PEO1, PEO4 and PEO14 cell lines only. These are the same cell lines that demonstrated
over-expression of En2 at the mRNA level and so we would have anticipated a higher protein
product expression also. In these cell lines, the protein may undergo increased rates of
transport across the cell including secretion and internalisation. The membrane localisation
may therefore reflect this process. EN2 is known to function as a transcription factor so we
would normally expect to detect it in the cell nucleus, however this was not the case in these
examples of EOC cell lines. Perhaps this role only takes place in non-cancerous embryonic
neural cells, and it is blocked from entering the nucleus in cancer cells, maintaining a latent
cytoplasmic state. Conversely, EN2 may carry out a different function in EOC tumours and
this will be further addressed in Chapter 4.
The Direct ELISA and Western Blotting techniques were carried out in order to quantify the
relative expression of EN2 at the protein level in the ovarian cancer cell lines and controls.
These experiments demonstrated similar findings, with high EN2 protein concentration in the
paired cell lines PEO1 and PEO4, PEO14 and PEO23, and PEA1 and PEA2, along with the
serous cell line OV90 and the clear cell line TOV21G. EN2 expression in the rEN2 positive
control was predictably high, whilst protein expression in the fibroblast cells was negligible.
These results were consistent with the En2 mRNA expression levels, however there were no
apparent differences in protein expression between platinum-sensitive and platinum-resistant
cell lines.
STAT1 is another transcription factor that has recently been studied in the EOC paired cell
line series. It is not a homeobox transcription factor like EN2, but forms dimers in response
to Interferon (IFN) stimulation, before binding to the promoter element of genes so enhancing
expression. Stronach et al identified STAT1 and HDAC4 as two genes potentially involved in
the re-sensitization of platinum-resistant ovarian cancer cells [340]. Histone proteins such as
81
HDAC4 are known to regulate transcription as a result of de-acetylation, and as previous
reports had indicated that STAT1 could be regulated by acetyl modifications [342, 343], it
was thought that these genes and their protein products may physically interact in EOC
tumours. Ultimately they demonstrated an increased expression of HDAC4 in platinumresistant cells, which appeared to promote nuclear translocation and activation of STAT1 in
response to platinum exposure, resulting in enhanced cell survival. However when the
HDAC4 gene was silenced, STAT1 activation did not occur following platinum treatment and
improved cell kill was observed, i.e. cisplatin sensitivity was restored. Although the role and
function of HDAC and STAT1 differ considerably from EN2, their findings in the platinum
sensitive/resistant paired EOC cell lines bare some similarities to the results shown with EN2.
They identified an increase in HDAC4 expression in the resistant cell lines compared with the
sensitive pairings, which was supported by Western Blot data of HDAC4 protein expression,
however the WB images do not entirely correlate. For instance, the HDAC4 expression is
higher in the resistant PEO4 line compared with PEA2, however the protein band is much
stronger for PEA2 compared with PEO4. It is also stronger in the PEO14 cell line compared
with PEO4, despite a greater than three-fold increase in gene expression in the PEO4 line.
This finding highlights the fact that the protein concentration does not always mirror the level
of gene expression as was seen with En2/EN2 expression. The STAT1 gene was also overexpressed in the platinum-resistant cell lines compared to their sensitive pairings, and the
resultant transcription factor protein was located in the cytoplasm. In the platinum-resistant
cells only, they observed translocation of STAT1 to the nucleus after activation by
phosphorylation in response to cisplatin treatment. In the case of EN2 protein, we observed
predominant cytoplasmic staining, with only one molecular weight band evident in the cell
lines on WB. It is unlikely that different isoforms of EN2 exist within these cell lines, as a
post-translational modification such as phosphorylation or glycosylation, would be expected
to alter the protein molecular weight, hence potentially showing multiple bands on a WB.
The predicted molecular weight of EN2 in normal tissue is 33kDa, and urine samples from
patients with prostate cancer did appear to demonstrate an EN2 band at this weight [303],
whereas in urine samples from bladder cancer patients, the molecular weight of detected EN2
is closer to 37kDa [304]. However, the EN2 protein detected in the cell lysates of the EOC
cell lines corresponded to a molecular weight of 50kDa. This was slightly greater than the
45kDa weight of the rEN2 protein. As the levels of protein expression were fairly consistent
between the three techniques, fluorescent immunohistochemistry, ELISA and WB, and
82
correlated with the En2 mRNA expression levels, it is highly unlikely that the 50kDa band is
a false band due to non-specific binding. The WB was also repeated to ensure that the
protein ladder was reading correctly. The observed difference in protein size between the
EOC cell line data and the secreted urinary protein from prostate cancer patients could be
explained by post-translational modifications which are present in the intracellular form but
not in the secreted form. There is also the possibility that the EN2 protein seen in these
immortalised cancer cell lines, differs from that in directly obtained human specimens.
However the authors of a recent publication studying the expression of EN2 in renal cell
carcinoma clearly demonstrated a strong band at approximately 50kDa in the En2 overexpressed HEK 293T cell lines, and referred to this as the full-length EN2 protein [344]. Our
subsequent analysis of human ovarian cancer urine samples and En2 over-expressed cell lines
may shed further light on this issue.
Cell lines derived from normal human tissue are notoriously difficult to grow and maintain,
hence a normal ovarian surface epithelial cell line was not available to us. Therefore, a
human skin fibroblast cell line was used but this has limitations as a negative control for
ovarian cancer, as it does not derive from the ovary and does not contain any epithelial
elements. Nevertheless it proved useful in this preliminary work to demonstrate that En2 is
not present in all human tissue and to ensure that there is no non-specific binding of antibody
in the various experiments. When similar experiments are conducted on human tissue,
normal ovary specimens will be available as a more accurate control sample.
83
3.4.
CONCLUSION
As hypothesised, En2 was over-expressed in a number of EOC cell lines, in comparison to
human fibroblast cells, however there was a wide degree of variability between cell lines
representative of the serous histological sub-type, and there were no significant differences in
expression between serous, endometrioid and clear cell histological sub-types.
When
analysing En2 expression in the platinum-sensitive and –resistant paired cell lines, there was
significantly higher expression in platinum-resistant cells, however this was not seen in other
examples of serous and endometrioid platinum-resistant cell lines. As previously discussed,
some of these discrepancies may be as a result of inaccurate histological sub-type labelling.
EN2 protein was detectable via immunohistochemistry in all of the cell lines to a varying
degree but when using quantifiable techniques, this was clearly of a greater concentration in
the platinum-sensitive and –resistant paired cell lines. However there was no clear difference
between them in terms of drug sensitivity. Although cell line work can be very informative
and guide on-going work, it is not always fully representative of what occurs in the human
tumour in vivo, especially if derived from ascitic fluid or metastatic tumour deposits rather
than the primary tumour, therefore similar work on human tumour samples is vital to draw
more comprehensive conclusions regarding EN2 expression in epithelial ovarian cancer (see
Chapter 4).
84
CHAPTER 4
EN2 EXPRESSION IN HUMAN
EPITHELIAL OVARIAN CANCER TISSUE
AND ASSOCIATED BODY FLUIDS
85
4.
EN2 EXPRESSION IN HUMAN EPITHELIAL OVARIAN CANCER TISSUE
AND ASSOCIATED BODY FLUIDS
4.1.
INTRODUCTION
Ovarian cancer accounts for around 3% of all female cancers, however it results in the
highest number of deaths within female reproductive tract cancers due to it usually being
diagnosed at an advanced stage of disease. The symptoms of ovarian cancer are often very
vague and may be misinterpreted as those of more common, benign conditions such as
Irritable Bowel Syndrome. As these symptoms are mostly related to the pressure effects of
the growing tumour mass, they may only present when the primary mass is already very
large, or when the disease has disseminated to the peritoneum. A small proportion of women
will be known to carry the BRCA1 or BRCA2 mutation which increases their risk of
developing ovarian cancer, and are therefore very closely monitored or in fact undergo a
prophylactic bilateral oophorectomy. However the majority of women will not partake in any
specific monitoring for the early detection of ovarian cancer as there is currently no National
Screening Programme in the UK. The only clinically utilised biomarker for epithelial ovarian
cancer is CA125, but this has not been approved for use in diagnosis as it may be elevated in
benign gynaecological conditions such as endometriosis, hence has a relatively low
sensitivity of 50-62% for early stage disease, although rising to 90% in advanced stage
ovarian cancer [128-130].
Many of the women with advanced disease will experience a good initial response to
platinum-based chemotherapy and surgery, and may have disease-free intervals of well over
12 months, however most patients will relapse either with local or distant disease. Although
initial relapse as well as subsequent relapsed disease can often be controlled with repeated
platinum-based chemotherapy regimens, the progression-free relapse time interval will
eventually become less than 6 months, indicating that the tumours have developed resistance
to platinum.
There is a great deal of interest in biomarker research in epithelial ovarian cancer especially
for early diagnosis, given that the disease is usually diagnosed at a late stage which reflects
the poor 5 year survival rates of only 30% for advanced-stage disease [46, 345]. However
86
there is increasing research into prognostic and treatment response biomarkers, especially
focussing on the identification of molecular signatures that may indicate the early
development of platinum-resistant disease. Urinary biomarkers have received particular
attention due to their non-invasive, straightforward acquisition, the deposition of cancerassociated material into urine, and the ability to serially monitor changes over the course of
the disease, including treatment responses.
Certain gene mutations and amplifications help to differentiate between the different
histological sub-types of EOC, for instance TP53, BRCA1 and BRCA2 mutations being
frequently identified in high grade serous cancers [39, 46].
TP53 mutations and over-
expression are also associated with poor outcome [47], whereas the BRCA1 and -2 mutations
convey a good prognosis with improved sensitivity to chemotherapy and targeted treatment
[107-109]. In contrast, mucinous tumours often express KRAS mutations, low-grade serous
tumours express BRAF mutations, and low-grade endometrioid tumours may express mutated
ß-catenin which confers a good prognosis or PIK3 mutations which correlate with
chemotherapy resistance [44, 101, 102].
CA125 is the most well-known serum protein biomarker measured in EOC, however its true
utility lies in monitoring of treatment response and relapsed disease. It has limitations as a
diagnostic marker and does not provide any prognostic information. Other mucin-related
glycoproteins such as MUC1 (CA15-3), HE4 and mesothelin have been extensively studied,
alone or in combination with CA125, but still lack the desired high sensitivity and specificity
necessary for a gold-standard clinical biomarker [136-138, 166].
Interestingly, only
increased MUC1 (CA15-3) levels have been shown to correlate with worse prognosis [133].
HE4 and mesothelin have also been evaluated in the urine of EOC patients. Hellstrom et al
reported sensitivities of 86.6% and 89% for HE4 protein in the urine of early and late stage
ovarian carcinoma patients respectively [346]. These values corresponded to a specificity of
94.4%. Badgwell D et al demonstrated elevated urinary mesothelin levels in 42% of early
stage patients, rising to 75% in advanced disease [166]. Also matrix metalloproteinase
(MMP)-2 and -9 are elevated in the urine of advanced ovarian cancer patients compared with
healthy controls, particularly in those with normal serum CA125 levels [165].
The human kallikrein (KLK) proteins can also be identified in various biological fluids
including serum and ascites, so have been investigated in EOC [148]. Elevated serum levels
of KLK-4, KLK-5, KLK-6, KLK-10 and KLK-15 have been associated with a poor prognosis
87
in ovarian cancer, with KLK-4 and -6 also linked to chemotherapy resistance [149, 151-156].
Conversely elevated KLK-8, KLK-9 and KLK-11 are associated with a favourable prognosis
[157-161].
Although more invasive to acquire than urine, the protein-rich ascitic fluid frequently
associated with advanced EOC, has also been examined for potential prognostic biomarkers.
Elevated interferon-γ (IFN-γ) levels are associated with a shorter progression-free and overall
survival [167], whereas elevated KLK-8 is associated with an improved progression-free
survival [159], mimicking the findings in serum.
Anti-tumour antibodies may also prove useful as diagnostic biomarkers as they enable earlier
and lower level detection of tumour antigen than may be possible with direct protein assays,
hence allowing diagnosis of tumours at an early stage. Antibodies to epithelial cell adhesion
molecule (Ep-CAM) [170], HSP-90 [171, 172], HER2 [174] and p53 [106, 347] have been
identified in ovarian cancer but may be present in multiple tumour types. Serum anti-MUC1
IgG antibody was elevated in relapsed EOC patients compared with treatment responders,
and the IgM serotype was inversely correlated with overall survival [133].
The search for novel clinical biomarkers in EOC has included analysis of various homeobox
genes and their protein products, including the 39 HOX genes. The expression of HOXA9,
HOXA10 and HOXA11 confers a serous, endometrioid and mucinous phenotype respectively
(reviewed in [348]), but as yet have not demonstrated an association with prognosis.
Compared with normal ovarian surface epithelium, two studies demonstrated elevated
expression of HOXB7 in ovarian carcinomas however Hong et al could not detect elevated
HOXB7 levels, but did show significantly elevated expression of HOXB4 [237, 349, 350].
Antibodies to HOXA7 and HOXB7 have also been identified in ovarian cancer but further
work with larger sized cohorts is required [173, 237, 349].
Another member of the
homeobox family, PAX8, is consistently over-expressed in high grade serous ovarian
carcinomas but negative in breast adenocarcinoma so is often used by the pathologist to help
determine the origin of certain pelvic serous tumours, especially if the primary tumour source
is not always clearly evident [351-353].
BRN-3A(l) is also a homeobox transcription factor
protein, which demonstrates enhanced cytoplasmic and nuclear expression in the epithelium
and stroma of EOC [354]. Of note, 12% of non-cancerous normal ovarian tissue epithelium
and 40% of normal stroma did show staining, however significant increases in extent and
intensity of staining were demonstrated in moderate and high grade EOC. This protein was
88
also detectable in tumour cells from ascites of ovarian cancer patients. Although it is unlikely
to prove useful as a diagnostic biomarker given its presence in some normal ovary specimens,
the intensity and extent of BRN-3A(l) staining may have clinical utility as a prognostic
marker, but may also prove to be a target for therapeutic intervention.
EN2 is another homeobox protein, typically only expressed in the normal adult Purkinje
neurons and kidney tubular epithelial cells, whose expression has been evaluated in the
tissue, serum and urine of a number of epithelial cancers. At present there is no published
data on its presence in EOC. En2 mRNA over-expression has been demonstrated in breast
adenocarcinoma but was not observed in normal breast epithelium [301], and was also
demonstrated in prostatic adenocarcinoma with absence in normal prostate tissue and benign
disease [303]. Such expression was also discovered at the protein level with evidence of
secretion into ductal lumen. This prompted further investigation of EN2 as a biomarker in
prostate cancer and demonstrated that the presence of EN2 in urine is highly predictive of
prostate cancer, with a sensitivity of 66% and specificity of 88.2% [303]. Subsequent work
has demonstrated a linear relationship between urinary EN2 and prostate cancer volume, as
well as elevated urinary EN2 correlation with advancing tumour stage [314, 355]. In a cohort
of 466 patients with urothelial bladder cancer, the mean urinary EN2 concentration was
significantly elevated compared to that for control subjects, with an overall sensitivity of 82%
and specificity of 75% [304].
High-grade tumours demonstrate higher mean EN2
concentration and higher sensitivity, namely 87%. Although EN2 can be detected in urine
from prostate and bladder cancer patients these two diseases usually present with different
symptoms, and the presence of a positive family history and elevated serum PSA level would
be more suggestive of prostate cancer. Urinary EN2 testing could therefore be used alone or
in conjunction with other tests to direct further clinical investigations in those with a high
suspicion of bladder or prostate cancer, and multi-centre tests are on-going.
The over-expression of En2 and its protein product in breast, prostate and bladder cancer
tissue specimens, along with the identification of EN2 in voided urine samples, raises the
possibility of its presence in other epithelial carcinomas, such as EOC. To date, the level of
EN2 expression in relation to tumour volume and stage has only been assessed in prostate
cancer. There are few published associations with other factors such as histological sub-type,
tumour grade, progression-free and overall survival. On the background of our research, En2
mRNA and the resultant EN2 protein are certainly over-expressed in a number of EOC cell
lines, but there is a wide degree of variability between the cell lines representative of the
89
serous histological sub-type, and there are no significant differences between serous,
endometrioid and clear cell sub-types. Nevertheless when analysing En2 mRNA expression
in the platinum-sensitive and –resistant paired cell lines, there is significantly higher
expression in platinum-resistant cells. Hence if En2 mRNA and EN2 protein are found to be
over-expressed in some or all of the histological sub-types of human EOC specimens,
perhaps this may prove useful as a diagnostic, prognostic or treatment response biomarker.
90
Study Objectives and Hypothesis
The objectives of this chapter were to evaluate the En2 gene expression as well as EN2
protein expression in human epithelial ovarian tumours. We also acquired urine and ascites
samples from patients with EOC, in order to study EN2 protein secretion into these fluids,
along with patient and healthy donor blood in order to study EN2 auto-antibody responses.
These clinical samples were evaluated and the findings presented under three distinct subheadings, as follows: 
EN2 gene and protein expression in human ovarian tissue

EN2 protein expression in urine from ovarian cancer patients

EN2 protein expression in ascites from ovarian cancer patients

EN2 autoantibody levels in the plasma from ovarian cancer patients
If En2 mRNA and EN2 protein, or EN2 auto-antibodies were found to be over-expressed in
such human EOC specimens, this may prove useful as a diagnostic, prognostic or treatment
response biomarker.
We hypothesised that En2 would be over-expressed in a number of EOC specimens,
especially high grade serous tumours, in comparison to normal ovarian surface epithelium,
and that the resultant protein product would also be detectable within the cells. The level of
expression may also differ with increasing disease grade and stage, and significantly vary
between platinum-sensitive and platinum-resistant tumours, making it useful as a prognostic
biomarker. We expected to detect EN2 protein more readily in ascitic fluid than in urine due
to the anatomical location of the ovaries allowing direct secretion into the abdominal cavity.
This may provide the clinician with prognostic information regarding the patient’s ovarian
tumour.
91
4.2.
RESULTS
4.2.1. EN2 EXPRESSION IN HUMAN OVARIAN TISSUE
4.2.1.1. En2 gene expression in human ovarian tissue and correlation with clinicopathological characteristics
RNA was extracted from 116 tumour samples preserved in RNAlaterTM. The tumours
represented a variety of EOC histological sub-types, namely 78 serous, 8 endometrioid, 2
mucinous, 4 clear cell, 8 borderline tumours and 5 benign tumours. The full demographic
data for these tumours is shown in Table 4.1. On review of the hospital pathology reports, 3
of the serous tumours were reported as arising from the fallopian tube, and 17 arising from
the peritoneum.
Although initially grouped separately, these epithelial tumours were
predominantly serous in nature and are thought to share a common origin, pathogenesis and
behaviour with serous ovarian tumours, therefore in this analysis they were subsequently
grouped together as “pelvic serous carcinoma”, a term suggested by Nik et al [33]. Six
malignant mixed Müllerian tumours (MMMT) were also analysed. Six normal ovary and 3
normal fallopian tube RNA samples were analysed for comparison, either using commercial
RNA or tissue specimens obtained from theatre. The cDNA template from all samples was
used to quantify En2 expression via quantitative PCR. The En2 relative expression value was
calculated as a ratio of the mean Cycle Threshold (CT) value relative to the housekeeping
gene ß-actin.
Table 4.2 summarises the demographic data for the tumours classified by histological subtype. The relative expression of En2 in normal ovary and fallopian tube was very low in
comparison to that of the malignant tumour samples. Although higher than the normal
specimens, the En2 expression in benign and borderline epithelial tumours was also much
lower than the malignant tumours. Figure 4.1 shows the mean CT value relative to ß-actin,
representing En2 relative expression, for the tumours and normal specimens, and is classified
by histological sub-type. Many of the groups were not normally distributed and some were
small in number, therefore the Kruskal Wallis test was used for analysis of variance between
the groups. In order to provide a larger control group, the normal ovary and fallopian tube
data was combined, as the mean CT value was consistent.
92
N
Age
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
68
71
70
81
57
66
58
56
53
61
48
61
75
63
62
45
78
74
83
72
70
78
61
73
66
81
72
71
70
62
61
46
60
49
69
63
42
73
62
56
62
47
66
64
63
Histology
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Tissue type
Grade
Stage
Time of
Surgery
Platinum
Sensitivity
PFS
(months)
OS
(months)
Followup result
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
3
3
3
3
3
3
2
3
3
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3c
3c
3c
3c
3c
3c
1c
4
3c
3c
3c
3c
4
3c
3c
3c
3c
4a
3c
4
3c
3c
3c
3c
3c
3c
3c
3c
3c
4a
3c
3c
4b
1c
3c
3c
3c
3c
4a
3c
4b
4
4a
4a
3c
Interval
Primary
Interval
Primary
Interval
Primary
Primary
Interval
Interval
Interval
Primary
Primary
Interval
Interval
Primary
Interval
Interval
Interval
Interval
Interval
Primary
Interval
Primary
Interval
Primary
Primary
Primary
Interval
Interval
Interval
Primary
Primary
Interval
Interval
Primary
Primary
Interval
Interval
Interval
Interval
Interval
Interval
Interval
Interval
Primary
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Resistant
Sensitive
Refractory
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Resistant
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
8
16
61
49
5
6
17
3
8
7
11
9
8
2
10
2
34
15
47
5
53
50
1
18
0
14
10
11
7
10
7
56
10
5
2
5
10
10
6
13
4
14
13
20
14
16
30
68
55
43
11
207
13
38
21
52
38
22
13
30
76
55
34
54
27
53
55
18
37
0
44
34
39
21
20
37
60
27
44
11
12
43
26
10
38
76
33
26
25
22
Dead
Dead
Alive
Alive
Alive
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Alive
Dead
Alive
Dead
Lost
Alive
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Alive
Alive
Dead
Dead
Dead
Dead
Alive
Dead
Dead
Dead
Dead
Alive
Alive
Alive
Alive
CT value
relative to
β-actin
4671.40
66434.30
56448.20
60081.80
9605.47
24064.90
814.43
1264.75
37892.90
36349.30
65975.40
573.89
469.39
22688.00
101748.00
6515.41
217.47
16.00
60920.50
44442.10
30672.10
447.16
1488.50
309.68
13917.80
57236.20
12586.90
3018.55
49654.60
151.654
26794.30
98623.27
55864.40
79829.80
49827.00
6815.67
4373.71
42190.80
1645.88
1286.86
16.33
308.61
348.41
1826.22
8.69
Table 4.1. Full demographic data from the93
human ovarian tumours in RNA later.
IHC EN2
Staining
Intensity
3
3
3
3
2
1
3
3
2
2
3
2
2
3
1
3
2
3
2
3
3
3
3
3
2
2
3
3
3
3
1
3
3
3
3
3
2
2
3
2
3
IHC EN2
Staining
%
60
40
90
100
90
75
100
90
80
65
85
60
80
100
60
60
80
50
100
100
100
65
75
60
65
90
100
90
80
75
40
75
60
75
100
40
60
100
60
75
100
IHC EN2
Overall
Score
9
6
12
12
8
4
12
12
8
6
12
6
8
12
3
9
8
9
8
12
12
9
12
9
6
8
12
12
12
12
2
12
9
12
12
6
6
8
9
8
12
IHC EN2
Status
Positive
Positive
Positive
Positive
Positive
Negative
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Negative
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Negative
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
79
62
75
56
86
78
53
81
71
78
79
62
56
60
48
45
84
57
48
78
68
61
52
40
48
62
46
40
62
71
71
46
56
69
74
73
61
70
73
63
70
73
53
71
74
70
47
64
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Clear cell carcinoma
Clear cell carcinoma
Clear cell carcinoma
Clear cell carcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Primary peritoneal adenocarcinoma
Fallopian tube adenocarcinoma
Fallopian tube adenocarcinoma
Fallopian tube adenocarcinoma
MMMT
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
1
3
2
2
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
2
3
3
3
3
3c
4
4
4
3c
3c
3c
3c
3c
3c
3c
3c
4
3c
3c
2b
1c
3c
1c
3c
3a
3c
3c
3c
3c
2c
3c
3c
3c
3c
3c
4
4b
3c
3c
3c
4b
3c
3c
4a
3c
3c
3c
3c
4b
3c
4
2c
Primary
Interval
Interval
Interval
Interval
Interval
Interval
Interval
Primary
Interval
Interval
Primary
Interval
Primary
Interval
Primary
Primary
Primary
Primary
Primary
Primary
Interval
Primary
Primary
Primary
Primary
Primary
Interval
Interval
Interval
Primary
Primary
Interval
Interval
Interval
Interval
Interval
Interval
Interval
Interval
Interval
Interval
Interval
Interval
Primary
Interval
Interval
Primary
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Resistant
Refractory
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Sensitive
Refractory
Sensitive
Resistant
Resistant
Resistant
94
7
1
30
6
6
16
5
12
2
14
18
15
26
0
2
46
41
16
38
0
6
2
0
0
15
2
0
8
34
21
5
9
7
27
6
21
41
38
10
6
5
14
21
0
17
4
2
0
13
5
35
75
14
56
12
44
17
31
31
44
31
2
11
51
46
36
44
4
13
25
5
12
31
16
5
60
50
34
13
96
55
42
20
47
56
43
29
19
12
27
27
6
29
23
19
6
Dead
Dead
Alive
Alive
Dead
Alive
Dead
Dead
Dead
Dead
Dead
Dead
Alive
Dead
Dead
Alive
Alive
Dead
Alive
Dead
Dead
Dead
Alive
Lost
Alive
Dead
Dead
Alive
Dead
Dead
Dead
Dead
Alive
Dead
Dead
Alive
Alive
Alive
Dead
Dead
Dead
Alive
Alive
Dead
Dead
Dead
Dead
Dead
9705.86
55286.53
90.17
9841.35
86453.73
20518.54
53034.39
27262.69
44288.38
6721.84
41322.52
21315.87
67.87
222.04
2591.62
80944.20
19210.90
28618.10
3479.44
9214.18
50874.00
1473.10
12896.04
145.98
115.73
10919.70
34388.55
41609.94
36602.10
45376.00
407.21
4769.56
4937.76
448.71
74.78
25.01
142.98
731.46
52304.20
1184.15
32421.00
3874.09
587.99
14458.60
575.89
466.15
866.85
53961.40
2
2
1
1
2
2
3
1
2
3
3
2
3
3
3
2
3
2
2
3
2
1
3
3
3
3
3
3
3
3
2
3
2
2
2
3
2
3
3
2
-
60
100
100
75
60
65
85
75
60
65
90
75
75
100
40
75
100
60
65
100
100
100
100
100
90
75
100
60
100
75
60
100
100
100
50
90
75
80
100
60
-
6
8
4
4
6
6
12
4
6
9
12
8
12
12
6
8
12
6
6
12
8
4
12
12
12
12
12
9
12
12
6
12
8
8
6
12
8
12
12
6
-
Positive
Positive
Negative
Negative
Positive
Positive
Positive
Negative
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Negative
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
-
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
66
62
65
90
88
65
46
65
48
59
60
69
36
48
68
48
58
54
75
68
74
84
54
76
70
59
61
MMMT
MMMT
MMMT
MMMT
MMMT
Borderline mucinous tumour
Borderline serous tumour
Borderline serous tumour
Borderline serous tumour
Borderline serous tumour
Borderline mucinous tumour
Borderline serous tumour
Borderline serous tumour
Serous cystadenoma
Mucinous cystadenoma
Mucinous cystadenoma
Serous cystadenofibroma
Serous cystadenoma
Normal ovary RNA (CR561072)
Normal ovary RNA (CR561228)
Normal ovary
Normal ovary
Normal ovary
Normal ovary
Normal fallopian tube
Normal F/tube RNA (CR559552)
Normal F/tube RNA (CR559553)
Cancer
Cancer
Cancer
Cancer
Cancer
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Benign
Benign
Benign
Benign
Benign
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
Normal
3
3
3
3
3
-
3b
4
3c
3a
3c
-
Primary
Primary
Interval
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
-
Resistant
Resistant
-
95
2
0
2
4
1
-
2
4
7
6
1
-
Dead
Dead
Dead
Dead
Dead
-
31864.00
292.98
25702.80
1535.66
18556.55
29.85
373.42
16.39
7.21
156.46
17.57
20.11
83.27
125.34
76.09
0.81
5.92
105.03
9.38
19.36
39.52
64.21
7.78
29.13
3.63
27.75
52.33
2
1
1
2
2
3
2
3
2
3
3
3
2
1
1
3
2
-
80
40
60
60
65
75
100
100
60
100
40
100
75
60
20
100
100
-
8
2
3
6
6
12
8
12
6
12
6
12
8
3
2
12
8
-
Positive
Negative
Negative
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Negative
Negative
Positive
Positive
-
Histology
Pelvic serous carcinoma
Serous adenocarcinoma
Primary peritoneal
adenocarcinoma
Fallopian tube adenocarcinoma
Endometrioid adenocarcinoma
Mucinous adenocarcinoma
Clear cell carcinoma
MMMT
Borderline epithelial tumour
Benign tumour
Normal ovary
Normal fallopian tube
N
Mean age
(range)
Median
grade
Median
stage
78
58
17
65.28 (40-86)
65.60 (42-86)
64.47 (40-74)
3
3
3
3
3
3
3
8
2
4
6
8
5
6
3
63.67 (47-74)
61.63 (45-84)
54 (48-60)
49(40-62)
72.50 (62-90)
56 (36-69)
55.2 (48-68)
71.83 (54-84)
63.33 (59-70)
3
3
3
3
3
-
4
3
3
3
3
-
Surgery
(Number of patients)
Primary
Interval
23
55
20
38
2
15
1
7
1
4
5
8
-
Platinum Status
(Number of patients)
Sensitive Resistant
57
19
43
14
13
3
2
1
1
0
1
0
-
1
5
0
1
0
-
Mean CT value
relative to
ß-actin
22,541
26,144
14,115
2
2
1
2
3
-
Table 4.2. Summary data from the human ovarian tumours stored in RNA later with
mean relative En2 mRNA expression.
M e a n C T v a lu e
r e la t iv e t o B - a c t in
40000
**
****
30000
***
20000
10000
3000
1500
T
u
b
e
n
e
o
p
B
o
ia
n
B
rd
e
e
n
rl
ig
in
M
e
M
M
a
le
c
u
C
n
a
ll
E
F
+
e
N
o
rm
a
l
P
O
e
v
lv
a
ic
ry
S
P
T
ll
s
r
in
C
o
io
d
M
m
e
tr
o
in
o
u
id
a
e
m
b
u
ro
u
s
ll
o
C
p
a
ia
rc
n
T
n
ri
e
P
F
a
ry
a
ri
S
m
e
ro
u
s
O
v
to
a
ri
e
a
a
n
l
0
Figure 4.1. En2 mRNA expression in human ovarian tumours, classified by histological
sub-type. Malignant, borderline, benign and normal tumours were analysed by rt-PCR. The mean En2
mRNA expression for each histological sub-type is shown relative to the house-keeping gene β-actin
(x100,000). Error bars represent the SE, and the Kruskal Wallis test with Dunn’s correction was used for
analysis (**=p<0.01; ***=p<0.001; ****=p<0.0001).
96
636
25,839
1,407
11,392
21,986
88
63
28
28
En2 expression in the serous ovarian group was significantly elevated in comparison to the
normal specimen group (p<0.001), and this significant over-expression was maintained when
the serous ovarian, peritoneal and fallopian tube groups were combined as the “pelvic serous
carcinoma” group (p<0.005). Interestingly, the mean En2 expression appeared to be much
lower in the fallopian tube adenocarcinoma group compared with the serous ovarian and
primary peritoneal groups, however there were only 3 tumour samples in this group. The
endometrioid EOC sub-group also demonstrated En2 over-expression compared with the
normal specimens (p<0.01). There are only 4 examples of clear cell and 2 examples of
mucinous tumour which limited the analysis, however there was a trend towards En2 overexpression compared with the normal samples, although clearly at a lower level than that of
the serous and endometrioid sub-types.
Figure 4.2 depicts the CT values relative to β-actin as a scatter plot for each of the epithelial
tumour sub-types, in order to demonstrate the high level of variability in En2 expression
within some of the groups. It was thought that some of this variability could reflect the
different stages of disease, or perhaps En2 expression could differ between platinum-sensitive
and platinum-resistant disease. It is unlikely to reflect differing disease grade as the tumours
are predominantly high-grade (Grade 3); there was only 1 example of low-grade and 5
examples of moderate-grade disease. Hence further analysis of disease stage, timing of
100000
50000
e
n
a
ll
E
o
p
B
o
ia
n
B
rd
e
T
n
u
b
ig
in
rl
e
r
a
C
le
c
u
M
e
ll
C
o
in
tr
e
m
o
n
d
e
u
id
io
m
o
in
rc
a
C
s
N
o
rm
a
l
P
O
e
v
lv
a
ic
ry
S
+
e
F
ro
u
s
0
a
C T v a lu e r e la t iv e t o B - a c t in
surgery relative to chemotherapy, and platinum-status was carried out.
Figure 4.2. A high degree of variability in En2 mRNA expression exists within the
histological sub-types of EOC. EOCs, borderline, benign and normal tumours were analysed by rt-PCR.
The CT values relative to β-actin (x100,000) are plotted for each individual tumour, within its histological subtype. The mean and SE are shown (red lines).
97
High grade serous ovarian carcinomas (HGSOC) are the most common histological
presentation in the clinic, and represented 97% of the pelvic serous tumours within this data
set. Hence the En2 relative expression of HGSOC was plotted in Figure 4.3 along with the
borderline, benign and normal specimens. En2 expression in the HGSOC group was again
significantly elevated in comparison to the normal specimen group (p<0.0001).
****
20000
10000
150
100
50
0
e
n
u
b
ig
T
n
e
ia
n
B
rd
o
ll
o
p
B
N
o
rm
a
l
O
v
a
ry
H
+
ig
F
h
a
G
ra
d
e
S
e
e
rl
ro
in
u
e
s
M e a n C T v a lu e
r e la t iv e t o B - a c t in
30000
Figure 4.3. En2 mRNA expression in human high grade serous tumours.
HGSOC,
borderline, benign and normal tumours were analysed by rt-PCR. The mean En2 mRNA expression for each
histological sub-type is shown relative to the house-keeping gene β-actin (x100,000). Error bars represent the
SE, and the Kruskal Wallis test with Dunn’s correction was used for analysis (****=p<0.0001).
Subsequent analysis of this HGSOC group demonstrated a statistically significant increased
level of En2 expression in tumours taken from Stage 3 versus Stage 4 patients (Figure 4.4
(A); p=0.0013). Patients with Stage 4 disease usually receive neoadjuvant chemotherapy in
the first instance and then proceed to interval surgical debulking. Hence the primary and
interval surgery cohorts of HGSOC were analysed separately to see whether the reduction in
En2 expression in Stage 4 patients merely reflected pre-treatment with chemotherapy, or
whether the level of expression was truly less with advancing disease stage. Again the data
was not normally distributed despite log transformation, so the non-parametric MannWhitney test was used for statistical analysis. There was certainly a trend towards lower En2
expression in tumours that have been exposed to neoadjuvant chemotherapy compared with
primary surgery specimens (Figure 4.5; p=0.0512), however Figure 4.4(B) and 4.4(C) show
98
that a lower En2 expression was still demonstrated in tumours from Stage 4 patients
compared with Stage 3, whether they had received neoadjuvant chemotherapy or not,
although this was only statistically significant in the interval surgery cohort (p=0.0096).
Platinum-based chemotherapy treatment data was available for 75 of the high grade serous
epithelial tumours, including progression-free survival figures, which allowed us to determine
the platinum-sensitivity-status of the tumours. If a patient’s disease relapsed within 6 months
of completing platinum-based chemotherapy, they were reported to have platinum-resistant
EOC. Conversely, patients were deemed to have platinum-sensitive disease if they had no
documented disease progression for 6 months or more after completion of platinum-based
chemotherapy. En2 mRNA expression was higher in tumours that were later shown to be
platinum-resistant compared to those that were platinum-sensitive, however this was only
statistically significant when considering those patients who had already received some
chemotherapy prior to surgery i.e. the interval surgery cohort (Figure 4.6(A) and 4.6(B);
p=0.0232).
Survival data was available for the majority of patient tumour samples, hence progressionfree and overall survival could be analysed. Assessment of the overall survival curves for all
of the ovarian histological sub-groups showed that MMMT, mucinous and clear cell tumours
(median survival of 6, 6.5 and 16 months respectively) had a much worse prognosis
compared with endometrioid and pelvic serous carcinoma (median survival of 36 and 38
months respectively), which was consistent with the published outcomes for these tumour
sub-types (Figure 4.7)[356-358].
Subsequent survival analyses of the HGSOCs were
performed based on low versus high En2 mRNA expression. As this was the first time that
En2 expression had been evaluated in ovarian cancer, we did not yet know what constituted
low or high expression, hence the analysis was repeated with two different cut-off values; the
first being a CT value relative to ß-actin<1000 (low expression) versus >1000 (high
expression), and the second being <10,000 (low expression) versus >10,000 (high
expression).
The overall survival analysis of all of the HGSOCs was not statistically
significant when considering either cut-off value. However when analysing the HGSOC
specimens from patients who had received prior chemotherapy i.e. “interval surgery
specimens”, there was a statistically significant reduction in median overall survival in the
high En2 expression (CT value >10,000) compared with the low En2 expression (CT value
<10,000) cohort (Figure 4.8(D)). The median overall survival was 28 months in the high En2
expression cohort compared with 42 months in those with low En2 expression (p=0.0329).
99
igthe nGsritayd e S e r o u s T u m o u r s ; P r i m a r y S u r g e r y - S t aHgig
e hv sG Irnatde en sSi e
t yr o u s T u m o u r s ; I n t e r v a l S u r g e r y - S t a g e v s In t e n s i t y
H i g h G r a d e S e r o u s T u m o u r s - S t a g e v sHIn
B
30000
20000
10000
10000
)
)
S
1
=
(n
4
E
G
A
T
S
T
T
A
A
G
G
E
E
3
4
(n
(n
=
=
3
9
5
)
3
)
S
S
T
T
A
A
G
G
E
E
4
3
(n
(n
=
=
2
1
2
9
)
)
4
5
=
(n
3
E
G
A
T
S
20000
0
0
0
**
S
10000
M e a n C T v a lu e
M e a n C T v a lu e
20000
r e la t iv e t o B - a c t in
30000
30000
C
40000
r e la t iv e t o B - a c t in
**
r e la t iv e t o B - a c t in
M e a n C T v a lu e
A
Figure 4.4. En2 mRNA expression is lower in human high grade serous tumours from Stage 4 versus Stage 3 patients. High grade serous
tumours from Stage 3 and Stage 4 patients were analysed by rt-PCR. The mean En2 mRNA expression is shown relative to the house-keeping gene β-actin (x100,000) for all
C T v a lu e s v s C h e m o ( H i g h g r a d e s e r o u s ) ( M a y - 1 4 )
HGSOCs versus disease stage (A) (p=0.0013), primary surgery HGSOCs versus stage (B), and interval surgery HGSOCs versus stage (C)(p=0.0096). Error bars represent
the SE, and the Mann-Whitney test was used for analysis; n=number of sample.
( p = 0 .0 5 1 2 )
M e a n C T v a lu e
r e la t iv e t o B - a c t in
40000
30000
20000
10000
rv
te
In
P
ri
m
a
a
l
ry
S
S
u
u
rg
rg
e
e
ry
ry
(n
(n
=
=
5
2
5
2
)
)
0
Figure 4.5. En2 mRNA expression is lower in human high grade serous tumours exposed to neoadjuvant chemotherapy. High grade
serous tumours from patients receiving primary or interval surgery were analysed by rt-PCR. The mean En2 mRNA expression is shown relative
to the house-keeping gene β-actin (x100,000)(p=0.0512). Error bars represent the SE, and the Mann-Whitney test was used for analysis; n=number of sample.
100
C T v a lu e s v s P la t in u m ( H ig h g r a d e s e r o u s ; I n t e r v a l s u r g e r y ) ( M a y - 1 4 )
C T v a lu e s v s P l a t i n u m
( H ig h g r a d e s e r o u s )
B
40000
*
40000
M e a n C T v a lu e
30000
20000
10000
r e la t iv e t o B - a c t in
30000
20000
10000
0
)
)
2
3
4
1
=
=
(n
(n
e
t
iv
n
it
ta
s
is
n
s
e
e
S
R
m
u
u
n
n
la
ti
ti
la
P
P
P
la
la
ti
ti
n
n
u
u
m
m
m
S
R
e
e
n
s
s
is
it
ta
iv
n
e
t
(n
(n
=
=
5
1
6
9
)
)
0
P
r e la t iv e t o B - a c t in
M e a n C T v a lu e
A
Figure 4.6. En2 mRNA expression is higher in human high grade serous tumours that
become platinum-resistant rather than platinum-sensitive. High grade serous tumours were
analysed by rt-PCR and follow-up data was studied to determine if the patient had platinum-resistant
(PFS<6months) or platinum-sensitive (PFS6months) disease. The mean En2 mRNA expression is shown
relative to the house-keeping gene β-actin (x100,000) for all HGSOCs versus platinum-status (A), and interval
surgery HGSOCs versus platinum status (B)(p=0.0232). Error bars represent the SE, and the Mann-Whitney
test was used for analysis; n=number of sample.
O v e r a ll S u r v iv a l v s H is t o lo g y ( A ll s u b t y p e s )
P e lv ic S e r o u s C a r c in o m a
P e r c e n t s u r v iv a l
100
E n d o m e tr io id
C le a r C e ll
M u c in o u s
MMMT
50
0
0
20
40
60
100
200
O v e r a ll S u r v iv a l (m o n t h s )
Figure 4.7. Overall survival analysis of the human ovarian tumours, classified by
histological sub-type. The survival curve for each of the ovarian tumour histological sub-types is
demonstrated, with overall survival measured in months. The median survival for pelvic serous (38 months)
and endometrioid carcinoma (36 months) is notably greater than that of clear cell (16 months) and mucinous
carcinoma (6.5 months), as well as MMMT (6 months). The Log-rank (Mantel Cox) test was used for analysis;
p<0.0001 for overall data.
101
B
50
100
P e r c e n t s u r v iv a l
P e r c e n t s u r v iv a l
C T > 1 ,0 0 0
C
C T < 1 ,0 0 0
50
C T > 1 0 ,0 0 0
50
50
100
0
O v e r a ll S u r v iv a l ( m o n t h s )
20
40
60
C T > 1 0 ,0 0 0
p = 0 .0 3 2 9
50
0
0
80
C T < 1 0 ,0 0 0
100
0
0
0
C T < 1 0 ,0 0 0
100
C T > 1 ,0 0 0
0
D
P e r c e n t s u r v iv a l
C T < 1 ,0 0 0
100
P e r c e n t s u r v iv a l
A
50
100
0
20
O v e r a ll S u r v iv a l ( m o n t h s )
O v e r a ll S u r v iv a l ( m o n t h s )
40
60
80
O v e r a ll S u r v iv a l ( m o n t h s )
Figure 4.8. Overall survival analysis of the high grade serous human ovarian tumours, comparing En2 mRNA expression. The survival curves of
high versus low En2 expression for the high grade serous ovarian tumours are demonstrated for all specimens (high En2 expression = CT value relative to ß-actin >1,000 (A) or
>10,000 (C)), and for interval surgery specimens (high En2 expression = CT value >1,000 (B) or >10,000 (D)). Overall survival is measured in months. The median survival for
interval surgery HGSOCs with high En2 (28 months) was significantly lower than those with low En2 (42 months) when using CT value >10,000 for high expression. The Log-rank
(Mantel Cox) test was used for analysis; p=0.0329 for interval surgery HGSOC with CT value >10,000 for high expression.
100
C T < 1 ,0 0 0
C T > 1 ,0 0 0
p = 0 .0 0 3 6
50
P e r c e n t s u r v iv a l
P e r c e n t s u r v iv a l
100
C T > 1 ,0 0 0
p = 0 .0 0 0 4
50
0
20
40
60
P r o g r e s s io n - f r e e S u r v iv a l ( m o n t h s )
C T < 1 0 ,0 0 0
100
C T < 1 ,0 0 0
0
0
D
C
C T > 1 0 ,0 0 0
50
0
0
20
40
60
P r o g r e s s io n - f r e e S u r v iv a l ( m o n t h s )
C T < 1 0 ,0 0 0
100
P e r c e n t s u r v iv a l
B
P e r c e n t s u r v iv a l
A
C T > 1 0 ,0 0 0
50
0
0
20
40
60
P r o g r e s s io n - f r e e S u r v iv a l ( m o n t h s )
0
20
40
60
P r o g r e s s io n - f r e e S u r v iv a l ( m o n t h s )
Figure 4.9. Progression-free survival analysis of the high grade serous human ovarian tumours, comparing En2 mRNA expression. The survival
curves of high versus low En2 expression for the high grade serous ovarian tumours are demonstrated for all specimens (high En2 expression = CT value relative to ß-actin >1,000
(A) or >10,000 (C)), and for interval surgery specimens (high En2 expression = CT value >1,000 (B) or >10,000 (D)). Progression-free survival is measured in months. The median
survival was significantly lower for all HGSOCs with high En2 (9 months) versus those102
with low En2 (18 months)(p=0.0036), and for interval surgery HGSOCs with high En2 (8
months) versus those with low En2 (27 months) (p=0.0004). The Log-rank (Mantel Cox) test was used for analysis.
The progression-free survival analysis was performed using the same cut-off values for low
versus high En2 expression and demonstrated a statistically significant reduction in
progression-free survival when analysing all of the HGSOCs with high En2 expression (CT
value >1,000) compared with low En2 expression (CT value <1,000) (Figure 4.9(A)). The
median progression-free survival was 9 months in the high En2 expression cohort compared
with 18 months in those with low En2 expression (p=0.0036). This difference was even more
pronounced when considering the interval surgical HGSOCs only, where median
progression-free survival was 8 months in the high En2 expression cohort compared with 27
months in those with low En2 expression (Figure 4.9(B); p=0.0004).
There were no
statistically significant differences observed when considering the CT value >10,000 cut-off
for high En2 expression, although the curves clearly separated in the interval surgical
HGSOC cohort, with median progression-free survival of 8 months in those with high En2
expression compared with 14 months in those with low expression (p=0.0519).
103
4.2.1.2. EN2 protein expression in human ovarian tissue and correlation with clinicopathological characteristics
In order to assess the prevalence and expression pattern of EN2 at the protein level in ovarian
cancer tissue, enzymatic immunohistochemical analysis was performed on two cohorts of
patient tumours.
The first cohort consisted of 116 specimens stored in RNAlaterTM
preservative. These were subsequently formalin-fixed, paraffin embedded, sectioned and
mounted onto slides in duplicate, so that EN2 and Haematoxylin & Eosin (H&E) staining
could be performed on adjacent sections. When analysing the H&E slides, it was clear that
several specimens were necrotic and did not contain any viable tissue, and others did not
contain any evidence of epithelial tumour or cells, so these were excluded from the EN2
scoring. It was originally thought that the cohort of tumours included 4 normal ovary
specimens, however when the H&E slides were examined, it was clear that these examples
were predominantly stromal tissue and did not include any normal surface epithelium. In the
absence of epithelial cells, it was felt that these samples could not be considered as
comparative normal ovary specimens for immunohistochemistry, and were excluded from
further analysis.
Hence the resultant cohort contained a total of 104 specimens, namely 88 epithelial ovarian
tumours, 7 borderline tumours, and 5 benign tumours. There were also 4 examples of
MMMT. Figure 4.10 demonstrates examples of the paired EN2 (brown staining) & H&E
immunohistochemical staining of malignant tissue, which helped to identify the epithelial
component of the tumour (deep pink/purple staining). An example of the normal kidney
positive control is shown, along with normal ovarian stromal tissue which demonstrates the
absence of non-specific EN2 staining. The EN2 staining intensity (0-3+) is illustrated. The
percentage of positively stained cells was also recorded and scored (0-4). The product of
these two scores was then calculated, with a resultant score of 0-4 representing EN2 negative
staining, and 5-12 representing EN2 positive staining [321]. The full demographic data for
these tumour slides including the EN2 intensity, percentage positivity and overall IHC EN2
score, is shown in Table 4.1 (Section 4.2.1.1.).
The immunohistochemical staining demonstrated positive cytoplasmic EN2 protein staining
in 88.6% of epithelial ovarian cancer specimens and 100% of borderline tumours, but only
40% of benign tumours. The epithelial component of the MMMT specimens only showed
positive EN2 staining in 2 out of the 4 examples. Figure 4.11 demonstrates examples of the
104
paired EN2 & H&E immunohistochemical staining of the histological sub-types of EOC.
Comparative analysis of the tumour sections using the Chi-squared statistical analysis
confirmed a significant difference between the EOC and borderline tumours compared to the
benign tumours (Table 4.3). However there were no discernible differences between the
histological sub-types of EOC, although the numbers of non-serous tumours were small. The
primary serous ovarian, fallopian tube and primary peritoneal specimens showed
predominantly positive EN2 staining so these were combined for all further sub-group
analysis.
Figure 4.10. Examples of EN2 protein expression in human ovarian tissue from Cohort
1, using enzymatic immunohistochemistry. Enzymatic immunohistochemistry examples of normal
kidney (A), normal ovarian stromal tissue (B) and ovarian adenocarcinoma stained for EN2 (C, E, G), or
adjacent sections stained for H&E (D, F, H), at 10x magnification. The assigned staining intensity (1-3+) is
indicated. EN2 staining (brown) is present in the cytoplasm, but not in the nucleus, and is not seen in the
surrounding stroma. On the H&E stained sections, pink staining represents cytoplasm and blue staining
represents the nucleus.
There were no significant differences in staining between the tumour grades however there
were only 7 examples of low/moderate grade disease. However there was a significant
difference between high-grade serous disease stage, with 94% EN2 positivity in Stage III
disease, falling to 75% in Stage IV (x2 =5.095, p=0.024). These findings mirrored the
reduction in En2 mRNA expression seen in Stage IV compared with Stage III disease, and
although it did not quite achieve statistical significance, this difference was maintained when
only considering those patients that had received neoadjuvant chemotherapy (x2 =3.616,
p=0.057). This suggested that the level of EN2 staining reduced with advancing disease
stage, and did not merely reflect pre-treatment with chemotherapy. No clear differences were
105
seen between tumours that had been exposed to neoadjuvant chemotherapy compared with
primary surgery specimens, or between the platinum sensitive and resistant groups.
Figure 4.11. Examples of EN2 protein expression in the different histological sub-types
of human ovarian tissue from Cohort 1, using enzymatic immunohistochemistry.
Enzymatic immunohistochemistry examples of EN2 and H&E staining in adjacent sections of serous, mucinous,
endometrioid and clear cell ovarian adenocarcinoma specimens, at 10x magnification. EN2 staining (brown) is
present in the cytoplasm, but not in the nucleus, and is not seen in the surrounding stroma. On the H&E stained
sections, pink staining represents cytoplasm and blue staining represents the nucleus.
106
N
Benign
Borderline
Epithelial carcinoma
MMMT
5
7
88
4
EN2 -ve
(0-4)
n
%
3
60.0%
0
0.0%
10
11.4%
2
50.0%
Histological sub-type
All Serous
76
9
x2
p
14.364
0.002
88.2%
87.9%
100.0%
87.5%
100.0%
100.0%
75.0%
1.781
NS
EN2 +ve
(5-12)
n
%
2
40.0%
7
100.0%
78
88.6%
2
50.0%
67
51
2
14
2
6
3
Mucinous
Endometrioid
Clear cell
2
6
4
0
0
1
11.8%
12.1%
0.0%
12.5%
0.0%
0.0%
25.0%
Tumour Grade (All Serous)
1
2
3
1
3
71
0
1
6
0.0%
33.3%
8.5%
1
2
65
100.0%
66.7%
91.5%
2.210
NS
Tumour Stage (High Grade Serous)
Early (I & II)
Advanced (III & IV)
1
70
0
8
0.0%
11.4%
1
62
100.0%
88.6%
0.129
NS
III
IV
50
20
3
5
6.0%
25.0%
47
15
94.0%
75.0%
5.095
0.024
III (neoadjuvant only)
IV (neoadjuvant only)
28
18
2
5
7.1%
27.8%
26
13
92.9%
72.2%
3.616
0.057
Surgery (High Grade Serous)
Primary
Interval
25
46
1
7
4.0%
15.2%
24
39
96.0%
84.8%
2.038
NS
Platinum Sensitivity (High grade serous)
Sensitive (≥6months)
Resistant (<6months)
50
18
7
1
14.0%
5.6%
43
17
86.0%
94.4%
0.909
NS
Platinum Sensitivity (High grade serous, Interval surgery)
Sensitive (≥6months)
Resistant (<6months)
36
10
7
0
19.4%
0.0%
29
10
80.6%
100.0%
2.293
NS
Serous ovarian
Serous fallopian tube
Serous primary peritoneal
7
0
2
58
2
16
Table 4.3. EN2 protein expression in human ovarian tissue from Cohort 1. The product of the EN2
staining intensity score (0-3+) and the percentage of positively stained cells score (0-4) was calculated. The resultant score of
0-4 represented EN2 negative staining, and 5-12 represented EN2 positive staining. Comparative analysis of the tumour
sections was performed using the chi-square test, which demonstrated significant differences between the EOC and borderline
tumours compared with benign tumours (x2 =14.364, p=0.002), and between high-grade serous disease stage III and IV (x2
=5.095, p=0.024). The larger percentage value for each variable is highlighted in red.
107
The second cohort of patient tumours included 90 pre-cut slides from formalin-fixed, paraffin
embedded tissues consisting of 63 epithelial ovarian tumours, 17 borderline tumours, 7
benign cases and 3 MMMTs. Slides from 4 normal ovaries were also donated by the
Pathology Department at RSCH, although no demographic details were available for these
patients. Figure 4.12 demonstrates examples of the EN2 immunohistochemical staining,
including normal tissue with ovarian surface epithelium, and malignant tissue, illustrating the
EN2 staining intensity (0-3+). The percentage positivity score was recorded as with the
previous cohort, and the product of these two scores was calculated, to determine EN2
negative or positive staining. The full demographic data for these tumour slides including the
EN2 intensity and percentage positivity scores is shown in Table 4.4).
Figure 4.12. Examples of EN2 protein expression in human ovarian tissue from Cohort
2, using enzymatic immunohistochemistry. Enzymatic immunohistochemistry examples of normal
ovary (A, B) and ovarian adenocarcinoma (C-E) stained for EN2, at 20x magnification. A control slide with no
secondary antibody is shown for comparison (A). Ovarian surface epithelium is indicated by the black arrow.
The assigned staining intensity (0-3+) is indicated. EN2 staining (brown) is present in the cytoplasm, but not in
the nucleus, and is not seen in the surrounding stroma.
The immunohistochemical staining demonstrated positive cytoplasmic EN2 protein staining
in 63.5% of epithelial ovarian cancer specimens and 52.9% of borderline tumours, but only
14.3% of benign tumours. Ovarian surface epithelium could be identified in all of the normal
ovary specimens and did not demonstrate EN2 staining. MMMT specimens showed positive
EN2 staining in 2 out of the 3 examples. Comparative analysis of the tumour sections using
the Chi-squared statistical analysis confirmed a significant difference between the malignant
and benign and normal samples (Table 4.5). However, there were no differences between
the cohorts in terms of histological sub-type of EOC, tumour grade or stage, primary versus
interval surgery, or platinum sensitive versus resistant disease.
108
N
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
Age
57
66
42
42
53
53
73
32
50
80
55
65
81
62
78
60
50
46
68
75
63
60
70
53
66
70
68
58
81
66
42
25
61
77
80
67
42
49
71
59
46
52
74
60
Histology
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Serous adenocarcinoma
Tissue type
Grade
Stage
Time of
Surgery
Platinum
Sensitivity
PFS
(months)
OS
(months)
Follow-up
result
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
3
2
3
3
3
3
3
3
3
2
3
3
3
3
3
3
3
3
4
3c
3c
2a
4
3a
3c
3c
3c
3c
4a
1c
2c
3c
3
3c
3c
4a
3c
2b
2b
3c
4a
3c
3c
3c
3c
3c
4a
3c
4a
3c
3c
4a
3b
3c
4a
1c
3a
3c
3c
2b
3c
1a
Interval
Primary
Primary
Primary
Primary
Interval
Primary
Primary
Primary
Primary
Interval
Primary
Primary
Primary
Primary
Primary
Interval
Primary
Primary
Primary
Primary
Interval
Primary
Primary
Interval
Interval
Interval
Interval
Interval
Primary
Interval
Primary
Primary
Primary
Interval
Primary
Primary
Primary
Interval
Primary
Primary
Primary
Interval
Primary
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
10
25
86
43
12
14
81
7
8
24
124
11
0
77
2
13
12
110
8
3
8
5
11
101
32
8
7
0
49
1
4
15
21
12
16
20
81
21
17
86
34
28
134
68
29
72
24
131
23
24
70
129
34
26
0
123
11
34
44
119
118
19
10
114
30
21
110
57
34
24
14
33
55
2
15
60
38
44
52
70
94
94
94
91
Dead
Dead
Alive
Dead
Dead
Dead
Dead
Alive
Dead
Dead
Dead
Alive
Dead
Dead
Dead
Alive
Dead
Dead
Dead
Alive
Alive
Lost
Dead
Alive
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Dead
Alive
Alive
Alive
Alive
109
IHC EN2
Staining
Intensity
1
2
2
3
2
2
1
0
3
3
2
2
2
2
2
1
2
1
0
2
1
1
2
2
2
3
3
2
2
1
3
2
1
1
1
3
1
1
2
2
1
1
2
2
Table 4.4. Full demographic data from the human ovarian tumours in Cohort 2.
IHC EN2
Staining
%
85
50
60
50
60
75
100
0
70
75
55
45
85
85
75
60
85
85
0
75
75
75
50
75
65
85
60
45
65
5
65
65
35
60
55
65
100
85
75
85
75
35
95
55
IHC EN2
Overall
Score
4
6
6
9
6
8
4
0
9
12
6
6
8
8
8
3
8
4
0
8
4
4
6
8
6
12
9
6
6
1
9
6
2
3
3
9
4
4
8
8
4
2
8
6
IHC EN2
Status
Negative
Positive
Positive
Positive
Positive
Positive
Negative
Negative
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Negative
Positive
Negative
Negative
Positive
Negative
Negative
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Negative
Positive
Positive
Negative
Negative
Negative
Positive
Negative
Negative
Positive
Positive
Negative
Negative
Positive
Positive
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
84
78
52
59
42
73
70
44
78
62
57
49
51
67
69
60
84
65
58
71
75
69
25
78
43
81
66
89
35
32
66
71
63
37
42
28
65
54
46
66
67
67
55
73
46
80
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Clear cell carcinoma
MMMT
MMMT
MMMT
Borderline mucinous tumour
Borderline serous tumour
Borderline mucinous tumour
Borderline serous tumour
Borderline mucinous tumour
Borderline serous tumour
Borderline mucinous tumour
Borderline mucinous tumour
Borderline mucinous & serous tumour
Borderline mucinous tumour
Borderline mucinous tumour
Borderline serous tumour
Borderline mucinous tumour
Borderline endometrioid tumour
Borderline mucinous tumour
Borderline mucinous tumour
Borderline serous tumour
Benign serous cystadenoma
Benign serous cystadenoma
Benign serous cystadenoma
Benign mucinous
Benign serous cystadenofibroma
Benign Serous cystadenoma
Benign mucinous cystadenoma
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Borderline
Benign
Benign
Benign
Benign
Benign
Benign
Benign
2
3
2
3
3
3
2
1
2
2
3
2
3
1
3
2
3
2
3
3
3
-
1a
3c
3b
3c
1a
3c
1a
1c
1a
4a
1a
1c
4
2a
3c
1a
3c
1a
3c
3c
3b
-
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Interval
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Primary
Resistant
Sensitive
Resistant
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Sensitive
Resistant
Resistant
-
110
1
1
88
0
99
129
11
25
87
114
112
103
9
97
61
1
1
2
-
94
1
14
5
94
5
99
134
16
127
123
119
118
108
31
102
100
84
1
3
5
9
-
Lost
Dead
Dead
Dead
Alive
Dead
Dead
Alive
Dead
Alive
Alive
Alive
Alive
Alive
Dead
Alive
Alive
Dead
Dead
Dead
Dead
Dead
-
2
0
2
2
2
2
1
2
1
1
2
2
2
3
3
3
2
2
1
2
1
3
2
1
1
1
1
2
3
1
0
2
2
1
2
1
1
2
2
1
1
2
2
0
1
1
75
0
75
85
40
40
65
60
85
85
85
65
85
95
85
85
85
45
75
75
85
35
45
65
85
85
60
50
20
75
0
90
50
60
75
85
65
45
65
40
85
35
85
0
75
50
8
0
8
8
4
4
3
6
4
4
8
6
8
12
12
12
8
6
4
8
4
6
6
3
4
4
3
6
6
6
0
8
6
3
8
4
3
6
6
2
4
4
8
0
4
3
Positive
Negative
Positive
Positive
Negative
Negative
Negative
Positive
Negative
Negative
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Positive
Negative
Positive
Negative
Positive
Positive
Negative
Negative
Negative
Negative
Positive
Positive
Positive
Negative
Positive
Positive
Negative
Positive
Negative
Negative
Positive
Positive
Negative
Negative
Negative
Positive
Negative
Negative
Negative
N
p
11.619
0.020
Normal Ovary
Benign
Borderline
Epithelial carcinoma
MMMT
4
7
17
63
3
Histological sub-type
Serous
Mucinous
Endometrioid
Clear cell
44
6
12
1
16
3
3
1
36.4%
50.0%
25.0%
100.0%
28
3
9
0
63.6%
50.0%
75.0%
0.0%
2.896
NS
Tumour Grade (All Serous)
2
3
3
40
1
14
33.3%
35.0%
2
26
66.7%
65.0%
0.003
NS
Tumour Stage (High Grade Serous)
Early (I & II)
Advanced (III & IV)
6
33
2
12
33.3%
36.4%
4
21
66.7%
63.6%
0.02
NS
III
IV
24
9
8
4
33.3%
44.4%
16
5
66.7%
55.6%
0.349
NS
III (neoadjuvant only)
IV (neoadjuvant only)
8
6
2
1
25.0%
16.7%
6
5
75.0%
83.3%
0.141
NS
Surgery (High Grade Serous)
Primary
Interval
28
13
12
3
42.9%
23.1%
16
10
57.1%
76.9%
1.497
NS
Platinum Sensitivity (High grade Serous)
Sensitive (≥6months)
Resistant (<6months)
30
5
12
1
40.0%
20.0%
18
4
60.0%
80.0%
0.734
NS
Platinum Sensitivity (High grade Serous; Interval Surgery)
Sensitive (≥6months)
Resistant (<6months)
9
3
2
1
7
2
77.8%
66.7%
0.148
NS
22.2%
33.3%
EN2 +ve
(5-12)
n
%
0
0%
1
14.3%
9
52.9%
40
63.5%
2
66.7%
x2
EN2 -ve
(0-4)
n
%
4
100%
6
85.7%
8
47.1%
23
36.5%
1
33.3%
Table 4.5. EN2 protein expression in human ovarian tissue from Cohort 2. The product of the EN2
staining intensity score (0-3+) and the percentage of positively stained cells score (0-4) was calculated. The resultant score of
0-4 represented EN2 negative staining, and 5-12 represented EN2 positive staining. Comparative analysis of the tumour
sections was performed using the chi-square test, which demonstrated a significant difference between the EOC tumours
compared with benign and normal tissue (x2 =11.619, p=0.020). The larger percentage value for each variable is highlighted in
red.
111
When data from Cohort 1 and Cohort 2 were combined, comparative analysis again
confirmed a significant difference between the EOC and borderline tumours compared to the
benign and normal samples, with a trend towards reduced staining with advanced tumour
stage in the high grade serous sub-type, however this did not achieve statistical significance
(Table 4.6). Figure 4.13 demonstrates the progression-free survival analysis of the HGSOC
tumours indicating that those with positive EN2 expression had significantly shorter
progression-free survival, namely a median PFS of only 10 months in the positive EN2 group
compared with 17.5 months in the negative EN2 group (p=0.0103). There was no significant
difference in overall survival between the groups (Figure 4.14).
112
N
EN2 +ve
(5-12)
n
%
0
0%
3
25.0%
16
66.7%
118
78.1%
4
57.1%
x2
p
26.854
<0.001
Normal
Benign
Borderline
Epithelial carcinoma
MMMT
4
12
24
151
7
n
4
9
8
33
3
EN2 -ve
(0-4)
%
100%
75.0%
33.3%
21.9%
42.9%
Histological sub-type
Serous
Mucinous
Endometrioid
Clear cell
120
8
18
5
25
3
3
2
20.8%
37.5%
16.7%
40.0%
95
5
15
3
79.2%
62.5%
83.3%
60.0%
2.468
NS
Tumour Grade (All Serous)
1
2
3
1
6
111
0
2
20
0.0%
33.3%
18.0%
1
4
91
100.0%
66.7%
82.0%
1.111
NS
Tumour Stage (High Grade Serous)
Early (I & II)
Advanced (III & IV)
7
103
2
20
28.6%
19.4%
5
83
71.4%
80.6%
0.343
NS
III
IV
74
29
11
9
14.9%
31.0%
63
20
85.1%
69.0%
3.481
0.062
III (neoadjuvant only)
IV (neoadjuvant only)
36
24
4
6
11.1%
25.0%
32
18
88.9%
75.0%
2.000
NS
Surgery (High Grade Serous)
Primary
Interval
53
59
13
10
24.5%
16.9%
40
49
75.5%
83.1%
0.983
NS
Platinum Sensitivity (High grade serous)
Sensitive (≥6months)
Resistant (<6months)
80
23
19
2
23.8%
8.7%
61
21
76.3%
91.3%
2.494
NS
Platinum Sensitivity (High Grade Serous; Interval Surgery)
Sensitive (≥6months)
Resistant (<6months)
45
13
9
1
20.0%
7.7%
36
12
80.0%
92.3%
1.071
NS
Table 4.6. EN2 protein expression in the combined human ovarian tissue cohorts. The product of the
EN2 staining intensity score (0-3+) and the percentage of positively stained cells score (0-4) was calculated. The resultant
score of 0-4 represented EN2 negative staining, and 5-12 represented EN2 positive staining. Comparative analysis of the
tumour sections was performed using the chi-square test, which demonstrated a significant difference between the EOC and
borderline tumours compared with the benign and normal tissues (x2 =26.854, p=<0.001). The larger percentage value for
each variable is highlighted in red.
113
E N 2 -v e
P e r c e n t s u r v iv a l
100
E N 2 +ve
p = 0 .0 1 0 3
50
0
0
50
100
150
P r o g r e s s io n - fr e e s u r v iv a l (m o n t h s )
Figure 4.13. Progression-free survival analysis of the combined high-grade serous
human ovarian tumours, comparing EN2 protein expression. The survival curves of EN2
negative versus EN2 positive protein expression for the high-grade serous ovarian tumours are demonstrated
(EN2 -ve = 0-4 IHC score; EN2 +ve = 5-12 IHC score). Progression-free survival is measured in months. The
median survival was significantly lower for all HGSOCs with EN2 +ve staining (10 months) versus those with
EN2 –ve staining (17.5 months)(p=0.0103). The Log-rank (Mantel Cox) test was used for analysis.
E N 2 -v e
P e r c e n t s u r v iv a l
100
E N 2 +ve
50
0
0
50
100
150
O v e r a ll s u r v iv a l (m o n t h s )
Figure 4.14. Overall survival analysis of the combined high-grade serous human
ovarian tumours, comparing EN2 protein expression. The survival curves of EN2 negative
versus EN2 positive protein expression for the high-grade serous ovarian tumours are demonstrated (EN2 -ve =
0-4 IHC score; EN2 +ve = 5-12 IHC score). Overall survival is measured in months. There were no significant
differences between EN2 staining (EN2 +ve median survival = 34 months; EN2 –ve median survival = 44
months; p=0.1813). The Log-rank (Mantel Cox) test was used for analysis.
114
4.2.1.3. EN2 protein expression in human ovarian tissue arrays
In order to confirm the EN2 protein expression findings in our patient cohorts, 2 ovarian
tissue arrays and 1 female reproductive tissue array were purchased and enzymatic
immunohistochemical staining and analysis were performed.
The first array, OV2082, consisted of 104 cases with two tissue cores per case, including
examples of serous papillary, mucinous, endometrioid, and clear cell epithelial carcinoma,
along with cancer adjacent normal ovary and normal ovary. When analysing the normal
ovary cores, it was evident that some only contained stromal tissue, with no epithelial cell
component. Hence, these were excluded from further analysis as they would not provide a
true assessment for EN2 expression.
Figure 4.15 demonstrates examples of the EN2
immunohistochemical staining, including normal tissue and malignant tissue, illustrating the
EN2 staining intensity (0-3+). The percentage positivity score was recorded as with the
patient tumour slides, however this could only be recorded for the area of tumour within the
two cores provided for each case. The product of the intensity and percentage positivity
score was calculated, to determine EN2 negative or positive staining. The specification sheet
from the OV2082 ovarian tissue array, including the EN2 intensity and percentage positivity
scores, is shown in Table 4.7.
Figure 4.15. Examples of EN2 protein expression from the OV2082 ovarian tissue array,
using enzymatic immunohistochemistry. Enzymatic immunohistochemistry examples of normal
ovary (A) and ovarian adenocarcinoma (B-E) stained for EN2, at 20x magnification. Ovarian surface
epithelium is indicated by the black arrow. The assigned staining intensity (0-3+) is indicated. EN2 positivity
detected by DAB staining (brown) is present in the cytoplasm, but not in the nucleus, and is not seen in the
surrounding stroma.
115
N
Age
Histology
Grade
Stage
IHC EN2
Staining
Intensity
IHC EN2
Staining
%
IHC EN2
Overall
Score
IHC
EN2
Status
A1
A2
A3
A4
A5
A6
A7
A8
A9
A10
A11
A12
A13
A14
A15
A16
B1
B2
B3
B4
B5
B6
B7
B8
B9
B10
B11
B12
B13
B14
B15
B16
C1
C2
C3
C4
C5
C6
C7
C8
C9
C10
C11
C12
C13
C14
C15
C16
D1
D2
D3
D4
D5
D6
D7
D8
D9
D10
D11
D12
D13
D14
D15
D16
E1
E2
E3
E4
E5
E6
E7
E8
40
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary cystadenocarcinoma
Serous papillary cystadenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary cystadenocarcinoma
Serous papillary cystadenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma (sparse)
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma with necrosis
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma (sparse)
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma (sparse)
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Adenocarcinoma (sparse)
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary cystadenocarcinoma (sparse)
Serous papillary cystadenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
1
1
1
100
4
Negative
1
1
3
100
12
Positive
1
1
3
100
12
Positive
2
1
2
100
8
Positive
2
1a
2
100
8
Positive
2
1b
3
100
12
Positive
2
1a
2
100
8
Positive
2
3c
3
100
12
Positive
2
3a
2
100
8
Positive
2
-
2
100
8
Positive
2
2
2
100
8
Positive
2
2b
3
100
12
Positive
2
1
3
100
12
Positive
2
3c
2
100
8
Positive
2
3c
100
12
Positive
2
1
No tissue
3
2
40
4
Negative
1
2b
2
100
8
Positive
2
4
2
70
6
Positive
3
4
1
100
4
Negative
2
3a
2
100
8
Positive
3
3c
2
60
6
Positive
2
3c
2
100
8
Positive
2
3c
3
100
12
Positive
2
1a
3
100
12
Positive
2
3c
2
75
8
Positive
2
3c
40
4
Negative
3
4
No tissue
2
2
100
8
Positive
2
2b
1
100
4
Negative
2
2c
2
60
6
Positive
2
1
3
100
12
Positive
2
4
100
12
Positive
3
1a
3
No tissue
2
100
8
Positive
2
3a
1
100
4
Negative
3
2
2
100
8
Positive
3
1
3
100
12
Positive
3
3c
2
100
8
Positive
48
30
68
36
35
74
50
52
65
50
70
43
43
61
46
38
64
46
59
47
49
54
41
37
47
59
56
76
67
40
44
55
50
45
53
Table 4.7. Full demographic data from the OV2082 ovarian tissue array, with detailed EN2
scoring. The array contained two cores for each case, therefore
116 EN2 staining was recorded for the combined area of tumour
within the two cores provided.
E9
E10
E11
E12
E13
E14
E15
E16
F1
F2
F3
F4
F5
F6
F7
F8
F9
F10
F11
F12
F13
F14
F15
F16
G1
G2
G3
G4
G5
G6
G7
G8
G9
G10
G11
G12
G13
G14
G15
G16
H1
H2
H3
H4
H5
H6
H7
H8
H9
H10
H11
H12
H13
H14
H15
H16
I1
I2
I3
I4
I5
I6
I7
I8
I9
I10
I11
I12
I13
I14
I15
I16
J1
J2
J3
J4
63
67
39
53
47
21
48
48
48
39
60
43
39
54
63
54
60
55
67
63
53
60
42
49
49
56
48
43
41
38
54
27
52
38
43
33
51
64
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Adenocarcinoma (sparse)
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma (ovarial tissue)
Serous papillary adenocarcinoma (ovarial tissue)
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma (sparse)
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma (sparse)
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Serous papillary adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous papillary adenocarcinoma
Mucinous papillary adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
2
1c
3
100
12
Positive
2
1a
3
100
12
Positive
2
1a
100
12
Positive
3
3c
3
No tissue
2
100
8
Positive
3
3b
2
100
8
Positive
3
1c
2
60
6
Positive
3
3a
2
100
8
Positive
3
2b
2
100
8
Positive
3
1b
2
100
8
Positive
3
1a
2
100
8
Positive
3
2
2
100
8
Positive
3
4
3
100
12
Positive
3
3c
2
70
6
Positive
3
1
2
100
8
Positive
3
4
3
100
12
Positive
3
3c
2
60
6
Positive
-
1
-
-
-
-
2
1a
1
20
2
Negative
3
1
2
100
8
Positive
3
3c
2
100
8
Positive
3
3c
1
100
4
Negative
3
2b
2
60
6
Positive
3
2b
2
100
8
Positive
3
3c
2
100
8
Positive
3
3c
2
100
8
Positive
3
3c
3
60
9
Positive
3
1
2
100
8
Positive
3
3c
2
100
8
Positive
3
1c
2
100
8
Positive
3
1
2
100
8
Positive
3
1a
3
100
12
Positive
1
1c
2
100
8
Positive
1
1
2
100
8
Positive
1
2b
2
100
8
Positive
1
3c
2
100
8
Positive
2
1a
2
100
8
Positive
1
3
2
100
8
Positive
1
1a
2
100
8
Positive
117
J5
J6
J7
J8
J9
J10
J11
J12
J13
J14
J15
J16
K1
K2
K3
K4
K5
K6
K7
K8
K9
K10
K11
K12
K13
K14
K15
K16
L1
L2
L3
L4
L5
L6
L7
L8
L9
L10
L11
L12
L13
L14
L15
L16
M1
M2
M3
M4
M5
M6
M7
M8
M9
M10
M11
M12
M13
M14
M15
M16
-
34
50
52
51
48
44
55
43
50
57
45
54
55
40
45
41
44
38
43
70
51
40
44
50
60
20
1 mon.
40
15
30
58
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Mucinous adenocarcinoma with necrosis
Mucinous adenocarcinoma
Mucinous adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Endometrioid adenocarcinoma
Clear cell carcinoma
Clear cell carcinoma
Clear cell carcinoma
Clear cell carcinoma
Clear cell carcinoma
Clear cell carcinoma
Clear cell carcinoma
Clear cell carcinoma
Transitional cell carcinoma
Transitional cell carcinoma
Cancer adjacent normal ovarial tissue
Cancer adjacent normal ovarial tissue
Cancer adjacent normal ovarial tissue
Cancer adjacent normal ovarial tissue
Cancer adjacent normal ovarial tissue
Cancer adjacent normal ovarial tissue
Cancer adjacent normal ovarial tissue
Cancer adjacent normal ovarial tissue
Cancer adjacent normal ovarial tissue
Cancer adjacent normal ovarial tissue
Normal ovarial tissue
Normal ovarial tissue
Normal ovarial tissue
Normal ovarial tissue
Normal ovarial tissue
Normal ovarial tissue
Normal ovarial tissue with follicle
Normal ovarial tissue with follicle
Normal ovarial tissue
Normal ovarial tissue
Malignant melanoma (tissue marker – skin)
2
3c
1
100
4
Negative
3
3c
1
100
4
Negative
3
1
2
60
6
Positive
3
1a
2
100
8
Positive
2
1
3
100
12
Positive
2
1a
1
100
4
Negative
2
1c
1
100
4
Negative
2
1
2
50
6
Positive
1
1
2
100
8
Positive
1
1
1
60
3
Negative
3
2
3
100
12
Positive
2
1c
2
100
8
Positive
3
1
1
100
4
Negative
2
3c
2
60
6
Positive
3
2a
100
8
Positive
-
1
2
No tissue
1
100
4
Negative
-
1c
2
30
4
Negative
-
1a
1
100
4
Negative
-
1a
2
20
4
Negative
2
1
1
100
4
Negative
-
-
-
-
-
-
-
-
-
-
-
-
-
-
0
100
0
Negative
-
-
1
40
2
Negative
-
-
1
30
2
Negative
-
-
-
-
-
-
-
-
-
-
-
-
-
-
0
100
0
Negative
-
-
-
-
-
-
-
-
-
-
-
-
-
-
3
100
12
Positive
118
The immunohistochemical staining demonstrated positive cytoplasmic EN2 protein staining
in 80.4% of epithelial ovarian cancer specimens, with no EN2 staining in the epithelial or
stromal components of the normal ovary and normal adjacent tissue. Comparative analysis of
the tumour sections using the Chi-squared statistical analysis confirmed a significant
difference between the malignant and normal tissues, as well as a significant difference
between the histological sub-types of EOC (Table 4.8). However there were no discernible
differences between tumour grade or disease stage. No treatment or follow-up data was
provided with this array.
N
EN2 +ve
(5-12)
n
%
0
0.0%
74
80.4%
x2
p
14.040
<0.001
Normal/Normal adjacent tissue
Epithelial carcinoma
4
92
EN2 -ve
(0-4)
n
%
4
100.0%
18
19.6%
Histological sub-type
Serous
Mucinous
Endometrioid
Clear cell
66
11
11
4
8
2
4
4
12.1%
18.2%
36.4%
100.0%
58
9
7
0
87.9%
81.8%
63.6%
0.0%
20.754
<0.001
Tumour Grade (All Serous)
1
2
3
4
28
33
1
5
2
25.0%
17.9%
6.1%
3
23
31
75.0%
82.1%
93.9%
2.589
NS
Tumour Stage (All Epithelial)
Early (I & II)
Advanced (III & IV)
56
34
12
6
21.4%
17.6%
44
28
78.6%
82.4%
0.189
NS
III
IV
28
6
5
1
17.9%
16.7%
23
5
82.1%
83.3%
0.005
NS
Tumour Stage (High Grade Serous)
Early (I & II)
Advanced (III & IV)
16
17
0
2
0.0%
11.8%
16
15
100.0%
88.2%
2.004
NS
III
IV
13
4
1
1
7.7%
25.0%
12
3
92.3%
75.0%
0.883
NS
Table 4.8. EN2 protein expression in the OV2082 ovarian tissue array. The product of the EN2
staining intensity score (0-3+) and the percentage of positively stained cells score (0-4) was calculated. The
resultant score of 0-4 represented EN2 negative staining, and 5-12 represented EN2 positive staining. Comparative
analysis of the tumour sections was performed using the chi-square test, which demonstrated significant differences
between the EOC and normal/normal adjacent tissue (x2 =14.040, p<0.001), and between the histological sub-types
of EOC (x2 =20.754, p=<0.001). The larger percentage value for each variable is highlighted in red.
119
EN2 immunohistochemical staining was subsequently performed on another ovarian cancer
tissue array, CJ2, which contained survival outcome data. This array consisted of 59 tissue
cores, including 25 epithelial ovarian cancers, 4 borderline tumours and 2 benign epithelial
tumours. There were also examples of MMMT, dysgerminoma and granulosa cell tumours,
however as our work focussed on epithelial ovarian cancer, only the EOCs, borderline and
benign epithelial cores were analysed further. Figure 4.16 shows examples of the EN2
immunohistochemical staining in malignant tissue, illustrating the EN2 staining intensity (03+). The percentage positivity score was recorded as with the patient tumour slides, however
this could only be recorded for the area of tumour within the core provided. The product of
these two scores was calculated, to determine EN2 negative or positive staining.
The
specification sheet from the CJ2 ovarian tissue array, including the EN2 intensity and
percentage positivity scores for each tissue core, is shown in Table 4.9.
The immunohistochemical staining demonstrated positive cytoplasmic EN2 protein staining
in 68% of epithelial ovarian cancer specimens, 25% of borderline tumours and 50% of benign
epithelial tumours. There were no normal ovary cores available for analysis in this array. Of
note, one of the benign tumours, a serous cystadenofibroma, showed strong EN2 expression,
however the follow-up data recorded that this patient died “from cancer” 72 months later, so
it is possible that the original diagnosis was incorrect.
Figure 4.16. Examples of EN2 protein expression from the CJ2 ovarian tissue array,
using enzymatic immunohistochemistry. Enzymatic immunohistochemistry examples of ovarian
adenocarcinoma (A-D) stained for EN2, at 20x magnification. The assigned staining intensity (0-3+) is
indicated. EN2 positivity detected by DAB staining (brown) is present in the cytoplasm, but not in the nucleus,
and is not seen in the surrounding stroma.
120
N
Age
Histology
1
2
3
4
5
6
7
8
9
10
11
12
13
14
65
26
35
65
48
44
49
67
75
50
44
50
47
38
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
58
70
65
79
57
43
68
50
23
51
70
21
34
61
54
53
54
34
31
51
61
24
27
41
56
41
27
55
23
64
43
46
47
48
49
50
48
50
25
46
37
51
52
53
54
55
56
57
58
59
60
55
55
45
37
58
65
39
70
58
.
Borderline serous papillary cystadenoma
Borderline mucinous cystadenoma
Borderline serous papillary cystadenoma
Serous cystadenocarcinoma
Endometrioid carcinoma
Granulosa cell tumour
Endometrioid carcinoma
Borderline mucinous cystadenoma
Serous cystadenofibroma
Serous papillary adenocarcinoma, Gr3
Fibroma
Serous papillary cystadenocarcinoma
Endometrioid carcinoma
Mixed Brenner tumour & mucinous
cystadenoma
Serous papillary cystadenocarcinoma
Serous papillary cystadenocarcinoma
Struma ovarii
Endometrioid adenoacanthofibroma
Serous surface papillary carcinoma
Serous papillary cystadenocarcinoma
Undifferentiated carcinoma
Clear cell carcinoma
Dysgerminoma
Common epithelial carcinoma, Gr3
Malignant muellerian mixed tumour
Dysgerminoma
Fibrothecoma
Serous papillary cystadenocarcinoma
Sertoli-Leydig cell tumour
Fibrothecoma
Fibrothecoma
Serous papillary cystadenocarcinoma
Dysgerminoma
Serous papillary cystadenocarcinoma
Granulosa-Theca cell tumour
Sertoli-Leydig cell tumour
Sclerosing stromal tumour
Clear cell carcinoma
Malignant muellerian mixed tumour
Clear cell carcinoma
Embryonal carcinoma
Granulosa cell tumour
Dysgerminoma
Serous papillary cystadenocarcinoma
Metastatic adenocarcinoma (likely
breast)
Malignant lymphoma
Adenocarcinoma from salpinx
Fibrothecoma
Metastatic undifferentiated carcinoma
Metastatic signet ring carcinoma from
stomach
Serous cystadenocarcinoma, Gr2
Fibrothecoma
Serous cystadenocarcinoma, Gr2
Endometrioid carcinoma
Fibrothecoma
Adenocarcinoma, Gr3
Clear cell carcinoma
Fibroma
Clear cell carcinoma
Carbon
Stage
Followup
(mths)
Followup
result
IHC EN2
Staining
Intensity
IHC EN2
Staining
%
IHC
EN2
Overall
Score
IHC
EN2
Status
Borderline
Borderline
Borderline
Cancer
Cancer
Borderline
Cancer
Borderline
Benign
Cancer
Benign
Cancer
Cancer
Benign
IA
IA
IA
II B
II B
IA
II B
II B
IA
II C
IA
IV
IV
IA
0
0
0
125
0
187
72
160
78
5
183
153
83
2
Lost
Lost
Lost
Dead
Lost
Alive
Dead
Dead
Dead
Lost
Alive
Dead
Dead
Lost
0
0
2
2
2
0
1
1
2
2
0
1
1
0
100
100
85
85
85
0
0
8
8
8
Negative
Negative
Positive
Positive
Positive
85
85
60
45
4
4
6
6
Negative
Negative
Positive
Positive
85
85
4
4
Negative
Negative
Cancer
Cancer
Benign
Benign
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Cancer
Benign
Cancer
Benign
Benign
Benign
Cancer
Cancer
Cancer
Borderline
Benign
Borderline
Cancer
Cancer
Cancer
Cancer
Borderline
Cancer
Cancer
Cancer
II B
III B
IA
IA
II B
III B
IB
IA
II B
II B
IA
IA
IA
III B
IB
IA
IA
II B
II C
III C
IA
IA
IA
II B
IA
IA
II A
IA
IA
III C
IV
42
26
180
156
16
48
164
30
6
174
174
174
173
172
82
172
170
21
0
5
136
0
0
161
18
159
14
157
150
34
141
Dead
Dead
Alive
Dead
Lost
Dead
Dead
Dead
Lost
Alive
Alive
Alive
Alive
Alive
Dead
Alive
Alive
Dead
Lost
Lost
Dead
Lost
Lost
Alive
Dead
Alive
Dead
Alive
Alive
Dead
Alive
2
1
1
2
2
3
0
2
0
2
0
0
0
2
3
0
0
2
0
2
1
2
0
3
2
2
0
1
0
1
0
85
35
8
2
Positive
Negative
40
85
85
4
8
12
Negative
Positive
Positive
60
6
Positive
40
100
4
0
Negative
Negative
60
6
Positive
65
6
Positive
85
8
Positive
75
75
75
12
8
8
Positive
Positive
Positive
85
4
Negative
Cancer
Cancer
Benign
Cancer
Cancer
II
II B
IA
IV
IV
133
136
135
135
134
Alive
Alive
Alive
Alive
Alive
3
2
0
2
1
85
8
Positive
Cancer
Benign
Cancer
Cancer
Benign
Cancer
Cancer
Benign
Cancer
III B
IA
II C
II B
IA
III C
III C
IA
IA
.
45
48
47
47
45
45
45
45
44
.
Dead
Alive
Alive
Alive
Alive
Alive
Alive
Alive
Alive
.
0
0
2
2
0
3
2
0
1
.
100
0
Negative
60
85
6
8
Positive
Positive
85
55
12
6
Positive
Positive
35
2
Negative
Tissue
type
Table 4.9. Full demographic data from the CJ2 ovarian tissue array, with detailed EN2
scoring for the malignant and benign tumours of epithelial cell origin.
121
Comparative analysis of the tumour sections using the Chi-squared statistical analysis is
shown in Table 4.10. In this tissue array, there were no discernible differences between the
benign and malignant tumours or between the histological sub-types of EOC, although the
numbers of non-serous tumours were small. There were no significant differences in staining
between the disease stages when considering all epithelial tumours, however there was a
trend towards reduced staining with advanced tumour stage. Information on tumour grade
was not available but when considering all of the serous sub-type of tumours, there was a
significant difference between the stages, where all early stage tumours stained positively for
EN2 but only 50% were positive in advanced stage disease (x2=4.773; p=0.029).
N
EN2 -ve
EN2 +ve
(0-4)
(5-12)
n
%
n
%
Benign
2
1
50.0%
1
50.0%
Borderline
4
3
75.0%
1
25.0%
Epithelial carcinoma
25
8
32.0%
17
68.0%
MMMT
2
1
50.0%
1
50.0%
Serous
15
4
26.7%
11
73.3%
Endometrioid
4
2
50.0%
2
50.0%
Clear cell
5
1
20.0%
4
80.0%
Early (I & II)
15
3
20.0%
12
80.0%
Advanced (III & IV)
10
5
50.0%
5
50.0%
III
8
3
37.5%
5
62.5%
IV
2
2
100.0%
0
0.0%
Early (I & II)
7
0
0.0%
7
100.0%
Advanced (III & IV)
8
4
50.0%
4
50.0%
III
7
3
42.9%
4
57.1%
IV
1
1
100.0%
0
0.0%
x2
p
2.885
NS
1.089
NS
2.482
NS
2.500
NS
4.773
0.029
1.143
NS
Pathology type
Tumour Stage (All Epithelial)
Tumour Stage (All Serous)
Table 4.10. EN2 protein expression in the CJ2 ovarian tissue array. The product of the EN2
staining intensity score (0-3+) and the percentage of positively stained cells score (0-4) was calculated. The
resultant score of 0-4 represented EN2 negative staining, and 5-12 represented EN2 positive staining. Comparative
analysis of the tumour sections was performed using the chi-square test, which demonstrated a significant
difference between the early and advanced stage serous tumours (x2 =4.773, p=0.029). The larger percentage value
for each variable is highlighted in red.
122
No treatment data was provided, however overall survival data was available for most of the
EOC cores. As information on tumour grade was not available, all of the serous tumours
were analysed together, but there was no observed significant difference in overall survival as
shown in Figure 4.17 (p=0.5532).
E N 2 -v e
P e r c e n t s u r v iv a l
100
E N 2 +ve
50
0
0
50
100
150
200
O v e r a ll S u r v iv a l ( m o n t h s )
Figure 4.17. Overall survival analysis of the serous tumours from the CJ2 ovarian tissue
array. The survival curves of EN2 negative versus EN2 positive protein expression for the serous ovarian
tumours are demonstrated (EN2 -ve = 0-4 IHC score; EN2 +ve = 5-12 IHC score). Overall survival is measured
in months. There were no significant differences between EN2 staining (p=0.5532). The Log-rank (Mantel
Cox) test was used for analysis.
123
The female reproductive tissue array, FRS801, was also purchased as it contained further
examples of normal ovary and EOC, but also normal fallopian tube and fallopian tube
adenocarcinoma. Figure 4.18 demonstrates examples of EN2 immunohistochemical staining
performed on the FRS801 female reproductive array, taken at 20x magnification. There was
strong EN2 staining in the serous adenocarcinoma cases compared with the normal ovary,
however it was interesting to see that the normal fallopian tube was also strongly EN2
positive, along with fallopian tube adenocarcinoma.
Figure 4.18. Examples of EN2 protein expression from the FRS801 female reproductive
tissue array, using enzymatic immunohistochemistry. Enzymatic immunochemistry examples of
normal ovary (A), ovarian adenocarcinoma (B, C), normal fallopian tube (D) and fallopian tube adenocarcinoma
(E) stained for EN2, at 20x magnification. Ovarian surface epithelium is indicated by the black arrow. EN2
positivity detected by DAB staining (brown) is present in the cytoplasm of the adenocarcinoma examples along
with the normal fallopian tube epithelium, but not in normal ovary surface epithelium.
124
4.2.2. EN2 EXPRESSION IN URINE FROM OVARIAN CANCER PATIENTS
EN2 has shown promise as a non-invasive urinary biomarker in both prostate and bladder
cancer, and demonstrates a positive correlation with prostate cancer volume and advancing
tumour stage.
In order to assess the prevalence of EN2 protein in the urine of patients with ovarian cancer,
midstream urine samples were either collected from patients undergoing debulking surgery
for suspected or confirmed ovarian cancer, or from known ovarian cancer patients attending
the Oncology Outpatients’ clinic. Twenty two urine samples were obtained from patients,
and 18 samples were obtained from female healthy volunteers and these were all evaluated
using the Direct ELISA.
The EN2 protein concentrations for the healthy volunteer control group were combined and a
cut-off value for EN2 positivity of 91.7ng/ml was calculated, using the control group “mean
plus 2x standard deviations”, as shown in Figure 4.19.
When reviewing the patients’
diagnostic and treatment information, we realised that three of the urine samples were from
patients with borderline tumours, one of which was EN2 positive, and one was from a patient
on surveillance with no serological or radiological evidence of active disease. Subsequent
analysis was therefore performed on the 18 samples from confirmed invasive EOC patients
with active disease. The EN2 protein concentration was deemed to be positive in the urine of
61% (11/18) of those invasive EOC patients, and the mean EN2 concentration was
significantly elevated in the urine of these patients compared with female healthy control
(Figure 4.20, p=0.0260). However, these samples were taken at varying stages of disease
and represented different histological sub-types of EOC, most notably 7 patients had recently
received chemotherapy and only 43% (3/7) of these samples were positive for EN2. When
just considering the 16 high-grade serous tumour patient samples, the most common
histological sub-type of EOC, 63% (10/16) were positive for EN2 and the mean EN2 protein
concentration was significantly elevated compared to the control group (p=0.0205). There
were only 7 urine samples from newly diagnosed patients with high-grade serous tumours,
however 86% (6/7) were positive for EN2 and their mean EN2 concentration was again
significantly elevated compared with the control group (p=0.0457).
125
E N 2 s e c r e t io n f r o m o v a r ia n c a n c e r u r in e s a m p le s u s in g t h e D ir e c t E L IS A
250
C o n tro ls
B o rd e rlin e tu m o u r
[ E N 2 ] n g /m l
200
N e w d ia g n o s is
150
R e c e n t c h e m o th e ra p y
R e la p s e d d is e a s e
100
N o a c tiv e d is e a s e
50
M e a n + 2 S D = 9 1 .7
C
o
n
tr
o
O ls
V
O 1
V 0
O 1 8
V 10
1
O 17
V
O 1
V 0
O 1 9
V 2
1 O 1 4
4 V 2
1 01 1 8
4 1 2
2 01 4H 9
1 1
2 01 4HG
8 1
2 01 4E B
5 1
0 4 M
3 Y
1 S
4
O SM
V
O 1
V 0
O 1 1
V 0
O 1 2
0
O OV1 5
V V 1
1 1 1
2 1
1
O /G 3
V D
1
O 22
V
O 1
V 25
1
O 32
V
1
3
0
0
Figure 4.19. EN2 protein concentration in urine samples from patients with EOC
compared with healthy controls, as determined by the Direct ELISA. Urine samples from
EOC patients at varying stages of disease were analysed by the Direct ELISA technique. The EN2 protein
concentration (ng/ml) for each sample was calculated relative to a PBS/0.1%Tween standard curve. The control
group represents the mean EN2 concentration from 18 female healthy volunteer urine samples (solid black bar).
The black line represents the positive cut-off value, calculated as the mean + 2SDs of the control cohort, where
values above the line are considered EN2 positive (11/18 invasive EOC samples).
E N 2 s e c r e t io n f r o m o v a r ia n c a n c e r u r in e s a m p le s u s in g t h e D ir e c t E L IS A
*
250
*
*
[ E N 2 ] n g /m l
200
150
100
50
s
is
)
is
s
u
w
D
d
ia
g
ia
n
g
o
n
s
o
ro
e
S
w
e
e
N
(N
h
s
ig
u
H
H
ig
h
G
ra
d
e
S
e
A
ro
ll
E
p
it
h
G
e
li
ra
a
d
l
e
T
C
u
o
m
n
o
tr
u
o
ls
rs
0
Figure 4.20. EN2 urinary protein concentration is significantly elevated in EOC
compared with healthy controls, as determined by the Direct ELISA. Scatter graphs of
individual values of urine EN2 concentration (ng/ml) tested by ELISA in patients with EOC. The results are
grouped by histology, or disease status, with the mean EN2 protein concentration (thick red bar) and standard
deviation (thin red line) depicted. The difference in mean EN2 concentration between patients and controls was
compared using the unpaired t-test with Welch’s correction (*=p<0.05).
126
4.2.3. EN2 EXPRESSION IN ASCITES FROM OVARIAN CANCER PATIENTS
Although more invasive to acquire than urine, the protein-rich ascitic fluid frequently
associated with advanced EOC, has shown promise as a potential prognostic biomarker.
In order to assess the prevalence of EN2 protein in the ascites of patients with ovarian cancer,
19 ascites samples were either collected at the time of debulking surgery, or from patients
undergoing ascitic drainage procedures on the wards. The 3 control ascites samples were
obtained from patients with benign gynaecological pathology and were kindly donated by Dr
Sandra Diebold at King’s College, London. These samples were all evaluated using the
Direct ELISA.
The EN2 protein concentrations for each control and patient ascites sample are shown in
Figure 4.21. Control samples 34 and 35 were from patients with benign ovarian cysts and
demonstrated very little EN2, whereas sample 36 was obtained from a patient with a benign
fimbrial cyst and contained 108ng/ml of EN2 protein. The fallopian tube was previously
shown to express EN2 on immunohistochemical staining (Figure 4.18; Section 4.2.1.3.), so
perhaps in the presence of a fimbrial cyst, these cells could be shed into the abdominal cavity,
especially if the cyst were to burst, explaining the elevated EN2 protein concentration in
ascites.
Three ascites samples were obtained from patients with metastatic breast
adenocarcinoma, but these all demonstrated low levels of EN2 protein (<50ng/ml).
We did not have enough control samples to allow us to combine the values and calculate a
cut-off value for EN2 positivity, however using an arbitrary value of 100ng/ml which was
similar to that used for the urine samples, only 25% (4/16) of the EOC ascites samples were
EN2 positive.
When reviewing the diagnostic and treatment information from these 4
patients, it was evident that the samples were all taken at the time of disease relapse and the
patients were not yet receiving chemotherapy. However there were other relapsed disease
patients who had negligible concentrations of EN2 protein in their ascites. Interestingly, 7 of
the EOC patients had either recently received or were currently receiving chemotherapy, and
the EN2 concentration was less than 100ng/ml in all of their ascites samples. There was only
one sample from a newly diagnosed patient (130912MC) and this contained very low levels
of EN2 protein.
127
E N 2 s e c r e t io n f r o m A s c it e s s a m p le s u s in g t h e D ir e c t E L IS A
800
[ E N 2 ] n g /m l
C o n t r o ls
600
A s c it e s ( O v a r ia n C a )
P e r ito n e a l w a s h in g s
400
( O v a r ia n C a )
A s c it e s ( B r e a s t C a )
200
C
o
C nt
o ro
C nt l 3
o ro 4
n l
2 tro 3 5
4
1 l
1 01 36
8
0 1
1 21 H
9
G
1 06 2J
2
0 12 M
1 71 V
9 2 L
0
2 71 M W
6
0 2G
0 71
7
D
1 08 2A
3 1 H
0 2
1 91 P
1 2 B
0
1 11 M C
1
0 3
1 11 C
7
0 3 F
1 11 H
2
3 S
0 09 S
1 1 P
1
1 0 3K
4 1 L
0
0 31 3A
3
0 4 P
0 41 CG
3
0 4
2 41 G O
6
1 4
2 0 SB
7 1
1 1
2 0 J
8 1 K
1 1
0 B
1 S
1
S
C
0
Figure 4.21. EN2 protein concentration in ascites samples from patients with EOC
compared with breast cancer ascites and benign disease controls, as determined by the
Direct ELISA. Ascites samples from EOC patients at varying stages of disease were analysed by the Direct
ELISA technique (red bars). Specimens from patients with benign gynaecological disease (black bars) and
breast cancer (pink bars) were obtained for comparison. The EN2 protein concentration (ng/ml) for each sample
was calculated relative to a PBS/0.1%Tween standard curve.
128
4.2.4. EN2 AUTOANTIBODY LEVELS IN THE PLASMA FROM OVARIAN
CANCER PATIENTS
Anti-tumour antibodies may prove useful as diagnostic biomarkers as they enable earlier and
lower level detection of tumour antigen than may be possible with direct protein assays,
hence allowing diagnosis of tumours at an early stage.
In order to assess the prevalence of EN2 autoantibodies in the blood of patients with
advanced epithelial ovarian cancer, plasma samples were collected from patients attending
the Oncology Outpatients’ clinic at The Royal Surrey County Hospital, most of whom had
received 3 cycles of neoadjuvant chemotherapy prior to their blood test. Ninety-eight plasma
samples were obtained from patients, and 123 samples were obtained from female healthy
volunteers. For control comparison of unrelated cancers, samples of plasma from 2 cohorts
of breast cancer patients and 1 cohort of prostate cancer patients were also analysed, along
with healthy control samples. The spontaneous IgG immune response against EN2 along
with the comparative tumour antigen NY-ESO-1, were evaluated in all of the samples using
the Indirect ELISA.
Scatter graphs depicting the individual values of plasma immunoglobulin IgG anti-EN2
antibodies in the EOC patients versus female healthy controls are shown in Figure 4.22.
Only 2.04% (2/98) ovarian cancer plasma samples demonstrated EN2 IgG responses, with
1.6% (2/123) in the control cohort, whereas 15.3% (15/98) demonstrated NYESO-1 IgG
responses which was significantly elevated compared with the control cohort (p<0.001).
The level of EN2 IgG response seen in ovarian cancer was equivalent to that found in breast
cancer (0-2%), however much higher EN2 IgG responses were found in prostate cancer
plasma samples (9.1%) which were significantly elevated compared with the control cohort
(p<0.001) (Figure 4.23).
129
Figure 4.22. There is no significant EN2 IgG
response in EOC patients compared with
healthy controls, as determined by ELISA.
Scatter graphs of individual values of plasma
immunoglobulin IgG anti-EN2 and anti-NYESO-1
antibodies tested by ELISA in patients with EOC, versus
healthy controls. The solid black lines indicate the mean.
The blue lines indicate the positive cut-off value,
calculated as the mean + 3SDs of the control cohorts.
Values above the cut-off were considered positive. The
levels of antibodies against the recombinant proteins are
expressed as optical density (OD). The proportion of IgG
responders to each antigen were compared between
patients and controls using a chi-squared test
(***=p<0.001).
Figure 4.23. The EN2 IgG response in Breast and Prostate cancer patients compared
with healthy controls, as determined by ELISA. Scatter graphs of individual values of plasma
immunoglobulin IgG anti-EN2 and anti-NYESO-1 antibodies tested by ELISA in 2 cohorts of patients with
breast cancer (A) and one cohort of prostate cancer patients (SUN study) (B), versus healthy controls. The solid
black lines indicate the mean. The blue lines indicate the positive cut-off value, calculated as the mean + 3SDs
of the control cohorts. Values above the cut-off were considered positive. The levels of antibodies against the
recombinant proteins are expressed as optical density (OD). The proportion of IgG responders to each antigen
were compared between patients and controls using a chi-squared test (*=p<0.05; **=p<0.01; ***=p<0.001).
130
4.3.
DISCUSSION
The aims of this chapter were to evaluate the En2 gene expression and EN2 protein in the
major histological sub-types of human epithelial ovarian cancer.
This evaluation was
predominantly conducted using preserved tumour specimens obtained at the time of
debulking surgery, however we were also able to evaluate EN2 expression in urine and
ascites samples, as well as quantify the expression of autoantibodies to EN2 in patient
plasma. In all of the cohorts of malignant specimens there was a preponderance of poorly
differentiated/high grade and advanced stage disease.
This reflected the typical patient
caseload at The Royal Surrey County Hospital, a tertiary referral centre for gynae-oncology,
where specialist surgeons mostly operate on cases with a high likelihood of malignancy
determined by imaging and CA125 serology.
Early stage disease is more commonly
diagnosed as an incidental finding when gynaecologists at District General hospitals remove
a benign-looking ovarian cyst.
We sought to acquire primary tumour specimens at the time of debulking surgery, rather than
peritoneal metastases or distant metastases, as EOC is known to be a heterogeneous disease
and can often demonstrate variable genetic mutations at different anatomical sites of spread
[332, 333]. High-grade serous tumours were predominant in all of our cohorts, and this
histological sub-type is the most common type in clinical practice, representing over 60% of
EOCs. Therefore the high-grade serous tumours were additionally evaluated within subset
analysis. Although these serous tumours may have been diagnosed in fallopian tube, ovary or
peritoneal specimens, they were combined for the purposes of our analysis, as they are
histologically similar, vary little in terms of treatment and survival outcome, and are
increasingly believed to have a common origin and pathogenesis [48, 49]. Many pathologists
believe that the origin of high-grade serous carcinoma actually lies in the fimbria of the
fallopian tube, with resultant incorporation of normal fimbrial epithelial cells into inclusion
cysts within the ovary, or direct spread of occult fallopian epithelial carcinoma to the ovary
[32, 33]. Hence normal fallopian tubes as well as normal ovary specimens were included in
our cohort analyses when possible. Some MMMT specimens were also present within our
tumour cohorts and these were initially analysed for En2 mRNA as well as EN2 protein,
along with the EOCs. However these were not included within subsequent analyses, such as
platinum sensitivity and survival outcome, as they are usually treated differently and have a
131
shorter overall survival compared with most other EOC tumours, which is related to the
sarcomatous component of the disease.
In our large cohort of specimens, En2 was shown to be significantly elevated at the mRNA
level in the pelvic serous and endometrioid histological sub-types of EOC, compared with
normal ovary and fallopian tube. The other EOC histological sub-types, mucinous and clear
cell, demonstrated lower level En2 expression compared with the serous sub-type, however
the numbers of these specimens were small. The non-invasive borderline tumours and benign
tumours demonstrated very little En2 expression. The relative expression of En2 was also
high in MMMT and this is likely to represent the epithelial cell component of these tumours.
Although they are still serous tumours, the relative En2 expression in primary peritoneal and
fallopian tube tumours was notably lower compared to serous ovarian tumours. In fact the
majority of these tumours were treated with chemotherapy prior to surgery (neoadjuvant
treatment) and a separate analysis comparing tumours exposed to neoadjuvant chemotherapy
compared with primary surgery specimens, showed a definite trend towards lower En2
expression in the neoadjuvant group. This may explain the lower overall expression levels in
the primary peritoneal and fallopian tube tumour groups.
When the serous tumours were combined as “pelvic serous carcinomas”, the significant
increase in En2 expression compared to normal specimens was maintained, likewise when
the 77 high grade serous ovarian carcinomas (HGSOCs) were separately analysed. However
there was a high level of variability in En2 expression within this group. Further analysis of
the HGSOCs showed that this variability was related to differing treatment schedules
(neoadjuvant chemotherapy versus primary surgery), stages of disease, and platinumsensitivity status.
En2 expression was lower in those who received neoadjuvant
chemotherapy as previously discussed, but also reduced in Stage IV disease, regardless of
prior treatment.
In the neoadjuvant treatment tumour specimens, En2 expression was
significantly higher in those that proved to have platinum-resistant disease on subsequent
follow-up, which is in keeping with our ovarian cancer cell line data that showed higher En2
mRNA expression in platinum-resistant compared with platinum-sensitive pairings. This
data was also supported by the survival curve analyses which demonstrated decreased
progression-free survival and overall survival in high En2 expressers. The level of statistical
significance varied depending on whether the CT value relative to ß-actin >1000 or >10,000
cut-off for high expression was utilised, but this may become clearer with longer follow-up as
20% of the PFS data and 40% of the OS data remained censored at the time of analysis.
132
Ultimately there appeared to be a PFS reduction of at least 6 months and an OS reduction of
at least 14 months in those with high En2 expression in their interval debulking surgery
specimens.
In current clinical practice, pathological analysis of the interval debulking surgical specimen
is performed in order to confirm diagnosis, disease stage and evaluate the viable disease
volume, but does not routinely include analysis of potential treatment response or prognostic
biomarkers. The gynae-oncology surgeon will record the visual amount of residual disease in
the abdomen and pelvis after surgery, as it is currently thought that this residual disease at
debulking surgery is related to PFS [55, 56, 359, 360]. At present, if the surgery is deemed to
be sub-optimal (>0.2cm residual disease), patients may receive more cycles of post-operative
chemotherapy or receive additional targeted agents such as Bevacizumab. However the
molecular profile of the viable cells within the surgical specimen may have more bearing on
subsequent survival, as our analysis of the En2 mRNA level suggests. Tumours from patients
developing resistance to platinum at an early time-point after diagnosis may harbour
inherently resistant cells within the heterogeneous primary tumour which highly express En2.
However in platinum-sensitive disease, the viable tumour cells that remain after neoadjuvant
chemotherapy have a significantly lower En2 expression. This observation supports the
potential use of En2 mRNA expression as a treatment response and prognostic biomarker, as
those patients with higher En2 expression levels at interval debulking surgery, may require an
alternative subsequent treatment protocol in view of their probable platinum-resistance.
The prevalence and expression pattern of EN2 was also assessed at the protein level in two
separate cohorts of patient tumours which enabled the combined analysis of a large number
of EOC specimens, with 111 HGSOCs in total. Positive cytoplasmic EN2 staining was
demonstrated in 78% of EOCs and over 65% of borderline tumours, however was negative in
the majority of benign tumours and in all normal ovary specimens. In keeping with the En2
mRNA expression levels, there was less EN2 positive staining seen in the mucinous and clear
cell histological sub-types, however there were only 13 examples of these. In HGSOC, there
was a trend towards reduced EN2 staining in the primary tumour of those with metastatic
(Stage 4) disease, as seen at the mRNA level, however this was only statistically significant
in Cohort 1. There was no significant increase in EN2 staining in HGSOC platinum-resistant
disease however those with positive EN2 expression had a 7.5 month reduction in PFS
compared with the negative EN2 group, which is similar to the PFS reduction seen in high
En2 mRNA expression compared with low expression.
133
Immunohistochemical staining of the ovarian cancer tissue arrays was performed to confirm
the findings in our patient cohorts, and again demonstrated high levels of EN2 protein
staining in EOC cores, particularly the serous histological sub-type, although they
demonstrated varying expression in the clear cell sub-type. As with our patient analysis,
there was no EN2 staining in the epithelial or stromal components of normal ovary or normal
adjacent ovary tissue. Again, reduced EN2 staining was observed in those with Stage 4
disease in both arrays, however this was only statistically significant in the CJ2 array.
Treatment data was not available for these arrays and only CJ2 had survival data, but no
significant differences in overall survival were demonstrated. When analysing tumour cores
from tissue arrays we must remember that the cores are usually very small and although they
are verified to ensure that tumour is present, it may not always be representative of the entire
ovarian tumour, especially as they may be quite heterogeneous in nature.
Interestingly, EN2 expression was positive in normal fallopian tube specimens within the
FRS801 tissue array, in contrast to the negative expression seen in ovarian surface epithelium
(OSE) of normal ovary examples. This is a similar finding to that of PAX8 expression,
which is another homeobox-containing transcription factor. PAX8 is not expressed in OSE
however it is present in the secretory epithelial cells of the normal fallopian tube, as well as
inclusion cysts within normal ovary, and is expressed in over 90% of serous ovarian
carcinomas [351]. This observation has helped to support the theory that high-grade serous
carcinoma originates in the fallopian tube [361]. We did not have examples of inclusion
cysts within normal ovary in our slides evaluated by immunohistochemistry, but it is possible
that EN2 is positive in such cells, much like PAX8 expression. In the literature, it appears
that PAX8 mRNA is also readily detectable in normal fallopian tube [362] however we
showed a very low level of En2 mRNA expression in our samples. We only had access to 3
examples of normal fallopian tube for mRNA analysis, two of which were commercial RNA
specimens and the other was obtained from theatre. It is difficult to know if these samples
definitely contained epithelial cells, and their absence could explain the low level of En2
expression, although the data sheet provided with one of the fallopian tube RNA samples
does quote “100% full cross-section seen” on the pathology verification notes. Likewise this
information is unclear for the normal ovary RNA commercial specimens, although we have
also demonstrated absence of EN2 protein staining in OSE on immunohistochemistry.
Although EN2 was not significantly expressed at the mRNA or protein level in normal ovary
specimens, the positive protein expression in normal fallopian tube makes it unlikely that
134
EN2 expression would prove clinically useful as a diagnostic biomarker of ovarian cancer in
human tissue as the specificity value would prove to be quite low. Nevertheless it has
significant potential as a prognostic and treatment response biomarker in tumours,
particularly at the mRNA level.
Elevated urinary EN2 protein was present in 58% of patients with invasive EOC, however
this rose to 63% when only considering those with high-grade serous EOC, and 86% when
evaluating newly diagnosed high-grade serous samples. In all of these cohorts, a significant
difference was demonstrated between their mean EN2 concentration and that of female
healthy controls. Our results are in keeping with the reported sensitivities of other urinary
proteins in advanced EOC patients, such as HE4 protein (89%) and mesothelin (75%) [166,
346].
When EN2 was investigated as a potential biomarker in prostate cancer, it
demonstrated a sensitivity of 66%, rising to 87% in high-grade tumours [303]. This suggests
that urinary EN2 levels could be utilised as a non-invasive diagnostic biomarker in highgrade serous patients, although a much larger sample size and similar series in other
institutions would need to be investigated in order to confirm this.
In addition, serial
measurement of EN2 protein in a patient’s urine from diagnosis through to treatment and
follow-up, may demonstrate decreasing levels in response to chemotherapy or surgery, with
subsequent increase in relapsed disease. This would make it useful as a treatment response
and disease relapse biomarker, especially as it would be less invasive than the currently used
serum CA125 biomarker, and may even supply prognostic information.
Ascitic fluid is more invasive to acquire than urine but is frequently associated with advanced
EOC. Therefore we provisionally analysed ascites samples in order to assess the prevalence
of EN2 in this protein-rich fluid. It proved difficult to acquire benign ascites samples for
comparison, especially working within a tertiary referral centre environment, so we could not
calculate a cut-off value for EN2 positivity in ascites.
Only 4 EOC ascites samples
demonstrated particularly high levels of EN2, compared to the control samples and breast
cancer ascites samples, however these were all from patients with relapsed disease. We were
unable to draw any comprehensive overall conclusions, as the sample size was small and the
samples were acquired at various stages of disease, including diagnosis, first and second
disease relapse. If a larger cohort of samples could be obtained either at the time of diagnosis
or at first relapse, progression-free survival and overall survival analysis could also be carried
out. If patients with elevated ascitic EN2 protein have a shorter PFS or OS, this could
potentially be used as a prognostic biomarker and help to guide treatment decisions.
135
Anti-tumour IgG antibodies provide a novel field in diagnostic biomarker research as they
allow recognition of potential tumour antigens at an earlier stage in tumour development
because the immune system responds to tumour emergence very early. This enables antigen
detection at very low levels, perhaps lower than can be reliably detected by direct protein
assay. As well as a potentially useful early diagnostic tool, they may also aid monitoring for
treatment response and tumour progression. As antibodies can facilitate antibody-dependent
cell-mediated cytotoxicity as well as complement-dependent cytotoxicity, it is thought that
such antibodies may also provide anti-tumour defence [169].
Serum autoantibodies directed against tumour associated antigens have been frequently
detected in the blood of patients with different types of cancer [169]. As yet, there are no
leading candidates in ovarian cancer although many have been studied, including epithelial
cell adhesion molecule (Ep-CAM) [170], HSP-90 [171, 172], HOXA7 [173], HER2 [174],
p53 [106, 347] and MUC1 [133]. Therefore, we hypothesised that expression of EN2 protein
in EOC could induce an autoantibody response which could be utilised as a diagnostic or
treatment response biomarker. Unfortunately our research demonstrated very low levels of
EN2 IgG antibody response in EOC patient plasma, at only 2.04% compared with 1.6% in
controls. However levels of plasma immunoglobulin IgG anti-NY-ESO-1 antibodies were
significantly higher than the control group at 15.3%. NY-ESO-1 is a cancer/testis antigen
which appears to be one of the most immunogenic tumour antigens. It has been shown to
induce a humoral immune response in around 50% of patients with advanced NY-ESO-1
expressing tumours [363, 364].
Admittedly we analysed plasma EOC samples rather than serum, which is usually used in
autoantibody studies, as these were the only EOC samples available within our Biobank,
however the findings were unlikely to differ considerably in serum especially as the NYESO-1 results were consistent with the published literature suggesting that our cohort was
representative of the EOC population [365]. The majority of the EOC patients had already
received 3 cycles of chemotherapy prior to their blood samples being taken, however similar
EN2 IgG antibody responses were seen in the serum of breast cancer patients whose samples
were taken prior to first definitive treatment, so would have clearly detectable disease. It was
interesting to see that prostate cancer patients demonstrated a significantly higher prevalence
of IgG responses to EN2 as well as NY-ESO-1 in serum, compared with healthy control
males (9.1% vs 1.4% for EN2; 4.8% vs 0.9% for NY-ESO-1). These were pre-treatment
samples from men with all stages of prostate cancer. This observed difference between male
136
and female cancer autoantibody responses has been previously reported and could relate to
the effect of sex hormones on the humoral immune response [366, 367].
Ultimately we demonstrated that serum/plasma IgG anti-EN2 antibodies were rarely detected
in female patients with EOC and breast cancer, and were not elevated compared to female
healthy controls. In contrast, IgG anti-NY-ESO-1 responses were clearly demonstrated in
both female cancer groups. This suggests that EN2 IgG antibody responses cannot be used as
a diagnostic tool and the lack of immunogenicity limits its use as an immunotherapeutic
target in the blood of EOC patients.
137
4.4.
CONCLUSION
Thorough evaluation of En2 gene expression as well as EN2 protein expression in human
tissue and biological fluids has enabled us to better define its role as a biomarker in epithelial
ovarian cancer. The En2 mRNA level in tumour tissue could be used as a diagnostic
biomarker as it was significantly elevated compared with normal ovary, however the
acquisition of tissue is an invasive procedure and does not help in the diagnosis of early stage
disease, when an ovarian mass may be undetectable on imaging. However, it holds more
promise as a treatment response biomarker in interval debulking surgery tumour tissue where
the En2 mRNA level falls in response to chemotherapy, or as a prognostic biomarker in such
tissue where an elevated level could predict platinum resistance and a shorter progressionfree survival.
EN2 protein shows promise as a diagnostic biomarker in urine and given its non-invasive
acquisition, it could be used for EOC screening, or to detect early relapsed disease. A larger
sample size is required to assess this further.
Although EN2 protein was detected in some
ascites samples, a larger cohort is required to evaluate its role as a diagnostic or prognostic
biomarker. Finally, EN2 protein analysis via immunohistochemistry could be utilised as a
prognostic marker in interval debulking surgery tumour tissue, as those with positive EN2
expression had a shorter progression-free survival. The expression of En2 mRNA or EN2
protein in this tissue could help the clinician to plan the patient’s subsequent treatment
regimen.
138
CHAPTER 5
DNA METHYLATION OF THE EN2
PROMOTER REGION
139
5.
DNA METHYLATION OF THE EN2 PROMOTER REGION
5.1.
INTRODUCTION
Epithelial ovarian cancer is commonly diagnosed at an advanced stage of disease which is
related to the non-specific presenting symptoms and the absence of a screening programme
with robust, standardized methods of early detection. It is assumed that such early detection
would result in diagnosis at an earlier stage of disease, which is commonly associated with an
improved outcome. Hence there is a great deal of interest in biomarker research to aid with
early diagnosis, and this often coincides with the discovery of prognostic and treatment
response biomarkers.
There has been increasing interest in the association between DNA methylation and gene
expression particularly in the field of oncology, where aberrant methylation has been
observed in several tumour types. DNA methylation describes the addition of methyl groups
to cytosine or adenine DNA nucleotides and usually occurs at CpG islands, which are clusters
of adjacent cytosine and guanine nucleotides [110, 111]. These CpG islands tend to lie
within the promoter region of a gene and are usually unmethylated.
In this state,
transcription can occur with resultant gene expression, whereas methylation of the promoter
CpG island appears to repress promoter activity and the gene is silenced [112]. DNA
promoter hypermethylation, along with global hypomethylation, is an important epigenetic
process involved in development and cell differentiation, however it has also been
demonstrated in cancer, and is thought to be an early event in tumour development [368,
369]. Therefore it could prove useful in the detection of precancerous lesions or early stage
disease [110, 113].
Methylation patterns in primary tumour specimens have also been
shown to mimic those in other biological fluids, such as plasma and sputum, which supports
its potential use as a diagnostic biomarker [114-124, 370-377].
In colorectal cancer, the promoter sequence of the septin 9 (SEPT9) gene, which belongs to a
class of GTPases, demonstrates epigenetic silencing via hypermethylation. Studies have
shown sensitivities of 67-90% and specificities of 88-90% for the detection of colorectal
cancer using plasma SEPT9 methylation [372, 378, 379]. It could also be detected in 30% of
patients with adenomas and hyperplastic polyps i.e. precancerous lesions [120]. However a
140
subsequent prospective screening trial of methylated SEPT9 in 7941 individuals, published in
early 2013, only demonstrated a sensitivity of 48.2% for the detection of colorectal cancer
[380]. Combinations of methylation markers in plasma or serum have improved sensitivities
to 87%, with specificities of 90-96% [381, 382]. Methylation of the vimentin (VIM) gene is
readily detectable in stool, and detects 86% of colon cancers and 76% of adenomas at a
diagnostic specificity cut-off of 95% [371].
CDKN2A (p16) and SHOX2 methylation have been investigated in lung cancer. Promoter
methylation of CDKN2A results in transcriptional silencing and appears to be an early event
in lung carcinogenesis [383]. It has been identified in smokers and non-smokers and in all
histological sub-types [384]. Of particular interest in early lung cancer detection, CDKN2A
methylation is detectable in sputum samples prior to clinical diagnosis, namely in squamous
cell carcinoma [118, 119], and has also been identified in plasma, serum and
bronchioalveolar lavage [115-117]. SHOX2 DNA methylation is of particular interest in
cases where there is diagnostic uncertainty from cytology results, i.e. they are cytologically
negative or there are too few cells for analysis. One study identified the presence of SHOX2
methylation in 62% of bronchial aspirates that were negative by cytology [121]. In lung
tumour tissue the sensitivity for SHOX2 methylation has been shown to be as high as 96%
[385]. In plasma samples, SHOX2 methylation gives a sensitivity of 60% and a specificity of
90% for lung cancer detection [373].
In prostate cancer, epigenetic silencing of GSTP1 due to promoter methylation is present in
over 90% of cases, and is rarely seen in normal prostate or benign conditions [386]. As with
other tumour sites, the hypermethylated gene can be detected in precancerous prostatic lesion,
and has also been isolated in serum, plasma and urine [122, 374-377].
Hypermethylation of one or more of BRCA1, CDKN2A and 14-3-3σ has been found in 95%
of sporadic breast cancers [387]. Primary breast tumour specimens and matched patient
serum blood samples have also shown concordance in gene methylation [123]. Wong et al
demonstrated that peripheral blood methylation of the BRCA1 promoter region was
associated with a 3.5-fold increased risk of developing breast cancer before the age of 40
years [370]. Conversely Bosviel et al showed a significant difference in BRCA1 methylation
in the sera of breast cancer patients compared with controls, in those over 70 years,
postmenopausal or with low body mass index [388]. Hypermethylation of a panel of six
141
genes (GSTP1, RAR-β2, CDKN2A, p14, RASSF1A, and DAP-kinase) was noted in 82% of
nipple fluids from cancerous breasts, but none was evident in normal breasts [124].
Up to 21% of epithelial ovarian carcinomas contain BRCA, FANCF or PALB2 promoter
methylation, which is associated with poor prognosis [125, 126]. Fiegl et al studied 71 genes
within ovarian cancers, and also found that methylation of the homeobox genes HOXA10 and
HOXA11 differed compared with non-cancerous tissue [127]. HOXA11 methylation showed
a particular association with poor outcome, raising the possibility for use as a prognostic
marker.
In 2007, Rauch et al observed hypermethylation of all four HOX gene clusters along with
Engrailed-1 and Engrailed-2 in lung cancer cell lines [213] and subsequent research has
identified hypermethylated En1 and En2 in a variety of tumour sites. Karpinski et al studied
the methylation status of three CpG islands at the 2q14.2 chromosomal band, including En1,
in 148 sporadic colorectal cancers [306]. Generally 18% to 25% of sporadic colorectal
cancers demonstrated CpG island methylator phenotype (CIMP), and the average number of
methylated sites was significantly higher in these tumours, namely 70% En1 methylation was
seen in CIMP positive tumours compared with 22% in CIMP negative tumours, however the
authors did not comment upon the corresponding En gene or protein expression.
Subsequently Mayor et al detected methylated En1 in stool DNA from patients with
colorectal carcinoma with 44% sensitivity and 97% specificity in patients with corresponding
tumour methylation, which could be developed as a diagnostic biomarker [307]. In serum
however, the sensitivity was only 11%. En1 was also methylated in 65% of primary prostate
tumours, significantly differentiating cancer from normal tissue when combined with SCTR
hypermethylation [309]. Combined methylation of these two genes may provide potential
novel biomarkers for prostate cancer detection.
Significant hypermethylation of the En1 gene has been observed in adenoid cystic carcinoma
(ACC) which is a rare and aggressive malignancy of the salivary glands, but interestingly
over-expressed EN1 protein has been demonstrated in such tumours [311, 312]. The authors
suggested a potential role for En1 methylation status as a biomarker in this disease, as there
were observed correlations with histological tumour grade, tumour location and patient
outcome. En2 hypermethylation, along with other homeobox genes, has been identified in
the follicular lymphoma cell line, RL, and in ten primary follicular lymphomas [310].
Similar to the observation of En1 hypermethylation in ACC, En2 hypermethylation in
142
follicular lymphoma did not coincide with transcriptional down-regulation of En2, indicating
aberrant epigenetic regulation in follicular lymphoma.
Although En2 hypermethylation is present in a number of malignancies its functional role is
still unknown, as it does not appear to contribute to gene silencing in all tumour sites. There
is no published evidence of its prevalence in epithelial ovarian cancer, or of its potential for
use as a biomarker in this setting.
143
Study Objectives and Hypothesis
The objectives of this chapter were to evaluate the methylation status of the En2 gene
promoter region in a broad selection of EOC cell lines as well as human epithelial ovarian
tumours and normal ovary.
The methylation status would also be compared to the En2
mRNA expression levels determined in cell lines in Chapter 3, and in human tissue in
Chapter 4. If hypermethylation was present in such specimens, this may prove useful as a
diagnostic, prognostic or treatment response biomarker.
We hypothesised that En2 promoter hypermethylation would be present in a number of the
EOC cell lines and human tumour specimens independent of En2 mRNA expression, as En2
hypermethylation has been observed in other malignancies and lacks association with gene
silencing. In keeping with the latter findings, perhaps benign tumour and normal tissue
would demonstrate En2 promoter hypomethylation. The methylation percentage may also
vary between platinum-sensitive and platinum-resistant tumours, making it useful as a
prognostic biomarker.
144
5.2.
RESULTS
5.2.1. DNA methylation of the En2 Promoter Region in Cell Lines
Genomic DNA was extracted from cultures of fourteen EOC cell lines representing a variety
of histological sub-types. PEO1 represented a cell line derived from a platinum-sensitive,
serous ovarian carcinoma. The PEO4 cell line was derived from the same patient, once they
had developed platinum-resistant disease. Similarly, PEO14 and PEO23, and PEA1 and
PEA2, were paired cell lines derived from the same patient before and after developing
platinum-resistant disease. The SKOV3 and OVCAR3 serous cell lines were also noted to be
platinum resistant. A human normal epithelial ovarian cell line was not available therefore a
human fibroblast cell line was used as a negative control, along with NIH3T3, a murine
fibroblast cell line, and WPMY1, a myofibroblast stromal cell line derived from normal adult
prostate. The positive control was CpG Methylated Jurkat gDNA, which is derived from
human acute T-Cell leukemia cells, and is reported to be 100% methylated gDNA.
The cell line and Jurkat gDNA were incubated overnight either with a McrBC endonuclease
to digest the methylated regions of the gene, or with control buffer. Quantitative rt-PCR was
then performed on the gDNA using En2 promoter primers, with the difference in CT value
between the paired samples (CT) representing the amount of methylated En2. In order to
calculate the percentage of methylated En2 in the promoter region, each cell line CT value
was compared with the CT value from the 100% methylated Jurkat gDNA. For each cell
line, the percentage of methylated En2 was calculated for three independent rt-PCR
experiments (Figure 5.1).
When En2 gene expression was previously measured in the 3 negative control cell lines,
human fibroblasts, NIH3T3 and WPMY1, (data also presented in Chapter 3), they all
demonstrated very low levels of En2 mRNA, however they also demonstrated very low
mean levels of percentage En2 promoter methylation (3.27%, 2.74% and 10.01%
respectively). This suggested that low En2 gene expression was not occurring as a result of
En2 hypermethylation-related epigenetic silencing.
145
Figure 5.1. En2 promoter methylation status and mRNA expression in EOC cell lines
and controls. (A) The percentage En2 promoter methylation status compared to CpG Jurkat gDNA was
analysed in fourteen ovarian cancer cell lines, using quantitative rt-PCR. Fibroblasts, NIH3T3 and WPMY1
cells were used as negative controls and CpG Jurkat gDNA as a positive control. (B) The En2 mRNA
expression is shown relative to the housekeeping gene β-actin (x100,000) for the same cell lines. Error bars
represent the SD (n=3) and the one-way ANOVA with Bonferroni correction was used for analysis (*=p<0.05;
****=p<0.0001).
Evaluation of the EOC cell lines revealed En2 promoter methylation of >40% in all lines
except for OV90, PEO14 and PEO23, with significant elevation compared to the fibroblast
and NIH3T3 cell lines (p<0.01). It was interesting to note that the hypermethylation of the
serous lines SKOV3, CaOV3, OVCAR3, COV318 and PEA1, along with the clear cell line
ES2 corresponded to low level En2 mRNA expression, which could indicate epigenetic
silencing of the gene in those cell lines. However PEO4, PEO23, PEA2 and TOV21G clearly
showed high percentage methylation and mRNA over-expression. Ultimately, there was no
146
overall correlation between percentage methylation of the En2 promoter and En2 gene
expression in the EOC and control cell lines, as demonstrated in Figure 5.2. In addition,
there were no significant differences between the histological sub-types of EOC or between
the platinum-sensitive/resistant paired serous cell lines.
% E n 2 M e th y la tio n
100
50
0
0
100
200
300
E n 2 m R N A e x p r e s s io n
Figure 5.2. En2 promoter methylation status does not correlate with En2 mRNA
expression in EOC cell lines. The percentage En2 promoter methylation value (relative to CpG Jurkat
gDNA) was plotted against the mean En2 mRNA expression (relative to the housekeeping gene β-actin
x100,000) for each EOC cell line. Linear regression revealed no clear correlation (r 2=0.02436; p=0.5497).
147
5.2.2. DNA methylation of the En2 Promoter Region in Ovarian Tumours
Genomic DNA was extracted from 28 tumour samples preserved in RNAlaterTM. The
tumours represented a variety of EOC histological sub-types, namely 16 serous, 2
endometrioid, 2 clear cell, 4 borderline tumours and 4 benign tumours. The full demographic
data for these tumours is shown in Table 5.1. Five normal ovary and one normal fallopian
tube gDNA sample were analysed for comparison, either using commercial gDNA or tissue
specimens obtained from theatre. The positive control was CpG Methylated Jurkat gDNA,
which is derived from human acute T-Cell leukemia cells, and is reported to be 100%
methylated gDNA.
Table 5.1. Full demographic data from selected human ovarian tumours, normal tissue
and commercial ovary gDNA, with the resultant percentage of methylated En2. The
corresponding En2 mRNA expression level is also shown for comparison (results taken from Chapter 4).
148
The tissue specimen and Jurkat gDNA were incubated overnight either with a McrBC
endonuclease to digest the methylated regions of the gene, or with control buffer.
Quantitative PCR was then performed on the gDNA using En2 promoter primers, with the
difference in CT value between the paired samples (CT) representing the amount of
methylated En2. In order to calculate the percentage of methylated En2 in the promoter
region, each specimen CT value was compared with the CT value from the 100%
methylated Jurkat gDNA. The percentage of methylated En2 for each tumour specimen and
normal ovary/fallopian tube control is demonstrated in Figure 5.3, and there were clearly
some examples of invasive tumour that had elevated levels of methylation compared to the
normal specimens.
% M e t h y la t io n o f E n 2 p r o m o t e r r e g io n in o v a r ia n t u m o u r s
% M e th y la t io n
100
C p G J u rk a t C o n tr o l
E p ith e lia l tu m o u rs
50
B o rd e rlin e tu m o u rs
B e n ig n tu m o u rs
N o rm a l o v a ry /
fa llo p ia n tu b e
0
Figure 5.3. En2 promoter methylation status in individual human EOC tumours, and
normal tissue. The percentage En2 promoter methylation status compared to CpG Jurkat gDNA was
analysed in twenty-eight ovarian tumours and six normal ovary/fallopian tube specimens, using quantitative rtO v a r ia n t u m o u r
PCR. CpG Jurkat gDNA was used as a positive control. Error bars represent the SE in cases where the sample
analysis was repeated.
Figure 5.4 shows the mean percentage En2 methylation for all of the EOC specimens along
with the borderline, benign and normal specimens. The mean percentage methylation of the
En2 promoter region in EOC tumours was significantly elevated compared with the normal
specimen group (27.6% versus 10.71%; p=0.0401).
Although higher than the normal
specimens, the mean percentage En2 methylation in benign and borderline epithelial tumours
was lower than the malignant tumours, however this was not statistically significant.
149
% M e t h y la t io n o f E n 2 p r o m o t e r r e g io n in o v a r ia n t u m o u r s
% E n 2 M e th y la tio n
100
*
50
b
tu
n
ia
p
n
ry
B
/f
e
a
n
ll
o
ig
e
in
rl
e
rd
a
o
o
v
B
l
a
rm
u
N
o
J
G
C
p
e
rs
o
m
tu
m
tu
tu
l
a
li
e
h
it
p
E
u
rs
o
o
m
la
y
th
e
m
%
0
0
(1
rk
a
t
u
u
te
d
rs
)
0
Figure 5.4. Mean percentage En2 promoter methylation in human EOC tumours,
classified by histological sub-type. The mean percentage En2 promoter methylation status compared to
CpG Jurkat gDNA of overall EOC, borderline, benign and normal tumours were grouped together for
comparison. Error bars represent the SE, and the Mann-Whitney test was used for analysis (*=p<0.05).
The En2 mRNA expression was previously measured in all of these tissue specimens (see
Chapter 4, Section 4.2.1.1.) and is detailed in Table 5.1 along with the corresponding
percentage En2 methylation. The normal and benign specimens demonstrated low levels of
En2 mRNA, along with low levels of percentage En2 promoter methylation, which suggested
that low En2 gene expression did not occur as a result of En2 hypermethylation-related
epigenetic silencing.
Evaluation of the EOC tumour specimens revealed a mean En2
promoter methylation of 27.6% with two samples in particular demonstrating 51.23% and
71.43% methylation, however there was no clear correlation with En2 gene expression
(Figure 5.5). The latter two samples highly expressed the En2 gene however another serous
tumour with 39.82% methylation, showed very low En2 gene expression.
150
% E n 2 M e th y la tio n
100
50
0
0
20000
40000
60000
80000
E n 2 m R N A e x p r e s s io n
Figure 5.5. En2 promoter methylation status does not correlate with En2 mRNA
expression in EOC human tumours. The percentage En2 promoter methylation value (relative to
CpG Jurkat gDNA) was plotted against the En2 mRNA expression (relative to the housekeeping gene β-actin
x100,000) for each EOC tumour. Linear regression revealed no clear correlation (r 2=0.09639; p=0.1828).
The endometrioid and clear cell histological sub-types of EOC did all show a higher level of
methylation (>29%) however there were only 4 examples of these non-serous tumours, so it
was difficult to draw conclusions from this. When analysing all of the tumours according to
treatment response, there was a trend towards increased En2 promoter methylation in those
with platinum-sensitive disease, however this was not statistically significant (Figure 5.6).
In order to carry out progression-free and overall survival comparisons of the data, a cut-off
value for low versus high percentage En2 methylation was calculated using the mean
methylation value of the normal specimens + 2 standard deviations of the data. This gave a
cut-off value of 21.26%, however using this to determine low versus high En2 methylation
did not result in any significant differences in progression-free or overall survival (Figure
5.7).
151
% E n 2 M e th y la tio n
100
50
P
la
ti
n
u
m
R
P
e
la
s
ti
is
n
ta
u
n
m
t/
S
R
e
e
n
fr
s
a
it
c
iv
to
e
ry
(n
(n
=
1
=
5
4
)
)
0
Figure 5.6. Mean percentage En2 promoter methylation in human EOC tumours,
classified by platinum sensitivity status. There was no statistically significant difference between the
mean percentage En2 promoter methylation status of the platinum-resistant/refractory (PFS<6months) and
platinum-sensitive (PFS6months) EOC tumours. Error bars represent the SE, and the Mann-Whitney test was
used for analysis; n=number of sample.
E n 2 m e th y la tio n < 2 1 %
E n 2 m e th y la tio n < 2 1 %
100
E n 2 m e th y la tio n > 2 1 %
50
0
P e r c e n t s u r v iv a l
P e r c e n t s u r v iv a l
100
E n 2 m e th y la tio n > 2 1 %
50
0
0
20
40
60
80
0
P r o g r e s s io n - f r e e S u r v iv a l ( m o n t h s )
20
40
60
80
O v e r a ll S u r v iv a l ( m o n t h s )
Figure 5.7. Survival analyses of the human epithelial ovarian tumours, comparing
percentage En2 promoter methylation. The progression-free survival (A) and overall survival (B)
curves of high versus low percentage En2 methylation are demonstrated for all specimens (the cut-off value for
low versus high En2 methylation was 21.26% which was equal to the mean percentage En2 methylation for the
normal ovary specimens + 2SDs). Survival is measured in months. There were no significant differences
between median progression-free survival or overall survival according to low versus high En2 methylation
status. The Log-rank (Mantel Cox) test was used for analysis.
152
5.3.
DISCUSSION
The aims of this chapter were to evaluate the methylation status of the promoter region of
En2 genomic DNA in control and epithelial ovarian cancer cell lines, as well as human
epithelial tumours and normal tissue. The methylation status would then be compared to the
En2 mRNA expression levels previously determined in Chapter 4.
Low mean levels of methylated En2 were demonstrated in normal ovary and fallopian tube as
well as benign tissue, compared with epithelial ovarian cancer specimens although the
specimen numbers were small. These findings were mirrored in the 3 negative control cell
lines as well as the majority of EOC cell lines. In keeping with the original hypothesis, the
normal and benign specimens, as well as the negative control cell lines, demonstrated En2
promoter hypomethylation coupled with low levels of En2 mRNA. This suggested that
hypomethylation of the En2 gene did not influence expression of the gene, at least in terms of
over-expression. Similarly, many of the EOC tumour specimens demonstrated En2 promoter
hypermethylation as well as En2 mRNA over-expression, suggesting that promoter
methylation lacks association with epigenetic silencing of this gene, as hypothesised. Of
note, there were a few examples of EOC cell lines and tumours that did show En2 promoter
hypermethylation alongside low expression of En2 mRNA, but it is difficult to know if the
latter resulted from epigenetic silencing due to the hypermethylation, or from alternate
genetic or epigenetic control. There were no significant differences in En2 methylation status
between the different histological sub-types or the platinum sensitivity status, and En2
hypermethylation did not appear to influence progression-free or overall survival.
In this small cohort of tumours there was no conclusive evidence that En2 methylation status
could be used as a diagnostic or prognostic biomarker in epithelial ovarian cancer, however it
was interesting to observe the apparent independence between En2 gene methylation and
mRNA expression. Traditionally it has been believed that hypermethylation of a gene’s
promoter region results in promoter activity repression and subsequent silencing of the gene
[112]. So it was hypothesised that this epigenetic modification may be occurring within
tumour suppressor genes in cancerous cells, or that global hypomethylation may contribute to
the over-expression of oncogenes, probably at an early stage in tumour development [389].
Methylation of homeobox genes and other developmental regulatory transcription factors is
particularly prevalent in lung [213, 390], breast [391, 392], colorectal cancer [393],
153
lymphoma [310], astrocytoma [308, 394] and glioblastoma mutliforme [395], however within
some of these studies, the authors have discovered certain hypermethylated genes that are not
associated with reduction of mRNA expression or protein levels, or in fact display overexpression at the mRNA level.
The expression of SOX11, a member of the SOXC family of transcription factors, is thought
to be regulated by mechanisms other than promoter methylation, as hypomethylation of
SOX11 was observed in SOX11 positive and SOX11 negative cases of mantle cell
lymphoma, demonstrating no clear correlation with mRNA expression. A study of SHOX2
methylation in lung tumour tissue demonstrated higher methylation levels in 96% of cases
compared with matched normal adjacent control tissue, however the authors did not observe a
downregulation of SHOX2 expression. In fact, the mRNA expression appeared slightly
elevated in tumour tissues although this was not statistically significant [385]. Bell et al.
observed En1 gene hypermethylation in adenoid cystic carcinoma and also demonstrated
EN1 protein over-expression in these tumours [312]. Bennett et al. identified extensive
hypermethylation of homeobox genes, including En2, in primary follicular lymphoma
specimens and cell lines, but this did not coincide with transcriptional down-regulation of
En2 and many other genes, suggesting aberrant epigenetic regulation in follicular lymphoma
[310].
In pilocytic astrocytomas, which are the most common childhood brain tumours, Lambert et
al. observed a strong positive correlation between En2 methylation and expression, along
with a number of other key neural developmental genes [394]. However this methylation
frequently involved the gene body rather than the promoter region. Gene body-specific
methylation has previously been identified on the active X chromosome in contrast to
promoter methylation with corresponding gene silencing of the inactive X chromosome
[396].
It has been hypothesised that such gene body methylation increases levels of
transcription via repression of internal gene promoters, suppressing repetitive elements and
switching promoter usage within a gene [396, 397].
Further detailed research involving En2 hypermethylation has been conducted in nonmalignant conditions, such as the autism spectrum disorders. Previous work highlighted
genetic linkage of En2 in families with autistic members, along with broad similarities
between the neuropathology seen in En2 knockout mice and autistic individuals, resulting in
it being considered as an autism susceptibility gene [293-300].
154
James et al. therefore
studied En2 promoter region methylation in post-mortem cerebellar cortex specimens from
autistic and matched unaffected individuals [398].
Within the context of global DNA
hypermethylation in autistic samples compared with controls, they also demonstrated
significant En2 promoter hypermethylation in the autistic examples compared with the
control samples (23±4% versus 10±2%; p=0.005), using the McrBC-PCR assay, the same
assay that we used in our methylation work. They confirmed these findings using a second
assay, MSR-PCR, which demonstrated similar En2 promoter region hypermethylation. The
authors were surprised to observe significantly elevated expression levels of En2 mRNA in
the autism samples, alongside the promoter region hypermethylation, demonstrating a
positive correlation between these variables in the case and control samples (p=0.007). They
also confirmed an increased level of EN2 protein in the autism specimens via western blot,
consistent with the increase in En2 gene expression.
155
5.4.
CONCLUSION
As hypothesised, En2 promoter hypermethylation was identified in a number of EOC cell
lines and human ovarian tumour specimens, compared with relative hypomethylation in
normal tissue. These findings indicate a possible role in diagnosis, however the invasive
nature of the necessary tissue biopsies may limit its clinical use. If such hypermethylation
could be detected in fluid samples such as blood, via the analysis of circulating tumour cells
or DNA, or urine, this may have improved clinical utility as a biomarker of early diagnosis.
There was no statistically significant difference in methylation status between platinumsensitive and –resistant tumours, precluding any role as a prognostic biomarker, however if
the cohort size were to be increased, this may result in a more significant difference.
In keeping with the findings of others in the context of malignant and non-malignant disease,
hypermethylation of the En2 promoter did not correlate with a decreased level of En2 gene
expression hence epigenetic silencing of the gene was not apparent. Several cases of human
EOC tumour in fact demonstrated En2 over-expression alongside En2 promoter
hypermethylation, which has also been noted in the autistic cerebellum. It is possible that
other epigenetic mechanisms exert more control over En2 transcriptional upregulation or that
En2 methylation and gene expression are completely independent processes.
Ultimately it
seems that there are very complex epigenetic mechanisms of molecular control involved in
the regulation of En2 expression.
156
CHAPTER 6
THE ROLE & FUNCTION OF EN2 IN
ONCOGENESIS
157
6.
THE ROLE AND FUNCTION OF EN2 IN ONCOGENESIS
6.1.
INTRODUCTION
The homeobox gene family includes more than 100 members, each of which encodes a
protein containing a 61 amino acid homeodomain. This specific region acts as a binding site
for other proteins to enable activation or repression of downstream target genes [399]. The
Engrailed-2 (En2) gene belongs to this homeobox family, and like many of the other
members of this gene superfamily, En2 plays a role in central nervous system development,
as well as skeletal and limb development [185]. En2 is located on chromosome 7 (7q36.3)
and encodes the 333 amino acid EN2 protein, first characterised in Drosophila, which
contains five homology regions acting as binding domains for facilitation of transcriptional
activation or repression, dependent upon the particular binding location [265-268, 278, 279].
Nedelec et al. also observed specific and high affinity binding of the Engrailed protein to the
eukaryotic translation initiation factor 4E (eIF4E), suggesting that it may also play a
regulatory role in translation [269].
The homeodomain sequences are also thought to
facilitate secretion of EN2 protein from the cell via vesicular transport, along with
internalization by the cell, which is an unusual characteristic for a transcription factor [271276].
The EN2 protein has been identified in the developing brains of human foetuses, particularly
in the dorsal cells of the midbrain/hindbrain border region, where it influences survival of
mesencephalic dopaminergic neurons [23, 280-282]. It is believed to guide axons to establish
rostro-caudal polarity and continues to exert such action even after internalization by the
axon [284, 285]. Expression has also been noted in the arcuate nucleus which is located at
the ventral surface of the medulla oblongata, and is involved in control of cardioventilatory
activity and blood pressure [289, 400]. High En2 expression is still evident in many of these
anatomical areas shortly after birth, however the levels reduce dramatically soon after [283,
289]. As a result, the only known sites of En2 gene expression in the normal adult are in the
nuclei of Purkinje neurones within the cerebellum [290], and in the cytoplasm of the tubular
epithelial cells of the kidney [344]. It is unclear why the gene persists in these regions or for
what purpose.
158
In a murine model of homozygous En2 gene mutation, disrupted formation of the
mesencephalon and metencephalon as well as cerebellar hypoplasia, were observed [287].
Later work showed that gene over-expression resulted in retardation of maturation of
Purkinje neurones [291].
In a human study of 13 cases of sudden infant death, 61%
demonstrated hypoplasia of the arcuate nucleus with predominantly negative EN2 protein
expression, suggesting a crucial role for EN2 in normal neuronal development and anatomic
organisation [289]. Beyond the peri-natal period, pathological expression of the gene has
been linked to young-onset Parkinson’s disease [292] but also Autism where there appears to
be a genetic linkage in families with autistic members, as well as anatomical changes which
mimic those in En2 knockout mice, including cerebellar hypoplasia and a reduction in the
number of Purkinje cells [293-300, 398].
En2 was originally identified as a potential oncogene in breast cancer, after gene overexpression in mammary cell lines was seen to promote malignant characteristics, such as a
reduction in cell cycling time and loss of cell-to-cell contact [301]. In a human breast cancer
cell line, suppression of En2 resulted in a significant decrease in cellular proliferation rate.
Protein over-expression was observed in both cell lines and human breast tumours.
Subsequent work on human bladder cancer cell lines and patient tumour specimens supported
this data [304]. However the En2 gene and its protein product has been studied most in the
human prostate, where it is expressed in prostate tumours, but absent in normal prostate
epithelial cells, and suppression of En2 again results in decreased cell proliferation [302].
Subsequent immunohistochemical staining of patient prostate tumour biopsies has shown that
EN2 protein expression is most intense in the ductal structures of tumours, along with
presence in the cytoplasm and basal membrane, with absence of nuclear staining [303]. EN2
containing blebs were identified in prostatic acini and ducts suggesting secretion into ductal
lumen, which was entirely consistent with the previously demonstrated secretory properties
of EN2. Equivalent secretion was also observed in three prostate cancer cell lines grown
(LNCaP, DU145 and PC3) where EN2 protein was detectable in the supernatant of the
culture medium used to grow these cell lines, however further analysis of this secretion was
not carried out. Secretion and deposition of EN2 into urine by men with prostate cancer was
subsequently confirmed by western blot and ELISA analysis in over 65% of patients [314].
Higher EN2 levels also correlated with advancing tumour stage [355].
Elevated urinary EN2 protein was observed in 61% of epithelial ovarian cancer patients
compared with healthy controls, however the sample size was insufficient to evaluate
159
increasing grade or stage of disease, or differing histological sub-types (Chapter 4). As the
ovaries are not in close anatomical proximity with the kidneys or urinary tract, the presence
of EN2 in the urine cannot be explained by release from cancer cells directly shed from the
primary tumour. There may be an active process of EN2 secretion from the cancer cell into
the tumour microenvironment, via vesicular transport, including into the vasculature and
lymphatics. As blood is then filtered through the kidneys, the protein may travel through the
glomerulus into the kidney tubules, enabling detection in the voided urine. Analysis of EN2
protein secretion from EOC cell lines may help to provide further information regarding the
methods of secretion and transportation.
Development of resistance to platinum chemotherapeutic agents affects most patients with
advanced EOC, and is the main reason for the poor 5-year survival rates. Initial response to
treatment in chemo-sensitive cancer is around 80%, but the prognosis is poor in those
developing resistance to chemotherapy at an early stage [64].
Most patients acquire
resistance during cycles of therapy and so their disease progression becomes evident within 6
months of chemotherapy completion i.e. “platinum resistant” [65]. The platinum drugs are
believed to promote cell death by covalently binding to purine bases on the DNA, interfering
with DNA repair mechanisms, resulting in DNA damage and eventual cellular apoptosis [6567].
However the cell may develop resistance to these drugs via the use of many
mechanisms, including increased DNA repair or increased tolerance to DNA damage. It is
not yet known whether the chemotherapy resistant cells derive from cells within the
heterogeneous tumour that are inherently chemotherapy resistant, or whether resistance
emerges as the cancer cells interact with the tumour microenvironment. Previous work
described in Chapter 3 and Chapter 4, demonstrated that En2 mRNA expression was
significantly higher in platinum-resistant EOC cell lines compared with paired platinumsensitive lines, and higher in neoadjuvant treatment tumour specimens from patients that
proved to have platinum-resistant disease on subsequent follow-up. This was coupled with
decreased progression-free survival and overall survival in high En2 expressers. It was
hypothesised that patients who later develop platinum-resistant disease, may harbour
inherently resistant cells within their primary tumour that highly express En2. However it is
not known whether the elevated En2 expression is a by-product of the mechanisms promoting
platinum-resistance, or whether it is a direct contributor to this phenotype. Within the context
of En2 expression in cancer, there has been no previous mention of any associations with
drug sensitivity or resistance, let alone platinum-resistance.
160
Gene over-expression and
silencing studies in paired platinum-sensitive and platinum–resistant EOC cell lines, with
subsequent platinum challenge, may help to further elucidate the role of En2 in cancer.
In normal cellular function, En2 is believed to be involved in the Wnt/β-catenin signalling
pathway (Figure 6.1). In the absence of Wnt pathway activation, β-catenin associates with a
protein complex consisting of axin, APC, CKI and GSK3β, which phosphorylates β-catenin
and targets it for ubiquitination and degradation. It may also associate with the cytoplasmic
tail of E-cadherin preventing any down-stream activation. However if the Wnt pathway is
activated, perhaps by the binding of Wnt to the Frizzled receptor with recruitment of coreceptors, axin dissociates from the degradation complex and preferentially binds to the
receptor complex, preventing degradation of β-catenin [401, 402] Accumulation of β-catenin
with subsequent translocation to the nucleus, promotes transcription of Wnt target genes
which includes the En2 gene [403, 404]. However the subsequent down-stream effects of
En2 and its protein product have not been studied extensively. Up-regulation of En2 has now
been demonstrated in ovarian cancer as well as breast, prostate and bladder cancer, however
it remains unclear whether En2 functions to directly promote tumour formation, or if its upregulation is a down-stream effect of another genetic pathway. As previously discussed, En2
expression may directly influence platinum-sensitivity in EOC, or its over-expression in
platinum-resistant tumours may occur as a result of other influential associated pathways.
Studying the up-regulation and down-regulation of genes in response to En2 over-expression
and silencing via microarray analysis, may provide a more detailed view of the role of En2 in
tumour development and its gene associations.
161
Figure 6.1. The Wnt signalling pathway in the absence and presence of ligand (taken
from [405]). In the absence of Wnt ligand the protein complex (CKIα, GSK3β, APC, Axin) phosphorylates βcatenin, which is then a target for ubiquitination and degradation. Binding of Wnt to the Frizzled/LRP-5/6
receptor complex promotes associated axin binding, preventing formation of the degradation complex. βcatenin can travel to the nucleus and interact with transcription factors resulting in activation of target genes,
including En2.
162
Study Objectives and Hypothesis
The objectives of this chapter were to establish the role and function of EN2 in EOC cancer
cells, with a particular focus on the potential involvement in platinum-resistance.
Cell supernatants would be collected from EOC cell lines with the aim of quantifying the
secreted protein concentration and comparing molecular weight of intracellular and secreted
EN2.
Forced over-expression of En2 in PEA1 cells and silencing of En2 in PEA2 cells would be
performed, in order to study platinum drug sensitivity and the possible association with EN2
expression.
Using microarray hybridisation, gene expression profiling would be performed analysing
differences between native PEA1 and PEA2 cell lines with En2 over-expressed PEA1 clones
and En2 silenced PEA2 clones. This may help to clarify the role of EN2 and its associated
proteins in oncogenesis.
We hypothesised that EN2 would be secreted from EOC cell lines, in particular those with a
high level of En2 mRNA and cellular protein expression. This protein may be secreted from
the cell via vesicular transport. If En2 expression plays a direct role in resistance to platinum
treatment, over-expression of En2 may promote platinum resistance in innately platinum
sensitive cells. Conversely, silencing of the En2 gene may re-sensitise cells to cisplatin
treatment. Microarray analysis of these cells could identify additional genes known to be
involved in the development of resistance to platinum-based treatment, confirming the
important role that EN2 plays in cancer.
163
6.2.
RESULTS
6.2.1. Secretion of EN2 protein from EOC cell lines
In order to quantify the relative expression of secreted EN2 protein from the ovarian cancer
cell lines, a Direct ELISA and Western Blot were performed on supernatants acquired from
the fourteen EOC cell lines. The recombinant EN2 (rEN2) protein developed in E. coli was
used as a positive control whilst supernatant from the normal skin fibroblast cell line was
used as a negative control. Supernatants from the mouse fibroblast cell line NIH3T3, the
prostate stromal cell line WPMY-1, and the melanoma cell line A375M were also included
for comparison. The secreted EN2 protein concentration from each cell line is illustrated in
Figure 6.2(A), with the histological sub-type and platinum status detailed.
Only the OV90 and TOV112D cell lines demonstrated EN2 secretion greater than 100ng/ml,
whilst the fibroblast cell lines and the majority of EOC cell lines exhibited less than 50ng/ml
of protein.
These results were supported by analysis of secreted EN2 protein within the same cell line
supernatants and rEN2 control, using gel electrophoresis and protein immunoblotting (Figure
6.2(B)). The rEN2 positive control demonstrated a strong band at approximately 45kDa,
whilst no band was seen in the supernatants from the fibroblast cell lines. Evaluation of the
EOC cell line supernatants revealed demonstrable protein bands at 50kDa in the OV90 and
TOV112D cell lines only.
It was a little surprising to see that the OV90 and TOV112D cells demonstrated the highest
concentration of secreted EN2 protein as these two cell lines demonstrated low En2 mRNA
expression and very little EN2 protein within the cell lysate (see Figures 3.1 and 3.7,
Chapter 3).
Further experiments were planned to evaluate EN2 secretion from cells under varying
conditions that may inhibit protein secretion, as well as studying whether the secreted protein
was conjugated or actively transported within vesicles. Given that secreted EN2 protein was
only confidently detected in 2 cell lines and that these did not highly express EN2 at the
genetic or protein level, further analysis of the supernatants was not carried out.
164
Figure 6.2. Secreted EN2 protein expression in EOC cell lines and controls, as
determined by the Direct ELISA and Western Blotting. Supernatants from fourteen ovarian
cancer cell lines were analysed by Direct ELISA (A) and Western Blotting (B) techniques. Fibroblast, NIH3T3
and WPMY cell line lysates were used as negative controls and rEN2 as a positive control. The melanoma cell
line, A375M, was included for interest. The histological sub-type and platinum-sensitivity status for each EOC
cell line is depicted. The EN2 protein concentration (ng/ml) for each cell line was calculated relative to the
relevant serum-free media standard curve. Uniform total protein loading was performed for the Western Blot,
according to the BCA Protein Assay, therefore the blot area and intensity reflect the relative EN2 protein
concentration between the cell lines.
165
6.2.2. En2 over-expression in the PEA1 EOC cell line
A transient transfection of the PEA1 cell line was initially performed using the Origene En2
plasmid, in order to over-express the En2 gene. Thirty hours after initial incubation with the
transfection reagents, the cells were imaged using light microscopy, prior to mRNA
extraction, or lysis for protein analysis. There were no discernible morphological differences
between the transfected and control cells as shown in Figure 6.3.
Figure 6.3. Images of the PEA1 EOC cell lines after transient forced over-expression of
En2. Light microscope images of transfected PEA1 cells, cells incubated in media, and Lipofectamine
transfection agent alone, at 20x magnification.
Figure 6.4 demonstrates the greater than 3500 fold increase in En2 mRNA expression
relative to β-actin achieved in the transfected cells compared with the low level of expression
in the transfection agent and media alone controls, which are consistent with the previous
PEA1 cell line En2 expression levels (Figure 3.1, Section 3.2.1.). Cell lysates were also
analysed for EN2 protein expression via Western Blotting and demonstrated clear bands of
equal intensity at 50kDa, however the En2 over-expressed cell lysate showed an additional
protein band at a lower molecular weight, around 43kDa, which was equivalent to that of the
recombinant EN2 protein band (Figure 6.5). These blots are shown at 90 seconds and 10
minutes exposure.
166
E n 2 m R N A e x p r e s s io n o f P E A 1 c e ll lin e
a f t e r f o r c e d o v e r - e x p r e s s io n
r e la t iv e t o B - a c t in
M e a n C T v a lu e
100000
10000
1000
100
10
m
s
la
P
2
n
E
L
ip
o
fe
c
ta
M
e
m
d
in
ia
e
O
o
n
n
ly
ly
id
1
Figure 6.4. En2 mRNA expression in the PEA1 EOC cell line after transient forced
over-expression of En2. En2 mRNA expression in transfected PEA1 cells (red bar), cells exposed to
media (blue bar) and Lipofectamine transfection agent alone (grey bar), was analysed by quantitative rt-PCR.
The En2 mRNA expression is shown on a log10 scale relative to the housekeeping gene β-actin (x100,000).
Figure 6.5. EN2 protein expression in the PEA1 EOC cell line after transient forced
over-expression of En2, as determined by Western Blotting. Whole cell lysates from
transfected PEA1 cells, cells exposed to Lipofectamine transfection agent and to media alone, were analysed by
Western Blotting. Recombinant EN2 (rEN2) was used as a positive control. Uniform total protein loading was
performed according to the BCA Protein Assay, therefore the blot area and intensity reflect the relative EN2
protein concentration between the cell lines. Blots exposed at 90 seconds and 10 minutes are shown.
167
A stable transfection was subsequently performed using the untagged Origene En2 plasmid,
containing the neomycin resistance gene, allowing antibiotic selection. Over 40 clones were
selected and the resultant En2 mRNA expression levels were compared with the PEA2 cell
line, the paired platinum-resistant cell line, which demonstrated a mean En2 mRNA
expression relative to β-actin of 177 (Figure 3.1, Section 3.2.1.). En2 expression equivalent
to, or greater than this, was achieved in 11 of the clones, a selection of which are shown in
Figure 6.6.
Cell lysates were also acquired from some of the clones and analysed for EN2
protein expression via Western Blotting. Clear bands of equal intensity at 50kDa were again
demonstrated, but only Clone H3 showed the additional protein band at the lower molecular
weight of around 43kDa (Figure 6.7). Clone H3 had a 7000 fold increase in En2 mRNA
expression compared with the wild-type PEA1 cell line (Figure 3.1, Section 3.2.1.). It was
hypothesised that the 50kDa band may represent EN2 protein that has undergone posttranslational modification, perhaps by glycosylation, whereas the lower weight band,
equivalent to the rEN2, may represent the native protein prior to modification. Further work
was carried out to investigate this theory and is described later in this Chapter (Section
6.2.3.).
E n 2 e x p r e s s io n in P E A 1 s t a b le o v e r - e x p r e s s io n c lo n e s ( A u g - 1 4 )
r e la t iv e t o B - a c t in
M e a n C T v a lu e
100000
10000
1000
100
10
H
e
e
lo
C
lo
C
n
n
e
n
lo
C
4
2
F
3
H
2
G
e
n
lo
C
n
lo
C
C
lo
n
e
e
E
E
1
5
2
1
Figure 6.6. En2 mRNA expression in PEA1 EOC cell line clones after stable forced
over-expression of En2. En2 mRNA expression in a selection of clones grown from transfected PEA1
cells was analysed by quantitative rt-PCR. The En2 mRNA expression is shown on a log10 scale relative to the
housekeeping gene β-actin (x100,000).
168
Figure 6.7. EN2 protein expression in PEA1 EOC cell line clones after stable forced
over-expression of En2, as determined by Western Blotting. Whole cell lysates from a selection
of clones grown from transfected PEA1 cells were analysed by Western Blotting. Recombinant EN2 (rEN2)
was used as a positive control. Uniform total protein loading was performed according to the BCA Protein
Assay, therefore the blot area and intensity reflect the relative EN2 protein concentration between the cell lines.
169
6.2.3. Post-translational modification of EN2
After observing the doublet EN2 protein bands in the transient and stable En2 over-expressed
cell lysates, further work was undertaken to investigate whether the lower weight band
represented an immature form of the EN2 protein prior to post-translational modification,
with the 50kDa band representing mature, functional EN2.
Enzymatic deglycosylation was performed using the PEA1 cell line lysates from the initial
transient transfection of the En2 plasmid, and the results were compared with the original
lysates using Western Blotting.
Initially the same protocol was performed using bovine
fetuin which is a glycoprotein containing sialylated N-linked and O-linked glycans that can
be used as a positive control for endoglycosidase enzymes. The molecular weight of this
protein differed in the literature from 48kDa to 64kDa, but was consistently reported to run as
a doublet on an SDS-PAGE gel, due to heterogeneous glycosylation. There was no clearly
visible reduction in the size of the 50kDa and 43kDa protein bands in the transfected or
control PEA1 cells following deglycosylation (Figure 6.8).
However a new band was
visualised at around 33kDa in all of the deglycosylated lysates. Rat whole brain lysate was
also included on this Western Blot analysis, as this had also been suggested as a positive
control for EN2 protein when using the Abcam goat anti-EN2 antibody. Interestingly, the rat
brain lysate demonstrated a band at 43kDa as well as a strong band at around 33kDa,
consistent with the newly observed band in the deglycosylated lysates.
170
A
B
Figure 6.8. Enzymatic deglycosylation of EN2 protein in PEA1 EOC cell lines after
transient forced over-expression of En2, as determined by Western Blotting. Enzymatic
deglycosylation was performed using bovine fetuin protein as a positive control for the endoglycosidase
enzymes and visualised using Coomassie wash solution, demonstrating the doublet band representing known
heterogenous glycosylation (A). The deglycosylation protocol was then performed on the PEA1 cell line lysates
from the transient En2 plasmid transfection experiment and analysed by Western Blotting, using the original
lysates for comparison (B). The deglycosylated lysates are labelled (dg). Rat whole brain lysate was used as a
positive control. The concentration of loaded protein was uniform therefore the blot area and intensity reflect the
relative EN2 protein concentration between the cell lines.
171
6.2.4. siRNA-mediated En2 silencing in the PEA2 EOC cell line
siRNA-mediated En2 silencing in the PEA2 EOC cell line was performed, using three
different siRNAs as well as a negative control siRNA. This technique was initially refined
using the PC3 prostate adenocarcinoma cell line, resulting in a 40-60% reduction in En2
mRNA expression, which was consistent with the previous results of Bose et al [302].
Forty-eight hours after initial incubation with the siRNAs, the cells were imaged using light
microscopy, prior to mRNA extraction, or lysis for protein analysis.
There were no
discernible morphological differences between the silenced and control cells as shown in
Figure 6.9.
Figure 6.9. Images of the PEA2 EOC cell lines after siRNA-mediated En2 silencing.
Light microscope images of En2 silenced PEA2 cells using three different siRNAs, compared with cells
incubated with negative siRNA, NeoFx transfection agent, and media alone, at 20x magnification.
Figure 6.10 demonstrates the reduction in relative En2 mRNA expression obtained using the
three different siRNAs, with siRNA 4676 producing the most discernible reduction (52.7%)
compared to the media alone control cells. Cell lysates were also analysed for EN2 protein
expression via Western Blotting and demonstrated a reduction in the intensity of the 50kDa
band in the three different siRNA-treated lysates compared with lysates from the media and
NeoFx transfection agent alone cells (Figure 6.11). Admittedly, the β-actin loading control
Western Blot demonstrated a larger band for the media alone lysate suggesting a greater
concentration of loaded protein, despite using the BCA protein assay for equilibration.
However there was still a notable decrease in intensity of the EN2 blot in the siRNA-treated
lysates compared with the NeoFx transfection agent alone, and the loading control was more
uniform in these samples.
172
M e a n C T v a lu e r e la tiv e to B -a c t in
400
300
200
100
ly
n
76
O
46
75
46
74
46
ia
ed
F
eo
M
N
N
eg
at
iv
e
si
x
R
N
on
A
ly
0
Figure 6.10. En2 mRNA expression in the PEA2 EOC cell line after siRNA-mediated
En2 silencing. En2 mRNA expression in transfected PEA2 cells (red bars), cells transfected with negative
siRNA (green bar), cells exposed to media (blue bar) and NeoFx transfection agent alone (grey bar), was
analysed by quantitative rt-PCR. The En2 mRNA expression is shown relative to the housekeeping gene β-actin
(x100,000).
Figure 6.11. EN2 protein expression in the PEA2 EOC cell line after siRNA-mediated
En2 silencing, as determined by Western Blotting. Whole cell lysates from transfected PEA2
cells, cells transfected with negative siRNA, cells exposed to NeoFx transfection agent and to media alone, were
analysed by Western Blotting. Recombinant EN2 (rEN2) was used as a positive control. Uniform total protein
loading was performed according to the BCA Protein Assay, and the blot was reprobed with B-actin to act as an
additional loading control.
173
Despite observing En2 mRNA silencing with a reduction in protein expression, we had hoped
to obtain an En2 expression level equivalent to that in the paired PEA1 cell line, i.e. a CT
value relative to β-actin of between 10 and 20. The siRNA-mediated En2 silencing protocol
was therefore repeated on several occasions adjusting concentrations of cells per well,
transfection agent and En2 siRNA, but at best only a 60% reduction in En2 mRNA
expression was achievable, using siRNA 4676. Therefore the CompoZr knockout zinc-finger
nuclease kit from Sigma-Aldrich was ordered, which would be custom made and validated by
the company for En2 gene knockout. Despite developing 16 pairs of zinc-finger nucleases,
none of these met the company’s specification during their extensive validation process, as
they could not achieve satisfactory excision of the En2 gene. This entire process took
approximately 20 weeks which did not leave enough time to contact different companies, or
to pursue alternative technologies, such as Cas9-nuclease RNA-guided genome editing [406408].
174
6.2.5. The effect of En2 over-expression on platinum-resistance in EOC cell lines
Having achieved successful stable En2 mRNA over-expression in a number of cell clones
from the PEA1 EOC cell line, the plan was to evaluate the sensitivity of these cells to
cisplatin treatment, in comparison to the wild-type PEA1 and PEA2 cell lines.
This
experiment should have also included the use of siRNA-mediated En2 silenced PEA2 cell
line clones, however as discussed in Section 6.2.4., En2 could not be silenced to a
satisfactory level to provide meaningful comparisons.
The cytotoxic effect of cisplatin on PEA1 cell line clones demonstrating varying levels of
En2 mRNA expression was determined by the MTS assay (Section 2.19.), along with two
different passages of wild-type PEA2 cells. The percentage cell survival was recorded
relative to increasing concentrations of cisplatin, with IC50 values used to compare the levels
of cytotoxicity between the different cell line groups (Figure 6.12). The majority of the
PEA1 clones produced IC50 values of 13-14 μM, with the level of En2 over-expression in
these clones ranging from a 0-850 fold increase compared with wild-type PEA1 cells. The
IC50 value of both PEA2 cell lines was much higher at 31 μM, which was consistent with the
known cisplatin-resistant property of the PEA2 cell line. Most notably the cell survival
curves for PEA1 clones H3 and F7, which comprised >7000 fold increased En2 mRNA
expression compared with wild-type PEA1, did appear to separate from the other PEA1
clones with a higher IC50 of 19 μM. However this was still considerably lower than the IC50
of the PEA2 cells and suggested that elevated En2 mRNA expression may contribute to, but
does not solely cause, resistance to cisplatin treatment.
Microarray analysis of gene up-regulation and down-regulation occurring as a result of forced
En2 over-expression, may provide further information as to which pathways involve En2 and
its protein product in oncogenesis, and whether there are links with other genes known to be
involved in cisplatin resistance.
175
Figure 6.12. A high level of En2 forced over-expression does promote resistance to
cisplatin in PEA1 EOC cancer cells. IC50 values of PEA2 cells from 2 different passages were notably
elevated in comparison to PEA1 cell line clones after treatment with cisplatin, as determined by the MTT assay.
This is in keeping with the known platinum-resistant status of the PEA2 cell line. Most PEA1 clones showed
IC50=13-14 μM irrespective of level of En2 over-expression (0-850 fold increase in En2 mRNA compared with
wild-type PEA1). However clones H3 and F7 showed IC50=19 μM (>7000 fold increase in En2 mRNA
compared with wild-type PEA1).
176
6.2.6. Microarray analysis of the effect of En2 over-expression on oncogenic pathways
Gene expression profiling and analysis was performed using microarray hybridisation to
compare gene expression between wild-type PEA1 cells, two PEA1 cell line clones with
stable En2 over-expression (E12 = >200 fold and H3 = >7000 fold En2 expression compared
with wild-type PEA1), and wild-type PEA2 cells (>13 fold En2 expression compared with
wild-type PEA1). The sample preparation, hybridisation and array scanning was performed
in-house, however the extracted data was subsequently analysed by Dr Carla Moller-Levet, a
Bioinformatics Experimental Officer at the University of Surrey.
Out of a total of 27,192 probes that passed the quality control filtering criteria, 131 were upregulated in the >200 fold PEA1 E12 clone, and 186 were up-regulated in the >7000 fold
PEA1 H3 clone, when using a cut-off of >3 fold up-regulation compared with wild-type
PEA1 (Figure 6.13). Sixty-seven probes were down-regulated in the >200 fold PEA1 E12
clone, and 109 were down-regulated in the >7000 fold PEA1 clone, when using the same cutoff value.
Figure 6.13. Heatmap visualization of significantly regulated probe sets. The heat map
shows expression values mapped to a colour gradient from low (blue) to high (red) expression; the data is not
scaled. Experiments are arranged according to a hierarchical clustering dendrogram. A filter criteria of at least
3-fold change was applied resulting in 131 up- and 67 down-regulated probes in PEA1 E12 compared with wildtype PEA1. There were 186 up- and 109 down-regulated probes in PEA1 H3 compared with wild-type PEA1.
The columns represent the samples, and the rows represent the gene probes.
177
Table 6.1 details the microarray probes with corresponding gene symbol, which have a >3
fold up-regulation. Where available, the gene name is listed for those probes up-regulated in
both PEA1 clones. Due to the inclusion of single samples of each cell line or clone only, i.e.
no replicates, we were advised to focus on the probes up-regulated in both PEA1 cell line
clones compared with wild-type PEA1, in order to improve the significance of subsequent
gene pathway analysis. There were 97 probes demonstrating a >3 fold up-regulation which
were common to both PEA1 clones, with a p-value of intersection of <1x10-16, which is the
smallest p-value that the R Bioconductor package can generate. These probes were uploaded
to the GeneGo from MetaCoreTM software to perform an enrichment analysis, which enables
the identification of relevant gene sets within the data. These genes can then be grouped
within significant signalling or metabolic pathway maps, or process networks. The 10 most
relevant pathway maps are listed in Table 6.2, including the most pertinent up-regulated
genes. Most notable pathways included TGF-β receptor signalling, Activin A signalling
regulation and Activin A involvement in cytoskeleton remodelling, with the up-regulated
Activin A, Plasminogen Activator Inhibitor-1 (PAI1) and Ski genes repeatedly involved in
such pathways. The 10 most relevant process networks are listed in Table 6.3, including the
associated up-regulated genes. Most notable process networks included G1-S growth factor
regulation, TGF-β, GDF and Activin signalling, connective tissue degradation, and apoptosis,
with the up-regulated Activin A, PAI1, Ski and GADD45β genes repeatedly involved in such
pathways.
178
Molecular
Probe
A_32_P189093
A_24_P158089
A_23_P37685
A_23_P122924
A_24_P211044
A_23_P133408
A_23_P30126
A_23_P394395
A_24_P535256
A_23_P92929
A_23_P87013
A_32_P179317
A_23_P250963
A_23_P62890
A_32_P42666
A_24_P328471
A_24_P695691
A_32_P71744
A_23_P143016
A_23_P398515
A_23_P354314
A_23_P338479
A_24_P115443
A_23_P329016
A_23_P83624
A_23_P48455
A_23_P31218
A_24_P349539
A_23_P111360
A_23_P171074
A_24_P247316
A_23_P59718
A_23_P21333
A_24_P508012
A_32_P877
A_24_P931443
A_32_P67266
A_23_P82162
A_23_P119311
A_23_P255331
A_24_P138912
A_24_P470809
A_23_P5785
A_24_P93798
A_24_P365975
A_24_P147849
A_32_P111235
A_23_P208325
A_24_P100277
A_23_P76992
A_24_P288754
A_23_P200396
Gene Symbol
AI090167
NM_000602
NM_024600
NM_002192
NM_022640
NM_000758
NM_005130
NM_020433
AK001903
NM_016644
NM_001001522
BF761348
NM_004172
NM_002053
A_32_P42666
NM_019091
A_24_P695691
BG695979
NM_006673
NM_004203
NM_153832
NM_014143
NM_018688
NM_138778
NM_021149
NM_030943
NM_032625
A_24_P349539
NM_004506
NM_004867
BC029796
NM_003130
NM_005664
ENST00000344293
BM999343
NM_003485
BC045163
NM_003080
NM_012315
NM_032623
NM_002917
A_24_P470809
NM_152384
NM_031480
NM_005202
A_24_P147849
BG612665
NM_004234
NM_007326
NM_002632
NM_002641
NM_019118
PEA1
>200x
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
PEA1
>7000x
1
1
1
1
1
0
0
1
1
1
1
0
1
1
1
1
0
1
0
0
1
0
0
0
1
1
1
0
1
1
1
1
1
1
1
1
1
1
0
1
1
1
1
1
1
1
1
1
1
1
1
1
Gene Name
AI090167
PAI1
TMEM204
INHA/ACTA
CSH1
JPH2
AK001903
PRR16
TAGLN
SLC1A3
GBP1
…
PLEKHA3
BG695979
GPR161
COTL1
AMN
C7orf13
HSF2
ITM2A
C5orf55
SRI
MKRN3
TAF3
BM999343
GPR68
SP_1102
SMPD2
MGARP
RFNG
…
BBS5
RIOK1
COL8A2
…
BG612665
ZNF235
CYB5R3
PGF
PIGA
TMEM234
Molecular
Probe
A_23_P204119
A_23_P167358
A_23_P129075
A_23_P119006
A_23_P87011
A_23_P251499
A_32_P202621
A_24_P161233
A_32_P130962
A_23_P371145
A_24_P39928
A_24_P79040
A_32_P174398
A_23_P408195
A_32_P231463
A_32_P704982
A_24_P398370
A_23_P31188
A_24_P608931
A_23_P170030
A_23_P207170
A_23_P204947
A_24_P150791
A_23_P204941
A_23_P62081
A_23_P207520
A_23_P13382
A_24_P397294
A_23_P329573
A_23_P123234
A_24_P270728
A_23_P127948
A_23_P142560
A_23_P433798
A_24_P131622
A_24_P150580
A_23_P170719
A_23_P390006
A_23_P75362
A_24_P228550
A_23_P41804
A_23_P62583
A_23_P1029
A_24_P326660
A_23_P83579
A_24_P926960
A_32_P68142
A_23_P103256
A_23_P312179
A_23_P150609
A_23_P208567
A_24_P319374
Gene Symbol
ENST00000261741
NM_001039717
NM_024908
THC2438889
NM_001001522
NM_002593
A_32_P202621
THC2334717
AF131777
NM_138430
ENST00000318438
NM_144691
AF085351
NM_152399
THC2350817
BE379389
A_24_P398370
AK091818
AW043836
NM_016564
NM_022640
NM_004004
NM_020655
NM_004004
NM_003020
Z74615
NM_001013254
NM_145867
NM_000211
A_23_P123234
NM_012385
NM_001124
NM_014795
NM_024825
NM_007177
NM_016563
A_23_P170719
NM_138325
ENST00000382123
NM_030773
NM_033120
NM_001409
NM_017459
NM_006500
NM_014862
NM_001409
AI470277
NM_021023
NM_015120
NM_001007139
THC2401542
NM_005814
PEA1
>200x
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
PEA1
>7000x
1
1
1
1
1
1
0
0
0
0
0
0
0
1
1
0
1
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Gene Name
RNA bp 19
C4orf29
WDR76
…
TAGLN
PCOLCE
TMEM155
THC2350817
…
CSH1
ZEB2
Table 6.1. Microarray probes up-regulated in two PEA1 En2 over-expression clones
compared with wild-type PEA1 cells. The listed microarray probes demonstrated a >3 fold upregulation in either of the PEA1 clones (>200 or >7000 fold En2 expression) compared with wild-type PEA1.
Up-regulation is represented by “1”. The gene symbols highlighted in grey (n=97) represent those that showed
up-regulation in both PEA1 clones and the relevant gene name is listed where available. Some of these genes
were identified within the enrichment analysis pathway maps (red), whilst others proved of interest during
subsequent literature searches (yellow).
179
A_23_P254079
A_23_P142506
A_24_P210513
A_23_P205449
A_24_P128085
A_24_P290799
A_24_P105747
A_32_P963
A_23_P368681
A_24_P418744
A_23_P302134
A_23_P153583
A_32_P226620
A_24_P281084
A_23_P54116
A_23_P130653
A_24_P365753
A_23_P21207
A_32_P128572
A_32_P10643
A_23_P350491
A_24_P687131
A_23_P320887
A_23_P21324
A_24_P357536
A_32_P7015
A_24_P628384
A_24_P410399
A_32_P120638
A_23_P131183
A_24_P338603
A_23_P311468
A_24_P177585
A_24_P813520
A_32_P170736
A_23_P69121
A_24_P270460
A_32_P177955
A_23_P377965
A_32_P69076
A_24_P212531
A_23_P422305
A_23_P149946
A_24_P926400
A_23_P100011
A_24_P84608
A_32_P44139
A_24_P137545
A_24_P846810
A_32_P42684
A_23_P33187
A_23_P83328
A_32_P199292
A_23_P147423
A_23_P168683
A_24_P328504
A_23_P336554
A_24_P942335
NM_003943
NM_015675
NM_018167
NM_017955
AL833177
NM_024419
NM_001039569
ENST00000282169
NM_015660
A_24_P418744
AK094772
NM_006247
THC2441546
BC016751
NM_014992
NM_031429
BX537594
NM_003335
NM_020948
A_32_P10643
NM_198794
CR619653
NM_020239
NM_057179
NM_012167
NM_012339
NM_001001655
NM_032165
NM_033547
NM_001485
NM_003036
NM_182498
NM_182625
CR626222
AK098422
NM_005067
NM_005532
BC030123
BC010538
BX113895
NM_152490
AK095295
NM_033100
NM_001001561
NM_005829
XM_929546
AA627222
NM_024949
A_24_P846810
NM_014331
NM_032558
NM_000118
ENST00000337682
NM_182920
NM_019082
NM_007237
NM_134470
BC002881
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
1
0
1
1
0
1
1
1
1
1
1
1
1
1
1
1
1
0
1
1
1
1
0
1
0
1
1
1
0
1
1
1
0
1
1
1
1
1
1
1
1
1
0
1
1
1
0
1
0
0
1
1
1
STBD1
GADD45β
BTBD7
CDCA4
PGS1
UHMK1
GIMAP2
FLJ37453
PPP5C/PP5
THC2441546
BC016751
DAAM1
RTBDN
NKAPP1
UBA7
MIER1
C7
MAP4K5
CR619653
TWIST2
FBXO11
TSPAN15
ALKBH2
INTS4
SKI
ZNF428
GEN1
CCDC71L
SIAH2
IFI27
C18orf18
Histone H3
B3GALNT2
AK095295
CDHR1
GGA1
AP3S2
LOC646609
Galectin1
…
SLC7A11
HIATL1
FAM60A
SP140
IL1RAP
TICRR
A_23_P58082
A_24_P77008
A_23_P162171
A_23_P67169
A_24_P144784
A_23_P84860
A_23_P96965
A_23_P382128
A_23_P345692
A_24_P563966
A_23_P64898
A_24_P229669
A_23_P307563
A_23_P17316
A_24_P743869
A_24_P273253
A_32_P7176
A_24_P270496
A_32_P808
A_23_P107684
A_32_P40667
A_23_P90696
A_32_P44831
A_24_P938293
A_24_P299685
A_24_P463989
A_24_P396753
A_24_P693946
A_23_P141248
A_32_P53183
A_23_P52336
A_23_P92903
A_32_P17525
A_32_P158723
A_32_P117026
A_23_P142878
A_23_P10605
A_23_P381577
A_32_P217773
A_24_P865
A_23_P147801
A_23_P432947
A_23_P386268
A_23_P398854
A_23_P40415
A_23_P51231
A_23_P169039
A_23_P29684
A_23_P206022
A_24_P912606
A_32_P84454
A_23_P61945
A_24_P332953
A_24_P754999
A_24_P485105
A_23_P37455
A_23_P346673
A_23_P205069
180
NM_199511
NM_000963
NM_006500
NM_000641
NM_194072
NM_007177
NM_030786
NM_173687
NM_138284
AL117454
NM_005810
ENST00000283760
NM_001024660
NM_152864
BC073157
BC090889
AW949170
NM_006491
BC031691
NM_014347
AV742170
NM_021643
A_32_P44831
NM_005524
NM_198389
THC2376306
NM_021643
THC2307989
NM_021947
AF165514
NM_170744
NM_031908
THC2438621
AK123861
AK094796
NM_032827
A_23_P10605
NM_145011
NM_032872
NM_206921
NM_018293
NM_013372
NM_178497
NM_173660
NM_007038
NM_004350
NM_003068
NM_015873
NM_001004439
S80931
AK057476
NM_198159
A_24_P332953
BU622073
THC2288214
NM_001037223
AK000249
X51602
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
SNAI2
Pathway map
p-value
Linked up-regulated genes from
expression analysis
1
Development: TGFβ receptor signalling
2.167 x10-04
Ski, PAI1, GADD45β
2
Development: Transcription factors in segregation of
hepatocytic lineage
2.283 x10-03
Activin A, Activin
3
Signal transduction: Activin A signalling regulation
2.759 x10-03
Ski, Activin A
4
Immune response: IL-1 signalling pathway
4.865 x10-03
IL1RAP, PAI1
5
Stimulation of TGFβ signalling in lung cancer
5.767 x10-03
Ski, PAI1
6
Development: Prolactin receptor signalling
8.330 x10-03
CSH1 (somatomammotropin A), Lactogen
7
Metabolic syndrome X (general schema)
3.532 x10-02
PAI1
8
Cytoskeleton remodelling: Role of Activin A in
cytoskeleton remodelling
4.683 x10-02
Activin A
9
Expression targets of Tissue factor signalling in
cancer
5.139 x10-02
PAI1
10
Development: Glucocorticoid receptor signalling
5.594 x10-02
PAI1
Table 6.2. Pathway maps obtained from enrichment analysis of the up-regulated genes
in two PEA1 En2 over-expression clones compared with wild-type PEA1 cells. These are
the 10 most relevant pathways obtained from the enrichment analysis of the 97 probes with >3 fold upregulation in both PEA1 clones compared with wild-type PEA1.
Process network
p-value
Linked up-regulated genes from
expression analysis
1
Cell cycle: G1-S growth factor regulation
6.589 x10-03
Ski, Activin A, ActivinβA, Activin
2
Signal transduction: TGFβ, GDF & Activin
signalling
2.168 x10-02
Ski, ActivinβA, PAI1
3
Reproduction: Feeding & neuro-hormone signalling
4.835 x10-02
CSH1, Lactogen, PAI1
4
Signal transduction: BMP & GDF signalling
4.894 x10-02
Ski, GADD45β
5
Cell cycle: Meiosis
6.418 x10-02
SIAH2, PP5
6
Inflammation: Interferon signalling
6.848 x10-02
GBP1, IFI27
7
Proteolysis: Connective tissue degradation
7.851 x10-02
PC7, PAI1
8
Apoptosis: Apoptosis stimulation by external signals
1.085 x10-01
Activin A, Neutral sphingomyelinase
9
Development: Ossification & bone remodelling
1.252 x10-01
Ski, Activin
10
Apoptosis: Apoptotic nucleus
1.278 x10-01
SIAH2, GADD45β
Table 6.3. Process networks obtained from enrichment analysis of the up-regulated
genes in two PEA1 En2 over-expression clones compared with wild-type PEA1 cells.
These are the 10 most relevant process networks obtained from the enrichment analysis of the 97 probes with >3
fold up-regulation in both PEA1 clones compared with wild-type PEA1.
181
Table 6.4 details the microarray probes with corresponding gene symbol, which have a >3
fold down-regulation in each of the two PEA1 cell line clones compared with wild-type PEA.
The gene name is listed for those probes down-regulated in both PEA1 clones, although this
information was not always accessible. There were 32 probes demonstrating a >3 fold downregulation which were common to both PEA1 clones, with a p-value of intersection of <1x1016
, which is the smallest p-value that the R Bioconductor package can generate. These probes
were uploaded to the GeneGo from MetaCoreTM software to perform an enrichment analysis.
The 10 most relevant pathway maps are listed in Table 6.5, including the related downregulated genes.
Most notable pathways included cytoskeleton remodelling and cell
adhesion, with the down-regulated Phospholipase-C beta (PLC-β), Protein kinase A
regulated type II beta (PRKAR2β), and NUR77 (also known as orphan nuclear receptor TR3
and NR4A1) genes repeatedly involved in such pathways. It was also interesting to observe
the down-regulation of Tenascin-C (TNC) and Matrix metalloproteinase-1 (MMP1) which
may play a role in epithelial to mesenchymal transition, a process that has been implicated in
the development of chemoresistance. The 10 most relevant process networks are listed in
Table 6.6, including the associated down-regulated genes. Most notable process networks
included Wnt signalling and inflammatory responses, with the down-regulated PLC-β,
PRKAR2β and IL-1α genes repeatedly involved.
When comparing the wild-type PEA2 cell line with wild-type PEA1, again using a cut-off of
>3 fold change in regulation between the two samples, 248 probes were up-regulated and 746
probes were down-regulated in PEA2 compared with PEA1. When the fold change cut-off
was increased to 5, 51 probes were up-regulated and 215 were down-regulated. Interestingly
this suggested that the PEA2 cell line, which expresses a higher level of En2 compared with
PEA1, contained more silenced genes, however it is unclear whether that is related to high
En2 expression or entirely independent. Due to the lack of replicate samples, further analysis
of the PEA2 data in the GeneGo software was not possible, as it was unlikely to provide
satisfactory statistical significance.
182
Molecular
Probe
A_23_P72096
A_23_P215744
A_23_P1691
A_23_P7642
A_23_P354805
A_23_P83498
A_32_P136295
A_32_P134007
A_24_P365515
A_23_P252432
A_23_P27332
A_24_P696761
A_24_P943922
A_23_P47340
A_24_P766865
A_23_P69537
A_32_P356316
A_23_P62901
A_23_P44674
A_23_P501831
A_23_P157865
A_23_P149613
A_23_P208747
A_24_P933059
A_23_P119040
A_23_P312752
A_23_P15174
A_23_P28507
A_24_P311679
A_24_P110601
A_24_P941643
A_23_P259362
A_32_P205303
A_23_P18713
A_23_P338134
A_23_P27306
A_32_P85042
A_23_P128230
A_23_P42975
A_23_P210158
A_23_P62387
A_24_P657695
A_23_P131566
A_24_P388528
A_24_P127950
A_23_P333150
A_23_P415401
A_24_P349196
A_23_P105442
A_23_P206585
A_23_P28927
A_23_P35995
A_24_P374863
A_24_P683905
A_23_P140450
A_32_P214503
A_32_P171923
A_23_P137173
A_32_P101031
A_32_P49054
A_24_P324250
A_32_P221822
A_24_P353103
A_23_P113471
A_32_P42705
A_32_P223551
A_32_P94667
A_23_P163227
A_23_P202334
A_23_P410233
A_23_P29939
A_32_P186731
Gene Symbol
NM_000575
NM_033427
NM_002421
NM_003118
NM_007249
NM_006546
NM_052847
NM_052898
NM_021784
NM_004617
NM_003199
NM_001001552
NM_020925
NM_020693
AK096498
NM_006681
NM_002119
NM_006763
NM_001311
NM_032385
NM_002160
NM_002021
NM_005091
A_24_P933059
NM_024935
NM_002242
NM_005949
NM_012214
NM_014421
A_24_P110601
NM_182734
AF156973
ENST00000376646
NM_004827
NM_052919
NM_030781
AK056856
NM_002135
NM_002736
THC2278254
NM_018159
A_24_P657695
NM_014746
NM_173216
AK092844
AK127177
NM_001206
ENST00000340612
NM_181711
NM_002738
A_23_P28927
NM_024769
AK090421
AK001829
NM_003645
BC038512
THC2429853
NM_021992
NM_144586
THC2385437
A_24_P324250
ENST00000366932
NM_001007248
NM_174912
BC017721
ENST00000380831
NM_014644
NM_020990
NM_022970
AL832009
NM_007308
AK074473
PEA1
>200x
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
0
0
0
0
0
PEA1
>7000x
1
1
1
0
0
1
0
1
0
1
1
0
0
0
1
0
0
0
0
1
1
0
0
0
1
1
1
0
0
0
1
1
1
0
1
1
1
1
1
0
0
1
1
0
1
1
0
0
0
0
0
0
1
0
1
0
0
0
1
1
0
1
1
1
0
0
0
1
1
1
1
1
Gene Name
IL-1α
CTTNBP2
MMP1
IGF2BP1
XKR4
TM4SF4
TCF4
…
FAXDC2
TNC
KIAA1772
KCNJ13
MT1F
PLCB1
NPCDRG
VPS13A
KIAA1920
COLEC12
IPO5P1
NR4A1/NUR77
PRKAR2β
…
RNF144A
U2AF1L4
LRRC28
FLJ00330
SLC27A2
LYPD1
…
RRP15
ZNF599
FAAH2
Molecular
Probe
A_24_P719
A_23_P910
04
A_23_P547
81
A_23_P428
36
A_24_P139
298
A_32_P228
152
A_32_P122
037
A_23_P374
226
A_23_P406
844
A_23_P216
385
A_32_P121
468
A_32_P116
079
A_24_P330
857
A_23_P413
518
A_32_P475
44
A_23_P255
38
A_23_P214
695
A_24_P773
208
A_32_P232
539
A_23_P315
214
A_23_P427
836
A_23_P421
645
A_23_P153
89
A_24_P217
971
A_23_P213
520
A_32_P134
050
A_23_P163
764
A_23_P509
336
A_32_P144
19
A_32_P103
599
A_32_P452
508
A_23_P148
29
A_23_P133
249
A_23_P721
036
A_24_P592
2
A_23_P635
400
A_23_P204
41
A_23_P300
847
A_23_P775
740
A_32_P130
2
A_23_P166
56
A_23_P158
467
A_23_P108
76
A_32_P241
823
A_32_P109
40
A_24_P217
653
A_23_P214
234
A_23_P500
026
A_23_P128
000
A_23_P215
6
A_23_P690
913
A_23_P156
30
A_23_P534
431
A_23_P213
17
A_23_P217
745
A_23_P393
428
A_23_P408
034
A_23_P135
376
A_23_P897
990
A_23_P163
A_23_P154
087
A_32_P249
338
A_23_P376
50
A_23_P369
727
A_24_P725
994
A_23_P240
998
A_32_P208
83
A_24_P409
076
A_23_P350
166
A_32_P158
689
A_23_P372
272
A_23_P356
234
585
Gene Symbol
NM_000860
NM_002354
NM_016541
NM_173561
AL359062
AK127194
NM_152435
NM_015973
NM_153350
NM_004170
AK090467
AK127194
NM_001218
NM_001432
BC037919
NM_006632
NM_033181
THC2437069
THC2316753
NM_017451
NM_012464
NM_005074
NM_152386
NM_001432
NM_000860
ENST00000382216
AK000158
NM_006216
AW979273
AK096022
BC042026
NM_024817
AF146796
NM_000204
AK124400
AK001019
NM_002298
NM_148896
NM_020796
BC013657
AK026235
NM_052947
NM_032523
NM_005256
THC2377128
NM_000341
NM_001999
NM_144777
NM_144661
NM_203339
NM_001850
NM_005907
NM_006741
NM_004887
NM_001174
NM_005329
AB007877
NM_005630
NM_023938
NM_007361
NM_025202
ENST00000330861
NM_138326
NM_004734
THC2460656
NM_005396
A_32_P208076
A_24_P409166
NM_173570
AL359055
NM_001218
NM_002126
PEA1
>200x
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
PEA1
>7000x
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
1
Table 6.4. Microarray probes down-regulated in two PEA1 En2 over-expression clones
compared with wild-type PEA1 cells. The listed microarray probes demonstrated a >3 fold downregulation in either of the PEA1 clones (>200 or >7000 fold En2 expression) compared with wild-type PEA1.
Down-regulation is represented by “1”. The gene symbols highlighted in grey (n=32) represent those that
showed down-regulation in both PEA1 clones and the relevant gene name is listed where available. Some of
these genes were identified within the enrichment analysis pathway maps (red), whilst others proved of interest
183
during subsequent literature searches (yellow).
Gene
Name
Pathway map
p-value
Linked down-regulated genes from
expression analysis
1
CFTR-dependent regulation of ion channels in CF
1.761 x10-05
PLCβ, PKA-reg type II, PKA-reg
2
Neurophysiological process: Corticoliberin
signalling via CRHR1
3.216 x10-05
PLCβ, PKA-reg, NUR77
3
Transport: The role of AVP in regulation of
Aquaporin 2 and renal water reabsorption
3.216 x10-05
PKA-reg type II, PKA-reg, PRKAR2β
4
Muscle contraction: Regulation of eNOS activity in
cardiomyocytes
4.528 x10-05
PLCβ, PKA-reg type II, PKA-reg
5
Development: Non-genomic action of Retinoic acid
in cell differentiation
4.528 x10-05
PKA-reg type II, PKA-reg, PRKAR2β
6
Reproduction: GnRH signalling
9.628 x10-05
PLCβ, PKA-reg, NUR77
7
Cytoskeleton remodelling: Thyroliberin in
cytoskeleton remodelling
8.061 x10-04
PLCβ1, PKA-reg
8
Action of lithium on synaptic transmission and
autophagy
9.071 x10-04
PLCβ, PKA-reg
9
Development: Hedgehog & PTH signalling
pathways in bone & cartilage
9.597 x10-04
PLCβ, PKA-reg
10
Cell adhesion: Cell-matrix glycoconjugates
1.069 x10-03
TNC, MMP-1
Table 6.5. Pathway maps obtained from enrichment analysis of the down-regulated
genes in two PEA1 En2 over-expression clones compared with wild-type PEA1 cells.
These are the 10 most relevant pathways obtained from the enrichment analysis of the 32 probes with >3 fold
down-regulation in both PEA1 clones compared with wild-type PEA1.
Process network
p-value
Linked down-regulated genes from
expression analysis
1
Inflammation: MIF signalling
6.259 x10-07
PLCβ1, PLCβ, PKA-reg, IL-1α, MMP-1
2
Neurophysiological process: Corticoliberin
signalling
3.516 x10-05
PLCβ, PKA-reg, NUR77
3
Signal transduction: WNT signalling
6.800 x10-05
PLCβ1, PLCβ, PKA-reg II, PKA-reg
4
Reproduction: Gonadotrophin regulation
1.074 x10-04
PLCβ1, PLCβ, PKA-reg, NUR77
5
Signal transduction: Cholecystokinin signalling
3.529 x10-04
PLCβ1, PLCβ, PKA-reg
6
Inflammation: Protein C signalling
3.729 x10-04
PLCβ1, PLCβ, IL-1α
7
Muscle contraction: Nitric oxide signalling in the
cardiovascular system
5.599 x10-04
PLCβ, PKA-reg II, PKA-reg
8
Cell adhesion: Platelet aggregation
1.135 x10-03
PLCβ1, PLCβ, PKA-reg
9
Reproduction: GnRH signalling pathway
1.310 x10-03
PLCβ1, PLCβ, PKA-reg
10
Inflammation: Innate inflammatory response
1.682 x10-03
PLCβ1, PLCβ, IL-1α
Table 6.6. Process networks obtained from enrichment analysis of the down-regulated
genes in two PEA1 En2 over-expression clones compared with wild-type PEA1 cells.
These are the 10 most relevant process networks obtained from the enrichment analysis of the 32 probes with >3
fold down-regulation in both PEA1 clones compared with wild-type PEA1.
184
Technical validation of selected up- and down-regulated genes identified from the pathway
maps and process networks was then performed, using quantitative rt-PCR, and included the
PEA1 clones along with the wild-type PEA1 and PEA2 cell lines. Although it was not
identified on one of the pathway maps or process networks, it was noted that the TWIST2
gene was up-regulated in both PEA1 clones compared with wild-type PEA1. This gene
encodes a transcription factor protein that has been implicated in epithelial-mesenchymal
transition (EMT) of cancer cells as well as cisplatin-resistance in EOC through the
AKT/GSK-3β signalling pathway, therefore primers were included for this gene within the rtPCR analysis.
Figure 6.14 demonstrates the relative expression of up-regulated genes
identified from the microarray analysis plotted on a log scale graph, and confirms the upregulation of the PAI1, GADD45β, and Activin A genes in the En2 over-expressed PEA1
clones compared with wild-type PEA1. An even greater level of up-regulation of these genes
was observed in the wild-type PEA2 cell line, even though the En2 expression was not as
high as in the cloned cells. However, up-regulation of the Ski, IL1RAP and CSH1 genes was
not confirmed, in fact CSH1 was down-regulated in the En2 over-expressed cell line and
clones. This could reflect an error in primer design. Interestingly, 2.5 fold up-regulation of
TWIST2 was identified in the PEA1 clones compared with wild-type PEA1, however very
1000000
PEA1
100000
P E A 1 E 1 2 (2 0 0 x )
10000
P E A 1 H 3 (7 0 0 0 x )
P E A 2 ( 1 3 .5 x )
1000
100
10
1
W
T
C
S
IS
T
H
2
1
P
R
1
IL
v
ti
A
c
D
A
G
A
A
in
5
D
P
4
A
k
S

I1
0 .1
i
C T v a lu e r e la tiv e t o B - a c t in
minimal increase in expression was demonstrated in PEA2.
Figure 6.14. Quantitative rt-PCR validation of microarray-identified up-regulated genes
in PEA1 En2 over-expressing clones. cDNA from wild-type PEA1, PEA1 clone E12 (200x En2
expression compared with w-t PEA1), PEA1 clone H3 (7000x En2 expression) and wild-type PEA2 cells (13.5x
En2 expression) was analysed by quantitative rt-PCR. The mRNA expression from seven selected genes is
shown on a log10 scale, relative to the housekeeping gene β-actin (x100,000).
185
Although it was not identified on one of the pathway maps or process networks, downregulation of the IL1A gene was analysed in addition to the other down-regulated genes
identified from the microarray. This gene seemed particularly interesting given the previous
findings of up-regulation of the IL1RAP gene.
Figure 6.15 demonstrates the relative
expression of down-regulated genes identified from the microarray analysis plotted on a log
scale graph, and confirms the down-regulation of the PLCβ1, PRKAR2β, NUR77/NR4A1,
TNC and MMP1 genes in the En2 over-expressed PEA1 clones compared with wild-type
PEA1. Down-regulation of these genes was also observed in the wild-type PEA2 cell line,
although this was not so notable with the TNC gene. Down-regulation of IL1A was identified
in the PEA1 clones compared with wild-type PEA1, however marked up-regulation was
10000
PEA1
P E A 1 E 1 2 (2 0 0 x )
1000
P E A 1 H 3 (7 0 0 0 x )
P E A 2 ( 1 3 .5 x )
100
10
1
M
A
1
M
IL
P
1
C
N
T
A
4
R
/N
N
U
R
P
7
7
R
K
P
A
L
R
C
2

B
1
0 .1
1
C T v a lu e r e la tiv e t o B - a c t in
demonstrated in PEA2.
Figure 6.15. Quantitative rt-PCR validation of microarray-identified down-regulated
genes in PEA1 En2 over-expressing clones. cDNA from wild-type PEA1, PEA1 clone E12 (200x
En2 expression compared with w-t PEA1), PEA1 clone H3 (7000x En2 expression) and wild-type PEA2 cells
(13.5x En2 expression) was analysed by quantitative rt-PCR. The mRNA expression from six selected genes is
shown on a log10 scale, relative to the housekeeping gene β-actin (x100,000).
186
The previous enrichment analysis of the up-regulated genes identified several components of
the TGFβ and Activin A signalling pathways, however many of the key components of these
pathways were not detected. Such genes may still have been up- or down-regulated in the
PEA1 En2 over-expressing clones, but did not achieve the >3 fold cut-off value. Therefore
additional primers were designed to help evaluate the mRNA expression of 5 other genes
which play important roles in TGFβ and Activin A signalling, namely genes encoding the
Activin receptors ActRIIa and ActRIIb, the intracellular signalling proteins SMAD3 and
SMAD4, and TGFβ1. Figure 6.16 demonstrates the relative expression of these genes
plotted on a log scale graph. Down-regulation of all of these genes was demonstrated in the
En2 over-expressed PEA1 clones compared with wild-type PEA1, with associated down-
100000
PEA1
P E A 1 E 1 2 (2 0 0 x )
10000
P E A 1 H 3 (7 0 0 0 x )
1000
P E A 2 ( 1 3 .5 x )
100
10
b
II
A
A
c
tR
tR
c
A
S
M
A
M
S
II
4
D
3
D

F
G
T
a
1
1
C T v a lu e r e la tiv e t o B - a c t in
regulation in the wild-type PEA2 cell line in all but the SMAD3 gene.
Figure 6.16. Quantitative rt-PCR validation of genes involved in TGFβ and Activin A
signalling, from PEA1 En2 over-expressing clones. cDNA from wild-type PEA1, PEA1 clone E12
(200x En2 expression compared with w-t PEA1), PEA1 clone H3 (7000x En2 expression) and wild-type PEA2
cells (13.5x En2 expression) was analysed by quantitative rt-PCR. The mRNA expression from five selected
genes is shown on a log10 scale, relative to the housekeeping gene β-actin (x100,000).
187
6.3.
DISCUSSION
The aims of this chapter were to establish the role and function of EN2 in EOC cancer cells,
with a particular focus on the potential involvement in platinum-resistance.
EN2 has previously been detected in the supernatant of prostate cancer cell lines with
subsequent repeated identification in the urine of men with prostate cancer [303, 314, 355].
Elevated urinary EN2 protein was observed in 61% of epithelial ovarian cancer patients, so it
was hoped that secreted EN2 protein would be observed in the supernatant from each of the
EOC cell lines, and that the concentration could be quantified. However only the OV90 and
TOV112D cell lines demonstrated EN2 secretion greater than 100ng/ml, whilst the fibroblast
cell lines and the majority of EOC cell lines exhibited less than 50ng/ml of protein. These
were also the only two cell lines that demonstrated 50kDa protein bands. The cells were left
for 24 hours in serum-free media in order to collect secreted protein, so perhaps a longer
incubation period may have produced greater protein concentration, however the lack of
serum for this period of time may impact on the growth and survival of the cells. Another
possible explanation for the minimal secretion could be that in vivo, elements within the
tumour microenvironment promote EN2 secretion from the tumour cells, however these
factors are absent within an in vitro model.
En2 was successfully transiently over-expressed in PEA1 cells and numerous stable cell line
clones were obtained, with En2 mRNA expression levels ranging from 0-7000 fold upregulation. Many of these PEA1 clones expressed En2 at a level well above that of the paired
PEA2 cell line, and Western blot analysis of the cloned cell lysates demonstrated an
additional protein band at 42kDa. The 50kDa band, which had previously been observed in
wild-type PEA1 cells, was still present and there was no clear increase in the size of that
band. It was speculated that the new 42kDa band may represent EN2 protein prior to the
completion of post-translational modification. The very high level of En2 mRNA expression
and resultant excess protein translation may have overwhelmed the normal modification
mechanisms. It was thought that one such post-translational modification could involve
glycosylation, and deglycosylation of the cell lysates did reveal a band at 33kDa, however
this latter band was also seen in the negative control PEA1 cells i.e. those that did not overexpress En2. There was also very minimal visual change in the size of the 50 or 42kDa
protein bands which could either suggest that the deglycosylation protocol is inefficient, or
188
that glycosylation is not the only post-translational modification occurring. It was interesting
to see that the 33kDa band corresponded to the largest band seen within the rat brain lysate
positive control, and is quoted in the literature as the size of EN2 protein in normal human
and murine cells. This suggests that EN2 protein in EOC cells is always modified from the
native state, explaining why we repeatedly detect it at 50kDa rather than 33kDa.
The original intention was to perform siRNA-mediated silencing of En2 in PEA2 cells but a
reduction in En2 expression equivalent to that of wild-type PEA1 cells could not be achieved.
Sigma-Aldrich were also unable to perform satisfactory gene knock-out using zinc-finger
nuclease technology. Several months later they did explain that they now had several predesigned Cas9-nucleases and nickases which target the first exon of the En2 gene to enable
RNA-guided genome editing, however there was insufficient time left to pursue this project.
In the meantime cisplatin challenge of the PEA1 clones and the wild-type PEA2 cell line was
conducted, establishing IC50 values in order to compare drug sensitivity. The IC50 for the
PEA1 clones with En2 over-expression of 0-1000 fold was 13-14μM, whilst the wild-type
PEA2 cell line IC50 was 31μM. Therefore, over-expression of En2 in PEA1 cells ≤1000 fold
did not alter the sensitivity of the cells to cisplatin, however when En2 mRNA demonstrated
>7000 fold increase, there was a reduction in sensitivity to 19μM, although still not to the
level of PEA2 cells. Ideally evaluation of the sensitivity of PEA2 cells with silenced En2
would have been carried out, to see if cisplatin resistance could be reversed, especially as this
could prove to be a promising treatment approach. Ultimately it seemed as if En2 overexpression may well contribute to platinum-resistance in EOC although it probably functions
in combination with other proteins to suppress apoptosis, promote DNA repair or increase
tolerance to DNA damage, in the presence of platinum-based therapy.
Gene expression profiling and analysis of wild-type PEA1 and PEA2 cells, along with PEA1
clones E12 and H3 with stable En2 over-expression, provided a more detailed insight into the
pathways that are influenced by En2 over-expression and its role in oncogenesis and
platinum-resistance. It was interesting to observe that almost four times as many genes were
down-regulated than up-regulated when comparing wild-type PEA2 with PEA1, suggesting
that more genes were silenced in association with platinum-resistance. Stronach et al. also
showed that more genes were down-regulated than up-regulated in the platinum-resistant
pairings of cell lines from high-grade serous ovarian cancer, when they evaluated the
PEO1/PEO4, PEA1/PEA2 and PEO14/PEO23 cell line pairings [340]. Unfortunately due to
189
the lack of replicate samples, further analysis of the PEA2 data could not be carried out in
order to identify individual genes, however by combining the gene expression profiles from
the two PEA1 clones, enrichment analysis could be performed comparing up- and downregulated genes in comparison with wild-type PEA1. In the absence of triplicate or greater
PEA1 cloned repeats, Dr Carla Moller-Levet advised the use of a minimum 3 fold cut-off
threshold. In comparison to other microarray analyses in the literature, this is quite a high
threshold to use and so the analysis may have excluded some pertinent genes. However the
number of sample replicates would have to be increased to allow the threshold to be reduced.
As a result, attention and literature reviews were focussed on the genes identified within the
pathway maps and process networks, with subsequent identification of all of the mutually upand down-regulated genes, using the molecular probe codes or gene symbols.
The TGFβ receptor signalling pathway was highlighted as being strongly linked to the overexpression of En2, with up-regulation of the component genes Ski, PAI1 and GADD45β. The
pathway map produced by the GeneGo software is shown in Figure 6.17. Ski is a known
nuclear proto-oncogene that suppresses TGFβ1 function therefore enhancing cell
proliferation [409-414], but its up-regulation could not be proven in the PEA1 clones and
wild-type PEA2 cell line in the subsequent rt-PCR validation. This could relate to the chosen
sequence of the primers rather than true absence of up-regulation, and so it would be
worthwhile repeating this validation in the future with differing primer sequences. However
PAI1 and GADD45β were confirmed as up-regulated in the two PEA1 clones compared with
wild-type PEA1, suggesting a direct effect of En2 over-expression on the up-regulation of
these genes. In addition, the wild-type PEA2 cells demonstrated elevated levels of PAI1 and
GADD45β, suggesting that over-expression of these genes may influence metastatic spread of
disease and platinum-resistance. In fact, PAI1 has been shown to be involved in cell invasion
and metastasis via regulation of the extracellular matrix [415-417], whilst numerous
publications have highlighted the anti-apoptotic role of GADD45β [418-422]. Although
other key components of this pathway, such as TGFβ1 itself and the SMAD proteins, had not
been identified as up- or down-regulated in the microarray analysis, this may have been due
to the choice of quite a high fold cut-off value i.e. 3 fold up- or down-regulation. Therefore
further analysis of their expression at the mRNA level was conducted, using rt-PCR. This
confirmed down-regulation of TGFβ1 in both of the PEA1 clones as well as PEA2, compared
with wild-type PEA1, which would be consistent with suppression by Ski as previously
discussed. SMAD4 expression was also down-regulated in response to En2 over-expression,
190
as well as in the PEA2 cell line, which may occur due to direct suppression by Ski or by En2
itself. Although it demonstrated clear down-regulation in the PEA1 clones, the SMAD3 gene
was in fact up-regulated in PEA2 cells, therefore expression of this gene may be more
strongly influenced by other genes in the wild-type PEA2 cell line. In addition, there are
known to be SMAD-independent pathways for TGFβ signalling, which utilise the
MAPK/PI3K pathways [423], perhaps linking GADD45β expression, therefore components
of this pathway could be further explored in the future and may help to clarify some of these
conflicting results.
Figure 6.17. The TGFβ receptor signalling pathway (taken from LS Research, Thomson
Reuters; lsresearch.thomsonreuters.com).
191
The Activin A signalling pathway was also implicated in response to En2 over-expression,
with involvement in signal transduction as well as cytoskeleton remodelling, and close
association with Ski and PAI1. Activin A is known to play a role in Epithelial-Mesenchymal
transition (EMT) which describes epithelial cell loss of differentiated characteristics, such as
cell-cell adhesion, apical-basal polarity and lack of cell motility [424-430]. Ultimately the
cells gain mesenchymal features that include motility, invasiveness and increased resistance
to apoptosis.
This process is thought to be prominent in the development of cancer
metastases [431, 432]. The Activin A gene was confirmed as up-regulated in the PEA1 clones
using rt-PCR, with a >100,000 fold increase in expression seen in the wild-type PEA2 cell
line compared with PEA1. After studying the Activin A signalling pathway map produced
by the GeneGo software (Figure 6.18), the mRNA expression levels of the ActRIIa and
ActRIIb genes were additionally evaluated. These encode for the cell membrane receptors
that bind the Activin A ligand.
Both of these receptors actually demonstrated down-
regulation in response to En2 over-expression, despite the significant up-regulation of Activin
A. Similar down-regulation was also observed in the wild-type PEA2 cell line. Perhaps there
is a negative feedback pathway which inhibits expression of the receptor genes in order to
limit the effects of Activin A, thereby preventing further signal transduction. It is also
possible that the functional Activin A protein in this setting is intracellular rather than
extracellular, therefore bypassing receptor activation. Activin A signal transduction is also
thought to utilise SMAD-independent pathways, involving MAPK/PI3K much like the TGFβ
signalling pathway [433, 434]. Again, future evaluation of additional components within this
pathway may provide further information.
192
Figure 6.18. The Activin A signalling transduction pathway (taken from LS Research,
Thomson Reuters; lsresearch.thomsonreuters.com).
When considering the pathway maps generated from the most significantly down-regulated
probes in both PEA1 clones i.e. “CFTR-dependent regulation of ion channels in cystic
fibrosis”, “Corticoliberin signalling via CRHR1” and “The role of AVP in regulation of
Aquaporin 2 and renal water reabsorption”, there did not appear to be any direct links to
known oncogenic pathways.
However when the process networks generated from the
enrichment analysis of the down-regulated probes were studied, involvement of Wnt
signalling along with various inflammatory processes was identified. The rt-PCR validation
confirmed the substantial down-regulation of the PLCβ1 and PRKAR2β genes in the PEA1
clones compared with wild-type PEA1, with an even greater down-regulation in wild-type
PEA2. PLCβ1 has been shown to play an important role in cell proliferation, with direct
targeting and activation of protein kinase C and release of calcium from endoplasmic
reticulum [435]. Several studies reported involvement of PLCβ1 in G0-G1/S transition and
193
G2/M progression in Friend murine erythroleukemia cells (FeLC), but there has been limited
focus on its role in humans, particularly in cancer [436, 437]. Interestingly, Poli et al.
observed that over-expression of PLCβ1 in K562 cells, which are representative of human
erythroleukaemia cells, promoted prolongation of S phase resulting in a delay in cell
proliferation [438]. PLCβ1 gene deletion has also been implicated in the progression of
myelodysplastic syndrome to acute myeloid leukaemia in humans, with a particularly poor
prognosis [439, 440]. Therefore in these EOC cell lines, the over-expression of En2 may
silence PLCβ1 resulting in promotion of cellular proliferation.
NUR77, also known as NR4A1, is a nuclear protein that has been shown to promote
apoptosis [441-447], however the microarray analysis and rt-PCR validation confirmed that
this gene was down-regulated by almost 1000 fold in the PEA1 clones and PEA2 cell lines.
It is likely that this gene silencing acts to enhance anti-apoptotic function, encouraging cancer
growth but also possibly promoting platinum resistance. Down-regulation of the TNC and
MMP1 genes in the En2 over-expressed PEA1 clones was also confirmed. Down-regulation
of these genes was also observed in the wild-type PEA2 cell line, most notably in the MMP1
gene. MMP1 is a member of the zinc-dependant matrix metalloproteinases that are involved
in extracellular matrix remodelling. Much of the literature describes up-regulation of MMP1
in advanced cancer, however down-regulation has been associated with metastatic tumours,
namely liver metastases in colon cancer [448, 449]. Suppression of TNC and MMP1 in EOC
cells may influence epithelial-mesenchymal transition with promotion of metastatic spread.
Although it was not identified on one of the pathway maps or process networks, downregulation of the IL1A gene was also evaluated, especially as the IL1RAP gene had been
identified as up-regulated within the enrichment analysis. The IL1RAP gene was not upregulated in the PEA1 clones although there was a 1.9 fold increase in expression in the
PEA2 cells, and although down-regulation of IL1A in the PEA1 clones was confirmed,
marked up-regulation was observed in PEA2. It would be prudent to repeat this experiment
with re-designed primers to ensure that these observations are correct, however it most likely
reflects the fact that En2 is not solely influencing these genes within the wild-type PEA2 cell
line.
There has been very little research related to over-expression of En2 either in the context of
developmental biology or disease. However one such publication conducted En2 overexpression experiments in Xenopus embryos to study brain patterning.
194
Koenig et al.
identified direct repression of Tcf-4, a transcription factor involved in anterior-posterior
patterning of the brain, in response to En2 over-expression [450]. Tcf-4 is also believed to be
involved in the Wnt-signalling pathway [451]. Although Tcf-4 was not identified in the
process networks or pathway maps of down-regulated genes in our microarray analysis, it
was one of the 32 genes down-regulated >3 fold in both PEA1 En2 over-expressed clones. It
is unclear at present whether this suppression of Tcf-4 plays a vital role in the progression of
ovarian cancer, although its involvement in Wnt-signalling makes it a definite gene of interest
in relation to En2. It is also reassuring to have been able to reproduce the findings of another
group studying En2.
When studying the table of all probes with >3 fold up-regulation in the PEA1 clones
compared with wild-type PEA1, TWIST2 was also present. This gene encodes a transcription
factor that has been implicated in cisplatin-resistance in ovarian cancer, through the
AKT/GSK-3β signalling pathway [452]. Along with TWIST2, other known EMT-related
transcription factors SNAIL, SLUG and ZEB2 that act as E-cadherin repressors, have
demonstrated up-regulation in cisplatin-resistant ovarian cell lines [432, 453]. Using their
gene symbols to search within the microarray table of up-regulated genes, SNAIL and ZEB2
were identified as up-regulated in the PEA1 H3 clone (>7000 fold En2 over-expression)
although not in the E12 clone, perhaps because they did not achieve the >3 fold threshold.
TWIST2 primers were included within the rt-PCR validation and a 2.5 fold up-regulation in
the PEA1 clones was observed, however there was minimal increased expression in PEA2.
Future research could include re-evaluation of this gene along with SNAIL, ZEB2 and SLUG
to determine whether they are down-stream targets of En2, promoting EMT and cisplatinresistance. As previously discussed, the Wnt/β-catenin pathway promotes En2 expression
due to the action of nuclear β-catenin. The latter protein has also been shown to activate
other target genes, some of which promote EMT [453, 454], therefore it is logical to assume
that En2 is also involved in this process.
195
6.4.
CONCLUSION
En2 appears to be up-regulated in epithelial ovarian cancer specimens compared with normal
tissue, and very high levels of mRNA and protein expression are associated with platinum
resistance and a shorter progression-free survival. However the results from the platinum
challenge of En2 over-expressed PEA1 cell lines suggest that En2 may directly influence the
development of platinum resistance, rather than being elevated as a result of up-stream
genetic factors.
Microarray hybridisation analysis comparing En2 over-expressed PEA1 clones with the wildtype PEA1 cell line has identified a number of over-expressed and down-regulated genes that
ultimately function to prevent apoptosis, and enhance cell invasion and metastasis, possibly
via the process of epithelial-mesenchymal transition. This may also directly influence the
development of resistance to platinum chemotherapy.
In the majority of cases, these
dysregulated genes were also identified in the wild-type PEA2 cell line, which represents
advanced, platinum-resistant disease.
This provides further evidence that En2 plays a
prominent role in the progression of epithelial ovarian cancer and resistance to treatment, and
it is likely to function through the involvement of multiple signalling pathways.
196
CHAPTER 7
DISCUSSION
197
7.
DISCUSSION
Despite significant advances in the medical management of solid tumours over the past
decade, there has not been a particular impact on the survival of advanced epithelial ovarian
carcinoma. CA125 remains the only approved serum biomarker in ovarian cancer, however
its routine use is restricted to monitoring ovarian cancer progression and treatment response.
The combination of other novel biomarkers with CA125 has also been extensively studied
however these are not routinely used in the clinic setting.
Many developmental homeodomain-containing transcription factors are aberrantly expressed
in cancer and have been shown to promote cancer development, progression, recurrence or
development of drug resistance [185, 455]. The homeobox gene, Engrailed-2 (En2), encodes
a transcription factor which plays an important role in embryonic development, but appears to
have limited function in the normal adult nervous system [270, 290]. EN2 over-expression
has been identified in a number of adult human cancers, namely breast, prostate, and bladder,
and appears to have a functional role in tumour development. There is a great deal of interest
in the potential of EN2 as a diagnostic biomarker in prostate cancer given that its secretion
into the urine was highly predictive of prostate cancer.
Preliminary work on cell lines and tissue arrays, suggested that EN2 may also be present in
ovarian cancer [305], raising the possibility of a role as a diagnostic or prognostic biomarker,
and suggesting functional significance. To confirm this hypothesis, the biology and function
of En2 gene expression as well as EN2 protein expression in EOC cell lines, tumour tissue
and bodily fluids was characterised, and correlated with clinico-pathological data. Further
information on protein size and localisation within the cell was gained, and over-expression
of En2 was studied focussing on the subsequent effects on treatment response and the
associated up- and down-regulated genes and their signalling pathways.
198
7.1.
EN2 AS A DIAGNOSTIC BIOMARKER IN EOC
The mRNA expression levels of En2 in ovarian tissue could be used as a diagnostic marker of
EOC, particularly in the high-grade serous histological sub-type, given that these levels were
negligible in normal ovarian specimens, and very low in non-invasive borderline and benign
serous ovarian tumours. It may also help to provide an early indication of whether the
original tumour derives from the ovary, peritoneum or fallopian tube, given that En2 mRNA
levels for the latter two groups tended to be lower. At present, the primary source of the
cancer is often not confirmed until primary or interval-debulking surgery has been completed.
EN2 protein staining was also detected in the majority of EOC and borderline tumour
specimens, and was negative in all normal ovary specimens, and in most benign tumours.
However the intensity of staining did not reliably distinguish between histological sub-types
or grade or stage of disease.
En2 promoter hypermethylation also showed promise as a diagnostic biomarker as low mean
levels of methylated En2 were demonstrated in normal ovary and benign tissue, compared
with epithelial ovarian cancer specimens. Nevertheless, acquisition of tissue for analysis is
an invasive technique requiring a core- or excision-biopsy and En2 mRNA or promoter
hypermethylation would not help in the diagnosis of early stage disease, when an ovarian
mass may be undetectable on imaging. Therefore many researchers have strived to identify
potential diagnostic biomarkers in fluids such as urine or blood, which are much more easily
obtainable and could be repeatedly measured for instance, within a screening programme.
Elevated EN2 protein was present in 86% of patients with newly diagnosed high-grade serous
EOC, compared with female healthy controls, suggesting that urinary EN2 levels could be
utilised as a non-invasive diagnostic biomarker in high-grade serous patients. The numbers
tested were low however, and its true clinical utility may lie in a panel test with other urinary
proteins such as HE4 and mesothelin. However the non-invasive nature of this test and the
potential for detection prior to any clinical signs or symptoms of disease, make it a worthy
candidate for ongoing research.
199
7.2.
EN2 AS A PROGNOSTIC BIOMARKER IN EOC
The En2 mRNA expression levels showed promise as a prognostic biomarker in interval
debulking surgery high-grade serous EOC tissue where an elevated level predicted a shorter
progression-free survival. However the exact survival differences and the level of statistical
significance depended on which CT value cut-off was utilised, although this would become
clearer with a longer period of follow-up. Elevated En2 mRNA levels also suggested a
shorter overall survival, further supporting its promise as a prognostic biomarker, however
40% of the overall survival data remained censored at the time of analysis.
EN2 protein analysis of interval debulking surgery high-grade serous tumour specimens, also
demonstrated utility as a prognostic biomarker. Those with positive EN2 expression had a
shorter progression-free survival, with durations in keeping with the mRNA data. However
there were no significant differences in overall survival.
Given that interval debulking surgery with pathological analysis is commonplace in the
management of high-grade serous EOC, evaluation of En2 mRNA or EN2 protein expression
could help the oncologist to decide on a more individualised post-operative treatment
regimen.
7.3.
EN2 AS A TREATMENT RESPONSE BIOMARKER IN EOC
When analysing high-grade serous EOC specimens in those patients that had received
neoadjuvant chemotherapy compared with those undergoing primary surgery, there was a
notable trend towards lower En2 mRNA levels following chemotherapy. This suggests a
direct effect of chemotherapy treatment on these levels. If gene expression could be analysed
at the time of initial diagnostic biopsy, and subsequently evaluated at interval debulking
surgery, a reduction in En2 mRNA levels could help to confirm treatment response.
Elevated En2 mRNA levels at interval debulking surgery in high-grade serous EOC patients,
were also associated with early disease relapse suggesting resistance to platinum
chemotherapy. In combination with the gynae-oncology surgeon’s assessment of the amount
of residual disease at debulking surgery, the level of En2 mRNA could be used to assess
treatment response, and guide the oncologist in whether to prescribe additional cycles of postoperative chemotherapy, or to change the combination of drugs.
200
7.4.
THE BIOLOGY AND FUNCTION OF EN2
EN2 protein was located in the cytoplasm of EOC cell lines and tumour tissue, and not the
nucleus, as seen in normal adult Purkinje neurones. In certain cell lines, it was visualised in
association with the cell membrane, which may relate to cellular secretion or uptake.
Although secreted EN2 was only detected in two EOC cell lines, its presence in EOC patient
urine suggested that EN2 can be actively secreted in to the bloodstream from tumour sites,
and subsequently filtered through the kidney tubules into urine.
Although the predicted molecular weight of EN2 in normal tissue is 33kDa, EN2 protein was
detected in EOC cell line lysates and supernatants at 50kDa, suggesting that in cancer it is
modified from the native state. When En2 was over-expressed in an EOC cell line, an
additional protein band was visualised at 42kDa, equivalent to that of the recombinant EN2
protein band. Deglycosylation of this EN2 protein resulted in identification of yet another
band at 33kDa, consistent with normal tissue EN2. It appears that EN2 protein in cancerous
tissue and biological fluids undergoes post-translational modification, which is likely to relate
to its functional role within the cell or its effect on neighbouring cells.
The regulation of En2 expression is likely to be controlled by complex epigenetic
mechanisms with little apparent involvement of En2 promoter hypermethylation.
A
reduction in En2 mRNA expression was not clearly associated with an increased percentage
methylation of the En2 promoter. Conversely En2 mRNA over-expression may occur in the
presence of En2 promoter hypermethylation in human EOC tumours, a phenomenon which
has been noted in the autistic cerebellum. Analysis of a larger cohort of tissue samples would
be required in order to further investigate such potential associations.
Increased En2 mRNA and EN2 protein expression were identified in platinum-resistant EOC
cell line models and human tissue samples compared with platinum-sensitive cases. In the
case of the paired EOC cell lines, this suggested that increased En2 expression may directly
influence the advancement of disease or development of platinum-resistance. Forced En2
mRNA over-expression in a platinum-sensitive cell line did increase resistance to cisplatin
treatment, but the effect was marginal compared with the paired, acquired platinum-resistant
cell line. Therefore EN2 probably plays a role in the development of platinum-resistance but
is unlikely to be the sole causative factor. Microarray analysis of En2 over-expressed cell
line clones has identified a number of pathway networks and individual genes which have
201
been implicated in cancer progression, or in fact, the development of platinum-resistance in
EOC. Suppression of apoptosis, promotion of DNA repair or increased tolerance to DNA
damage are potential mechanisms promoting disease growth, progression and chemotherapy
resistance. In addition, genes involved in cell invasion and metastasis, via the process of
epithelial-mesenchymal transition, were repeatedly identified. These dysregulated genes of
interest were also identified in the native, platinum-resistant cell line pairing, confirming that
the effects of forced En2 over-expression mimic what occurs in the development of
advanced, platinum-resistance EOC. This further demonstrates that En2 plays a prominent
role in the progression of epithelial ovarian cancer and resistance to treatment, and it is likely
to function through the involvement of multiple signalling pathways.
7.5.
FUTURE WORK
In order to further investigate the potential for urinary EN2 as a diagnostic, prognostic or
treatment response biomarker in EOC patient urine, more urine samples need to be analysed
at the time of diagnosis, prior to commencement of treatment. Serial urine samples could
then be obtained during and after treatment to monitor changes in EN2 protein expression
level, and correlated with demographic and follow-up data. In recent urinary EN2 studies in
prostate cancer, cohorts of 50-100 patients were analysed. Similar numbers of samples
should be easily obtainable from The Royal Surrey County Hospital, and the ethical approval
is already in place.
A repeat analysis of survival data in relation to tumour En2 mRNA and EN2 protein
expression should be conducted periodically to ensure that all currently censored data has
been included, where possible. This should ensure that platinum-sensitivity data is available
for all patients, and a minimum follow-up period of one year has been recorded.
It would be interesting to increase the cohort of tumour samples tested for En2 promoter
methylation to further investigate the association between En2 mRNA expression and
methylation status, and to correlate with demographic information such as disease stage,
platinum-sensitivity and survival data. If such hypermethylation could also be detected in
fluid samples such as blood or urine, via the analysis of circulating tumour cells or DNA, this
may have improved clinical utility as a biomarker of early diagnosis.
202
Having identified EN2 protein in EOC specimens at a different molecular weight to that
published in the literature for normal tissue, and having observed multiple bands in certain
cases, it would be prudent to further investigate other potential post-translational
modifications. EN2 protein present in EOC cell lines, human tumour tissue and urine may
undergo different modifications related to its function or localisation. Techniques such as dephosphorylation could be performed, along with more comprehensive accurate mapping of
protein configuration by mass spectrometry.
Over-expression of En2 may not necessarily initiate or transform molecular events in tumour
pathogenesis, however it may facilitate the further development and distant spread of the
tumour, through the effects of the En2 gene on apoptosis resistance, epithelial-mesenchymal
transition, and tumour cell migration as previously discussed. This phenomenon has been
suggested in relation to the transcription factor, PAX8, in the tumourigenic phenotype of
ovarian cancer cells [456]. Interestingly, PAX8 is also not expressed in ovarian surface
epithelium, like EN2, but is found in the majority of serous, endometrioid and clear cell
ovarian carcinomas [351]. PAX8 is also positive in fallopian tubal secretory epithelial cells
[361]. Di Palma et al. stably transfected a normal ovarian cell line (IOSE-80PC) with PAX8,
analysing the expression of epithelial and mesenchymal markers in two representative clones
[456].
SNAIL, TWIST and ZEB2, known E-cadherin repressors, were significantly
upregulated, as were certain mesenchymal markers. Similarly, it would be interesting to
overexpress En2 in this normal ovarian cell line, confirming the over-expression and downregulation of the genes identified in our microarray of En2-overexpressed PEA1 cells, but to
also carry out cell proliferation as well as migration assays to further assess its contribution to
tumour progression.
Ultimately, it would be prudent to repeat the microarray hybridisation analysis using replicate
biological samples. This would then enable a 2-fold cut-off for up- or down-regulation of
genes to be utilised, which would likely increase the numbers of associated dysregulated
genes. It would be interesting to see if the previously discussed epithelial-mesenchymal
transition related genes such as SNAIL, SLUG and ZEB2 are dysregulated in En2 overexpressed clones, and to further analyse the TGFβ receptor and Wnt signalling pathways.
203
Hopefully such additional work should help to promote the use of urinary EN2 protein as a
diagnostic biomarker in epithelial ovarian cancer, as well as consolidate our knowledge about
the direct role of En2 and its protein product in the development, progression and spread of
cancer.
204
REFERENCES
205
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
CRUK. Statistics and outlook for ovarian cancer. Ovarian cancer 2015 30th April 2014 [cited
2015 21st January 2015]; Available from:
http://cancerhelp.cancerresearchuk.org/type/ovarian-cancer/treatment/statistics-andoutlook-for-ovarian-cancer.
Pfisterer, J. and J.A. Ledermann, Management of platinum-sensitive recurrent ovarian
cancer. Semin Oncol, 2006. 33(2 Suppl 6): p. S12-6.
Gonzalez-Martin, A., Treatment of recurrent disease: randomized trials of monotherapy
versus combination chemotherapy. Int J Gynecol Cancer, 2005. 15 Suppl 3: p. 241-6.
Jemal, A., et al., Cancer statistics, 2009. CA Cancer J Clin, 2009. 59(4): p. 225-49.
Institute, N.C. SEER Stat Fact Sheets: Ovary Cancer. Surveillance, Epidemiology, and End
Results Program - Turning Cancer Data Into Discovery 2014 [cited 2014 28th August 2014];
Available from: http://seer.cancer.gov/statfacts/html/ovary.html.
Chen, S. and G. Parmigiani, Meta-analysis of BRCA1 and BRCA2 penetrance. J Clin Oncol,
2007. 25(11): p. 1329-33.
Koornstra, J.J., et al., Management of extracolonic tumours in patients with Lynch syndrome.
Lancet Oncol, 2009. 10(4): p. 400-8.
Barrow, E., et al., Cumulative lifetime incidence of extracolonic cancers in Lynch syndrome: a
report of 121 families with proven mutations. Clin Genet, 2009. 75(2): p. 141-9.
Ness, R.B., et al., Infertility, fertility drugs, and ovarian cancer: a pooled analysis of casecontrol studies. Am J Epidemiol, 2002. 155(3): p. 217-24.
Chittenden, B.G., et al., Polycystic ovary syndrome and the risk of gynaecological cancer: a
systematic review. Reprod Biomed Online, 2009. 19(3): p. 398-405.
Pearce, C.L., et al., Association between endometriosis and risk of histological subtypes of
ovarian cancer: a pooled analysis of case-control studies. Lancet Oncol, 2012. 13(4): p. 38594.
Jordan, S.J., et al., Does smoking increase risk of ovarian cancer? A systematic review.
Gynecol Oncol, 2006. 103(3): p. 1122-9.
Tworoger, S.S., et al., Association of oral contraceptive use, other contraceptive methods,
and infertility with ovarian cancer risk. Am J Epidemiol, 2007. 166(8): p. 894-901.
Collaborative Group on Epidemiological Studies of Ovarian, C., et al., Ovarian cancer and oral
contraceptives: collaborative reanalysis of data from 45 epidemiological studies including
23,257 women with ovarian cancer and 87,303 controls. Lancet, 2008. 371(9609): p. 303-14.
Ip, S., et al., A summary of the Agency for Healthcare Research and Quality's evidence report
on breastfeeding in developed countries. Breastfeed Med, 2009. 4 Suppl 1: p. S17-30.
Cibula, D., et al., Tubal ligation and the risk of ovarian cancer: review and meta-analysis.
Hum Reprod Update, 2011. 17(1): p. 55-67.
Gwinn, M.L., et al., Pregnancy, breast feeding, and oral contraceptives and the risk of
epithelial ovarian cancer. J Clin Epidemiol, 1990. 43(6): p. 559-68.
UpToDate. Risk Factors for Ovarian Cancer; Graphic 59585, v3.0. UpToDate [Graphic] 2014
[cited 2014 28th August 2014]; v3.0: Available from:
http://www.uptodate.com/contents/image?imageKey=OBGYN%2F59585&topicKey=PC%2F7
563&rank=1%7E150&source=see_link&search=ovarian+cancer+risk+factors+table&utdPopu
p=true].
Salehi, F., et al., Risk factors for ovarian cancer: an overview with emphasis on hormonal
factors. J Toxicol Environ Health B Crit Rev, 2008. 11(3-4): p. 301-21.
Beral, V., et al., Ovarian cancer and hormone replacement therapy in the Million Women
Study. Lancet, 2007. 369(9574): p. 1703-10.
Lalwani, N., et al., Histologic, molecular, and cytogenetic features of ovarian cancers:
implications for diagnosis and treatment. Radiographics, 2011. 31(3): p. 625-46.
Domchek, S.M., et al., Association of risk-reducing surgery in BRCA1 or BRCA2 mutation
carriers with cancer risk and mortality. JAMA, 2010. 304(9): p. 967-75.
206
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
41.
42.
43.
44.
45.
Simon, H.H., et al., Fate of midbrain dopaminergic neurons controlled by the engrailed genes.
J Neurosci, 2001. 21(9): p. 3126-34.
Blank, M.M., et al., Dietary fat intake and risk of ovarian cancer in the NIH-AARP Diet and
Health Study. Br J Cancer, 2012. 106(3): p. 596-602.
Huncharek, M., J.F. Geschwind, and B. Kupelnick, Perineal application of cosmetic talc and
risk of invasive epithelial ovarian cancer: a meta-analysis of 11,933 subjects from sixteen
observational studies. Anticancer Res, 2003. 23(2C): p. 1955-60.
Gates, M.A., et al., Risk factors for epithelial ovarian cancer by histologic subtype. Am J
Epidemiol, 2010. 171(1): p. 45-53.
Riska, A., et al., Occupation and risk of primary fallopian tube carcinoma in Nordic countries.
Int J Cancer, 2012. 131(1): p. 186-92.
Cramer, D.W. and W.R. Welch, Determinants of ovarian cancer risk. II. Inferences regarding
pathogenesis. J Natl Cancer Inst, 1983. 71(4): p. 717-21.
Fathalla, M.F., Incessant ovulation--a factor in ovarian neoplasia? Lancet, 1971. 2(7716): p.
163.
Cramer, D.W., et al., Determinants of ovarian cancer risk. I. Reproductive experiences and
family history. J Natl Cancer Inst, 1983. 71(4): p. 711-6.
Risch, H.A., Hormonal etiology of epithelial ovarian cancer, with a hypothesis concerning the
role of androgens and progesterone. J Natl Cancer Inst, 1998. 90(23): p. 1774-86.
Piek, J.M., et al., BRCA1/2-related ovarian cancers are of tubal origin: a hypothesis. Gynecol
Oncol, 2003. 90(2): p. 491.
Nik, N.N., et al., Origin and pathogenesis of pelvic (ovarian, tubal, and primary peritoneal)
serous carcinoma. Annu Rev Pathol, 2014. 9: p. 27-45.
Akdeniz, N., et al., Risk of malignancy index for adnexal masses. Eur J Gynaecol Oncol, 2009.
30(2): p. 178-80.
Vang, R., M. Shih Ie, and R.J. Kurman, Ovarian low-grade and high-grade serous carcinoma:
pathogenesis, clinicopathologic and molecular biologic features, and diagnostic problems.
Adv Anat Pathol, 2009. 16(5): p. 267-82.
Romero, I., Minireview: human ovarian cancer: biology, current management, and paths to
personalizing therapy. Endocrinology, 2012. 153(4): p. 1593-602.
Soslow, R.A., Histologic subtypes of ovarian carcinoma: an overview. Int J Gynecol Pathol,
2008. 27(2): p. 161-74.
CRUK. Types of Ovarian Cancer. Ovarian Cancer 2014 16th January 2014 [cited 2014 28th
August 2014]; Available from: http://www.cancerresearchuk.org/cancer-help/type/ovariancancer/about/types-of-ovarian-cancer.
Kurman, R.J. and M. Shih Ie, Pathogenesis of ovarian cancer: lessons from morphology and
molecular biology and their clinical implications. Int J Gynecol Pathol, 2008. 27(2): p. 151-60.
Kuo, K.T., et al., Analysis of DNA copy number alterations in ovarian serous tumors identifies
new molecular genetic changes in low-grade and high-grade carcinomas. Cancer Res, 2009.
69(9): p. 4036-42.
Singer, G., et al., Mutations in BRAF and KRAS characterize the development of low-grade
ovarian serous carcinoma. J Natl Cancer Inst, 2003. 95(6): p. 484-6.
Bonome, T., et al., Expression profiling of serous low malignant potential, low-grade, and
high-grade tumors of the ovary. Cancer Res, 2005. 65(22): p. 10602-12.
Malpica, A., et al., Grading ovarian serous carcinoma using a two-tier system. Am J Surg
Pathol, 2004. 28(4): p. 496-504.
Willner, J., et al., Alternate molecular genetic pathways in ovarian carcinomas of common
histological types. Hum Pathol, 2007. 38(4): p. 607-13.
Madore, J., et al., Characterization of the molecular differences between ovarian
endometrioid carcinoma and ovarian serous carcinoma. J Pathol, 2010. 220(3): p. 392-400.
207
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
Bast, R.C., Jr., B. Hennessy, and G.B. Mills, The biology of ovarian cancer: new opportunities
for translation. Nat Rev Cancer, 2009. 9(6): p. 415-28.
de Graeff, P., et al., Factors influencing p53 expression in ovarian cancer as a biomarker of
clinical outcome in multicentre studies. Br J Cancer, 2006. 95(5): p. 627-33.
Jaaback, K.S., et al., Primary peritoneal carcinoma in a UK cancer center: comparison with
advanced ovarian carcinoma over a 5-year period. Int J Gynecol Cancer, 2006. 16 Suppl 1: p.
123-8.
Barda, G., et al., Comparison between primary peritoneal and epithelial ovarian carcinoma: a
population-based study. Am J Obstet Gynecol, 2004. 190(4): p. 1039-45.
Network., N.C.C. NCCN Clinical Practice Guidelines in Oncology: Ovarian Cancer. 2013 [cited
2014 28th August 2014]; V 1.2013.: Available from:
http://www.nccn.org/professionals/physician_gls/pdf/ovarian.pdf.
Young, R.C., et al., Staging laparotomy in early ovarian cancer. JAMA, 1983. 250(22): p. 30726.
Dembo, A.J., et al., Prognostic factors in patients with stage I epithelial ovarian cancer.
Obstet Gynecol, 1990. 75(2): p. 263-73.
Ahmed, F.Y., et al., Natural history and prognosis of untreated stage I epithelial ovarian
carcinoma. J Clin Oncol, 1996. 14(11): p. 2968-75.
Kolomainen, D.F., et al., Can patients with relapsed, previously untreated, stage I epithelial
ovarian cancer be successfully treated with salvage therapy? J Clin Oncol, 2003. 21(16): p.
3113-8.
Hoskins, W.J., et al., The effect of diameter of largest residual disease on survival after
primary cytoreductive surgery in patients with suboptimal residual epithelial ovarian
carcinoma. Am J Obstet Gynecol, 1994. 170(4): p. 974-9; discussion 979-80.
Bristow, R.E., et al., Survival effect of maximal cytoreductive surgery for advanced ovarian
carcinoma during the platinum era: a meta-analysis. J Clin Oncol, 2002. 20(5): p. 1248-59.
Bookman, M.A., et al., Evaluation of new platinum-based treatment regimens in advancedstage ovarian cancer: a Phase III Trial of the Gynecologic Cancer Intergroup. J Clin Oncol,
2009. 27(9): p. 1419-25.
du Bois, A., et al., A randomized clinical trial of cisplatin/paclitaxel versus
carboplatin/paclitaxel as first-line treatment of ovarian cancer. J Natl Cancer Inst, 2003.
95(17): p. 1320-9.
Neijt, J.P., et al., Exploratory phase III study of paclitaxel and cisplatin versus paclitaxel and
carboplatin in advanced ovarian cancer. J Clin Oncol, 2000. 18(17): p. 3084-92.
Ozols, R.F., et al., Phase III trial of carboplatin and paclitaxel compared with cisplatin and
paclitaxel in patients with optimally resected stage III ovarian cancer: a Gynecologic
Oncology Group study. J Clin Oncol, 2003. 21(17): p. 3194-200.
Vasey, P.A., et al., Phase III randomized trial of docetaxel-carboplatin versus paclitaxelcarboplatin as first-line chemotherapy for ovarian carcinoma. J Natl Cancer Inst, 2004.
96(22): p. 1682-91.
Burger, R.A., et al., Incorporation of bevacizumab in the primary treatment of ovarian cancer.
N Engl J Med, 2011. 365(26): p. 2473-83.
Perren, T.J., et al., A phase 3 trial of bevacizumab in ovarian cancer. N Engl J Med, 2011.
365(26): p. 2484-96.
Yap, T.A., C.P. Carden, and S.B. Kaye, Beyond chemotherapy: targeted therapies in ovarian
cancer. Nat Rev Cancer, 2009. 9(3): p. 167-81.
Kelland, L., The resurgence of platinum-based cancer chemotherapy. Nat Rev Cancer, 2007.
7(8): p. 573-84.
Dasari, S. and P. Bernard Tchounwou, Cisplatin in cancer therapy: Molecular mechanisms of
action. Eur J Pharmacol, 2014. 740C: p. 364-378.
208
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
83.
84.
85.
86.
87.
88.
Davis, A., A.V. Tinker, and M. Friedlander, "Platinum resistant" ovarian cancer: what is it,
who to treat and how to measure benefit? Gynecol Oncol, 2014. 133(3): p. 624-31.
Ferry, K.V., T.C. Hamilton, and S.W. Johnson, Increased nucleotide excision repair in cisplatinresistant ovarian cancer cells: role of ERCC1-XPF. Biochem Pharmacol, 2000. 60(9): p. 130513.
Dabholkar, M., et al., ERCC1 and ERCC2 expression in malignant tissues from ovarian cancer
patients. J Natl Cancer Inst, 1992. 84(19): p. 1512-7.
Kuhlmann, J.D., et al., ERCC1-Positive Circulating Tumor Cells in the Blood of Ovarian Cancer
Patients as a Predictive Biomarker for Platinum Resistance. Clin Chem, 2014.
Gifford, G., et al., The acquisition of hMLH1 methylation in plasma DNA after chemotherapy
predicts poor survival for ovarian cancer patients. Clin Cancer Res, 2004. 10(13): p. 4420-6.
Pujade-Lauraine, E., et al., Bevacizumab combined with chemotherapy for platinum-resistant
recurrent ovarian cancer: The AURELIA open-label randomized phase III trial. J Clin Oncol,
2014. 32(13): p. 1302-8.
Chura, J.C., et al., Bevacizumab plus cyclophosphamide in heavily pretreated patients with
recurrent ovarian cancer. Gynecol Oncol, 2007. 107(2): p. 326-30.
Sanchez-Munoz, A., et al., Bevacizumab plus low-dose metronomic oral cyclophosphamide in
heavily pretreated patients with recurrent ovarian cancer. Oncology, 2010. 79(1-2): p. 98104.
Markman, M., Addition of bevacizumab to weekly paclitaxel significantly improves
progression-free survival in heavily pretreated recurrent epithelial ovarian cancer. Gynecol
Oncol, 2012. 124(1): p. 171; author reply 171-2.
Monk, B.J., et al., Anti-angiopoietin therapy with trebananib for recurrent ovarian cancer
(TRINOVA-1): a randomised, multicentre, double-blind, placebo-controlled phase 3 trial.
Lancet Oncol, 2014. 15(8): p. 799-808.
Naumann, R.W., et al., PRECEDENT: a randomized phase II trial comparing vintafolide
(EC145) and pegylated liposomal doxorubicin (PLD) in combination versus PLD alone in
patients with platinum-resistant ovarian cancer. J Clin Oncol, 2013. 31(35): p. 4400-6.
Sawers, L., et al., Glutathione S-transferase P1 (GSTP1) directly influences platinum drug
chemosensitivity in ovarian tumour cell lines. Br J Cancer, 2014.
Cass, I., et al., Improved survival in women with BRCA-associated ovarian carcinoma. Cancer,
2003. 97(9): p. 2187-95.
Vencken, P.M., et al., Chemosensitivity and outcome of BRCA1- and BRCA2-associated
ovarian cancer patients after first-line chemotherapy compared with sporadic ovarian cancer
patients. Ann Oncol, 2011. 22(6): p. 1346-52.
Konstantinopoulos, P.A., et al., Gene expression profile of BRCAness that correlates with
responsiveness to chemotherapy and with outcome in patients with epithelial ovarian cancer.
J Clin Oncol, 2010. 28(22): p. 3555-61.
Liu, T., et al., TNFAIP8 overexpression is associated with platinum resistance in epithelial
ovarian cancers with optimal cytoreduction. Hum Pathol, 2014. 45(6): p. 1251-7.
Etzioni, R., et al., The case for early detection. Nat Rev Cancer, 2003. 3(4): p. 243-52.
Database, S.E.a.E.R.S. [cited 2014 01.09.14]; Available from: http://seer.cancer.gov.
Institute, N.C. [cited 2014 01.09.14]; Available from: http://www.cancer.gov/dictionary.
Lin, K., et al., Benefits and harms of prostate-specific antigen screening for prostate cancer:
an evidence update for the U.S. Preventive Services Task Force. Ann Intern Med, 2008.
149(3): p. 192-9.
Paik, S., et al., A multigene assay to predict recurrence of tamoxifen-treated, node-negative
breast cancer. N Engl J Med, 2004. 351(27): p. 2817-26.
Allegra, C.J., et al., American Society of Clinical Oncology provisional clinical opinion: testing
for KRAS gene mutations in patients with metastatic colorectal carcinoma to predict
209
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
101.
102.
103.
104.
105.
106.
107.
108.
109.
response to anti-epidermal growth factor receptor monoclonal antibody therapy. J Clin
Oncol, 2009. 27(12): p. 2091-6.
Bang, Y.J., et al., Trastuzumab in combination with chemotherapy versus chemotherapy
alone for treatment of HER2-positive advanced gastric or gastro-oesophageal junction
cancer (ToGA): a phase 3, open-label, randomised controlled trial. Lancet, 2010. 376(9742):
p. 687-97.
Piccart-Gebhart, M.J., et al., Trastuzumab after adjuvant chemotherapy in HER2-positive
breast cancer. N Engl J Med, 2005. 353(16): p. 1659-72.
Romond, E.H., et al., Trastuzumab plus adjuvant chemotherapy for operable HER2-positive
breast cancer. N Engl J Med, 2005. 353(16): p. 1673-84.
Early Breast Cancer Trialists' Collaborative, G., et al., Relevance of breast cancer hormone
receptors and other factors to the efficacy of adjuvant tamoxifen: patient-level meta-analysis
of randomised trials. Lancet, 2011. 378(9793): p. 771-84.
Locker, G.Y., et al., ASCO 2006 update of recommendations for the use of tumor markers in
gastrointestinal cancer. J Clin Oncol, 2006. 24(33): p. 5313-27.
Rustin, G.J., et al., Use of CA-125 in clinical trial evaluation of new therapeutic drugs for
ovarian cancer. Clin Cancer Res, 2004. 10(11): p. 3919-26.
Gilligan, T.D., et al., American Society of Clinical Oncology Clinical Practice Guideline on uses
of serum tumor markers in adult males with germ cell tumors. J Clin Oncol, 2010. 28(20): p.
3388-404.
Harris, L., et al., American Society of Clinical Oncology 2007 update of recommendations for
the use of tumor markers in breast cancer. J Clin Oncol, 2007. 25(33): p. 5287-312.
Kishi, S., et al., Effects of prednisone and genetic polymorphisms on etoposide disposition in
children with acute lymphoblastic leukemia. Blood, 2004. 103(1): p. 67-72.
Lee, W., et al., Cancer pharmacogenomics: powerful tools in cancer chemotherapy and drug
development. Oncologist, 2005. 10(2): p. 104-11.
Henry, N.L. and D.F. Hayes, Cancer biomarkers. Mol Oncol, 2012. 6(2): p. 140-6.
Institute, N.C. Tumor markers. National Cancer Institute 2011 12/07/2011 [cited 2014
06/01/2014]; Available from:
http://www.cancer.gov/cancertopics/factsheet/Detection/tumor-markers.
Gamallo, C., et al., beta-catenin expression pattern in stage I and II ovarian carcinomas :
relationship with beta-catenin gene mutations, clinicopathological features, and clinical
outcome. Am J Pathol, 1999. 155(2): p. 527-36.
Kolasa, I.K., et al., PIK3CA amplification associates with resistance to chemotherapy in
ovarian cancer patients. Cancer Biol Ther, 2009. 8(1): p. 21-6.
Helleman, J., et al., Mismatch repair and treatment resistance in ovarian cancer. BMC
Cancer, 2006. 6: p. 201.
Soussi, T., p53 Antibodies in the sera of patients with various types of cancer: a review.
Cancer Res, 2000. 60(7): p. 1777-88.
Tsai-Turton, M., et al., p53 autoantibodies, cytokine levels and ovarian carcinogenesis.
Gynecol Oncol, 2009. 114(1): p. 12-7.
Lu, D., et al., Comparison of candidate serologic markers for type I and type II ovarian cancer.
Gynecol Oncol, 2011. 122(3): p. 560-6.
Boyd, J., et al., Clinicopathologic features of BRCA-linked and sporadic ovarian cancer. Jama,
2000. 283(17): p. 2260-5.
Tan, D.S., et al., "BRCAness" syndrome in ovarian cancer: a case-control study describing the
clinical features and outcome of patients with epithelial ovarian cancer associated with
BRCA1 and BRCA2 mutations. J Clin Oncol, 2008. 26(34): p. 5530-6.
Fong, P.C., et al., Inhibition of poly(ADP-ribose) polymerase in tumors from BRCA mutation
carriers. N Engl J Med, 2009. 361(2): p. 123-34.
210
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
Mikeska, T., et al., DNA methylation biomarkers in cancer: progress towards clinical
implementation. Expert Rev Mol Diagn, 2012. 12(5): p. 473-87.
Bird, A.P., CpG-rich islands and the function of DNA methylation. Nature, 1986. 321(6067): p.
209-13.
Herman, J.G. and S.B. Baylin, Gene silencing in cancer in association with promoter
hypermethylation. N Engl J Med, 2003. 349(21): p. 2042-54.
Kanai, Y., Genome-wide DNA methylation profiles in precancerous conditions and cancers.
Cancer Sci, 2010. 101(1): p. 36-45.
Radpour, R., et al., Hypermethylation of tumor suppressor genes involved in critical
regulatory pathways for developing a blood-based test in breast cancer. PLoS One, 2011.
6(1): p. e16080.
Topaloglu, O., et al., Detection of promoter hypermethylation of multiple genes in the tumor
and bronchoalveolar lavage of patients with lung cancer. Clin Cancer Res, 2004. 10(7): p.
2284-8.
Fujiwara, K., et al., Identification of epigenetic aberrant promoter methylation in serum DNA
is useful for early detection of lung cancer. Clin Cancer Res, 2005. 11(3): p. 1219-25.
Belinsky, S.A., et al., Gene promoter methylation in plasma and sputum increases with lung
cancer risk. Clin Cancer Res, 2005. 11(18): p. 6505-11.
Palmisano, W.A., et al., Predicting lung cancer by detecting aberrant promoter methylation in
sputum. Cancer Res, 2000. 60(21): p. 5954-8.
Kersting, M., et al., Differential frequencies of p16(INK4a) promoter hypermethylation, p53
mutation, and K-ras mutation in exfoliative material mark the development of lung cancer in
symptomatic chronic smokers. J Clin Oncol, 2000. 18(18): p. 3221-9.
Tanzer, M., et al., Performance of epigenetic markers SEPT9 and ALX4 in plasma for
detection of colorectal precancerous lesions. PLoS One, 2010. 5(2): p. e9061.
Schmidt, B., et al., SHOX2 DNA methylation is a biomarker for the diagnosis of lung cancer
based on bronchial aspirates. BMC Cancer, 2010. 10: p. 600.
Goessl, C., et al., Fluorescent methylation-specific polymerase chain reaction for DNA-based
detection of prostate cancer in bodily fluids. Cancer Res, 2000. 60(21): p. 5941-5.
Sharma, G., et al., Promoter hypermethylation of p16INK4A, p14ARF, CyclinD2 and Slit2 in
serum and tumor DNA from breast cancer patients. Life Sci, 2007. 80(20): p. 1873-81.
Krassenstein, R., et al., Detection of breast cancer in nipple aspirate fluid by CpG island
hypermethylation. Clin Cancer Res, 2004. 10(1 Pt 1): p. 28-32.
Lim, S.L., et al., Promoter hypermethylation of FANCF and outcome in advanced ovarian
cancer. Br J Cancer, 2008. 98(8): p. 1452-6.
Potapova, A., et al., Promoter hypermethylation of the PALB2 susceptibility gene in inherited
and sporadic breast and ovarian cancer. Cancer Res, 2008. 68(4): p. 998-1002.
Fiegl, H., et al., HOXA11 DNA methylation--a novel prognostic biomarker in ovarian cancer.
Int J Cancer, 2008. 123(3): p. 725-9.
Helzlsouer, K.J., et al., Prospective study of serum CA-125 levels as markers of ovarian cancer.
Jama, 1993. 269(9): p. 1123-6.
Einhorn, N., et al., Prospective evaluation of serum CA 125 levels for early detection of
ovarian cancer. Obstet Gynecol, 1992. 80(1): p. 14-8.
Woolas, R.P., et al., Elevation of multiple serum markers in patients with stage I ovarian
cancer. J Natl Cancer Inst, 1993. 85(21): p. 1748-51.
Rockhill, B., Proteomic patterns in serum and identification of ovarian cancer. Lancet, 2002.
360(9327): p. 169; author reply 170-1.
Menon, U., et al., Sensitivity and specificity of multimodal and ultrasound screening for
ovarian cancer, and stage distribution of detected cancers: results of the prevalence screen of
the UK Collaborative Trial of Ovarian Cancer Screening (UKCTOCS). Lancet Oncol, 2009.
10(4): p. 327-40.
211
133.
134.
135.
136.
137.
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
Budiu, R.A., et al., Soluble MUC1 and serum MUC1-specific antibodies are potential
prognostic biomarkers for platinum-resistant ovarian cancer. Cancer Immunol Immunother,
2011. 60(7): p. 975-84.
Xu, F.J., et al., Development of two new monoclonal antibodies reactive to a surface antigen
present on human ovarian epithelial cancer cells. Cancer Res, 1991. 51(15): p. 4012-9.
van Haaften-Day, C., et al., OVX1, macrophage-colony stimulating factor, and CA-125-II as
tumor markers for epithelial ovarian carcinoma: a critical appraisal. Cancer, 2001. 92(11): p.
2837-44.
McIntosh, M.W., et al., Combining CA 125 and SMR serum markers for diagnosis and early
detection of ovarian carcinoma. Gynecol Oncol, 2004. 95(1): p. 9-15.
Drapkin, R., et al., Human epididymis protein 4 (HE4) is a secreted glycoprotein that is
overexpressed by serous and endometrioid ovarian carcinomas. Cancer Res, 2005. 65(6): p.
2162-9.
Moore, R.G., et al., A novel multiple marker bioassay utilizing HE4 and CA125 for the
prediction of ovarian cancer in patients with a pelvic mass. Gynecol Oncol, 2009. 112(1): p.
40-6.
Huhtinen, K., et al., Serum HE4 concentration differentiates malignant ovarian tumours from
ovarian endometriotic cysts. Br J Cancer, 2009. 100(8): p. 1315-9.
Moore, R.G., et al., The use of multiple novel tumor biomarkers for the detection of ovarian
carcinoma in patients with a pelvic mass. Gynecol Oncol, 2008. 108(2): p. 402-8.
Rosen, D.G., et al., Potential markers that complement expression of CA125 in epithelial
ovarian cancer. Gynecol Oncol, 2005. 99(2): p. 267-77.
Healy, D.L., et al., Elevated serum inhibin concentrations in postmenopausal women with
ovarian tumors. N Engl J Med, 1993. 329(21): p. 1539-42.
Tsigkou, A., et al., Total inhibin is a potential serum marker for epithelial ovarian cancer. J
Clin Endocrinol Metab, 2007. 92(7): p. 2526-31.
Hogdall, E.V., et al., Protein expression levels of carcinoembryonic antigen (CEA) in Danish
ovarian cancer patients: from the Danish 'MALOVA'ovarian cancer study. Pathology, 2008.
40(5): p. 487-92.
Tholander, B., et al., Pretreatment serum levels of CA-125, carcinoembryonic antigen, tissue
polypeptide antigen, and placental alkaline phosphatase in patients with ovarian carcinoma:
influence of histological type, grade of differentiation, and clinical stage of disease. Gynecol
Oncol, 1990. 39(1): p. 26-33.
Visintin, I., et al., Diagnostic markers for early detection of ovarian cancer. Clin Cancer Res,
2008. 14(4): p. 1065-72.
Cooper, B.C., et al., Preoperative serum vascular endothelial growth factor levels:
significance in ovarian cancer. Clin Cancer Res, 2002. 8(10): p. 3193-7.
Paliouras, M., C. Borgono, and E.P. Diamandis, Human tissue kallikreins: the cancer
biomarker family. Cancer Lett, 2007. 249(1): p. 61-79.
Diamandis, E.P., et al., Human kallikrein 6 (hK6): a new potential serum biomarker for
diagnosis and prognosis of ovarian carcinoma. J Clin Oncol, 2003. 21(6): p. 1035-43.
Luo, L.Y., et al., Prognostic value of human kallikrein 10 expression in epithelial ovarian
carcinoma. Clin Cancer Res, 2001. 7(8): p. 2372-9.
Obiezu, C.V., et al., Higher human kallikrein gene 4 (KLK4) expression indicates poor
prognosis of ovarian cancer patients. Clin Cancer Res, 2001. 7(8): p. 2380-6.
Kim, H., et al., Human kallikrein gene 5 (KLK5) expression is an indicator of poor prognosis in
ovarian cancer. Br J Cancer, 2001. 84(5): p. 643-50.
Hoffman, B.R., et al., Immunofluorometric quantitation and histochemical localisation of
kallikrein 6 protein in ovarian cancer tissue: a new independent unfavourable prognostic
biomarker. Br J Cancer, 2002. 87(7): p. 763-71.
212
154.
155.
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
172.
173.
174.
Luo, L.Y., et al., The serum concentration of human kallikrein 10 represents a novel biomarker
for ovarian cancer diagnosis and prognosis. Cancer Res, 2003. 63(4): p. 807-11.
Yousef, G.M., et al., Prognostic value of the human kallikrein gene 15 expression in ovarian
cancer. J Clin Oncol, 2003. 21(16): p. 3119-26.
Xi, Z., et al., Kallikrein 4 is associated with paclitaxel resistance in ovarian cancer. Gynecol
Oncol, 2004. 94(1): p. 80-5.
Magklara, A., et al., The human KLK8 (neuropsin/ovasin) gene: identification of two novel
splice variants and its prognostic value in ovarian cancer. Clin Cancer Res, 2001. 7(4): p. 80611.
Shigemasa, K., et al., Human kallikrein 8 (hK8/TADG-14) expression is associated with an
early clinical stage and favorable prognosis in ovarian cancer. Oncol Rep, 2004. 11(6): p.
1153-9.
Kishi, T., et al., Human kallikrein 8, a novel biomarker for ovarian carcinoma. Cancer Res,
2003. 63(11): p. 2771-4.
Yousef, G.M., et al., Quantitative expression of the human kallikrein gene 9 (KLK9) in ovarian
cancer: a new independent and favorable prognostic marker. Cancer Res, 2001. 61(21): p.
7811-8.
Borgono, C.A., et al., Favorable prognostic value of tissue human kallikrein 11 (hK11) in
patients with ovarian carcinoma. Int J Cancer, 2003. 106(4): p. 605-10.
Rife, C.C., G.M. Farrow, and D.C. Utz, Urine cytology of transitional cell neoplasms. Urol Clin
North Am, 1979. 6(3): p. 599-612.
Cajulis, R.S., et al., Cytology, flow cytometry, image analysis, and interphase cytogenetics by
fluorescence in situ hybridization in the diagnosis of transitional cell carcinoma in bladder
washes: a comparative study. Diagn Cytopathol, 1995. 13(3): p. 214-23; discussion 224.
Budman, L.I., W. Kassouf, and J.R. Steinberg, Biomarkers for detection and surveillance of
bladder cancer. Can Urol Assoc J, 2008. 2(3): p. 212-21.
Coticchia, C.M., et al., Urinary MMP-2 and MMP-9 predict the presence of ovarian cancer in
women with normal CA125 levels. Gynecol Oncol, 2011. 123(2): p. 295-300.
Badgwell, D., et al., Urinary mesothelin provides greater sensitivity for early stage ovarian
cancer than serum mesothelin, urinary hCG free beta subunit and urinary hCG beta core
fragment. Gynecol Oncol, 2007. 106(3): p. 490-7.
Chen, Y.L., et al., Interferon-gamma in ascites could be a predictive biomarker of outcome in
ovarian carcinoma. Gynecol Oncol, 2013. 131(1): p. 63-8.
Huang, H., et al., Screening and identification of biomarkers in ascites related to intrinsic
chemoresistance of serous epithelial ovarian cancers. PLoS One, 2012. 7(12): p. e51256.
Sahin, U., et al., Human neoplasms elicit multiple specific immune responses in the
autologous host. Proc Natl Acad Sci U S A, 1995. 92(25): p. 11810-3.
Kim, J.H., et al., Identification of epithelial cell adhesion molecule autoantibody in patients
with ovarian cancer. Clin Cancer Res, 2003. 9(13): p. 4782-91.
Luo, L.Y., et al., Identification of heat shock protein 90 and other proteins as tumour antigens
by serological screening of an ovarian carcinoma expression library. Br J Cancer, 2002. 87(3):
p. 339-43.
Korneeva, I., et al., Serum antibodies to the 27-kd heat shock protein in women with
gynecologic cancers. Am J Obstet Gynecol, 2000. 183(1): p. 18-21.
Naora, H., et al., Aberrant expression of homeobox gene HOXA7 is associated with mullerianlike differentiation of epithelial ovarian tumors and the generation of a specific autologous
antibody response. Proc Natl Acad Sci U S A, 2001. 98(26): p. 15209-14.
Disis, M.L., et al., Pre-existent immunity to the HER-2/neu oncogenic protein in patients with
HER-2/neu overexpressing breast and ovarian cancer. Breast Cancer Res Treat, 2000. 62(3):
p. 245-52.
213
175.
176.
177.
178.
179.
180.
181.
182.
183.
184.
185.
186.
187.
188.
189.
190.
191.
192.
193.
194.
195.
196.
197.
198.
199.
Gourevitch, M.M., et al., Polymorphic epithelial mucin (MUC-1)-containing circulating
immune complexes in carcinoma patients. Br J Cancer, 1995. 72(4): p. 934-8.
Lewis, E.B., A gene complex controlling segmentation in Drosophila. Nature, 1978.
276(5688): p. 565-70.
Boersma, C.J., et al., Homeobox proteins as signal transduction intermediates in regulation of
NCAM expression by recombinant human bone morphogenetic protein-2 in osteoblast-like
cells. Mol Cell Biol Res Commun, 1999. 1(2): p. 117-24.
Shen, W.F., et al., The HOX homeodomain proteins block CBP histone acetyltransferase
activity. Mol Cell Biol, 2001. 21(21): p. 7509-22.
Stein, S., et al., Checklist: vertebrate homeobox genes. Mech Dev, 1996. 55(1): p. 91-108.
Manak, J.R. and M.P. Scott, A class act: conservation of homeodomain protein functions. Dev
Suppl, 1994: p. 61-77.
Pattyn, A., C. Goridis, and J.F. Brunet, Specification of the central noradrenergic phenotype
by the homeobox gene Phox2b. Mol Cell Neurosci, 2000. 15(3): p. 235-43.
Magli, M.C., et al., Coordinate regulation of HOX genes in human hematopoietic cells. Proc
Natl Acad Sci U S A, 1991. 88(14): p. 6348-52.
Srebrow, A., et al., Expression of Hoxa-1 and Hoxb-7 is regulated by extracellular matrixdependent signals in mammary epithelial cells. J Cell Biochem, 1998. 69(4): p. 377-91.
Abate-Shen, C., Deregulated homeobox gene expression in cancer: cause or consequence?
Nat Rev Cancer, 2002. 2(10): p. 777-85.
Shah, N. and S. Sukumar, The Hox genes and their roles in oncogenesis. Nat Rev Cancer,
2010. 10(5): p. 361-71.
Zacchetti, G., D. Duboule, and J. Zakany, Hox gene function in vertebrate gut morphogenesis:
the case of the caecum. Development, 2007. 134(22): p. 3967-73.
Cardoso, W.V., Transcription factors and pattern formation in the developing lung. Am J
Physiol, 1995. 269(4 Pt 1): p. L429-42.
Simpson, J.L., Genetics of the female reproductive ducts. Am J Med Genet, 1999. 89(4): p.
224-39.
Wellik, D.M., Hox genes and vertebrate axial pattern. Curr Top Dev Biol, 2009. 88: p. 257-78.
Myers, C., A. Charboneau, and N. Boudreau, Homeobox B3 promotes capillary
morphogenesis and angiogenesis. J Cell Biol, 2000. 148(2): p. 343-51.
Huang, L., et al., Posterior Hox gene expression and differential androgen regulation in the
developing and adult rat prostate lobes. Endocrinology, 2007. 148(3): p. 1235-45.
Sauvageau, G., et al., Differential expression of homeobox genes in functionally distinct
CD34+ subpopulations of human bone marrow cells. Proc Natl Acad Sci U S A, 1994. 91(25):
p. 12223-7.
Lawrence, H.J., et al., Expression of HOX C homeobox genes in lymphoid cells. Cell Growth
Differ, 1993. 4(8): p. 665-9.
Taylor, H.S., G.B. Vanden Heuvel, and P. Igarashi, A conserved Hox axis in the mouse and
human female reproductive system: late establishment and persistent adult expression of the
Hoxa cluster genes. Biol Reprod, 1997. 57(6): p. 1338-45.
Taylor, H.S., et al., HOXA10 is expressed in response to sex steroids at the time of
implantation in the human endometrium. J Clin Invest, 1998. 101(7): p. 1379-84.
Taylor, H.S., et al., Sex steroids mediate HOXA11 expression in the human peri-implantation
endometrium. J Clin Endocrinol Metab, 1999. 84(3): p. 1129-35.
Muragaki, Y., et al., Altered growth and branching patterns in synpolydactyly caused by
mutations in HOXD13. Science, 1996. 272(5261): p. 548-51.
Akarsu, A.N., et al., A large Turkish kindred with syndactyly type II (synpolydactyly). 2.
Homozygous phenotype? J Med Genet, 1995. 32(6): p. 435-41.
Goodman, F., et al., Deletions in HOXD13 segregate with an identical, novel foot
malformation in two unrelated families. Am J Hum Genet, 1998. 63(4): p. 992-1000.
214
200.
201.
202.
203.
204.
205.
206.
207.
208.
209.
210.
211.
212.
213.
214.
215.
216.
217.
218.
219.
220.
221.
Thompson, A.A. and L.T. Nguyen, Amegakaryocytic thrombocytopenia and radio-ulnar
synostosis are associated with HOXA11 mutation. Nat Genet, 2000. 26(4): p. 397-8.
Mortlock, D.P. and J.W. Innis, Mutation of HOXA13 in hand-foot-genital syndrome. Nat
Genet, 1997. 15(2): p. 179-80.
Goodman, F.R., et al., Novel HOXA13 mutations and the phenotypic spectrum of hand-footgenital syndrome. Am J Hum Genet, 2000. 67(1): p. 197-202.
Nakamura, T., et al., Fusion of the nucleoporin gene NUP98 to HOXA9 by the chromosome
translocation t(7;11)(p15;p15) in human myeloid leukaemia. Nat Genet, 1996. 12(2): p. 1548.
Borrow, J., et al., The t(7;11)(p15;p15) translocation in acute myeloid leukaemia fuses the
genes for nucleoporin NUP98 and class I homeoprotein HOXA9. Nat Genet, 1996. 12(2): p.
159-67.
Faber, J., et al., HOXA9 is required for survival in human MLL-rearranged acute leukemias.
Blood, 2009. 113(11): p. 2375-85.
Golub, T.R., et al., Molecular classification of cancer: class discovery and class prediction by
gene expression monitoring. Science, 1999. 286(5439): p. 531-7.
Bijl, J.J., et al., HOXC4, HOXC5, and HOXC6 expression in non-Hodgkin's lymphoma:
preferential expression of the HOXC5 gene in primary cutaneous anaplastic T-cell and orogastrointestinal tract mucosa-associated B-cell lymphomas. Blood, 1997. 90(10): p. 4116-25.
Redline, R.W., et al., Expression of AbdB-type homeobox genes in human tumors. Lab Invest,
1994. 71(5): p. 663-70.
Cillo, C., et al., Expression and structure of hox genes in wilms-tumor. Int J Oncol, 1995. 7(5):
p. 1145-50.
Shears, L., et al., Disrupting the interaction between HOX and PBX causes necrotic and
apoptotic cell death in the renal cancer lines CaKi-2 and 769-P. J Urol, 2008. 180(5): p. 2196201.
De Vita, G., et al., Expression of homeobox-containing genes in primary and metastatic
colorectal cancer. Eur J Cancer, 1993. 29A(6): p. 887-93.
Vider, B.Z., et al., Human colorectal carcinogenesis is associated with deregulation of
homeobox gene expression. Biochem Biophys Res Commun, 1997. 232(3): p. 742-8.
Rauch, T., et al., Homeobox gene methylation in lung cancer studied by genome-wide
analysis with a microarray-based methylated CpG island recovery assay. Proc Natl Acad Sci U
S A, 2007. 104(13): p. 5527-32.
Abe, M., et al., Disordered expression of HOX genes in human non-small cell lung cancer.
Oncol Rep, 2006. 15(4): p. 797-802.
Plowright, L., et al., HOX transcription factors are potential therapeutic targets in non-smallcell lung cancer (targeting HOX genes in lung cancer). Br J Cancer, 2009. 100(3): p. 470-5.
Raman, V., et al., Compromised HOXA5 function can limit p53 expression in human breast
tumours. Nature, 2000. 405(6789): p. 974-8.
Chen, H., S. Chung, and S. Sukumar, HOXA5-induced apoptosis in breast cancer cells is
mediated by caspases 2 and 8. Mol Cell Biol, 2004. 24(2): p. 924-35.
Chu, M.C., F.B. Selam, and H.S. Taylor, HOXA10 regulates p53 expression and matrigel
invasion in human breast cancer cells. Cancer Biol Ther, 2004. 3(6): p. 568-72.
Ma, X.J., et al., A two-gene expression ratio predicts clinical outcome in breast cancer
patients treated with tamoxifen. Cancer Cell, 2004. 5(6): p. 607-16.
Wu, X., et al., HOXB7, a homeodomain protein, is overexpressed in breast cancer and confers
epithelial-mesenchymal transition. Cancer Res, 2006. 66(19): p. 9527-34.
Chen, H., et al., HOXA5 acts directly downstream of retinoic acid receptor beta and
contributes to retinoic acid-induced apoptosis and growth inhibition. Cancer Res, 2007.
67(17): p. 8007-13.
215
222.
223.
224.
225.
226.
227.
228.
229.
230.
231.
232.
233.
234.
235.
236.
237.
238.
239.
240.
241.
242.
243.
244.
245.
Rubin, E., et al., A role for the HOXB7 homeodomain protein in DNA repair. Cancer Res, 2007.
67(4): p. 1527-35.
Chen, H., et al., Hoxb7 inhibits transgenic HER-2/neu-induced mouse mammary tumor onset
but promotes progression and lung metastasis. Cancer Res, 2008. 68(10): p. 3637-44.
Wang, Z., et al., The prognostic biomarkers HOXB13, IL17BR, and CHDH are regulated by
estrogen in breast cancer. Clin Cancer Res, 2007. 13(21): p. 6327-34.
Jerevall, P.L., et al., Exploring the two-gene ratio in breast cancer--independent roles for
HOXB13 and IL17BR in prediction of clinical outcome. Breast Cancer Res Treat, 2008. 107(2):
p. 225-34.
Sciavolino, P.J. and C. Abate-Shen, Molecular biology of prostate development and prostate
cancer. Ann Med, 1998. 30(4): p. 357-68.
Waltregny, D., et al., Overexpression of the homeobox gene HOXC8 in human prostate cancer
correlates with loss of tumor differentiation. Prostate, 2002. 50(3): p. 162-9.
Jung, C., et al., HOXB13 homeodomain protein suppresses the growth of prostate cancer cells
by the negative regulation of T-cell factor 4. Cancer Res, 2004. 64(9): p. 3046-51.
Jung, C., et al., HOXB13 induces growth suppression of prostate cancer cells as a repressor of
hormone-activated androgen receptor signaling. Cancer Res, 2004. 64(24): p. 9185-92.
Care, A., et al., HOXB7 constitutively activates basic fibroblast growth factor in melanomas.
Mol Cell Biol, 1996. 16(9): p. 4842-51.
Cillo, C., et al., Differential patterns of HOX gene expression are associated with specific
integrin and ICAM profiles in clonal populations isolated from a single human melanoma
metastasis. Int J Cancer, 1996. 66(5): p. 692-7.
Svingen, T. and K.F. Tonissen, Altered HOX gene expression in human skin and breast cancer
cells. Cancer Biol Ther, 2003. 2(5): p. 518-23.
Maeda, K., et al., Altered expressions of HOX genes in human cutaneous malignant
melanoma. Int J Cancer, 2005. 114(3): p. 436-41.
Morgan, R., et al., Antagonism of HOX/PBX dimer formation blocks the in vivo proliferation of
melanoma. Cancer Res, 2007. 67(12): p. 5806-13.
Osborne, J., et al., Expression of HOXD10 gene in normal endometrium and endometrial
adenocarcinoma. J Soc Gynecol Investig, 1998. 5(5): p. 277-80.
Cheng, W., et al., Lineage infidelity of epithelial ovarian cancers is controlled by HOX genes
that specify regional identity in the reproductive tract. Nat Med, 2005. 11(5): p. 531-7.
Yamashita, T., et al., Suppression of invasive characteristics by antisense introduction of
overexpressed HOX genes in ovarian cancer cells. Int J Oncol, 2006. 28(4): p. 931-8.
Miao, J., et al., HOXB13 promotes ovarian cancer progression. Proc Natl Acad Sci U S A, 2007.
104(43): p. 17093-8.
Morgan, R., et al., Targeting HOX and PBX transcription factors in ovarian cancer. BMC
Cancer, 2010. 10: p. 89.
Ota, T., et al., Expression and function of HOXA genes in normal and neoplastic ovarian
epithelial cells. Differentiation, 2009. 77(2): p. 162-71.
Brunelli, S., et al., Germline mutations in the homeobox gene EMX2 in patients with severe
schizencephaly. Nat Genet, 1996. 12(1): p. 94-6.
Granata, T., et al., Familial schizencephaly associated with EMX2 mutation. Neurology, 1997.
48(5): p. 1403-6.
Zhao, Y. and H. Westphal, Homeobox genes and human genetic disorders. Curr Mol Med,
2002. 2(1): p. 13-23.
Davidson, D., The function and evolution of Msx genes: pointers and paradoxes. Trends
Genet, 1995. 11(10): p. 405-11.
Gremel, G., et al., Functional and prognostic relevance of the homeobox protein MSX2 in
malignant melanoma. Br J Cancer, 2011. 105(4): p. 565-74.
216
246.
247.
248.
249.
250.
251.
252.
253.
254.
255.
256.
257.
258.
259.
260.
261.
262.
263.
264.
265.
266.
267.
Stuart, E.T., C. Kioussi, and P. Gruss, Mammalian Pax genes. Annu Rev Genet, 1994. 28: p.
219-36.
Buttiglieri, S., et al., Role of Pax2 in apoptosis resistance and proinvasive phenotype of
Kaposi's sarcoma cells. J Biol Chem, 2004. 279(6): p. 4136-43.
Dressler, G.R. and E.C. Douglass, Pax-2 is a DNA-binding protein expressed in embryonic
kidney and Wilms tumor. Proc Natl Acad Sci U S A, 1992. 89(4): p. 1179-83.
Poleev, A., et al., PAX8, a human paired box gene: isolation and expression in developing
thyroid, kidney and Wilms' tumors. Development, 1992. 116(3): p. 611-23.
Urbanek, P., et al., Complete block of early B cell differentiation and altered patterning of the
posterior midbrain in mice lacking Pax5/BSAP. Cell, 1994. 79(5): p. 901-12.
Stoykova, A. and P. Gruss, Roles of Pax-genes in developing and adult brain as suggested by
expression patterns. J Neurosci, 1994. 14(3 Pt 2): p. 1395-412.
Glaser, T., et al., PAX6 gene dosage effect in a family with congenital cataracts, aniridia,
anophthalmia and central nervous system defects. Nat Genet, 1994. 7(4): p. 463-71.
Stuart, E.T., et al., PAX5 expression correlates with increasing malignancy in human
astrocytomas. Clin Cancer Res, 1995. 1(2): p. 207-14.
Galili, N., et al., Fusion of a fork head domain gene to PAX3 in the solid tumour alveolar
rhabdomyosarcoma. Nat Genet, 1993. 5(3): p. 230-5.
Prosser, J. and V. van Heyningen, PAX6 mutations reviewed. Hum Mutat, 1998. 11(2): p. 93108.
Gehring, W.J. and K. Ikeo, Pax 6: mastering eye morphogenesis and eye evolution. Trends
Genet, 1999. 15(9): p. 371-7.
Gibson, W., et al., Inhibition of PAX2 expression results in alternate cell death pathways in
prostate cancer cells differing in p53 status. Cancer Lett, 2007. 248(2): p. 251-61.
De Vita, G., et al., Expression of the RET/PTC1 oncogene impairs the activity of TTF-1 and Pax8 thyroid transcription factors. Cell Growth Differ, 1998. 9(1): p. 97-103.
Wedeen, C.J. and D.A. Weisblat, Segmental expression of an engrailed-class gene during
early development and neurogenesis in an annelid. Development, 1991. 113(3): p. 805-14.
Wanninger, A. and G. Haszprunar, The expression of an engrailed protein during embryonic
shell formation of the tusk-shell, Antalis entalis (Mollusca, Scaphopoda). Evol Dev, 2001. 3(5):
p. 312-21.
Fjose, A., W.J. McGinnis, and W.J. Gehring, Isolation of a homoeo box-containing gene from
the engrailed region of Drosophila and the spatial distribution of its transcripts. Nature,
1985. 313(6000): p. 284-9.
Lowe, C.J. and G.A. Wray, Radical alterations in the roles of homeobox genes during
echinoderm evolution. Nature, 1997. 389(6652): p. 718-21.
Holland, L.Z., et al., Sequence and embryonic expression of the amphioxus engrailed gene
(AmphiEn): the metameric pattern of transcription resembles that of its segment-polarity
homolog in Drosophila. Development, 1997. 124(9): p. 1723-32.
Joyner, A.L., et al., Expression during embryogenesis of a mouse gene with sequence
homology to the Drosophila engrailed gene. Cell, 1985. 43(1): p. 29-37.
Logan, C., et al., Cloning and sequence comparison of the mouse, human, and chicken
engrailed genes reveal potential functional domains and regulatory regions. Dev Genet,
1992. 13(5): p. 345-58.
Tolkunova, E.N., et al., Two distinct types of repression domain in engrailed: one interacts
with the groucho corepressor and is preferentially active on integrated target genes. Mol Cell
Biol, 1998. 18(5): p. 2804-14.
Peltenburg, L.T. and C. Murre, Specific residues in the Pbx homeodomain differentially
modulate the DNA-binding activity of Hox and Engrailed proteins. Development, 1997.
124(5): p. 1089-98.
217
268.
269.
270.
271.
272.
273.
274.
275.
276.
277.
278.
279.
280.
281.
282.
283.
284.
285.
286.
287.
288.
289.
290.
van Dijk, M.A. and C. Murre, Extradenticle raises the DNA binding specificity of homeotic
selector gene products. Cell, 1994. 78(4): p. 617-24.
Nedelec, S., et al., Emx2 homeodomain transcription factor interacts with eukaryotic
translation initiation factor 4E (eIF4E) in the axons of olfactory sensory neurons. Proc Natl
Acad Sci U S A, 2004. 101(29): p. 10815-20.
Morgan, R., Engrailed: complexity and economy of a multi-functional transcription factor.
FEBS Lett, 2006. 580(11): p. 2531-3.
Joliot, A., et al., Identification of a signal sequence necessary for the unconventional
secretion of Engrailed homeoprotein. Curr Biol, 1998. 8(15): p. 856-63.
Joliot, A., et al., Association of Engrailed homeoproteins with vesicles presenting caveolaelike properties. Development, 1997. 124(10): p. 1865-75.
Maizel, A., et al., A short region of its homeodomain is necessary for engrailed nuclear export
and secretion. Development, 1999. 126(14): p. 3183-90.
Derossi, D., et al., The third helix of the Antennapedia homeodomain translocates through
biological membranes. J Biol Chem, 1994. 269(14): p. 10444-50.
Chatelin, L., et al., Transcription factor hoxa-5 is taken up by cells in culture and conveyed to
their nuclei. Mech Dev, 1996. 55(2): p. 111-7.
Derossi, D., et al., Cell internalization of the third helix of the Antennapedia homeodomain is
receptor-independent. J Biol Chem, 1996. 271(30): p. 18188-93.
Garcia-Bellido, A. and P. Santamaria, Developmental analysis of the wing disc in the mutant
engrailed of Drosophila melanogaster. Genetics, 1972. 72(1): p. 87-104.
Hanks, M., et al., Rescue of the En-1 mutant phenotype by replacement of En-1 with En-2.
Science, 1995. 269(5224): p. 679-82.
Hanks, M.C., et al., Drosophila engrailed can substitute for mouse Engrailed1 function in midhindbrain, but not limb development. Development, 1998. 125(22): p. 4521-30.
Martinez, S. and R.M. Alvarado-Mallart, Expression of the homeobox Chick-en gene in
chick/quail chimeras with inverted mes-metencephalic grafts. Dev Biol, 1990. 139(2): p. 4326.
Martinez, S., M. Wassef, and R.M. Alvarado-Mallart, Induction of a mesencephalic phenotype
in the 2-day-old chick prosencephalon is preceded by the early expression of the homeobox
gene en. Neuron, 1991. 6(6): p. 971-81.
Alberi, L., P. Sgado, and H.H. Simon, Engrailed genes are cell-autonomously required to
prevent apoptosis in mesencephalic dopaminergic neurons. Development, 2004. 131(13): p.
3229-36.
Wilson, S.L., et al., Spatially restricted and developmentally dynamic expression of engrailed
genes in multiple cerebellar cell types. Cerebellum, 2011. 10(3): p. 356-72.
Logan, C., et al., Rostral optic tectum acquires caudal characteristics following ectopic
engrailed expression. Curr Biol, 1996. 6(8): p. 1006-14.
Brunet, I., et al., The transcription factor Engrailed-2 guides retinal axons. Nature, 2005.
438(7064): p. 94-8.
Zec, N., et al., Expression of the homeobox-containing genes EN1 and EN2 in human fetal
midgestational medulla and cerebellum. J Neuropathol Exp Neurol, 1997. 56(3): p. 236-42.
Sarnat, H.B., et al., Agenesis of the mesencephalon and metencephalon with cerebellar
hypoplasia: putative mutation in the EN2 gene--report of 2 cases in early infancy. Pediatr Dev
Pathol, 2002. 5(1): p. 54-68.
Millhorn, D.E. and F.L. Eldridge, Role of ventrolateral medulla in regulation of respiratory and
cardiovascular systems. J Appl Physiol, 1986. 61(4): p. 1249-63.
Lavezzi, A.M., et al., Involvement of the EN-2 gene in normal and abnormal development of
the human arcuate nucleus. Int J Exp Pathol, 2005. 86(1): p. 25-31.
Sillitoe, R.V., et al., Engrailed homeobox genes determine the organization of Purkinje cell
sagittal stripe gene expression in the adult cerebellum. J Neurosci, 2008. 28(47): p. 12150-62.
218
291.
292.
293.
294.
295.
296.
297.
298.
299.
300.
301.
302.
303.
304.
305.
306.
307.
308.
309.
310.
311.
312.
313.
Jankowski, J., et al., Engrailed-2 negatively regulates the onset of perinatal Purkinje cell
differentiation. J Comp Neurol, 2004. 472(1): p. 87-99.
Rissling, I., et al., Haplotype analysis of the engrailed-2 gene in young-onset Parkinson's
disease. Neurodegener Dis, 2009. 6(3): p. 102-5.
Petit, E., et al., Association study with two markers of a human homeogene in infantile
autism. J Med Genet, 1995. 32(4): p. 269-74.
Baader, S.L., et al., Ectopic overexpression of engrailed-2 in cerebellar Purkinje cells causes
restricted cell loss and retarded external germinal layer development at lobule junctions. J
Neurosci, 1998. 18(5): p. 1763-73.
Gharani, N., et al., Association of the homeobox transcription factor, ENGRAILED 2, 3, with
autism spectrum disorder. Mol Psychiatry, 2004. 9(5): p. 474-84.
Benayed, R., et al., Support for the homeobox transcription factor gene ENGRAILED 2 as an
autism spectrum disorder susceptibility locus. Am J Hum Genet, 2005. 77(5): p. 851-68.
Kuemerle, B., et al., The mouse Engrailed genes: a window into autism. Behav Brain Res,
2007. 176(1): p. 121-32.
Wang, L., et al., Association of the ENGRAILED 2 (EN2) gene with autism in Chinese Han
population. Am J Med Genet B Neuropsychiatr Genet, 2008. 147B(4): p. 434-8.
Sen, B., et al., Family-based studies indicate association of Engrailed 2 gene with autism in an
Indian population. Genes Brain Behav, 2010. 9(2): p. 248-55.
Yang, P., et al., Intronic single nucleotide polymorphisms of engrailed homeobox 2 modulate
the disease vulnerability of autism in a han chinese population. Neuropsychobiology, 2010.
62(2): p. 104-15.
Martin, N.L., et al., EN2 is a candidate oncogene in human breast cancer. Oncogene, 2005.
24(46): p. 6890-901.
Bose, S.K., R.S. Bullard, and C.D. Donald, Oncogenic role of engrailed-2 (en-2) in prostate
cancer cell growth and survival. Transl Oncogenomics, 2008. 3: p. 37-43.
Morgan, R., et al., Engrailed-2 (EN2): a tumor specific urinary biomarker for the early
diagnosis of prostate cancer. Clin Cancer Res, 2011. 17(5): p. 1090-8.
Morgan, R., et al., Expression of Engrailed-2 (EN2) protein in bladder cancer and its potential
utility as a urinary diagnostic biomarker. Eur J Cancer, 2013. 49(9): p. 2214-22.
Michael, A., Riley, C., Boakee, S., Denyer, M., Pandha, HS., Annels, NE., EN2: A candidate
antigen for the development of targeted therapies in ovarian cancer. J Clin Oncol, 2011.
29(suppl; abstr e15528).
Karpinski, P., et al., The CpG island methylator phenotype correlates with long-range
epigenetic silencing in colorectal cancer. Mol Cancer Res, 2008. 6(4): p. 585-91.
Mayor, R., et al., Long-range epigenetic silencing at 2q14.2 affects most human colorectal
cancers and may have application as a non-invasive biomarker of disease. Br J Cancer, 2009.
100(10): p. 1534-9.
Wu, X., et al., CpG island hypermethylation in human astrocytomas. Cancer Res, 2010. 70(7):
p. 2718-27.
Devaney, J., et al., Epigenetic deregulation across chromosome 2q14.2 differentiates normal
from prostate cancer and provides a regional panel of novel DNA methylation cancer
biomarkers. Cancer Epidemiol Biomarkers Prev, 2011. 20(1): p. 148-59.
Bennett, L.B., et al., DNA hypermethylation accompanied by transcriptional repression in
follicular lymphoma. Genes Chromosomes Cancer, 2009. 48(9): p. 828-41.
Bell, D., et al., Developmental transcription factor EN1-a novel biomarker in human salivary
gland adenoid cystic carcinoma. Cancer, 2011. 118(5): p. 1288-92.
Bell, A., et al., CpG island methylation profiling in human salivary gland adenoid cystic
carcinoma. Cancer, 2011. 117(13): p. 2898-909.
Bell, D., et al., Developmental transcription factor EN1--a novel biomarker in human salivary
gland adenoid cystic carcinoma. Cancer, 2012. 118(5): p. 1288-92.
219
314.
315.
316.
317.
318.
319.
320.
321.
322.
323.
324.
325.
326.
327.
328.
329.
330.
331.
332.
333.
334.
335.
Pandha, H., et al., Urinary engrailed-2 (EN2) levels predict tumour volume in men undergoing
radical prostatectomy for prostate cancer. BJU Int, 2012. 110(6 Pt B): p. E287-92.
McGrath, S.E., et al., Engrailed homeobox transcription factors as potential markers and
targets in cancer. FEBS Lett, 2013. 587(6): p. 549-54.
Chou, W.C., et al., Acute myeloid leukemia bearing t(7;11)(p15;p15) is a distinct cytogenetic
entity with poor outcome and a distinct mutation profile: comparative analysis of 493 adult
patients. Leukemia, 2009. 23(7): p. 1303-10.
Goetz, M.P., et al., A two-gene expression ratio of homeobox 13 and interleukin-17B receptor
for prediction of recurrence and survival in women receiving adjuvant tamoxifen. Clin Cancer
Res, 2006. 12(7 Pt 1): p. 2080-7.
Jansen, M.P., et al., HOXB13-to-IL17BR expression ratio is related with tumor aggressiveness
and response to tamoxifen of recurrent breast cancer: a retrospective study. J Clin Oncol,
2007. 25(6): p. 662-8.
Thompson, I.M., et al., Assessing prostate cancer risk: results from the Prostate Cancer
Prevention Trial. J Natl Cancer Inst, 2006. 98(8): p. 529-34.
Pandha, H.J., S.; Sooriakumaran, P.; Bott, S.; Montgomery, B.; Hutton, A.; Eden, C.; Langley,
S.E.; Morgan, R., Correlation of Urinary Engrailed-2 Levels to Tumour Volume and
Pathological Stage in Men Undergoing Radical Prostatectomy. Journal of Cancer Therapy,
2013. 4: p. 726-733.
d'Adhemar, C.J., et al., The MyD88+ phenotype is an adverse prognostic factor in epithelial
ovarian cancer. PLoS One, 2014. 9(6): p. e100816.
Annels, N.E., et al., Spontaneous antibodies against Engrailed-2 (EN2) protein in patients
with prostate cancer. Clin Exp Immunol, 2014. 177(2): p. 428-38.
Team, R.C. R: A language and environment for statistical computing. . 2013 [cited 2015 21st
January 2015]; Available from: http://www.R-project.org/.
Smyth, G.K., Limma: Linear models for microarray data. Bioinformatics and Computational
Biology Solution Using R and Bioconductor, 2005.
DiFeo, A., et al., Roles of KLF6 and KLF6-SV1 in ovarian cancer progression and
intraperitoneal dissemination. Clin Cancer Res, 2006. 12(12): p. 3730-9.
Boerboom, D., et al., Misregulated Wnt/beta-catenin signaling leads to ovarian granulosa
cell tumor development. Cancer Res, 2005. 65(20): p. 9206-15.
Chen, J., et al., Overexpression of EFEMP1 correlates with tumor progression and poor
prognosis in human ovarian carcinoma. PLoS One, 2013. 8(11): p. e78783.
Wu, Q., et al., DNA methylation profiling of ovarian carcinomas and their in vitro models
identifies HOXA9, HOXB5, SCGB3A1, and CRABP1 as novel targets. Mol Cancer, 2007. 6: p.
45.
Langdon, S.P., et al., Characterization and properties of nine human ovarian adenocarcinoma
cell lines. Cancer Res, 1988. 48(21): p. 6166-72.
Seidman, J.D., et al., The histologic type and stage distribution of ovarian carcinomas of
surface epithelial origin. Int J Gynecol Pathol, 2004. 23(1): p. 41-4.
Bowtell, D.D., The genesis and evolution of high-grade serous ovarian cancer. Nat Rev
Cancer, 2010. 10(11): p. 803-8.
Vaughan, S., et al., Rethinking ovarian cancer: recommendations for improving outcomes.
Nat Rev Cancer, 2011. 11(10): p. 719-25.
Berns, E.M. and D.D. Bowtell, The changing view of high-grade serous ovarian cancer. Cancer
Res, 2012. 72(11): p. 2701-4.
Domcke, S., et al., Evaluating cell lines as tumour models by comparison of genomic profiles.
Nat Commun, 2013. 4: p. 2126.
Wiegand, K.C., et al., ARID1A mutations in endometriosis-associated ovarian carcinomas. N
Engl J Med, 2010. 363(16): p. 1532-43.
220
336.
337.
338.
339.
340.
341.
342.
343.
344.
345.
346.
347.
348.
349.
350.
351.
352.
353.
354.
355.
356.
357.
358.
Campbell, I.G., et al., Mutation of the PIK3CA gene in ovarian and breast cancer. Cancer Res,
2004. 64(21): p. 7678-81.
Sieben, N.L., et al., In ovarian neoplasms, BRAF, but not KRAS, mutations are restricted to
low-grade serous tumours. J Pathol, 2004. 202(3): p. 336-40.
Bartlett, J.M., et al., Transforming growth factor-beta mRNA expression and growth control
of human ovarian carcinoma cells. Br J Cancer, 1992. 65(5): p. 655-60.
Cooke, S.L., et al., Genomic analysis of genetic heterogeneity and evolution in high-grade
serous ovarian carcinoma. Oncogene, 2010. 29(35): p. 4905-13.
Stronach, E.A., et al., HDAC4-regulated STAT1 activation mediates platinum resistance in
ovarian cancer. Cancer Res, 2011. 71(13): p. 4412-22.
Lai, C.Y., et al., Engrailed-2 is down-regulated but also ectopically expressed in clear cell renal
cell carcinoma. Mol Biol Rep, 2014. 41(6): p. 3651-7.
Kramer, O.H., et al., Acetylation of Stat1 modulates NF-kappaB activity. Genes Dev, 2006.
20(4): p. 473-85.
Kramer, O.H., et al., A phosphorylation-acetylation switch regulates STAT1 signaling. Genes
Dev, 2009. 23(2): p. 223-35.
Guan, R., et al., A novel protein is lower expressed in renal cell carcinoma. Int J Mol Sci, 2014.
15(5): p. 7398-408.
Ozols, R.F., Systemic therapy for ovarian cancer: current status and new treatments. Semin
Oncol, 2006. 33(2 Suppl 6): p. S3-11.
Hellstrom, I., et al., Detection of the HE4 protein in urine as a biomarker for ovarian
neoplasms. Cancer Lett, 2010. 296(1): p. 43-8.
Gadducci, A., et al., Assessment of the prognostic relevance of serum anti-p53 antibodies in
epithelial ovarian cancer. Gynecol Oncol, 1999. 72(1): p. 76-81.
Kelly, Z.L., et al., HOX genes in ovarian cancer. J Ovarian Res, 2011. 4: p. 16.
Naora, H., et al., A serologically identified tumor antigen encoded by a homeobox gene
promotes growth of ovarian epithelial cells. Proc Natl Acad Sci U S A, 2001. 98(7): p. 4060-5.
Hong, J.H., et al., Expression pattern of the class I homeobox genes in ovarian carcinoma. J
Gynecol Oncol, 2010. 21(1): p. 29-37.
Nonaka, D., L. Chiriboga, and R.A. Soslow, Expression of pax8 as a useful marker in
distinguishing ovarian carcinomas from mammary carcinomas. Am J Surg Pathol, 2008.
32(10): p. 1566-71.
Laury, A.R., et al., PAX8 reliably distinguishes ovarian serous tumors from malignant
mesothelioma. Am J Surg Pathol, 2010. 34(5): p. 627-35.
Laury, A.R., et al., A comprehensive analysis of PAX8 expression in human epithelial tumors.
Am J Surg Pathol, 2011. 35(6): p. 816-26.
Ahmed, N., et al., Neuronal transcription factor Brn-3a(l) is over expressed in high-grade
ovarian carcinomas and tumor cells from ascites of patients with advanced-stage ovarian
cancer. J Ovarian Res, 2010. 3: p. 17.
Pandha, H.J., S. Sooriakumaran, P. Bott, S. Montgomery, B. Hutton, A. Eden, C. Langley, S.
Morgan, R., Correlation of Urinary Engrailed-2 Levels to Tumour Volume and Pathological
Stage in Men Undergoing Radical Prostatectomy. Journal of Cancer Therapy, 2013. 4(3): p.
726-733.
Omura, G.A., et al., Long-term follow-up and prognostic factor analysis in advanced ovarian
carcinoma: the Gynecologic Oncology Group experience. J Clin Oncol, 1991. 9(7): p. 1138-50.
Brown, E., et al., Carcinosarcoma of the ovary: 19 years of prospective data from a single
center. Cancer, 2004. 100(10): p. 2148-53.
Pectasides, D., et al., Treatment issues in clear cell carcinoma of the ovary: a different entity?
Oncologist, 2006. 11(10): p. 1089-94.
221
359.
360.
361.
362.
363.
364.
365.
366.
367.
368.
369.
370.
371.
372.
373.
374.
375.
376.
377.
378.
379.
380.
Hoskins, W.J., et al., The influence of cytoreductive surgery on recurrence-free interval and
survival in small-volume stage III epithelial ovarian cancer: a Gynecologic Oncology Group
study. Gynecol Oncol, 1992. 47(2): p. 159-66.
Hoskins, W.J., Surgical staging and cytoreductive surgery of epithelial ovarian cancer. Cancer,
1993. 71(4 Suppl): p. 1534-40.
Perets, R., et al., Transformation of the fallopian tube secretory epithelium leads to highgrade serous ovarian cancer in Brca;Tp53;Pten models. Cancer Cell, 2013. 24(6): p. 751-65.
Bowen, N.J., et al., Emerging roles for PAX8 in ovarian cancer and endosalpingeal
development. Gynecol Oncol, 2007. 104(2): p. 331-7.
Jager, E., et al., Induction of primary NY-ESO-1 immunity: CD8+ T lymphocyte and antibody
responses in peptide-vaccinated patients with NY-ESO-1+ cancers. Proc Natl Acad Sci U S A,
2000. 97(22): p. 12198-203.
Jager, E., et al., Recombinant vaccinia/fowlpox NY-ESO-1 vaccines induce both humoral and
cellular NY-ESO-1-specific immune responses in cancer patients. Proc Natl Acad Sci U S A,
2006. 103(39): p. 14453-8.
Stockert, E., et al., A survey of the humoral immune response of cancer patients to a panel of
human tumor antigens. J Exp Med, 1998. 187(8): p. 1349-54.
Liu, L., et al., Are circulating autoantibodies to ABCC3 transporter a potential biomarker for
lung cancer? J Cancer Res Clin Oncol, 2012. 138(10): p. 1737-42.
Cheng, Y., et al., Circulating autoantibody to ABCC3 may be a potential biomarker for
esophageal squamous cell carcinoma. Clin Transl Oncol, 2013. 15(5): p. 398-402.
Robertson, K.D., DNA methylation and human disease. Nat Rev Genet, 2005. 6(8): p. 597610.
Ting, A.H., K.M. McGarvey, and S.B. Baylin, The cancer epigenome--components and
functional correlates. Genes Dev, 2006. 20(23): p. 3215-31.
Wong, E.M., et al., Constitutional methylation of the BRCA1 promoter is specifically
associated with BRCA1 mutation-associated pathology in early-onset breast cancer. Cancer
Prev Res (Phila), 2011. 4(1): p. 23-33.
Zou, H., et al., Quantification of methylated markers with a multiplex methylation-specific
technology. Clin Chem, 2012. 58(2): p. 375-83.
deVos, T., et al., Circulating methylated SEPT9 DNA in plasma is a biomarker for colorectal
cancer. Clin Chem, 2009. 55(7): p. 1337-46.
Kneip, C., et al., SHOX2 DNA methylation is a biomarker for the diagnosis of lung cancer in
plasma. J Thorac Oncol, 2011. 6(10): p. 1632-8.
Ahmed, H., et al., Evidence of heavy methylation in the galectin 3 promoter in early stages of
prostate adenocarcinoma: development and validation of a methylated marker for early
diagnosis of prostate cancer. Transl Oncol, 2009. 2(3): p. 146-56.
Goessl, C., et al., DNA-based detection of prostate cancer in blood, urine, and ejaculates. Ann
N Y Acad Sci, 2001. 945: p. 51-8.
Goessl, C., et al., DNA-based detection of prostate cancer in urine after prostatic massage.
Urology, 2001. 58(3): p. 335-8.
Goessl, C., M. Muller, and K. Miller, Methylation-specific PCR (MSP) for detection of tumour
DNA in the blood plasma and serum of patients with prostate cancer. Prostate Cancer
Prostatic Dis, 2000. 3(S1): p. S17.
Grutzmann, R., et al., Sensitive detection of colorectal cancer in peripheral blood by septin 9
DNA methylation assay. PLoS One, 2008. 3(11): p. e3759.
Warren, J.D., et al., Septin 9 methylated DNA is a sensitive and specific blood test for
colorectal cancer. BMC Med, 2011. 9: p. 133.
Church, T.R., et al., Prospective evaluation of methylated SEPT9 in plasma for detection of
asymptomatic colorectal cancer. Gut, 2014. 63(2): p. 317-25.
222
381.
382.
383.
384.
385.
386.
387.
388.
389.
390.
391.
392.
393.
394.
395.
396.
397.
398.
399.
400.
401.
402.
Leung, W.K., et al., Quantitative detection of promoter hypermethylation in multiple genes in
the serum of patients with colorectal cancer. Am J Gastroenterol, 2005. 100(10): p. 2274-9.
Lee, B.B., et al., Aberrant methylation of APC, MGMT, RASSF2A, and Wif-1 genes in plasma
as a biomarker for early detection of colorectal cancer. Clin Cancer Res, 2009. 15(19): p.
6185-91.
Merlo, A., et al., 5' CpG island methylation is associated with transcriptional silencing of the
tumour suppressor p16/CDKN2/MTS1 in human cancers. Nat Med, 1995. 1(7): p. 686-92.
Tessema, M., et al., Concomitant promoter methylation of multiple genes in lung
adenocarcinomas from current, former and never smokers. Carcinogenesis, 2009. 30(7): p.
1132-8.
Schneider, K.U., et al., Correlation of SHOX2 gene amplification and DNA methylation in lung
cancer tumors. BMC Cancer, 2011. 11: p. 102.
Nakayama, M., et al., Hypermethylation of the human glutathione S-transferase-pi gene
(GSTP1) CpG island is present in a subset of proliferative inflammatory atrophy lesions but
not in normal or hyperplastic epithelium of the prostate: a detailed study using laser-capture
microdissection. Am J Pathol, 2003. 163(3): p. 923-33.
Jing, F., et al., Hypermethylation of tumor suppressor genes BRCA1, p16 and 14-3-3sigma in
serum of sporadic breast cancer patients. Onkologie, 2007. 30(1-2): p. 14-9.
Bosviel, R., et al., BRCA1 promoter methylation in peripheral blood DNA was identified in
sporadic breast cancer and controls. Cancer Epidemiol, 2012. 36(3): p. e177-82.
Strathdee, G. and R. Brown, Aberrant DNA methylation in cancer: potential clinical
interventions. Expert Rev Mol Med, 2002. 4(4): p. 1-17.
Pfeifer, G.P. and T.A. Rauch, DNA methylation patterns in lung carcinomas. Semin Cancer
Biol, 2009. 19(3): p. 181-7.
Fourkala, E.O., et al., DNA methylation of polycomb group target genes in cores taken from
breast cancer centre and periphery. Breast Cancer Res Treat, 2010. 120(2): p. 345-55.
Tommasi, S., et al., Methylation of homeobox genes is a frequent and early epigenetic event
in breast cancer. Breast Cancer Res, 2009. 11(1): p. R14.
Rada-Iglesias, A., et al., Histone H3 lysine 27 trimethylation in adult differentiated colon
associated to cancer DNA hypermethylation. Epigenetics, 2009. 4(2): p. 107-13.
Lambert, S.R., et al., Differential expression and methylation of brain developmental genes
define location-specific subsets of pilocytic astrocytoma. Acta Neuropathol, 2013. 126(2): p.
291-301.
Martinez, R., et al., A microarray-based DNA methylation study of glioblastoma multiforme.
Epigenetics, 2009. 4(4): p. 255-64.
Hellman, A. and A. Chess, Gene body-specific methylation on the active X chromosome.
Science, 2007. 315(5815): p. 1141-3.
Bert, S.A., et al., Regional activation of the cancer genome by long-range epigenetic
remodeling. Cancer Cell, 2013. 23(1): p. 9-22.
James, S.J., et al., Complex epigenetic regulation of engrailed-2 (EN-2) homeobox gene in the
autism cerebellum. Transl Psychiatry, 2013. 3: p. e232.
McGinnis, W. and R. Krumlauf, Homeobox genes and axial patterning. Cell, 1992. 68(2): p.
283-302.
Millhorn, D.E. and F.L. Eldridge, Role of ventrolateral medulla in regulation of respiratory and
cardiovascular systems. J Appl Physiol (1985), 1986. 61(4): p. 1249-63.
MacDonald, B.T., K. Tamai, and X. He, Wnt/beta-catenin signaling: components,
mechanisms, and diseases. Dev Cell, 2009. 17(1): p. 9-26.
Gordon, M.D. and R. Nusse, Wnt signaling: multiple pathways, multiple receptors, and
multiple transcription factors. J Biol Chem, 2006. 281(32): p. 22429-33.
223
403.
404.
405.
406.
407.
408.
409.
410.
411.
412.
413.
414.
415.
416.
417.
418.
419.
420.
421.
422.
423.
424.
McGrew, L.L., et al., Direct regulation of the Xenopus engrailed-2 promoter by the Wnt
signaling pathway, and a molecular screen for Wnt-responsive genes, confirm a role for Wnt
signaling during neural patterning in Xenopus. Mech Dev, 1999. 87(1-2): p. 21-32.
Kalkman, H.O., A review of the evidence for the canonical Wnt pathway in autism spectrum
disorders. Mol Autism, 2012. 3(1): p. 10.
Eisenmann, D.M., Wnt signaling, in The C. elegans Research Community, WormBook, e.
WormBook, Editor., WormBook.
Mali, P., et al., RNA-guided human genome engineering via Cas9. Science, 2013. 339(6121):
p. 823-6.
Cong, L., et al., Multiplex genome engineering using CRISPR/Cas systems. Science, 2013.
339(6121): p. 819-23.
Friedland, A.E., et al., Heritable genome editing in C. elegans via a CRISPR-Cas9 system. Nat
Methods, 2013. 10(8): p. 741-3.
Akiyoshi, S., et al., c-Ski acts as a transcriptional co-repressor in transforming growth factorbeta signaling through interaction with smads. J Biol Chem, 1999. 274(49): p. 35269-77.
Luo, K., et al., The Ski oncoprotein interacts with the Smad proteins to repress TGFbeta
signaling. Genes Dev, 1999. 13(17): p. 2196-206.
Reed, J.A., et al., Cytoplasmic localization of the oncogenic protein Ski in human cutaneous
melanomas in vivo: functional implications for transforming growth factor beta signaling.
Cancer Res, 2001. 61(22): p. 8074-8.
Prunier, C., et al., The oncoprotein Ski acts as an antagonist of transforming growth factorbeta signaling by suppressing Smad2 phosphorylation. J Biol Chem, 2003. 278(28): p. 2624957.
Luo, K., Ski and SnoN: negative regulators of TGF-beta signaling. Curr Opin Genet Dev, 2004.
14(1): p. 65-70.
Bravou, V., et al., TGF-beta repressors SnoN and Ski are implicated in human colorectal
carcinogenesis. Cell Oncol, 2009. 31(1): p. 41-51.
Schmalfeldt, B., et al., Increased expression of matrix metalloproteinases (MMP)-2, MMP-9,
and the urokinase-type plasminogen activator is associated with progression from benign to
advanced ovarian cancer. Clin Cancer Res, 2001. 7(8): p. 2396-404.
Koensgen, D., et al., Overexpression of the plasminogen activator inhibitor type-1 in
epithelial ovarian cancer. Anticancer Res, 2006. 26(2C): p. 1683-9.
Lin, S.W., et al., Critical involvement of ILK in TGFbeta1-stimulated invasion/migration of
human ovarian cancer cells is associated with urokinase plasminogen activator system. Exp
Cell Res, 2007. 313(3): p. 602-13.
De Smaele, E., et al., Induction of gadd45beta by NF-kappaB downregulates pro-apoptotic
JNK signalling. Nature, 2001. 414(6861): p. 308-13.
Zazzeroni, F., et al., Gadd45 beta mediates the protective effects of CD40 costimulation
against Fas-induced apoptosis. Blood, 2003. 102(9): p. 3270-9.
Zerbini, L.F. and T.A. Libermann, GADD45 deregulation in cancer: frequently methylated
tumor suppressors and potential therapeutic targets. Clin Cancer Res, 2005. 11(18): p. 640913.
Gupta, M., et al., Hematopoietic cells from Gadd45a- and Gadd45b-deficient mice are
sensitized to genotoxic-stress-induced apoptosis. Oncogene, 2005. 24(48): p. 7170-9.
Engelmann, A., et al., Gadd45 beta is a pro-survival factor associated with stress-resistant
tumors. Oncogene, 2008. 27(10): p. 1429-38.
ten Dijke, P. and C.S. Hill, New insights into TGF-beta-Smad signalling. Trends Biochem Sci,
2004. 29(5): p. 265-73.
Petraglia, F., et al., Inhibin and activin modulate human monocyte chemotaxis and human
lymphocyte interferon-gamma production. J Clin Endocrinol Metab, 1991. 72(2): p. 496-502.
224
425.
426.
427.
428.
429.
430.
431.
432.
433.
434.
435.
436.
437.
438.
439.
440.
441.
442.
443.
444.
445.
446.
Hyuga, S., et al., Possible role of hepatocyte growth factor/scatter factor and activin A
produced by the target organ in liver metastasis. Cancer Lett, 2000. 153(1-2): p. 137-43.
Grunert, S., M. Jechlinger, and H. Beug, Diverse cellular and molecular mechanisms
contribute to epithelial plasticity and metastasis. Nat Rev Mol Cell Biol, 2003. 4(8): p. 657-65.
Geiger, T.R. and D.S. Peeper, Metastasis mechanisms. Biochim Biophys Acta, 2009. 1796(2):
p. 293-308.
Kalluri, R., EMT: when epithelial cells decide to become mesenchymal-like cells. J Clin Invest,
2009. 119(6): p. 1417-9.
Kalluri, R. and R.A. Weinberg, The basics of epithelial-mesenchymal transition. J Clin Invest,
2009. 119(6): p. 1420-8.
Micalizzi, D.S., S.M. Farabaugh, and H.L. Ford, Epithelial-mesenchymal transition in cancer:
parallels between normal development and tumor progression. J Mammary Gland Biol
Neoplasia, 2010. 15(2): p. 117-34.
Zlobec, I. and A. Lugli, Epithelial mesenchymal transition and tumor budding in aggressive
colorectal cancer: tumor budding as oncotarget. Oncotarget, 2010. 1(7): p. 651-61.
Haslehurst, A.M., et al., EMT transcription factors snail and slug directly contribute to
cisplatin resistance in ovarian cancer. BMC Cancer, 2012. 12: p. 91.
Dupont, J., et al., Activin signaling pathways in ovine pituitary and LbetaT2 gonadotrope
cells. Biol Reprod, 2003. 68(5): p. 1877-87.
Zhang, L., et al., MEKK1 transduces activin signals in keratinocytes to induce actin stress fiber
formation and migration. Mol Cell Biol, 2005. 25(1): p. 60-5.
Raimondi, C. and M. Falasca, Phosphoinositides signalling in cancer: focus on PI3K and PLC.
Adv Biol Regul, 2012. 52(1): p. 166-82.
Faenza, I., et al., A role for nuclear phospholipase Cbeta 1 in cell cycle control. J Biol Chem,
2000. 275(39): p. 30520-4.
Fiume, R., et al., Involvement of nuclear PLCbeta1 in lamin B1 phosphorylation and G2/M cell
cycle progression. FASEB J, 2009. 23(3): p. 957-66.
Poli, A., et al., K562 cell proliferation is modulated by PLCbeta1 through a PKCalphamediated pathway. Cell Cycle, 2013. 12(11): p. 1713-21.
Follo, M.Y., et al., Phosphoinositide-phospholipase C beta1 mono-allelic deletion is associated
with myelodysplastic syndromes evolution into acute myeloid leukemia. J Clin Oncol, 2009.
27(5): p. 782-90.
Follo, M.Y., et al., Synergistic induction of PI-PLCbeta1 signaling by azacitidine and valproic
acid in high-risk myelodysplastic syndromes. Leukemia, 2011. 25(2): p. 271-80.
Lee, S.O., et al., The orphan nuclear receptor NR4A1 (Nur77) regulates oxidative and
endoplasmic reticulum stress in pancreatic cancer cells. Mol Cancer Res, 2014. 12(4): p. 52738.
Lee, S.O., et al., Targeting NR4A1 (TR3) in cancer cells and tumors. Expert Opin Ther Targets,
2011. 15(2): p. 195-206.
Yoon, K., et al., Activation of nuclear TR3 (NR4A1) by a diindolylmethane analog induces
apoptosis and proapoptotic genes in pancreatic cancer cells and tumors. Carcinogenesis,
2011. 32(6): p. 836-42.
Thompson, J. and A. Winoto, During negative selection, Nur77 family proteins translocate to
mitochondria where they associate with Bcl-2 and expose its proapoptotic BH3 domain. J Exp
Med, 2008. 205(5): p. 1029-36.
Wilson, A.J., et al., TR3/Nur77 in colon cancer cell apoptosis. Cancer Res, 2003. 63(17): p.
5401-7.
Wu, Q., et al., Dual roles of Nur77 in selective regulation of apoptosis and cell cycle by TPA
and ATRA in gastric cancer cells. Carcinogenesis, 2002. 23(10): p. 1583-92.
225
447.
448.
449.
450.
451.
452.
453.
454.
Sibayama-Imazu, T., et al., Induction of apoptosis in PA-1 ovarian cancer cells by vitamin K2 is
associated with an increase in the level of TR3/Nur77 and its accumulation in mitochondria
and nuclei. J Cancer Res Clin Oncol, 2008. 134(7): p. 803-12.
Gentner, B., et al., Differences in the gene expression profile of matrix metalloproteinases
(MMPs) and their inhibitors (TIMPs) in primary colorectal tumors and their synchronous liver
metastases. Anticancer Res, 2009. 29(1): p. 67-74.
Wong, J.C., et al., Absence of MMP2 expression correlates with poor clinical outcomes in
rectal cancer, and is distinct from MMP1-related outcomes in colon cancer. Clin Cancer Res,
2011. 17(12): p. 4167-76.
Koenig, S.F., et al., En2, Pax2/5 and Tcf-4 transcription factors cooperate in patterning the
Xenopus brain. Dev Biol, 2010. 340(2): p. 318-28.
Kunz, M., et al., Autoregulation of canonical Wnt signaling controls midbrain development.
Dev Biol, 2004. 273(2): p. 390-401.
Wang, T., et al., Twist2 contributes to cisplatin-resistance of ovarian cancer through the
AKT/GSK-3beta signaling pathway. Oncol Lett, 2014. 7(4): p. 1102-1108.
Mao, Y., et al., The role of nuclear beta-catenin accumulation in the Twist2-induced ovarian
cancer EMT. PLoS One, 2013. 8(11): p. e78200.
Schmalhofer, O., S. Brabletz, and T. Brabletz, E-cadherin, beta-catenin, and ZEB1 in
malignant progression of cancer. Cancer Metastasis Rev, 2009. 28(1-2): p. 151-66.
226
PUBLICATIONS
227

Spontaneous antibodies against Engrailed-2 (EN2) protein in patients with prostate
cancer. Annels NE, Simpson GR, Denyer M, McGrath SE, Falgari G, Killick E, Eeles R,
Stebbing J, Pchejetski D, Cutress R, Murray N, Michael A, Pandha H. Clin Exp Immuol.
2014 Aug; 177(2):428-38.

EN2: a novel prostate cancer biomarker. McGrath SE, Michael A, Morgan R, Pandha
H. Biomark Med. 2013 Dec; 7(6):893-901. Review.

Engrailed homeobox transcription factors as potential markers and targets in
cancer. McGrath SE, Michael A, Pandha H, Morgan R. FEBS Lett. 2013 Mar;
587(6):549-54. Review.
228