New Insights in Methodology of Screening For Cervical Cancer

University of Groningen
New insights in methodology of screening for cervical cancer

Wang, Rong
IMPORTANT NOTE: You are advised to consult the publisher's version (publisher's PDF) if you wish to cite from
it. Please check the document version below.
Document Version
Publisher's PDF, also known as Version of record
Publication date:
2015
Link to publication in University of Groningen/UMCG research database
Citation for published version (APA):

Wang, R. (2015). New insights in methodology of screening for cervical cancer: Nieuwe inzichten in de
methodologie van baarmoederhalskanker screening. [Groningen]: University of Groningen.
Copyright
Other than for strictly personal use, it is not permitted to download or to forward/distribute the text or part of it without the consent of the
author(s) and/or copyright holder(s), unless the work is under an open content license (like Creative Commons).
Take-down policy
If you believe that this document breaches copyright please contact us providing details, and we will remove access to the work immediately
and investigate your claim.
Downloaded from the University of Groningen/UMCG research database (Pure): http://www.rug.nl/research/portal. For technical reasons the
number of authors shown on this cover page is limited to 10 maximum.
Download date: 20-09-2019

New insights in methodology of
screening for cervical cancer
5RQJ:DQJ
Rong Wang ⦻㫹
Ph.D. thesis, University of Groningen, the Netherlands
ISBN: 978-90-367-7704-9 (printed version)

978-90-367-7703-2 (digital version)
Cover design: Rong Wang
Cover image: De brug in the Netherlands (front). Tianjin Liberation bridge (back).
Cover layout: Gildeprint-The Netherlands
Text Layout: Rong Wang
Printed by: Gildeprint-The Netherlands
Copyright © 2015 RONG WANG, Groningen, the Netherlands

All rights reserved.No part of this thesis may be reproduced, stored in a retrieval
system or transmitted in any form or by any means, without prior permission of the
author and publisher holding the copyright of the published articles.
The research in this thesis was partly financially supported by the Dutch Cancer
Society (RUG 2009-4577) and Natural Science Foundation of Tianjin
(12JCYBJC33700).
The publication of this thesis was financially supported by Graduate School of Medical
Sciences (GSMS) University of Groningen, University Medical Centrum Groningen,
Natural Science Foundation of Tianjin.
New insights in methodology of
screening for cervical cancer
PhD thesis
to obtain the degree of PhD at the

University of Groningen
on the authority of the
Rector Magnificus Prof. E. Sterken
and in accordance with
the decision by the College of Deans.
This thesis will be defended in public on
Wednesday 15 April 2015 at 11.00 hours
by
Rong Wang
born on 25 August 1977

in Tianjin, China
Promotors
Prof dr A. G. J. van der Zee
Prof dr E. Schuuring
Co-promotor
Dr G. B. A. Wisman
Assessment committee
Prof dr P. J. F. Snijders
Prof dr LF. A. M. Massuger
Prof dr H. W. Nijman
Paranimfen:
R. W. van Leeuwen
P. M. Schoonen
R. Wu
To my family
ࢳচ‫؞ࣩۊ‬ъ
Contents
Chapter 1 General Introductionಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ11
Chapter 2 Nationwide prevalence of human papillomavirus infection
in 37 cities in China ಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ37
Submitted for publication
Chapter 3 Clinical validation of the Cervista HPV HR test according to the
international guidelines for human papillomavirus test requirements
for cervical cancer screening ಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ65
J. Clin. Microbiol. 2014,52(12):4391.
Chapter 4 Discovery of new methylation markers to improve screening for
cervical intraepithelial neoplasia grade 2/3 ಹಹಹಹಹಹಹಹಹಹ73
Chapter 5 Genome-wide methylome analysis to discovery of (novel)
hypermethylation biomarkers for both cervical adenocarcinoma and
squamous cell carcinomaಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ109
Chapter 6 Summary and future perspectives ಹಹಹಹಹಹಹಹಹಹಹಹಹಹ141
Chapter 7 Nederlandse samenvatting ಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ157
About the author ಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ167
Acknowledgementsಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ169
General introduction
Chapter 1

11
Chapter 1
I. Cervical cancer
1.1. Epidemiology
Cervical cancer is the third most common cancer among women and the fourth
leading cause of female cancer deaths. Each year, there are more than 529,000 new
cases and around 275,000 deaths globally1. In the Netherlands, a low cervical
cancer incidence and mortality rate country2, cervical cancer ranks as 11th most
frequent cancer among women, but the 3rd among women between 15 ~ 44 yrs. The
latency data from World Health Organization/Institut Català d'Oncologia (WHO/ICO)
Information Centre (2014) reported that ~750 women are diagnosed with cervical
cancer and ~240 women die from the disease in the Netherlands each year3. In China,
a populous and diverse country which covers more than 1.3 billion people and 9.6
million square kilometers, cervical cancer ranks as the 8th most frequent cancer among
women and the 2nd in the age of 15~44 yrs. Data from WHO/ICO Information Centre
(2014) showed, that each year ~61,500 new cases are reported and ~29,500 patients
die of cervical cancer in China4. An explanation why China is a cervical cancer high
incidence country, is the fact that a nationwide organized cervical cancer screening
program is still lacking. The incidence and mortality of cervical cancer between
different regions within China are variable (Table 1)5. Moreover, the existing cancer
registries are geographically limited and the coverage of population is quite low,
indicating that these data are probably not representative for the whole country.
1.2. Clinicopathology of cervical cancer

The most common cervical cancers are squamous cell carcinomas (SCC) and
adenocarcinomas (ADC), which account for 75–90% and 10–25% of cervical cancers
respectively6,7, 8
. A small proportion of cervical cancers (~3-5%)9 represents
adenosquamous carcinomas and other rare histological types including
neuroendocrine carcinomas10.
The cervix uteri consists of the ectocervix and endocervix. The ectocervix is mainly
lined with non-keratinizing stratified squamous epithelium and the endocervix with
mucus producing columnar epithelium. The cells in the squamocolumnar junction
(SCJ) termed as transformation zone are less stable and particularly susceptible to
viral infections11. SCC most often occurs at the SCJ between the ecto- and endocervix
12
Table 1. The database of the incidence and mortality of cervical cancer
in China (data adapted from Shi et al5)

Cover Rate (Per Sites/Regi
Database national 100,000) ons Remark
population
Shanghai,
Jiashan,
Tianjin,
International Agency for Beijing,
Incidence Research on Cancer’s 2% 1.2-4.6 Wuhan,
(IARC’s) 1993-2002 females Qidong,
Harbin,
Zhongshan
and
Guangzhou
>10 /105
30~37 Yangcheng in
Chinese National Center for
6% 1.9-8.4 registry Shanxi
Cancer Registries (NCCR)
sites province
system1998-2003
&Shenzhen in
Guangdong
NCCR 2004-2006 6% 6.0-9.1
national retrospective survey 14.6 11 sites (overestimated)

1975
national retrospective survey ~10% 4.3 11 sites (underestimated)

1990–1992
national retrospective survey ~10% 2.5 11 sites (underestimated)

2004–2005
Mortality
Chinese National Center for
6% 1.2 ~ 7.0 30~37 sites
Cancer Registries (NCCR)
system1998-2003
2.7 age-
Center of Health Information (overall) standardized
~1%
and Statistics (CHIS) 4.2 (rural)
reporting system,WHO 1.8
( urban )
CDC 71 ~10%
13
Chapter 1
and is therefore detected easier by cytologic screening. In contrast, ADC develops

from the gland cells within the endocervical canal. Consequently, ADC is mainly
diagnosed at an advanced stage. Compared to SCC, ADC associates with a worse
prognosis, is less sensitive to radiation therapy and chemotherapy, and easily tends
to metastasize. Therefore, overall patients with ADC have a 5-years survival rate that
is 10-20% lower than SCC8,12.
1.3. Treatment of cervical cancer

According to the International Federation of Gynecology and Obstetrics (FIGO) stage,
microinvasive cervical cancer can be treated by large loop excision of the
transformation zone (LLETZ) or cone biopsy. Early stage tumors can be managed by
radical hysterectomy plus pelvic lymph node dissection or (chemo)radiotherapy.
Advanced stage tumors are almost always treated by (chemo)radiotherapy. Five year
survival approaches 100% for patients with tumors of stage IA and averages 70–85%
for those with stage IB1 and small IIA lesions, 50–70% for stages IB2 and IIB, 30–
50% for stage III and 5–15% for stage IV13. Therefore, early detection is one of the
most effective strategies for saving cervical cancer patient lives.
1.4. Cervical precursor lesions

1.4.1 Histologic and cytologic classification
Histologically, the premalignant lesion of SCC is referred to as cervical intraepithelial
neoplasia (CIN). Based on the severity of the changes and especially on the
proportion of the epithelial layer with neoplastic changes, CIN is stratified in 3 grades,
CIN1, CIN2 and CIN3 representing 1/3, 2/3 and almost the total layer of epithelium
containing atypical cells14. Routine cytological classifications are the Pap and
Bethesda systems (Table 2, Figure 1). According to the severity of abnormal smears,
the European Pap classification is graded into 5 classes, PAP1 to PAP515 (Table 2,
Figure 1). The American Bethesda classification was developed to differentiate
between lesions probable to progress to cervical cancer (high-grade squamous
intraepithelial lesions (HSIL)) and lesions less probable to progress (low-grade
squamous intraepithelial lesions (LSIL))16. In Table 2 and Figure 1, the
correspondence between the histologic and cytologic classifications is illustrated.
14
Table 2: Overview of the different nomenclatures and classifications for

histologic and cytologic cervical abnormalities (These data were adapted from the thesis
of Nan Yang: Detection of DNA Hypermethylation as a Diagnostic Tool in Cervical Neoplasia (ISBN978-90-367-
4169-9))
Histology Cytology
Dysplasia CIN Bethesda Papanicolaou
Normal Normal Within normal limits

Pap 1
Benign atypia Inflammatory atypia Benign cellular changes
Atypical cells Squamous atypia ASCUS Pap 2
Mild Dysplasia CIN1 Low-grade SIL (LSIL) Pap 3A1
Moderate Dysplasia CIN 2 Pap 3A2
Severe Dysplasia High-grade SIL (HSIL) Pap 3B

CIN 3
Carcinoma in situ Pap 4
(Microinvasive) cancer (Microinvasive) cancer (Microinvasive) cancer Pap 5

Figure 1: The development of cervical carcinogenesis(AdaptedfromSnijdersetal

200614).
The precursor of invasive ADC was first described by Friedell and McKay17 in 1953
as adenocarcinoma in situ (AdCIS). Histologic features of AdCIS include preservation
of normal glandular architecture and partial or complete involvement of endocervical
15
Chapter 1
glands and abrupt transition to normal endocervical epithelium. The cytoplasm can
be depleted or abundant, vacuolated, granular, and basophilic or eosinophilic18,19.
The 3 most frequent AdCIS subtypes are endocervical (usual), intestinal, and
endometrioid19. Unlike CIN, AdCIS is much less frequent and also more difficult to
detect effectively. AdCIS is not well visualized colposcopically as it can arise high up
in the endocervical canal20 and the cytomorphologic criteria for identifying neoplastic
glandular lesions are not as well defined as for CIN. Additionally, AdCIS frequently
coexists with squamous intraepithelial lesions or squamous cell carcinoma in 50% of
cases18. Therefore, the detection of premalignant ADC lesions in scrapings is more
difficult compared to premalignant SCC lesions.
1.4.2 Treatment of precursor lesions

Most cases of CIN1 will regress spontaneously, while up to 50% of the CIN3 lesions
may progress to cervical cancer, when left untreated. Both in the Netherlands as well
as in China, all identified CIN2 and CIN3 lesions are treated, because it is still not
possible to identify those lesions that are most likely to progress. Treatment options
for CIN2 and CIN3 are LLETZ, cryocoagulation, laser evaporation, cone biopsy or in
rare cases hysterectomy. In general, LLETZ is the preferred treatment for CIN2 and
CIN3, because it is simple, cheap, effective and allows pathologic examination of the
specimens21.
II. Cervical cancer screening program
Cervical cancer development takes a long time in most patients. Normally, HSIL
(CIN2 and CIN3) develop within 3–5 years following a high-risk HPV (hrHPV)
infection, whereas further progression to invasive cancer can take up to 20–30
years10. This long period offers many opportunities for intervention and prevention of
cervical cancer. Based on previous large-scale studies, a systematic routine
screening program can reduce the incidence of cervical cancer by at least 60% and
has been recommended by WHO, particularly in developing countries22. Recently, a
meta-review collected all the available evidence in literature, which also supports the
notion that cervical screening does offer protective benefits and is associated with a
reduction in the incidence of invasive cervical cancer and cervical cancer mortality23.
16
In the Netherlands, pilot-studies on population- based screening by cytomorphologic

classification were started in the 1970s. A nationwide screening program aimed at
specific age categories was initiated in 1988. Between 1988 and 1996, women aged
34–54 years were screened once every 3 years; and from 1996 onwards, women
aged 30–60 were screened every 5 years24. The current Dutch screening program is
primarily based on cytologic examination of cervical smears with hrHPV testing as a
triage test for abnormal cytologic results (ASCUS/LSIL)25. In the Netherlands, the
population-based screening program will change to primary hrHPV screening in
201626 .
In China, cervical cancer screening remains opportunistic screening. Some women

have access to population-based or employee-based screening through large
corporations or organizations. A first pilot screening program named the Cervical
Cancer Screening Study (SPOCCS) project was performed in Shanxi Province, a
region with a heavy burden of cervical cancer27,28. After that, followed by the
“Program of Cancer Prevention and Control in China (2004-2010)”, the government
took the “Top Down” planning approach, i.e. set-up of two demonstration sites as
potential models for cervical cancer control. One was in a poverty-stricken county in
Shangxi29 with a high cervical cancer incidence and applies simple, inexpensive
technologies to screen for precancerous lesions. Another one was in Shenzhen30, a
prosperous city with high-resource settings. With the experience from the two regions,
cervical cancer screening sites were increasingly expanded to a total of 31 cities, 43
sites, autonomous regions included as well31. One of the challenges for China is that
70% of the population is rural and 85% of the cervical cancers occur in the less
developed regions. Therefore, a government-sponsored program proposed by the
All-China Women's Federation for free cervical cancer screening in women age 35-
59 years has been setup in July 200932. During 2009-2011, 10 million rural women
were screened by either Pap smear or visual inspection with acetic acid (VIA).
However, in light of an estimated 500 million women in China, the coverage is still too
low. Therefore, in 2012, the program was extended over the next three years and
was offered to more rural women33. Among other hurdles, China lacks a sufficient
number of cytopathologists or trained health-care workers to screen all of these
women by Pap smear or VIA, respectively. A nationwide screening round needs to
17
Chapter 1
be followed by development of more appropriate health-care policies and services for

all Chinese women.
Currently, The Ministry of Health of China has launched the Guideline for Screening
and Early Detection and Treatment of Cervical Cancer34,35. According to the guideline,
the target population is women > 21 years old or with sexual intercourse experience
for more than three years. Depending on the diverse geographical socioeconomic
status and levels of exposure to the risks of the population, the guideline has
recommended three protocols with a different combination of methodologies: i)
primary screening by liquid-based cytology test plus HPV DNA test in developed
regions and/or women with good economic status; ii) primary screening by Pap
smear cytology test plus HPV DNA test in moderately developed regions; iii) primary
screening by VIA in low- resource settings35.
III. Cervical cancer screening approach

3.1 Visual inspection with acetic acid (VIA)
VIA, a “screen-and-treat” method, has been suggested to be performed in the less-
developed countries and regions. The principle of VIA is that most CIN2 and CIN3
lesions are acetowhite (i.e. they develop a white color when vinegar or 5% acetic
acid is applied to the cervical epithelium that harbors preinvasive lesions), while the
normal cervical epithelium maintains its pink color after application of 5% acetic
acid36. VIA is a point-of-care test, producing immediately results, also allowing
consecutive treatment eg. by cryocoagulation. Most importantly, VIA does not require
a cytopathology laboratory and therefore can be performed with little infrastructure
and low costs. Although VIA sounds quite simple, it is actually more complicated than
it appears. Acetowhitening is a relatively non-specific cervical finding. Many women
have acetowhitening of the cervix not because of CIN2 and CIN3 lesions but
immature squamous metaplasia or reparative conditions etc. In case of insufficient
quality control, VIA will lead to over-treatment of many women who do not have
CIN2+ lesions36.
18
3.2 Morphology-based testing

Generally, all the morphology-based testing methods such as automated Pap-smear
analysis or liquid-based cytology show the same limitations, i.e. less sensitivity
(50~60%) and subjective interpretations. In addition, these techniques require
training, the development of laboratory infrastructure, standardization and quality
control measures. Furthermore, there is a disputable problem that a relevant amount
of high-risk cases from patients living in low resource regions may be missed due to
low sensitivity techniques of the screening protocols. Therefore, there is ample room
for improvement.
3.3 HPV as a marker for cervical cancer screening

The strong association of hrHPV infection with cervical cancer is widely known. HPV
is a genus of the family Papovaviridae. The HPV virions are non-enveloped and
icosahedral with a circular double stranded DNA genome of almost 8kb in length. Its
genome can be divided into three different regions: (i) a coding region containing the
seven early genes E1~ E7; (ii) a region containing the late genes encoding capsid
proteins; and (iii) a non-coding region, termed the long control region (LCR)37.
Expression of the viral proteins E6 and E7 is crucial for cervical carcinogenesis, E6
binds to p53 and herewith inactivates p53-mediated growth suppression and
apoptosis, whereas E7 binds to Rb which inactivates this negative regulator of the
cell cycle37. It is well established that a series of subsequent steps after HPV infection
results in progression to cancer: HPV persistence, HPV-mediated epithelial
transformation, development of precancerous lesions (CIN1-3) and then invasive
cervical cancer.
So far, more than 150 HPV types have been isolated and sequenced38 ,39
. Table
3 illustrates the main genera and their association with human diseases. According to
their oncogenic potential, the mucosal (alpha) HPV types are divided into two groups:
low-risk HPV (lrHPV), which are mainly associated with benign genital warts, and
high-risk HPV (hrHPV), which are the etiological agents of cervical cancer39. So far
within the discovered cancer related HPVs, 12 types of HPV (HPV16, HPV18,
HPV31, HPV33, HPV35, HPV39, HPV45, HPV51, HPV52, HPV56, HPV58 and
HPV59) have now been consistently classified as hrHPV (also known as International
19
Chapter 1
Table 3. The main diseases associated with different HPV types

(AdaptedfromTommasinoetal201339).
HPV Disease (% attributed

Pathogenicity HPV type
family cases)
Cervical squamous cell

carcinoma (~50%)
Cervical adenocarcinoma
HPV16
(~35%)
Oropharyngeal cancer
(~25%)
Mucosal high-
risk Cervical squamous cell
carcinoma (~20%)
HPV18
Cervical adenocarcinoma
Į (~35%)
Cervical squamous cell

HPV31,33,35,39,45,51,52,56,58,59
carcinoma (~30%)
Benign genital lesions

HPV6,11
Mucosal low- Respiratory papillomatosis
risk
Oral focal epithelial
HPV13,32
hyperplasia
Cutaneous HPV2,3,27,57 Skin warts
μ benign HPV1 Skin warts
First ȕHPV types isolated

HPV5 and 8
from SCC of EV indiviuals
HPV9,12,14,15,17,19-25,36-38, Likely associated with SCC in

ȕ 47,49,75,76,80,92,93,96,98- EV patients as well as
100,104,105,107,110,111,113, immuno-compromised and
Cutaneous
115,118,120,122,124,143,145, immune-competent
150-152,159 individuals
HPV4,48,50,60,65,88,95,101,103,
Ȗ 108,109,112,115,115,119,121,123,126- Unknown
142,144,146-149,153-158,161-170
20
Agency for Research on Cancer (IARC) class I), HPV68 has been classified as
probable high-risk (also known as IARC class 2A), and another seven types have
been classified as possible high-risk (HPV26, HPV53, HPV-66, HPV67, HPV70,
HPV73 and HPV82, also known as IARC class 2B)40.
Globally, HPV16 and -18 are the most oncogenic genotypes, which are related to
about 70% of invasive cervical carcinoma, 50% of high-grade lesions (CIN2/3) and
35% of CIN1. However, the prevalence and distribution of HPV genotypes show
considerable geographic and ethnic variation, especially for the less common types.
In the Netherlands, about 4% of women in the general population are estimated to
harbor cervical HPV infection in their life time, and 82.3%, 70.2% and 22.7% of
respectively invasive cervical cancers, HSIL and LSIL are attributed to HPV16 or
HPV18. The most common HPV types in CIN are HPV16, HPV31 and HPV183. In
China, in the general population about 13.7% of women4 (ranging from 6.7% to
45.6% in different studies (Table 4)), are estimated to harbor HPV infection in
cervical scrapings, of which 75.5%, 44.2% and 23.1% of invasive cervical cancers,
HSIL and LSIL, respectively, are attributed to HPV16 or HPV18. In contrast to the
Netherlands, in China the most common types in CIN are HPV16, HPV52, HPV584.
The HPV prevalence in women from population-based screening studies in China in
different time periods are listed in Table 4. Although so far there are several
programs including HPV testing for screening of cervical cancer and their
precancerous lesions, at present these programs cover only a small part of China
presently representing ~9 provinces/cities and a population of ~134,000 women
(Figure 2). HPV testing is playing an increasing role in cervical screening.
Currently, 7 tests have been approved by the United States Food and Drug
Administration (FDA). The first reliable, quality standardized HPV DNA test is the
Hybrid Capture (HC) assay, as developed by Digene Corporation (Gaithersburg, MD,
USA), which gained FDA approval in 1999. In 2003, FDA approval was extended to
the use of Hybrid Capture 2 (HC2) in Pap testing. Later, the HC2 test, which detects
13 hrHPV was recognized as “golden standard” and widely applied today. However,
some limitations, for instance, cross-reactivity and lack of control for input DNA,
21
Chapter 1
Table 4 . HPV prevalence in women from population-based screening studies in

different time period
(AdaptedfromLietal,201372)
Author Study Location Number Age Lab-assays overall
/Pub year year (County/city, (N) Range HPV
Province) (yrs) prevale
nce (%)
(Belinson et 1999 Xiangyuan 1,997 35-45 HC2 18.2
al., 2001) 73
(Shen et al., 2001- Xiangyuan&yangcheng 9,683 30-50 HC2 27.5
2003) 74 2002 ,shanxi
(Zhao et al., 2001- Xiangyuan& 8,798 35-50 HC2 23.6
2006)75 2002 yangcheng,shanxi
(Zou et al., 2004 Yangcheng 745 15-59 HC2 16.0
2011)76
(Li et al., 2004 Xiushui, Jiangxi 2,432 30-49 HC2 18.5
2007)77
(Li et al., 2004- Shenyang, 685 15-59 GP5+/6+ 16.8
2006)78 2005 Liaoning mediated
PCR
(Dai et al., 2004- Yangcheng 662 15-59 GP5+/6+ 14.8*
2006)29 2005 mediated
PCR
(Wu et al., 2004- Shenzhen, 1,027 15-59 GP5+/6+ 16.6*
2007)30 2005 Guangdong mediated
PCR
(Wu et al., 2006- Shanxi,Beijing, 4215 17-54 HC2 14.3
2013)79 2007 Xinjiang,
Henan,Shanghai
(Li et 2006 Chaozhou Guangdong 1705 20-68 PCR 9.03
al.,2008)80 (MY09/11
primer)
(Zhao et al., 2006- Beijing 5,552 25-54 PCR 6.7
2009)81 2008 (MY09/11
primer)
(Li et al., 2006- Beijing 6,185 25-54 HC2 9.9
2010)82 2009
(Ye et 2007- Zhejiang 5,058 20-79 PCR 13.3
al.,201083 2008 (MY09/11
primer)
(Wang et 2007- Liaoning 24,041 18-60 PCR 45.6
al.,2012)84 2010 (MY09/11
primer)
(Wu et al., 2008- Fujian 2,338 20-70 PCR 22.5
2010)85 2009 (MY09/11
primer)
(Hu et al., 2009 Jiangsu 316 18-25 HC2 17.1
2011)86
(Chen et 2009- Chaozhou, 48,559 35-60 Multiplex 7.89
al.,2012)87 2010 Guangdong realtimeRCR
(Zhang et 2011 Shanghai suburb 10,000 17-89 PCR 12.6
al.,2013)88 (MY09/11
primer)
*HPVoverallprevalenceincludinghighͲriskHPVandlowͲriskHPV
22
Figure 2. National map of China showing all the geographical sites. In this map
the regions in which population-based screening studies including HPV testing is
performed are indicated72.
results in false negative and positive results41. The second HPV testing platform,
Cervista HPV HR test (Hologic, WI, USA) detects the same 13 hrHPV subtypes as
HC2 and includes HPV66 as well. The Cervista HPV16/18 test was the first FDA
approved assay which permits additional genotyping. In 2014, the Cobas 4800 HPV
test (Roche Molecular Diagnostics, Pleasanton, CA, USA) detecting not only the
same 13 hrHPV but also discriminating between HPV16, HPV18 and the other
hrHPV, was FDA-approved for primary HPV screening. Finally, the APTIMA HPV
assay (Hologic formerly GenProbe Inc., San Diego, CA, USA) for the detection of E6
and E7 hrHPV mRNA was approved by FDA41. In addition to the FDA approved HPV
detection tests, more than 50 other in-house and commercial HPV-tests are
available41. Therefore, it is a real challenge to choose the most reliable HPV assay
for primary cervical HPV screening. In 2009, in an international guideline the criteria
were reported for a candidate hrHPV test to be used for primary HPV screening42. A
new test should be ‘non-inferior’ with respect to clinical sensitivity (t90%) and
23
Chapter 1
specificity (t98%) for the detection of CIN2+ when compared with the clinically
validated HC2 assay in women aged 30 yrs. The guideline also describes the
requirements of the technical robustness of new assays through the measurement of
intra- and inter-laboratory reproducibility.
Because of the HPV natural history, most of HPV infections are usually temporary, in
around 91% of women with an HPV infection, HPV is already cleared within two
years. However, despite the clearance of HPV infections in most women, especially
in the younger, sexual active population, HPV is present in virtually all of these
women at a certain time. Since HPV testing does not discriminate between
temporary and persistent HPV infections, the specificity of the test to detect CIN2+
lesions is low. Therefore, despite of the superior sensitivity of HPV testing, the lower
specificity is an inevitable problem leading to a need to discover novel biomarkers for
cervical cancer screening for triage testing. P16INK4a and Ki67 are established cell
cycling biomarkers and described as a direct marker of HPV infection43, 44. A recent
study reported that a combined dual-stained cytology test for both p16INK4a and Ki67
had a sensitivity of 91.9% for detecting CIN2+ and 96.4% for CIN3+. This test was
also highly specific: 82.1% for CIN2+ and 76.9% for CIN3+45, 46. Another biomarker,
ProExC, is a cocktail of MCM2 and TOP2A proteins47, which might be a sensitive and
specific marker for distinguishing CIN2/3 from metaplastic squamous epithelium47.
However, there is insufficient evidence to integrate these strategies into the standard
of care for cervical cancer screening and large screening trials are still needed to
validate these biomarkers48. More importantly, these immunohistochemistry-based
triage tests cannot be performed on the DNA extracted for HPV testing and thus
need different handling in the lab, are microscopy-dependent, require a well-fixed
specimen with preserved morphology and skilled cytologic/histologic pathologists10.
3.4. DNA methylation as a biomarker for cervical cancer screening

In 1942, C. H. Waddington named a word ”epigenetic” when he was studying the
causal interactions between the genotype and the phenotype49. Now epigenetics
refers to heritable modifications of the genome without any changes in primary DNA
sequences including DNA methylation, histone modifications, chromatin remodeling,
and noncoding RNAs50. Among them, DNA methylation is the best characterized so
24
far. The modification is a covalent modification of DNA, in which a methyl group is

transferred from S-adenosylmethionine (SAM) to the fifth position of the pyrimidine
ring of cytosine catalyzed by DNA methyltransferases (DNMTs)51 52 (Figure 3A). DNA
methylation occurs only at cytosines located 5’ to guanosine in the CpG
dinucleotide53 (Figure 3B). Genomic CpG islands are nucleotide regions where the
percentage of the CpG dinucleotides is higher. Hypermethylation of CpG islands
within the promoter region is correlated with transcriptional silencing54 55 (Figure 3C).
Abnormal DNA methylation is a well-recognized epigenetic hallmark of cancer cells

and has been observed in most malignancies. The first aberrant methylation
alterations in human cancer was discovered by Feinberg and Vogelstein in 198356.
Cancer often exhibits hypermethylation at the promoter region with simultaneous
widespread hypomethylation from normal methylated sites (Figure 3C). Promoter
DNA methylation is an early event in cancer development57. Furthermore, its patterns
are regulated in developmental stage specific, cell type specific and tissue-specific
manner supporting the idea that DNA methylation analysis might be a valuable
biomarker for cancer screening and early detection.
Although infection with hrHPV is a necessary feature, HPV in itself is not sufficient for
development of cervical cancer. It is known that integration of the viral DNA into the
cellular genome causes not only genetic but also epigenetic alterations. These result
in the silencing of tumor suppressor genes (TSG) and the overexpression of
oncogenes58. As for HPV59, also aberrant methylation patterns of cancer-associated
genes have been observed throughout the process of cervical carcinogenesis59,
60,61
.Therefore, in order to satisfy the requirements from clinical practice, the research
on DNA methylation biomarkers for molecular diagnostics encourages the translation
of this field from the bench to clinical practice.
By 2009, more than 68 genes had been analyzed for methylation in cervical tissues
and/or scrapings representing various stages of cervical cancer development as
reviewed by Wentzensen et al62. From this compilation of markers, the authors
concluded that three markers (DAPK, CADM1 and RARB2) consistently showed
elevated methylation in cervical cancers. The low concordance between studies for
25
Chapter 1
Figure 3. A) Cytosine methylation: Methylation of the 5th position on cytosine reveals the
most common methylated residual described as 5-methyl-cytosine or 5-mC. Catalyzed by
DNMTs, a methyl-group is at the 5th position of cytosine in the presence of SAM, then SAM
is converted to SAH. (Adapted from http://www.intechopen.com/books/methylation-from-dna-rna-
and-histones-to-diseases-and-treatment/dna-methylation-stem-cells-and-cancer)
B) DNA methylation occurs only at cytosines located 5’ to guanosine in the CpG

dinucleotide sequences. (Adapted from http://jonlieffmd.com/blog/networks-of-genes-respond-to-
social-experiences)
C) Methylation of CpG islands within the promoter region is associated with gene
inactivation.(Adapted from http://missinglink.ucsf.edu/lm/genes_and_genomes/methylation.html)
the other genes most likely reflects the use of different assays, assay thresholds
and/or selected promoter regions that were analyzed. However, methylation results
obtained from tissue samples may not be directly extrapolated to cervical scrapings.
The difference in cell type composition may display distinct levels of background
methylation63. Apart from samples error, some other limitations still hold back the
process for the DNA methylation markers applied in the clinical practice. First of all,
by application of single methylation markers on cervical scrapings, sensitivities for
cervical cancer of 90% or more has been achieved; however ,the positivity rate for
HSIL and CIN2 is generally lower64 63
. Secondly, the combined methylation analysis
of more genes in a panel has resulted in markedly increased sensitivity for HSIL.
Using cervical scrapings of referral populations, sensitivities of over 80% for CIN3+
were obtained with marker panels, JAM3/EPB41L3/TERT/C13ORF1865,
26
SOX1/PAX1/LMX1A/NKX6-166 and CADM1/MAL67, but these findings need further

validation10,63.Third, because of the difficulty in early detection of cervical AdCIS by
morphologic-based method, methylation of certain genes might be more specific for
cervical AdCIS. So far, few studies have focused on the DNA methylation associated
with ADC. Recently one study reported methylation of PAX1, PTPRR, SOX1 and
ZNF582 in ADC68. However, validation in a large set of samples is lacking.
These years, advances in whole genome profiling technologies have revolutionized

the field of cancer research. In our previous studies69, our approach to identify new
methylation markers were based on a pharmacological unmasking expression
microarray approach to enrich for genes that are silenced and re-expressed during
functional reversal of DNA methylation upon treatment with demethylation agents.
Nevertheless, antibody-based MeDIP is always variable and microarray-based
screening has the drawbacks including their design, production and somehow the
inaccurate hybridization signals. The use of next-generation-sequencing platforms to
identify novel methylation markers with more sensitivity and accuracy compared to
traditional microarray profiling seems very promising70. These technologies have
facilitated the discovery of potential biomarkers for cervical cancer development and
progression as demonstrated in this thesis.
Outline of this thesis
Currently, the HPV-based cervical cancer screening program is going to be as

primary testing in order to substitute the lower sensitivity and subjective interpretation
of cytomorphology-based screening in several European countries including the
Netherlands. However, the method applied for HPV DNA detection should fulfill the
criteria that are recommended in the international guideline. In addition, although
high sensitivity is obviously advantage of HPV-based screening, the lower specificity
is not, because it will result in relatively more referrals for colposcopy despite the lack
of CIN2+ lesions. Therefore, there is still room to improve the early diagnosis of
(pre)malignant disease in cervical scrapings, especially in the discovery of
biomarkers with both high sensitivity and specificity. Compared to microarray-based
27
Chapter 1
methods with defined potential methylation regions, the use of MethylCap-Seq, a

new innovative method based on a genome-wide screening approach resulting in a
specific methylome of the analyzed sample, could be promising to identify novel
biomarkers.
In China studies on the prevention of HPV-related disease, especially cervical cancer

is becoming more important over the last few years. At this moment, most of the
reports regarding the prevalence of HPV infection and the distribution of genotypes in
China were generated with scrapings obtained from either hospital-based
populations or collected in few provinces/cities. The precise information on the
occurrence of cervical HPV infection on the national level is still absent. In order to
lay the basis for the future HPV-based screening and vaccination, in Chapter 2 the
epidemiological data regarding the correlation of the HPV prevalence and genotype
with the age and regional distribution in China were analyzed on total 120,772 clinical
cervical samples collected in 37 cities of China in 2012.
The Cervista HR HPV test was the second hrHPV assay approved by the FDA. In
comparative studies, Cervista showed similar sensitivity and specificity as the HC2
test. However, Cervista has not previously been formally validated for the use in
primary cervical screening according the international guideline, especially in
Chinese population. Hereby, in Chapter 3, the clinical sensitivity and specificity of
the Cervista HR HPV test were compared to that of HC2 for detection of high-grade
cervical lesions (CIN2+) in women aged 30 years in >7,000 screening samples
selected from a Chinese population-based SHENCCAST II dataset by non-inferiority
analysis following the international guideline. In addition, the intra- and inter-
laboratory reproducibility of Cervista in 510 scraping following the international
guideline was determined.
In clinical practice, early diagnosis in the precancerous stage is most important for
cervical cancer prevention. Therefore, the aim of population-based screening for
cervical cancer is to detect all CIN 2/3 lesions. However, so far, no methylation
markers, including those identified by our own group previously, have been reported
to be sufficient sensitive and specific for the detection of CIN 2/3 lesions. In Chapter
28
4, a new approach to discover new methylation markers specific for high-grade

cervical intraepithelial neoplasia (CIN2/3) was used based on innovative genome-
wide methylation analysis (MethylCap-Seq) by comparison of the methylome of 20
normal cervices with 18 CIN2/3 lesions, followed by the validation of the diagnostic
performance of several new candidate methylation markers in cervical scrapings.
Currently, the incidence of SCC is declining in most developed countries. In contrast,

there is a rise in the absolute and relative incidence of cervical ADC, especially in
younger women. One explanation for the increase of ADC is the less effective
cytomorphologicl-based detection of ADC in population-based screening programs.
DNA methylation markers might be of help to improve the diagnostic limitations.
However, no specific methylation markers have been described for the detection of
ADC. In Chapter 5, MethylCap-Seq analysis was performed by comparison of the
methylome of 20 normal cervices with 6 ADC and 6 SCC to identify methylation
markers with a high sensitivity for both ADC and SCC, followed by the validation of
the diagnostic performance of several new candidate methylation markers in cervical
scrapings.
The results of chapters 2 to 5 will be summarized in Chapter 6 including some future

perspectives.
Reference
1 Ferlay J, Shin H-RR, Bray F, Forman D, Mathers C, Parkin DM. Estimates of worldwide burden
of cancer in 2008: GLOBOCAN 2008. Int J Cancer 2010; 127: 2893–917.
2 Castle PE, de Sanjosé S, Qiao Y-L, Belinson JL, Lazcano-Ponce E, Kinney W. Introduction of
human papillomavirus DNA screening in the world: 15 years of experience. Vaccine 2012; 30
Suppl 5: F117–22.
3 ICO (Institut Català d’Oncologia) Information Centre on HPV and cancer. Netherlands:Human
Papillomavirus and Related Cancers. 2014; 2013. http://www.hpvcentre.net/statistics/
reports/NLD_FS.pdf.
4 ICO (Institut Català d’Oncologia) Information Centre on HPV and Cancer. China: Human
Papillomavirus and Related Cancers. 2014; 2013. http://www.hpvcentre.net/statistics/
reports/CHN_FS.pdf.
29
Chapter 1
5 Shi J-F, Canfell K, Lew J-B, Qiao Y-L. The burden of cervical cancer in China: synthesis of the
evidence. Int J Cancer 2012; 130: 641–52.
6 Vizcaino AP, Moreno V, Bosch FX, Muñoz N, Barros-Dios XM, Parkin DM. International trends
in the incidence of cervical cancer: I. Adenocarcinoma and adenosquamous cell carcinomas.
Int J cancer 1998; 75: 536–45.
7 Vizcaino AP, Moreno V, Bosch FX, et al. International trends in incidence of cervical cancer: II.
Squamous-cell carcinoma. Int J cancer 2000; 86: 429–35.
8 Seoud M, Tjalma W, Ronsse V. Cervical adenocarcinoma: moving towards better prevention.

Vaccine 2011; 29: 9148–58.
9 Cervical cancer: Epidemiology, aetiology, pathogenesis and main histological types. .
10 Steenbergen RDM, Snijders PJF, Heideman DAM, Meijer CJLM. Clinical implications of
(epi)genetic changes in HPV-induced cervical precancerous lesions. Nat Rev Cancer 2014; 14:
395–405.
11 Wright TC, Ferenczy A and KR. Carcinoma and other tumors of the cervix. Blauwstein’s
Pathology of the Female Genital Tract. Now York: Springer-Verlag, 1994.
12 Gien L, Beauchemin M, Thomas G. Adenocarcinoma: a unique cervical cancer. Gynecol Oncol

2010; 116: 140–6.
13 Waggoner SE. Cervical cancer. Lancet 2003; 361: 2217–25.
14 Snijders PJF, Steenbergen RDM, Heideman D a M, Meijer CJLM. HPV-mediated cervical

carcinogenesis: concepts and clinical implications. J Pathol 2006; 208: 152–64.
15 Papanicolaou GN, Traut HF. The diagnostic value of vaginal smears in carcinoma of the uterus.
1941. Arch Pathol Lab Med 1997; 121: 211–24.
16 Solomon D, Davey D, Kurman R, et al. The 2001 Bethesda System: terminology for reporting
results of cervical cytology. JAMA 2002; 287: 2114–9.
17 McCluggage W. Premalignant lesions of the lower female genital tract: cervix, vagina and vulva.
Pathol RCPA 2013; 45: 214–28.
18 Park KJ, Soslow RA. Current concepts in cervical pathology. Arch Pathol Lab Med 2009; 133:
729–38.
19 Loureiro J, Oliva E. The spectrum of cervical glandular neoplasia and issues in differential
diagnosis. Arch Pathol Lab Med 2014; 138: 453–83.
20 Bertrand M, Lickrish GM, Colgan TJ. The anatomic distribution of cervical adenocarcinoma in
situ: implications for treatment. Am J Obstet Gynecol 1987; 157: 21–5.
21 Bigrigg A, Haffenden DK, Sheehan AL, Codling BW, Read MD. Efficacy and safety of large-
loop excision of the transformation zone. Lancet 1994; 343: 32–4.
30
22 Patra S, Panda D. Cervical cancer screening in developing countries. Indian J Cancer; 47:
344–5.
23 Peirson L, Fitzpatrick-Lewis D, Ciliska D, Warren R. Screening for cervical cancer: a systematic

review and meta-analysis. Syst Rev 2013; 2: 35.
24 Van Ballegooijen M, Hermens R. Cervical cancer screening in the Netherlands. Eur J Cancer
2000; 36: 2244–6.
25 Boers A, Slagter-Menkema L, van Hemel BM, et al. Comparing the Cervista HPV HR Test and
Hybrid Capture 2 Assay in a Dutch Screening Population: Improved Specificity of the Cervista
HPV HR Test by Changing the Cut-Off. PLoS One 2014; 9: e101930.
26 Cervical cancer screening in the Netherlands.

http://www.rivm.nl/en/Documents_and_publications/Common_and_Present/Newsmessages/20
14/Cervical_cancer_screening_in_the_Netherlands.
27 Belinson J, Qiao Y, Pretorius R, et al. Prevalence of cervical cancer and feasibility of screening
in rural China: a pilot study for the Shanxi Province Cervical Cancer Screening Study. Int J
Gynecol Cancer 1999; 9: 411–7.
28 Belinson JL, Qiao YL, Pretorius RG, et al. Shanxi Province cervical cancer screening study II:
self-sampling for high-risk human papillomavirus compared to direct sampling for human
papillomavirus and liquid based cervical cytology. Int J Gynecol Cancer; 13: 819–26.
29 Dai M, Bao YP, Li N, et al. Human papillomavirus infection in Shanxi Province, People’s
Republic of China: a population-based study. Br J Cancer 2006; 95: 96–101.
30 Wu R, Dai M, Qiao Y, et al. Human papillomavirus infection in women in Shenzhen City,

People’s Republic of China, a population typical of recent Chinese urbanisation. Int J Cancer
2007; 121: 1306–11.
31 Li J, Kang LN, Qiao YL. Review of the cervical cancer disease burden in mainland China.
Asian Pac J Cancer Prev 2011; 12: 1149–53.
32 Women’s health in rural China. Lancet 2009; 374: 358.
33 Cervical Cancer Prevention Program in China: Innovation in Action - Project HOPE.

http://www.projecthope.org/news-blogs/stories/cervical-cancer-prevention.html.
34 Wang, LD, Peng Y DZ. Ministry of Health of The People’s Republic of China. Technical plan for
early detection and treatment of cancer screening in China, 2011th edn. Beijing:People’s
Medical Publishing House.
35 Liu XY, Feng AH, Cui YM, Tobe RG. Prevention of human papillomavirus (HPV) infection and
cervical cancer in China: How does HPV vaccination bring about benefits to Chinese women?
Biosci Trends 2013; 7: 159–67.
36 Wright TC, Kuhn L. Alternative approaches to cervical cancer screening for developing
countries. Best Pract Res Clin Obstet Gynaecol 2012; 26: 197–208.
31
Chapter 1
37 Zur Hausen H. Papillomaviruses and cancer: from basic studies to clinical application. Nat Rev
Cancer 2002; 2: 342–50.
38 Bernard H-U, Burk RD, Chen Z, van Doorslaer K, zur Hausen H, de Villiers E-M. Classification
of papillomaviruses (PVs) based on 189 PV types and proposal of taxonomic amendments.
Virology 2010; 401: 70–9.
39 Tommasino M. The human papillomavirus family and its role in carcinogenesis. Semin Cancer
Biol 2014; 26C: 13–21.
40 Bouvard V, Baan R, Straif K, et al. A review of human carcinogens--Part B: biological agents.

Lancet Oncol 2009; 10: 321–2.
41 Cubie HA, Cuschieri K. Understanding HPV tests and their appropriate applications.
Cytopathology 2013; 24: 289–308.
42 Meijer CJLM, Berkhof J, Castle PE, et al. Guidelines for human papillomavirus DNA test
requirements for primary cervical cancer screening in women 30 years and older. Int J Cancer
2009; 124: 516–20.
43 Scholzen T, Gerdes J. The Ki-67 protein: from the known and the unknown. J Cell Physiol
2000; 182: 311–22.
44 De Freitas AC, Coimbra EC, Leitão M da CG. Molecular targets of HPV oncoproteins: potential
biomarkers for cervical carcinogenesis. Biochim Biophys Acta 2014; 1845: 91–103.
45 Petry KU, Schmidt D, Scherbring S, et al. Triaging Pap cytology negative, HPV positive cervical
cancer screening results with p16/Ki-67 Dual-stained cytology. Gynecol Oncol 2011; 121: 505–
9.
46 Jin XW, Lipold L, McKenzie M, Sikon A. Cervical cancer screening: what’s new and what's
coming? Cleve Clin J Med 2013; 80: 153–60.
47 Sanati S, Huettner P, Ylagan LR. Role of ProExC: a novel immunoperoxidase marker in the
evaluation of dysplastic squamous and glandular lesions in cervical specimens. Int J Gynecol
Pathol 2010; 29: 79–87.
48 Bosch F. Human papillomavirus: science and technologies for the elimination of cervical cancer.
Expert Opin Pharmacother 2011; : 2189–204.
49 Waddington CH. The epigenotype. 1942. Int J Epidemiol 2012; 41: 10–3.
50 Berger SL, Kouzarides T, Shiekhattar R, Shilatifard A. An operational definition of epigenetics.

Genes Dev 2009; 23: 781–3.
51 Feinberg AP, Tycko B. The history of cancer epigenetics. Nat Rev Cancer 2004; 4: 143–53.
52 Bestor TH. The DNA methyltransferases of mammals. Hum Mol Genet 2000; 9: 2395–402.
53 Bird A. DNA methylation patterns and epigenetic memory. Genes Dev 2002; 16: 6–21.
32
54 Jones PA. Functions of DNA methylation: islands, start sites, gene bodies and beyond. Nat
Rev Genet 2012; 13: 484–92.
55 Li J, Jin H, Wang X. Epigenetic biomarkers: potential applications in gastrointestinal cancers.

ISRN Gastroenterol 2014; 2014: 464015.
56 Feinberg AP, Vogelstein B. Hypomethylation distinguishes genes of some human cancers from
their normal counterparts. Nature 1983; 301: 89–92.
57 Baylin SB, Ohm JE. Epigenetic gene silencing in cancer - a mechanism for early oncogenic
pathway addiction? Nat Rev Cancer 2006; 6: 107–16.
58 Jiménez-Wences H, Peralta-Zaragoza O, Fernández-Tilapa G. Human papilloma virus, DNA

methylation and microRNA expression in cervical cancer (Review). Oncol Rep 2014; 31: 2467–
76.
59 Steenbergen RDM, Kramer D, Braakhuis BJM, et al. TSLC1 Gene Silencing in Cervical Cancer
Cell Lines and Cervical Neoplasia. JNCI J Natl Cancer Inst 2004; 96: 294–305.
60 Vasiljeviü N, Scibior-Bentkowska D, Brentnall AR, Cuzick J, Lorincz AT. Credentialing of DNA

methylation assays for human genes as diagnostic biomarkers of cervical intraepithelial
neoplasia in high-risk HPV positive women. Gynecol Oncol 2014; 132: 709–14.
61 Yang N, Nijhuis ER, Volders HH, et al. Gene promoter methylation patterns throughout the
process of cervical carcinogenesis. Cell Oncol 2010; 32: 131–43.
62 Wentzensen N, Sherman ME, Schiffman M, Wang SS. Utility of methylation markers in cervical
cancer early detection: appraisal of the state-of-the-science. Gynecol Oncol 2009; 112: 293–9.
63 Cuzick J, Bergeron C, von Knebel Doeberitz M, et al. New technologies and procedures for
cervical cancer screening. Vaccine 2012; 30 Suppl 5: F107–16.
64 Wang SS, Smiraglia DJ, Wu Y-Z, et al. Identification of novel methylation markers in cervical
cancer using restriction landmark genomic scanning. Cancer Res 2008; 68: 2489–97.
65 Eijsink JJH, Lendvai Á, Deregowski V, et al. A four-gene methylation marker panel as triage
test in high- risk human papillomavirus positive patients. Int J cancer 2012; 1869: 1861–9.
66 Lai H-C, Lin Y-W, Huang THM, et al. Identification of novel DNA methylation markers in
cervical cancer. Int J Cancer 2008; 123: 161–7.
67 De Strooper LM a, Hesselink BAT, Berkhof J, et al. Combined CADM1/MAL methylation and

cytology testing for colposcopy triage of high-risk HPV-positive women. Cancer Epidemiol
Biomarkers Prev 2014; published online June 24. DOI:10.1158/1055-9965.EPI-14-0347.
68 Chang C-C, Huang R-L, Wang H-C, Liao Y-P, Yu M-H, Lai H-C. High methylation rate of
LMX1A, NKX6-1, PAX1, PTPRR, SOX1, and ZNF582 genes in cervical adenocarcinoma. Int J
Gynecol Cancer 2014; 24: 201–9.
33
Chapter 1
69 Ongenaert M, Wisman GBA, Volders HH, et al. Discovery of DNA methylation markers in
cervical cancer using relaxation ranking. BMC Med Genomics 2008; 1: 57.
70 Hurd PJ, Nelson CJ. Advantages of next-generation sequencing versus the microarray in
epigenetic research. Brief Funct Genomic Proteomic 2009; 8: 174–83.
71 Yang L, Parkin DM, Li L, Chen Y. Sources of information on the burden of cancer in China.
Asian Pac J Cancer Prev 2003; 4: 23–30.
72 Li J, Huang R, Schmidt JE, Qiao YL. Epidemiological Features of Human Papillomavirus (HPV)
Infection among Women Living in Mainland China. Asian Pac J Cancer Prev 2013; 14: 4015–
23.
73 Belinson J, Qiao YL, Pretorius R, et al. Shanxi Province Cervical Cancer Screening Study: a
cross-sectional comparative trial of multiple techniques to detect cervical neoplasia. Gynecol
Oncol 2001; 83: 439–44.
74 Shen Y, Chen F, Huang M, et al. [Population-based study of human papillomavirus infection in

high-risk area for cervical cancer in Shanxi Province, China]. Zhongguo Yi Xue Ke Xue Yuan
Xue Bao 2003; 25: 381–5.
75 Zhao F-H, Forman MR, Belinson J, et al. Risk factors for HPV infection and cervical cancer
among unscreened women in a high-risk rural area of China. Int J Cancer 2006; 118: 442–8.
76 Zou L, Bao Y-P, Li N, et al. Life-style and genital human papillomavirus in a cross-sectional
survey in Shanxi Province, China. Asian Pac J Cancer Prev 2011; 12: 781–6.
77 Li L, Qiao Z, Zhang M, et al. [Study on the value assessment of various screening programs
regarding cervical cancer screening strategy in the rural areas of China]. Zhonghua Liu Xing
Bing Xue Za Zhi 2007; 28: 964–7.
78 Li LK, Dai M, Clifford GM, et al. Human papillomavirus infection in Shenyang City, People’s
Republic of China: A population-based study. Br J Cancer 2006; 95: 1593–7.
79 Wu E-Q, Liu B, Cui J-F, et al. Prevalence of type-specific human papillomavirus and pap
results in Chinese women: a multi-center, population-based cross-sectional study. Cancer
Causes Control 2013; 24: 795–803.
80 Lin M, Yang L-Y, Li L-J, Wu J-R, Peng Y-P, Luo Z-Y. Genital human papillomavirus screening
by gene chip in Chinese women of Guangdong province. Aust N Z J Obstet Gynaecol 2008; 48:
189–94.
81 Zhao R, Zhang WY, Wu MH, et al. Human papillomavirus infection in Beijing, People’s
Republic of China: a population-based study. Br J Cancer 2009; 101: 1635–40.
82 Li C, Zhang W, Zhang S, Wu M, Wang J. [Analysis of correlation factors for cervical

intraepithelial neoplasia in married women in Beijing, China]. Zhonghua Yi Xue Za Zhi 2010; 90:
196–200.
34
83 Ye J, Cheng X, Chen X, Ye F, Lü W, Xie X. Prevalence and risk profile of cervical Human

papillomavirus infection in Zhejiang Province, southeast China: a population-based study. Virol
J 2010; 7: 66.
84 Wang S, Wei H, Wang N, et al. The prevalence and role of human papillomavirus genotypes in
primary cervical screening in the northeast of China. BMC Cancer 2012; 12: 160.
85 Wu D, Cai L, Huang M, Zheng Y, Yu J. Prevalence of genital human papillomavirus infection

and genotypes among women from Fujian province, PR China. Eur J Obstet Gynecol Reprod
Biol 2010; 151: 86–90.
86 Hu S-Y, Hong Y, Zhao F-H, et al. Prevalence of HPV Infection And Cervical Intraepithelial
Neoplasia And Attitudes towards HPV Vaccination among Chinese Women Aged 18-25 in
Jiangsu Province. Chin J Cancer Res 2011; 23: 25–32.
87 Chen Q, Xie L-X, Qing Z-R, et al. Epidemiologic characterization of human papillomavirus
infection in rural Chaozhou, eastern Guangdong Province of China. PLoS One 2012; 7: e32149.
88 Zhang R, Shi T-Y, Ren Y, et al. Risk factors for human papillomavirus infection in Shanghai
suburbs: a population-based study with 10,000 women. J Clin Virol 2013; 58: 144–8.
35
Chapter 1
36
Nationwide prevalence of HPV infection in China
Chapter 2
Nationwide prevalence of human papillomavirus infection

in 37 cities in China
Rong Wang1,2* , Xiao-lei Guo3 *, G. Bea. A. Wisman2, Ed Schuuring4, Wen-feng Wang3,
Zheng-yu Zeng3, Hong Zhu5, Shangwei Wu3,1§
1
Division of Clinical Microbiology, School of Laboratory Medicine,
Tianjin Medical University, China
2
Department of Gynecologic Oncology, University of Groningen,
University Medical Center Groningen, Groningen, The Netherlands
3
Department of Microbiology, Kingmed Center For Clinical Laboratory, China
4
Department of Pathology, University of Groningen, University Medical Center Groningen,
Groningen, The Netherlands
5
Department of Epidemiology & Biostatistics, Tianjin Medical University, China
*These authors contributed equally to this work

§
Corresponding author
37
Chapter 2
Abstract
Background: Type-specific high-risk HPV (hrHPV) infection is related to cervical
carcinogenesis. The prevalence of hrHPV infection varies geographically, which may
reflect the epidemiological characteristics of cervical cancer in different populations.
To lay the basis for HPV-based screening and vaccination programs in China, we
investigated the latest HPV prevalence and genotypic distributions in different female
age groups and geographical regions in China.
Methods: In 2012, a total of 120,772 liquid-based cytological samples from women
enrolled for population- or employee-based cervical screening in 37 Chinese cities
were obtained by the Laboratory of Molecular Infectious Diseases of Guangzhou
KingMed. Among those samples, 111,131 were tested by Hybrid Capture II and the
other 9,641 were genotyped by TellgenplexTM HPV DNA Assay.
Results: The total positive rate of hrHPV was 21.07%, ranging from 18.42% to
31.94% depending on regions. Age-specific prevalence showed a “two peak” pattern,
with the youngest age group (15-19 yrs) presenting the highest hrHPV infection rate
(30.55%) followed by the second peak for the old age group of 50-60 yrs. Overall, the
most prevalent genotypes were HPV16 (4.82%) and 52 (4.52%), followed by HPV58
(2.74%). Two genotypes HPV6 (4.01%) and 11 (2.29%) were predominant in the low
risk HPV (lrHPV) type while mixed genotypes HPV16+52 and HPV52+58 were most
common in women with multiple infections.
Conclusions: This study shows that HPV infection in China has increased to the
level of an HPV-heavy-burden country zone, with the prevalence rates varying
significantly depending on ages and regions. The data from this study represents the
most recent update on the nationwide prevalence of HPV infection in China, which
can serve as valuable reference to guide nationwide cervical cancer screening and
HPV vaccination programs.
Keywords: Human papillomavirus, hrHPV prevalence, HPV genotyping, cervical

cancer, China
38
Background
HPV infection may cause a variety of genital diseases, and type-specific persistence
infection of high-risk HPV (hrHPV) is significantly relevant with the occurrence of
cervical carcinogenesis 1. Cervical cancer is the third most common cancer in women
worldwide 2. Effective implementation of cervical screening programs in developed
countries has resulted in a steady decline in the incidence of cervical cancer 3.
However, in China, the most populated country, cervical cancer remains the second
leading cause of cancer deaths among the females aged from 15 to 44-year old 4. It
is estimated that 75,434 women are diagnosed with cervical cancer yearly
(11.3/100,000) and 33,914 (45.0%) of those die of the cancer 5,6.
To date, more than 200 HPV genotypes have been identified, and approximately 40
HPV genotypes have been detected in the female genital tract. HPV16 and 18 are
well known as oncogenic genotypes, in addition, HPV31, 33, 35, 39, 45, 51, 52, 56,
58, 59, 68, 69 and 82 are also closely associated with cervical cancer, therefore, all
of those are termed as “high-risk” HPV. On the other hand, as ‘‘low-risk’’ genotypes,
HPV6, 11, 42, 43 and 44 are the causative agents for benign or low-grade changes
in the cervical cells, for instance, genital warts 7,8.
The HPV-based screening program is supposed to be an additional approach for

early diagnosis of cervical cancer in order to complement the less sensitivity and
3
non-objectiveness of cytology-based screening method . The high negative
predictive value of hrHPV testing is applicable for the indication of low-risk population,
in which the screening interval can be safely extended 3. Based on the results of
clinical trials, several European countries will implement hrHPV testing as the primary
screening modality 9. In addition to HPV screening, HPV vaccination has been
certified as an effective strategy against HPV infection and has been implemented in
most western countries recently. Although Cervarix (HPV16/18) and Gardasil
(HPV6/11/16/18) protect against infection with HPV16 and 18, these vaccines
provide no effect on some of the hrHPV types found in at least 25% of cervical
10
cancers . Furthermore, the role of non-vaccine HPV types in the development of
lesions is still unknown, it is possible that non-vaccine HPV types may replace the
39
Chapter 2
vaccine types as the causative agents for cervical precancerous and cancer in
vaccinated cohorts without sufficient broad cross-protection 10.
The prevalence of HPV infection and type-specific distribution vary greatly by

11 12
geographical areas , as well as in the different regions within a country . In
addition, several other risk factors may influence the prevalence of HPV, such as
genetic variation, sexual behavior (age at the first sexual intercourse and the
individual with multiple sex partners), biological predisposition of the immature cervix
13
and immunodeficiency . Hence, surveillance in the general population is needed to
assess the clinical benefit from screening and vaccination strategies.
To obtain large scale information on epidemiological feature of HPV in China is

crucial for the prevention strategies worldwide because of the diversities of
geography and age characteristics. Although a pilot project regarding the hrHPV
infection rate and cervical cancer screening were conducted in few Chinese areas in
14
1999 , and several similar investigations in larger scale and covering more regions
15-18
were respectively achieved in 2003, 2008, 2009 and 2012 , the data are still far
9
from a nationwide standard and the information should be updated with time in
order to provide reference for effective screening and vaccination as well. We hereby
perform a national wide cross-sectional investigation in a large scale for the following
purposes: 1) to reveal the HPV prevalence in the regions not yet investigated; 2) to
follow up the potential changes of HPV infection in the regions previously studied; 3)
to clarify the genotypic distribution of HPV in different regions and age-grouped
populations; thereby enriching the informative resources of HPV related with cervical
cancer for further study, screening and vaccination.
Methods
Ethics Statement
The study was approved by the Ethics Committee of Tianjin medical University in
accordance with the Ethical principles for Biomedical Research Involving Human
Subjects (Ministry of health of the people's republic of china) and Declaration of
Helsinki for Human Research of 1974 (last modified in 2000). Samples were
40
originally obtained from clinical settings for laboratory diagnosis. After diagnostic
testing, the excess samples were anonymized and kept for this study. An informed
consent was obtained from each of the women. For the individuals under 18 years,
the consent was signed by the parents.
Study population
Kingmed Diagnostics is the largest reference laboratory in China and provides
diagnostic testing services for over 13,000 hospitals in 18 provinces and 4
municipalities in the national wide. From January to December of 2012, in the total of
120,772 samples was obtained for population- or employee-based screening from 37
cities, belonging to 18 regions (Guangdong province was regarded as one region
since 20 cities in this investigation are located in that province). Figure 1 shows the
map of all the geographical sites in China and the 18 regions were further grouped
into 4 macro-geographical regions (East, West, South, and North).
41
Chapter 2
The women enrolled from population- or employee-based screening programs had

an age range from 15 to 60 year-old, were sexually active and no history of cervical
treatment before the screening. The exclusion criteria consisted of current pregnancy,
<3 months post-partum, HIV-seropositivity and a history of either hysterectomy or
treatment for cervical cancer.
Out of all the samples, 111,131 collected from 15 regions (Hainan, Chongqing, Jinan,
Jilin, Shenyang, Tianjin, Shanghai, Nanning, Guangdong, Guiyang, Fuzhou,
Hangzhou, Chengdu, Changsha and Jiangxi) (Fig. 1) were detected by use of Hybrid
Capture II (HCII) and 105,069 (94.5%) could be grouped by the ages. The other
9,641 samples from 10 regions (Shanghai, Guiyang, Xi’an, Guangdong, Nanning,
Changsha, Anhui, Kunming, Shenyang and Jilin) (Fig. 1) were genotyped using
Tellgenplex™ HPV DNA Test and 9,194 (95.4%) had age information.
Specimen collection
According to the protocols of practice, the cervico-vaginal cells at the transformation
zone of the uterine cervix were collected by a gynecologist or a trained gynecologist
assistant with a standard cytobrush (with spatula), then suspended into a standard
19
transport medium (STM) and stored at 4ć . All specimens coded without
knowledge of the subjects. Subsequently, all the samples were shipped to laboratory
of Kingmed Diagnostics for HPV tests within 24hrs.
Hybrid Capture II (HCII)

Liquid-based samples were processed by following the instruction included in the
Digene sample conversion kit and the hrHPV DNA panel of 13 pooled types (HPV16,
18, 31, 33, 35, 39, 45, 51, 52, 56, 58, 59, and 68) was examined with the HCII HPV
DNA test (Digene Corporation, Gaithersburg, Maryland). The results were
determined as a ratio of mean relative light unit (RLU) for the sample to the mean
RLU values for the assay of a positive calibrator (PC). A RLU-to-PC ratio >1 (~1.08
pg DNA/ml) was defined as a positive result.
42
TellgenplexTM HPV DNA Test

HPV genotyping was performed using the Tellgenplex™HPV DNA Test (Tellgen Life
Science Co. Ltd. Shanghai, China). The assay is able to detect and genotype 26
HPV genotypes including 19 hrHPV genotypes (HPV16, 18, 26,31, 33, 35, 39, 45,
51, 52, 53, 55, 56, 58, 59, 66, 68, 82, and 83) and 7 lrHPV genotypes (HPV6, 11, 40,
42, 44, 61 and 73). The Tellgenplex HPV ™ kit was applied to examine the presence
or absence of the most common 26 HPV genotypes simultaneously in a single test
20,21
by a multiplex PCR combined with Luminex technology . The 3 steps of DNA
extraction, PCR amplification and hybridization, were included in the procedure and
the template of 10~20 pg/ml HPV DNA was needed for each assay.
Statistical analysis
Region-specific prevalence of HPV: The HPV infection rate in each region was
calculated by dividing the number of HPV-positive samples by the total number of
samples successfully tested for HPV. A binomial 95% confidence interval (95% CI)
was estimated for each calculation of the HPV prevalence. Chi-square (Ȥ²) tests were
used to compare the differences among all the regions and each two regions.
Age-specific prevalence: The HPV infection rate was estimated within 5 age groups
(15-, 20-, 30-, 40- and 50-60). A binomial 95% confidence interval (95% CI) was
estimated, P value for age trend of HPV infection was analyzed by using the linear-
by-linear association test. The difference between each two age group was
compared by Chi-square (Ȥ²) test. Multiple comparisons were performed using the
22
Bonferoni step-down procedure to minimize the inÀated risk of type 1 error and
P<0.05 as statistically significant.
HPV-type-specific prevalence: The frequency of each hrHPV and lrHPV genotype

was presented in hrHPV positive samples and lrHPV positive samples respectively.
All statistical analyses were conducted using SPSS20.0 software (SPSS20, lnc.,
Chicago, IL).
43
Chapter 2
Results
The total prevalence of hrHPV infection
The total hrHPV infection rate was 21.07% (95%Cl 20.83%~21.31%) and the range
was between 18.42% and 31.94% upon different regions. The prevalence of hrHPV
infection differed significant among the various regions (P<0.001). The regions with
highest hrHPV prevalence were Hainan (31.94%) and Chongqing (27.29%), followed
by Jinan, Shenyang, Jilin and Tianjin. Relatively, Jiangxi, Changsha, Hangzhou,
Chengdu, Fuzhou, Guangdong and Guiyang could be classified into the low burden
regions (Table 1).
Table 1 Region-specific prevalence of hrHPV infection by HCII
Positive Total Infection 95%CI of infection

Regions
samples samples Rate(%) Rate(%)
Hainan 221 692 31.94 28.47 - 35.41

Chongqing 191 700 27.29 23.99- 30.59
Jinan 2651 10306 25.72 24.88- 26.57
Shenyang 98 387 25.32 20.99- 29.65
Jilin 360 1423 25.3 23.04- 27.56
Tianjin 807 3220 25.06 23.57- 26.56
Shanghai 118 522 22.61 19.02- 26.20
Nanning 1976 8869 22.28 21.41- 23.15
Guiyang 597 2919 20.45 18.99- 21.91
Guangdong 14567 72763 20.02 19.73- 20.31
Fuzhou 441 2213 19.93 18.26- 21.59
Chengdu 375 1886 19.88 18.08- 21.68
Hangzhou 649 3269 19.85 18.49- 21.22
Jiangxi 290 1574 18.42 16.51- 20.34
Total 23413 111131 21.07 20.83- 21.31
HrHPV infection rate was different among all the regions using Chi-square test ( P<0.001). Subsequently,
multiple comparisons were further performed using the Bonferoni step-down procedure : marginal
difference between two most heavy-burden cities of Hainan and Chongqing (P=0.057), no difference
among the second group cities including Jinan, Shenyang, Jilin and Tianjin (P=0.892), as well as the less
infection rate group with Jiangxi, Changsha, Hangzhou, Chengdu, Fuzhou, Guangdong and Guiyang
(P=0.758).
44
As depicted in Figure 2, hrHPV infection was likely relevant with age. The group of
15-year old showed the highest prevalence (30.55%), followed by the group of 50-60
(23.30%). The total infection rate of hrHPV was associated with age (P<0.001, Table
S1). Next to the highest peak, the prevalence declined from 22.17% in the 20-year
group to 19.71% in the 30-year group. Subsequently, the infection rate significantly
increased again in 50-60year group over the 40-year group (P<0.001, Table S1).
Among the four macro-regions, only the south showed the similar trend with the
overall age-specific hrHPV prevalence, whereas the other three macro-geographic
regions did not show significant differences in all five age groups (Figure 2, Table S1).
The distribution of variant HPV genotypes

Among the well-recognized 26 HPV genotypes, all of the genotypes but HPV73 were
examined using Tellgenplex technique. The total infection rate of each genotype in all
of the samples and the distributions of each genotype in the HPV positive portion
were respectively evaluated. In descending order, the infection rate and distribution
are shown in Table 2: the most common hrHPV types were HPV16 (18.02%) and
HPV52 (16.9%) (P=0.288), followed by HPV58 (10.23%), then the third group,
including HPV59, 56, 39, 18 and 68 (P=0.373) shared the similar proportion.
45
Chapter 2
Relatively, HPV6 (45.80%), HPV11 (26.15%) and HPV61 (14.20%) were common in
lrHPV.
Table 2 The prevalence of each HPV genotype by Tellgenplex HPV DNA Test
HrHPV Positive Infection Rate Proportion

Genotypes samples (in 9641 total samples)% (in 2580 hrHPV positive samples)%
HPV16 465 4.82 18.02
HPV52 436 4.52 16.90
HPV58 264 2.74 10.23
HPV59 158 1.64 6.12
HPV56 157 1.63 6.09
HPV39 154 1.60 5.97
HPV18 143 1.48 5.54
HPV68 135 1.40 5.23
HPV51 111 1.15 4.30
HPV33 105 1.09 4.07
HPV31 81 0.84 3.14
HPV66 74 0.77 2.87
HPV82 69 0.72 2.67
HPV55 60 0.62 2.33
HPV53 59 0.61 2.29
HPV45 44 0.46 1.71
HPV35 38 0.39 1.47
HPV83 19 0.20 0.74
HPV26 8 0.08 0.31
LrHPV Positive Infection Rate Proportion
Genotypes samples (in 9641 total samples)% (in 844 lrHPV positive samples)
HPV6 387 4.01 45.80
HPV11 221 2.29 26.15
HPV61 120 1.24 14.20
HPV40 46 0.48 5.56
HPV44 46 0.48 5.44
HPV42 24 0.25 2.84
Significant difference among the proportion of all the genotype using Chi-square (Ȥ²) test (P<0.0001).
Multiple comparisons were further performed using the Bonferoni step-down procedure, the proportion of
the most common genotype HPV16 and 52 (P=0.288), as well as the proportion of HPV59, 56, 39,18 and
68 (P=0.583) were no significant difference.
46
For the region-specific distribution of hrHPV, the top three genotypes were analyzed
in each of the regions. HPV16, 58 and 52 were dominant in six regions, i.e., Guiyang,
Xi’an, Guangdong, Nanning, Changsha and Shenyang although the orders of the
three genotypes may vary in different regions (Table S2a). However, different top
three patterns were observed: such as HPV16, 18 and 83 in Shanghai, HPV16, 33
and 82 in Anhui, HPV16, 56 and 59 in Kunming, as well as HPV16, 52 and 58 in Jilin
(Table S2a). For lrHPV, HPV11 was the most common genotype in Shanghai and
Kunming, while HPV6 was the most frequent genotype in all the other regions (Table
S2b). The distribution of top three HPV genotypes was also determined on the age
basis. For hrHPV, HPV16, 52 and 58 were dominant among all of the age groups
except the group of 15-year old, in which HPV52, 16 and 59 were the major
genotypes (Table S3). As to lrHPV, HPV6 was the leading genotype in all age groups,
and the second commonly detected genotype was HPV11 in the younger age groups
(i.e., the groups of 15-, 20-, 30-), while in the older groups (40-, 50-60) HPV61 was
the most prevalent lrHPV type (Table S3).
Infection with multiple HPVs was detected in a total of 486 specimens (5.04%),
among which 434 (16.82%) and 52 (6.16%) were infected with hrHPVs and lrHPVs
respectively. In the multiple hrHPV infections, the frequencies of 6, 5, 4, 3 and 2
genotypes were 0.23%, 1.84%, 4.61%, 17.74% and 75.58%, respectively, and three
genotypes showed higher positive rates, i.e., HPV16 (35.02%), 52 (32.26%) and 58
(21.20%) (Fig3.). The top two double-agents were detected in the decline order of
HPV16+52 (26 cases) and HPV52+58 (14 cases) (Table S3.). On the regional basis,
Shanghai had the highest incidence of multiple infections (19.51%), followed by Jilin
(12.34%) and Nanning (6.25%) (Fig 4).Multiple infections were examined in 52
multiple lrHPV infections. The proportion of 4, 3 and 2 genotypes were 1.92%, 3.85%,
and 94.23%, respectively. Nanning showed the highest incidence of multiple
infections (3.13%), followed by Jilin (2.92%) and Xi’an (1.40%) (Fig4). Similar to the
overall age trend, within 472 cases with age information, after the first peak in the
group of 15- years old, another high peak was observed in the group of 50-60years
old both for hrHPV and lrHPV (P<0.001) ( Fig S1.)
47
Chapter 2
FIG3.The prevalence of each genotype hrHPV in multiple infections by Tellgenplex HPV DNA Test
Discussion
This is one of the few nation-wide investigations on high and low risk HPV in a large
scale of Chinese screening population. Because of the regional difference, the
population composition and the sampling period around the world, the prevalence
varies study by study, the heavily burdened HPV regions are Sub-saharan Africa
(24.0%), Eastern Europe (21.4%), Latin America (16.1%), and Southeastern Asia
23
(14%) . In this surveillance, the overall hrHPV positive rate was 21.07% (95%Cl
48
20.83%~21.31%), which has increased into the levels of HPV-heavy-burden

countries and higher than the global average level.
According to the population-based screening results previously reported, the overall

prevalence of hrHPV varies from 9.9%~27.5%, respectively15. The highest infection
rate is in Shanxi, a cervical cancer heavy burden region in China. And the lowest one
is in Beijing, the capital of China with prosperous economic and better healthcare
system. In addition to the “Top and down”, others are around 15%~20%, and a pool
analysis including 17 populations from 9 regions shows the positive rates of hrHPV
24
were 17.7% . Together, the overall hrHPV positive rate in this surveillance is
increasing slightly.
Compared with the region-based data, the results obtained in the present study are
25,26 25
higher than those previously reported for Shanghai , Shenyang , Guangdong
27,28 29
and Hangzhou (Table S5). In addition, some newly studied regions in this
surveillance showed high hrHPV incidences, for instance, Hainan (31.94%) and
Chongqing (27.29%), and several cities in north, including Jinan, Jilin and Tianjin.
The data revealed that the hrHPV infection is becoming serious in both the increase
of infection rate in the same regions and some of those reaching to the level of heavy
burden regions as well. Of course, the great improvement in screening strategy and
laboratory methods could also partially contribute to the increased prevalence.
As to the age-relevant hrHPV prevalence, a meta-analysis conducted by Bruni et al 23

showed a bimodal age distribution, with the peak of HPV infection occurring within a
younger age group (just after sexual debut). Globally, a lower prevalence plateau in
the middle age group, followed by a gradual reduction of incidence in developing
23
countries or a second peak in developed countries . The trend of hrHPV infection
showing high in younger and low in middle age groups reflects the natural history of
HPV infection. Young females are sensitive to HPV soon after beginning of sexual
activity due to immature immune protection, nevertheless most of HPV infection are
usually temporary; thus in 70% and 91% of women infected with HPV, the virus
30
may be cleared within one or two years respectively . The slight increase of the
HPV infection rate in older females might reflect the viral persistence or the
49
Chapter 2
reactivation of latent HPV assumingly because of the physiological and

immunological disorders resulted from hormone fluctuations during the menopausal
transition 31. In this investigation, the general age distribution showed the first peak of
hrHPV in the age group of 15 years (30.55%), then gradually decreasing in the
23
middle age, which is consistent with the data from Bruni et al . However, hrHPV
infection rate was significantly increased in the age group of 50-60 years compared
to the women in 40s, a similar trend seen in most developed countries. The
prosperous economy in most of the metropolis may have influence on the culture and
sexual behavior of Chinese people. Nonetheless, the exact mechanisms for the
increase of HPV prevalence remain unclear at this moment.
Information concerning the distribution of HPV genotypes is important not only for
vaccine development but also for HPV-based screening design, particularly for
32
selecting the testing spectrum of HPV genotypes and multiple HPV infection
detection. It is a prerequisite for the genotyping assays in cervical cancer screening
programs. However, HCII, the only Food and Drug Administration (FDA) approved
test cannot provide any genotype-specific information33. Therefore, Tellgenplex™
HPV DNA Test, a reported genotyping method21, was performed in this surveillance.
In the in-house clinical validation according to CAP(College of American Pathologists)
standards, the correlation between Tellgenplex™ HPV DNA Test and HCII was
shown acceptable (Coincident rate 90.5%, Kappa=0.88).
Consistent with the results yielded from some of Chinese population-specific

34
investigations, HPV16, 52 and 58 are the dominant hrHPV types . Compared with
25
the data by Wu et al , a population-based investigation from 5 regions (Beijing,
Shanghai, Xinjiang, Henan and Shangxi) in China, the infection rates of HPV16
(4.82%), 52 (4.52%) and 58 (2.74%) were all higher than the HPV16 (2.9%), 52
(1.7%) and HPV58 (1.5%) obtained in this study. Whereas, as to the proportion rate
of these three genotypes, besides HPV52 (16.9% &11.9%: P=0.006), which is higher
in the present study and there are no significant differences in the proportion of
HPV16 (20.2%&18.02%: P=0.245) and 58 (10.8 %&10.23%: P=0.686). The
proportion of HPV16 in this surveillance was even lower than that in the normal
cytology samples reported by Guan et al and Bruni et al (20.4% and 22.5%,
50
23,35
respectively) . Relatively, HPV52 and 58 accounted for 27.13%, which is
35
dramatically higher than the global level of 14.37% . Interestingly, although both
HPV52 and 58 were all common in Asian population, the significance of the two
36
genotypes is still unknown. Zhao et al reported that HPV52 infections are more
common among healthy individuals, whereas HPV58 is reported to be related to
cervical cancer. Some studies conducted in the South and West regions in China,
only including CIN or cancer samples, have demonstrated that HPV58 is more
prevalent than HPV52 18,27,37,38. HPV18, next to HPV16, is known to be most important
39
for cervical carcinogenesis . However, in our study it was at the seventh position,
and the infection rate (1.48%) is in line with the study of Wu et al. (P=0.871) 25.
In addition to carcinogenesis, many benign cutaneous warts, mucosal lesions and

low-grade cervical intraepithelial lesions generate a considerable health burden
40
associated with lrHPV infection . Specifically, HPV types 6 and 11 cause 90% of
genital warts, over 95% of recurrent respiratory papillomatosis cases, and
41
approximately 10% of early cervical lesions . The infection will become more
40
serious in immune-compromised individuals . In our surveillance, the incidence of
23
HPV6 (4.01%) infection was inconsistent with the results of Bruni et al , showing
that HPV6 is most frequent in Americans (2.9%) and is less frequent in Asian
individuals (0.2%), followed by HPV11 and 61.
Characterization of the prevalence of multiple HPV infections might be important for

10
the effect on cervical carcinogenesis. Herrero et al reported that women infected
with HPV16 alone were at similar or higher risk for cervical cancer than those
42
infected with both HPV16 and another HPV type. Lee et al reported an association
between infection with multiple HPV types and the increased risk of cervical cancer.
In a recent study, Schmitt et al. confirmed that co-infection will increase the infection
duration. Furthermore, patients with multiple high viral loads showed a 4- to 6-fold
increased risk of cervical precancerous cytological lesions than patients with single
43
high viral loads . In the present study, 434 hrHPV-positive samples (4.5%) were
multiple-infections. The incidence is in the same range ( 3.5%-5.3%) as reported in
China34, but higher than the global average (3.2%) 23
and the proportion (14.19%) is
25 23
lower than both domestic (25.8%) and international (20%) . The most common
51
Chapter 2
combinations of two types were HPV16+52 (26 cases) and HPV 52+58 (14 cases),
and the genotypes 52 and 58 were more likely involved in co-infections. From the
region-specific surveillance, some geographical features were observed, for instance
in Shanghai, a metropolis with internalized population, showing a situation close to
the world in the HPV prevalence, the genotypic distribution and multiple infections.
This study has confirmed the high incidence of HPV in overall China and strongly
argues the necessity for developing the national population-based screening
programs. However, the appropriate management of this HPV screening program for
the large number of women with HPV-positive specimens and no cytological
44
evidence of cervical pre-cancer or cancer remains a major concern . HPV
genotyping could be an option to stratify the HPV positive women. Simultaneously, a
stainable HPV detection and closely follow-up for the hrHPV carried women should
be implemented, thus more cost-effective techniques, for instance, CervistaTM HR
HPV test, COBAS HPV test and some other genotyping platforms might be the
alternatives for molecular HPV detection 45.
Conclusions
In conclusion, China has large population and a variety of territories; meanwhile,
economic conditions, cultural habits and population migrations have affected on
Chinese daily life and health situation dramatically, for instance, the cancers related
with sexual transmitted diseases. Therefore, attention should be paid to the
prevalence of HPV infection on the timely and regional basis because it has been
commonly recognized that the HPV infection plays a crucial role in the occurrence of
cervical cancer and the increase incidence. This surveillance results have indicated
that a national plan for cervical screening program is urgently needed, not only
because the increase in infection rate in some previously reported regions, but also
the high infection rate in most of newly investigated regions. Shortly, the most
significant discoveries by this investigation are following: (1). the prevalence of
hrHPV infection has reached a level that could not be ignored, and the prevalence
increase is ongoing with time obviously; (2). the prevalence of hrHPV infection
showed population variations on age and region as well as reflected economic,
cultural and lifestyle relevance; (3). the HPV16, 52 and 58 constituted the three
52
dominant genotypes consistently in different Chinese populations, a characteristic

pattern significantly different from the epidemiological features in most of industrial
countries, which defined the principles for cervical screening and vaccination in
China should be proposed distinguishably.
List of abbreviations
HPV: Human papillomavirus; hrHPV: high-risk HPV; lrHPV: low risk HPV; HCII:
Hybrid Capture II; CAP: College of American Pathologists
Competing interests
Shangwei Wu is Medical Director of Kingmed Diagnostics. E. Schuuring is a member
of the scientific advisory board of Roche, Hologic and QCMD, received travel
reimbursements from Roche, Abbott, Hologic Inc. and QCMD.
Authors’ contributions
RW, XLG WFW and ZZY carried out the molecular genetic studies, participated in the
sequence alignment and drafted the manuscript. GBAW, ES and HZ participated in
the design of the study and performed the statistical analysis. SWW and RW
conceived of the study, and participated in its design and coordination and helped to
draft the manuscript. All authors read and approved the final manuscript.
Acknowledgements
Rong Wang is appointed to a collaborative project between University of Groningen
in the Netherlands and Tianjin medical University of China. This study was supported
by the grant from natural science foundation of Tianjin (12JCYBJC33700).
References
1. Zur Hausen H: Papillomaviruses in human cancers. Mol Carcinog 1:147-50, 1988
2. Jemal A, Bray F, Center MM, et al: Global cancer statistics. CA Cancer J Clin 61:69-90, 2011
3. Priebe AM: 2012 cervical cancer screening guidelines and the future role of HPV testing. Clin
Obstet Gynecol 56:44-50, 2013
53
Chapter 2
4. Human Papillomavirus and Related Cancers. WHO/ICO HPV information centre summary
report
5. Ferlay J, Shin HR, Bray F, et al: Estimates of worldwide burden of cancer in 2008: GLOBOCAN
2008. Int J Cancer 127:2893-917, 2010
6. Internation Agency for Research on Cancer(IRAC). IARC handbooks of cancer prevention:

cervix cancer screening,chapter 2. Lyon:IARC Press
7. de Villiers EM, Fauquet C, Broker TR, et al: Classification of papillomaviruses. Virology 324:17-
27, 2004
8. Munoz N, Bosch FX, de Sanjose S, et al: Epidemiologic classification of human papillomavirus

types associated with cervical cancer. N Engl J Med 348:518-27, 2003
9. Castle PE, de Sanjose S, Qiao YL, et al: Introduction of human papillomavirus DNA screening
in the world: 15 years of experience. Vaccine 30 Suppl 5:F117-22, 2012
10. Herrero R: Human papillomavirus (HPV) vaccines: limited cross-protection against additional
HPV types. J Infect Dis 199:919-22, 2009
11. de Sanjose S, Diaz M, Castellsague X, et al: Worldwide prevalence and genotype distribution
of cervical human papillomavirus DNA in women with normal cytology: a meta-analysis.
Lancet Infect Dis 7:453-9, 2007
12. Krul EJ, Van De Vijver MJ, Schuuring E, et al: Human papillomavirus in malignant cervical
lesions in Surinam, a high-risk country, compared to the Netherlands, a low-risk country. Int J
Gynecol Cancer 9:206-211, 1999
13. Vinodhini K, Shanmughapriya S, Das BC, et al: Prevalence and risk factors of HPV infection
among women from various provinces of the world. Arch Gynecol Obstet 285:771-7, 2012
14. Belinson J, Qiao Y, Pretorius R, et al: Prevalence of cervical cancer and feasibility of
screening in rural China: a pilot study for the Shanxi Province Cervical Cancer Screening
Study. Int J Gynecol Cancer 9:411-417, 1999
15. Li J, Huang R, Schmidt JE, et al: Epidemiological features of Human Papillomavirus (HPV)
infection among women living in Mainland China. Asian Pac J Cancer Prev 14:4015-23, 2013
16. Zhao R, Zhang WY, Zhang SW, et al: [Study on subtype of human papillomavirus infection
among aged 25 - 54 reproductive women in Beijing from 2006 to 2008]. Zhonghua Fu Chan
Ke Za Zhi 46:184-7, 2011
17. Wu D, Cai L, Huang M, et al: Prevalence of genital human papillomavirus infection and
genotypes among women from Fujian province, PR China. Eur J Obstet Gynecol Reprod Biol
151:86-90, 2010
18. Li J, Mei J, Wang X, et al: Human papillomavirus type-specific prevalence in women with
cervical intraepithelial neoplasm in Western China. J Clin Microbiol 50:1079-81, 2012
54
19. Han J, Swan DC, Smith SJ, et al: Simultaneous amplification and identification of 25 human
papillomavirus types with Templex technology. J Clin Microbiol 44:4157-62, 2006
20. Wei W, Shi Q, Guo F, et al: The distribution of human papillomavirus in tissues from patients
with head and neck squamous cell carcinoma. Oncol Rep 28:1750-6, 2012
21. Hu Y, Qian HZ, Sun J, et al: Anal human papillomavirus infection among HIV-infected and
uninfected men who have sex with men in Beijing, China. J Acquir Immune Defic Syndr
64:103-14, 2013
22. Ludbrook J: Multiple comparison procedures updated. Clin Exp Pharmacol Physiol 25:1032-7,
1998
23. Bruni L, Diaz M, Castellsague X, et al: Cervical human papillomavirus prevalence in 5

continents: meta-analysis of 1 million women with normal cytological findings. J Infect Dis
202:1789-99, 2010
24. Zhao FH, Lewkowitz AK, Hu SY, et al: Prevalence of human papillomavirus and cervical
intraepithelial neoplasia in China: a pooled analysis of 17 population-based studies. Int J
Cancer 131:2929-38, 2012
25. Wu EQ, Liu B, Cui JF, et al: Prevalence of type-specific human papillomavirus and pap results
in Chinese women: a multi-center, population-based cross-sectional study. Cancer Causes
Control 24:795-803, 2013
26. Zhang R, Shi TY, Ren Y, et al: Risk factors for human papillomavirus infection in Shanghai
suburbs: A population-based study with 10,000 women. J Clin Virol, 2013
27. Chen Q, Xie LX, Qing ZR, et al: Epidemiologic characterization of human papillomavirus
infection in rural Chaozhou, eastern Guangdong Province of China. PLoS One 7:e32149, 2012
28. Wu RF, Dai M, Qiao YL, et al: Human papillomavirus infection in women in Shenzhen City,
People's Republic of China, a population typical of recent Chinese urbanisation. Int J Cancer
121:1306-11, 2007
29. Ye J, Cheng X, Chen X, et al: Prevalence and risk profile of cervical Human papillomavirus
infection in Zhejiang Province, southeast China: a population-based study. Virol J 7:66, 2010
30. Baseman JG, Koutsky LA: The epidemiology of human papillomavirus infections. J Clin Virol
32 Suppl 1:S16-24, 2005
31. Hildesheim A, Wang SS: Host and viral genetics and risk of cervical cancer: a review. Virus
Res 89:229-40, 2002
32. Smith JS, Lindsay L, Hoots B, et al: Human papillomavirus type distribution in invasive cervical
cancer and high-grade cervical lesions: a meta-analysis update. Int J Cancer 121:621-32,
2007
33. Cho EJ, Do JH, Kim YS, et al: Evaluation of a liquid bead array system for high-risk human
papillomavirus detection and genotyping in comparison with Hybrid Capture II, DNA chip and
sequencing methods. J Med Microbiol 60:162-71, 2011
55
Chapter 2
34. chung-yung chen c-hw: Human papillomavirus type distribution in southern china and taiwan,.
human papillomavirus and related diseases-from bench to bedside-research aspects, 2012
35. Guan P, Howell-Jones R, Li N, et al: Human papillomavirus types in 115,789 HPV-positive

women: a meta-analysis from cervical infection to cancer. Int J Cancer 131:2349-59, 2012
36. Zhao FH, Tiggelaar SM, Hu SY, et al: A multi-center survey of age of sexual debut and sexual
behavior in Chinese women: suggestions for optimal age of human papillomavirus vaccination
in China. Cancer Epidemiol 36:384-90, 2012
37. Wang S, Wei H, Wang N, et al: The prevalence and role of human papillomavirus genotypes in
primary cervical screening in the northeast of China. BMC Cancer 12:160, 2012
38.Liu SS, Chan KY, Leung RC, et al: Prevalence and risk factors of Human Papillomavirus (HPV)
infection in southern Chinese women - a population-based study. PLoS One 6:e19244, 2011
39. zur Hausen H: Papillomaviruses in the causation of human cancers - a brief historical account.
Virology 384:260-5, 2009
40. Leto M, Santos Junior GF, Porro AM, et al: Human papillomavirus infection: etiopathogenesis,
molecular biology and clinical manifestations. An Bras Dermatol 86:306-17, 2011
41. Preaud E, Largeron N: Economic burden of non-cervical cancers attributable to human

papillomavirus: a European scoping review. J Med Econ, 2013
42. Lee SA, Kang D, Seo SS, et al: Multiple HPV infection in cervical cancer screened by
HPVDNAChip. Cancer Lett 198:187-92, 2003
43. Schmitt M, Depuydt C, Benoy I, et al: Multiple HPV infections with high viral loads are
associated with cervical lesions but do not differentiate grades of cervical abnormalities. J Clin
Microbiol, 2013
44. Gage JC, Ajenifuja KO, Wentzensen NA, et al: The age-specific prevalence of human
papillomavirus and risk of cytologic abnormalities in rural Nigeria: implications for screen-and-
treat strategies. Int J Cancer 130:2111-7, 2012
45. Munson E, Du Chateau BK, Bellerose B, et al: Clinical laboratory evaluation of Invader(R)
chemistry and hybrid capture for detection of high-risk human papillomavirus in liquid-based
cytology specimens. Diagn Microbiol Infect Dis 71:230-5, 2011
56
Table S1. Age-specific HPV infection in each region by HCII
regions 15-yrs 20-yrs 30-yrs 40-yrs 50-60yrs
Shanghai 0.00% 22.81% 20.33% 27.27% 0.00%

Jiangxi 25.00% 21.79% 16.60% 18.15% 23.66%
Hangzhou 19.44% 18.83% 19.88% 19.82% 21.45%
East 18.75% 20.62% 18.71% 19.25% 22.18%
Chengdu 11.63% 22.55% 19.69% 18.32% 15.71%
Guiyang 29.52% 17.88% 21.86% 20.86% 17.09%
Chongqing 30.77% 29.02% 24.43% 26.35% 34.29%
West 25.29% 20.98% 21.27% 20.59% 19.01%
Guangdong 33.11% 21.76% 18.15% 18.85% 23.26%
Nanning 21.15% 22.25% 22.22% 22.32% 23.48%
Hainan 36.36% 22.13% 29.49% 37.30% 34.85%
Changsha 50.00% 23.15% 13.60% 15.46% 25.00%
Fuzhou 19.44% 20.91% 20.08% 20.13% 13.95%
South 31.67% 21.81% 18.72% 19.51% 23.19%
Tianjin 28.30% 26.78% 23.19% 25.10% 23.67%
Jilin 37.50% 25.30% 24.21% 24.28% 30.16%
Jinan 18.64% 25.34% 26.36% 26.18% 24.40%
Shenyang 71.43% 17.11% 25.42% 24.19% 26.32%
North 25.51% 25.58% 25.58% 25.73% 24.81%
Total 30.55% 22.17% 19.71% 20.55% 23.30%
In Total. P value for age trend of HPV infection was analyzed by using the linear-by-linear
association test (P<0.001). Furthermore, Infection rate is difference in each age group in addition
to the margin difference in the age group of 20-yrs and 50-60yrs (P=0.051).
In the Four marco-geographic regions, in addition to the South(p<0.001), the other three including
the East, West and North were all no significant.(P=0.401, 0.595, 0.974) in each age group.
57
Chapter 2
58
59
Chapter 2
Table S2b. Region-specific lr HPV subtype prevalence

by Tellgenplex™ HPV DNA Test
Total positive HPV6 HPV11 HPV40 HPV42 HPV44 HPV61
Shanghai 41 12 4 6 0 0 0 2
Positiverate 29.27% 9.76% 14.63% 0.00% 0.00% 0.00% 4.88%
100.00% 33.33% 50.00% 0.00% 0.00% 0.00% 16.67%
East 41 12 4 6 0 0 0 2
29.27% 9.76% 14.63% 0.00% 0.00% 0.00% 4.88%
100.00% 33.33% 50.00% 0.00% 0.00% 0.00% 16.67%
Guiyang 1069 70 26 20 5 2 6 11
Infectionrate 6.55% 2.43% 1.87% 0.47% 0.19% 0.56% 1.03%
proportion 100.00% 37.14% 28.57% 7.14% 2.86% 8.57% 15.71%
Xi'an 357 59 32 15 2 1 6 3
Infectionrate 16.53% 8.96% 4.20% 0.56% 0.28% 1.68% 0.84%
proportion 100.00% 54.24% 25.42% 3.39% 1.69% 10.17% 5.08%
West 1426 129 58 35 14 12 7 3
9.05% 4.07% 2.45% 0.98% 0.84% 0.49% 0.21%
100.00% 44.96% 27.13% 10.85% 9.30% 5.43% 2.33%
Guangdong 4516 224 90 54 13 5 19 43
Infectionrate 4.96% 1.99% 1.20% 0.29% 0.11% 0.42% 0.95%
proportion 100.00% 40.18% 24.11% 5.80% 2.23% 8.48% 19.20%
Nanning 64 7 5 2 0 0 0 0
Infectionrate 10.94% 7.81% 3.13% 0.00% 0.00% 0.00% 0.00%
proportion 100.00% 71.43% 28.57% 0.00% 0.00% 0.00% 0.00%
Changsha 1398 98 48 23 5 2 7 13
Infectionrate 7.01% 3.43% 1.65% 0.36% 0.14% 0.50% 0.93%
proportion 100.00% 48.98% 23.47% 5.10% 2.04% 7.14% 13.27%
Anhui 42 3 1 2 0 0 0 0
Infectionrate 7.14% 2.38% 4.76% 0.00% 0.00% 0.00% 0.00%
proportion 100.00% 33.33% 66.67% 0.00% 0.00% 0.00% 0.00%
Kunming 155 13 2 4 0 1 0 6
Infectionrate 8.39% 1.29% 2.58% 0.00% 0.65% 0.00% 3.87%
proportion 100.00% 15.38% 30.77% 0.00% 7.69% 0.00% 46.15%
South 6175 345 146 85 18 8 26 62
5.59% 2.36% 1.38% 0.29% 0.13% 0.42% 1.00%
100.00% 42.32% 24.64% 5.22% 2.32% 7.54% 17.97%
Shenyang 1075 288 146 81 19 9 5 28
Infectionrate 26.79% 13.58% 7.53% 1.77% 0.84% 0.47% 2.60%
proportion 153.19% 77.66% 43.09% 10.11% 4.79% 2.66% 14.89%
Jilin 924 70 33 14 2 4 3 14
Infectionrate 7.58% 3.57% 1.52% 0.22% 0.43% 0.32% 1.52%
proportion 100.00% 47.14% 20.00% 2.86% 5.71% 4.29% 20.00%
North 1999 358 179 95 21 13 8 42
17.91% 8.95% 4.75% 1.05% 0.65% 0.40% 2.10%
100.00% 50.00% 26.54% 5.87% 3.63% 2.23% 11.73%
60
61
Chapter 2
62
Table S4. The common combinations of double infections by Tellgenplex™

HPV DNA Test
Type1 Type2 Positivecases

HPV16 HPV52 26
HPV52 HPV58 14
HPV16 HPV59 13
HPV52 HPV39 11
HPV16 HPV58 10
HPV16 HPV33 10
HPV52 HPV68 9
HPV16 HPV56 8
HPV16 HPV39 7
HPV16 HPV68 7
HPV16 HPV51 7
HPV16 HPV66 6
HPV52 HPV59 6
HPV52 HPV18 6
HPV58 HPV56 6
HPV16 HPV18 5
HPV16 HPV82 5
HPV52 HPV56 5
HPV58 HPV18 5
HPV58 HPV33 5
HPV58 HPV66 5
HPV59 HPV56 5
HPV39 HPV68 5
HPV39 HPV53 5
HPV68 HPV31 5
FIG S1. Age-specific multiple infection by Tellgenplex™ HPV DNA Test
63
Chapter 2
Table S5 . HPV Prevalence in Women from population-based Screening Studies
Author Study Location Number(N) Age Lab- overall

/Pub year year Range assays HPV
(yrs) prevalence
(%)
(Belinson 1999 Xiangyuan 1,997 35-45 HC2 18.2
et
al.2001)
(Zhao et 2001- Xiangyuan and 8,798 35-50 HC2 23.6
al., 2001) 2002 Yangcheng,shanxi
(Shen et 2001- Xiangyuan and 9,683 30-50 HC2 27.5
al., 2003) 2002 Yangcheng,shanxi
(Zou et 2004 Yangcheng 745 15-59 HC2 16.0
al., 2011)
(Li et al., 2004 Xiushui, Jiangxi 2,432 30-49 HC2 18.5
2007)
(Li et al., 2004- Shenyang 685 15-59 GP5+/6+ 16.8
2006) 2005 mediated
PCR
(Dai et 2004- Yangcheng 662 15-59 GP5+/6+ 14.8
al., 2006) 2005 mediated
PCR
(Wu et 2004- Shenzhen 1,027 15-59 GP5+/6+ 16.6
al., 2007) 2005 mediated
PCR
(Zhao et 1999- Shanxi, 13,004 16-54 HC2 17.7
al., 2008 Beijing(18.8),
2012b) Xinjiang(8.2),
Henan(13.03),
Shanghai(18.66)
(Li et al., 2006- Beijing 6,185 25-54 HC2 9.9
2010) 2009
(Ye et 2007- zhejiang 5,058 20-79 PCR 13.3
al.,2010 2008 (MY09/11
primer)
(Hu et al., 2009 Jiangsu 316 18-25 HC2 17.1
2011)
(Zhang et 2011 Shanghai suburb 10,000 17-89 PCR 12.6
al.,2013) (MY09/11
primer)
Compared with the region-based data, the results obtained in the present study are higher than those
previously reported for Shanghai (this cohort & previously: 22.61% & 18.66% in urban and 12.6% in
suburbs)25,26, Shenyang (25.32% & 16.8%)25, Guangdong (20.02% & 16.6% in urban and 9.03% in
rural)27,28 and Hangzhou (19.85% & 13.3%)29.
64
Clinical validation of the Cervista HPV HR test according to the international guidelines
Chapter 3
Clinical validation of the Cervista HPV HR test according to

the international guidelines for human papillomavirus test
requirements for cervical cancer screening
Aniek Boers1, Rong Wang1, Lorian Slagter-Menkema2, Bettien M. van Hemel2, Hilde
Ghyssaert3, Ate G.J van der Zee1, G. Bea A. Wisman1, Ed Schuuring2#
1
Department of Obstetrics and Gynecology, Division of Gynecologic Oncology,
University of Groningen, University Medical Center Groningen, the Netherlands.
2
Department of Pathology and Medical Biology, University of Groningen,
University Medical Center Groningen, the Netherlands
3
Department of Pathology, AZ St Jan Brugge-Oostende, Brugge, Belgium
J. Clin. Microbiol. 2014,52(12):4391.
65
Chapter 3
Abstract
This study demonstrates that both the clinical sensitivity and specificity of the Cervista
HPV HR test for high-risk human papillomavirus (HPV) detection are not inferior to those
of the Hybrid Capture 2 (HC2) test. The intra- and interlaboratory reproducibilities of
Cervista were 92.0% (kappa, 0.83) and 90.4% (kappa, 0.80), respectively. The Cervista HPV
HR test fulfills all the international HPV test requirements for cervical primary screening
purposes.
It is well established that cervical cancer is caused by the persistent infection of

cervical epithelial cells by any of the ~14 genotypes of high-risk HPV (hrHPV). This
knowledge prompted the development of in vitro diagnostic tests for hrHPV testing in
clinical specimens. Generally, these tests have a high sensitivity and high negative
predictive value, making them potentially valuable tools for primary screening
strategies. Systematic reviews have shown that primary screening using hrHPV
testing has a higher sensitivity than that of cytology for detecting cervical
intraepithelial neoplasia (CIN) grade 2 or greater (CIN2+) (1, 2). Although many
current international guidelines limit HPV testing to the triage of borderline lesions
and to post-CIN follow-up, it is believed that in the near future, HPV testing will be
included as a viable strategy for primary screening (3). In line with the international
guidelines for HPV DNA testing in primary cervical cancer screening in women 30
years described by Meijer et al. (4), the recently updated guidelines from the
American Society for Colposcopy and Cervical Pathology (ASCCP) emphasize the
importance of using a validated HPV test, i.e., an HPV test that has proven
acceptable reproducibility, clinical sensitivity, specificity, and positive and negative
predictive values for cervical cancer screening of CIN2+ lesions (5).
The Cervista HPV HR test (Hologic Inc., Madison, WI) was the second hrHPV assay
approved by the FDA in 2009, 10 years after the approval of the Hybrid Capture 2
hrHPV DNA (HC2) test. The Cervista HPV HR assay uses Invader chemistry, a
signal amplification method to qualitatively detect specific nucleic acid sequences of
14 hrHPV types (HPV16,-18,-31,-33,-35,-39,-45,-51,-52,-56,-58,-59,-66 and -68),as
described previously (6), and it utilizes a primary reaction that produces a fluorescent
66
signal. Both types of reactions rely on oligonucleotide hybridization, invasive

structure formation, and cleavage by the Cleavase enzyme. Interpretation of the HPV
results was done in accordance with the Cervista HPV HR test package insert (7).
The Cervista test has a few advantages over the HC2 test: it detects the same
hrHPV types as the HC2 test plus HPV66, it requires half the sample volume of that
of the HC2 test, it includes the human histone 2 gene as an internal control for
sample adequacy, it demonstrates no cross-reactivity with common nononcogenic
HPV types, and it has a shorter processing time. In comparative studies, the Cervista
HPV HR test shows sensitivity and specificity results that are similar to those of the
HC2 test (8–17), but studies comparing both assays on the same samples in a
population-based screening setting are limited (9–11). Here, we evaluated the
Cervista HPV HR test according the international guidelines for HPV DNA testing in
primary cervical cancer screening in women30 years by performing the validation
process in strict accordance with that recommended by Meijer et al. (4)
The clinical performance of the Cervista HPV HR test was assessed relative to that of
the HC2 test using data from the SHENCCAST II study, a large cohort of screening
participants originally screened by cotesting using ThinPrep cytology and hrHPV
testing, applying both the Cervista HR HPV and HC2 HPV tests, all performed on the
same sample. A detailed description of the SHENCCAST II study, a multisite,
population-based, and cross-sectional study conducted in Guangdong Province in
China, which enrolled approximately 10,000 women, 25 to 59 years old, was reported
previously (10). To calculate the relative clinical specificity and sensitivity, 7,218
samples without CIN2+ lesions and 109 samples with CIN2+ lesions, respectively,
were used for a noninferiority analysis of the Cervista assay versus the HC2 assay.
The overall clinical specificities of the HC2 and the Cervista assays in 7,218 women
aged 30 years without CIN2+ (controls) were similar, at 89% (95% confidence
interval [CI], 88.0 to 89.5) and 91% (95% CI, 90.5 to 91.8), respectively (Table1). To
calculate the relative sensitivity on a representative population-based screening
cohort, 78 randomly selected samples with abnormal cytology (atypical cells of
undetermined significance [ASCUS]) and 31 samples with normal cytology (negative
for intraepithelial lesion or malignancy [NILM]), all with histologically proven CIN2+
lesions (45 with CIN2, 61 with CIN3, and 3 with carcinoma), were selected (cases)
67
Chapter 3
from the SHENCCAST II data set. The overall clinical sensitivities of the HC2 and
Cervista assays for detecting CIN2+ in women age30 years were 94% (95% CI,
87.2 to 97.4) and 89% (95% CI, 81.6 to 94.2), respectively (Table 1). The
noninferiority of the relative sensitivity and specificity of the Cervista HPV HR test
versus the HC2 test was confirmed, as the null hypothesis of inferiority was rejected
(t = 17.73, P˘0.0001 for specificity; t = 1.76 and P = 0.043 for sensitivity). Therefore,
the Cervista HPV HR test met the criterion of noninferiority set forth by the
international guidelines, i.e., it had a clinical sensitivity not less than 90% of the
sensitivity of the HC2 test and a clinical specificity not less than 98% of the specificity
of the HC2 test for detecting CIN2+ in women age 30 years.
Table 1. Comparison of the Cervista HPV HR and HC2 test findings among
7,327 scrapings collected in the multisite, population-based, and
cross-sectional SHENCCAST II study
HC2 HPV test result

(no.of samples)
Sample type Cervista HPV
Total no. of
(CIN score) HR result Positive Negative samples
Controls(<2) Positive 518 119 637
Negative 291 6,290 6,581a
Total 809 6,409b 7,218
Cases(2+) Positive 97 0 97c

Negative 5 7 12
Total 102d 7 109
a
The overall clinical specificity for the Cervista HPV HR test was 91% (95%CI, 90.5 to 91.8).
b
The overall clinical specificity for the HC2 test was 89%(95%CI,88.0 to 89.5)
c
The overall clinical sevsitivity for the Cervista HPV HR tests was 89% (95%CI, 81.6 to 94.2)
d
The overall clinical sensitivity for the HC2 test was 94% (95% CI,87.2 to 97.4)
The intra- and interlaboratory reproducibilities of the Cervista assay were evaluated
on 510 cervical scraping samples selected from women age 30 to 60 years
participating in the routine national population-based cervical screening program in
the Netherlands. A detailed description of the sample selection and analysis of the
intra- and interlaboratory reproducibilities is reported elsewhere (13). These samples
comprised 186 HC2-positive and 324 HC2-negative randomly selected scrapings
68
(36% hrHPV positivity), according to the international guidelines for HPV DNA testing.
To determine the intralaboratory reproducibility, all 510 samples were tested twice
(University Medical Center Groningen [UMCG] test 1 and UMCG test 2) with the
Cervista assay, according to the manufacturer’s product insert (7), at a 1- to 3-week
interval by the same experienced technician on the same Cervista system at the
Department of Pathology of the University Medical Center Groningen (UMCG). The
agreement between the two test results was 92.0% (lower bound of the 95% CI,
89.7%; kappa, 0.83; P˘0.001) (Table 2). For the interlaboratory agreement, an
aliquot (2 ml of PreservCyt) of the same samples was sent to an independent
reference laboratory in Bruges (Department of Pathology, AZ St. Jan Brugge-
Oostende, Bruges, Belgium), which routinely uses the Cervista assay. All samples
were randomly renumbered and provided to the reference laboratory without any
cytology or hrHPV test results. The agreement between the two laboratories was
90.4% (lower bound of the 95% CI, 88.4%; kappa, 0.80; P˘0.001) (Table 2). Thus,
the intra- and interlaboratory agreements of the Cervista assay met the requirement
of having a lower bound of the 95% CI of ˚87% and a kappa value of ˚0.5.
Table 2. Results of the intralaboratory reproducibility and interlaboratory

agreement of the Cervista HPV HR test among 510 samplesa
No.with UMCG test I result

Cervista HPV Low
Test HR result Positive Negative gDNAb Total no.
UMCG test 2 Positive 174 17 0 191
results Negative 24 293 1 318
Low gDNA 0 0 1 1
Total 198 310 2 510
Bruges test Positive 179 35 0 214

results Negative 12 281 1 294
Low gDNA 0 1 1 2
Total 191 317 2 510
a
The agreement between the two test results(UMCG test 1 and UMCG test2) was 92.0%(lower bound
of 95% CI, 89.7%; kappa,0.83; P<0.001). The agreement between the two independent laboratories
(UMCG test 1 compared with the Bruges test) was 90.4% (lower bound of 95% CI, 88.4%;
kappa,0.80;P<0.001). bgDNA,genomic DNA.
69
Chapter 3
The present study validates the Cervista HPV HR test for use in primary screening
for the detection of CIN2+ lesions in women age 30 years, in accordance with the
international guidelines for HPV DNA testing in primary screening, and it includes a
determination of: (i) the noninferiority of the Cervista test to the reference HC2 HPV
test and (ii) the intra- and interlaboratory reproducibilities of the Cervista assay. The
Cervista test met the criteria for noninferiority to the HC2 test (noninferiority test) set
forth in the international guidelines (4). Thus, the Cervista HPV HR test fulfills all the
requirements of the international guidelines and can be considered formally validated
for the use of primary cervical cancer screening in women of age 30 years.
Recently, we found that the specificity of the Cervista HPV HR test could be even
further improved when the standard second cutoff (default setting of the
manufacturer) was adapted (13).
ACKNOWLEDGMENTS
E.S. is on the scientific advisory board of Roche, Hologic, Inc., and QCMD and
received travel reimbursements from Roche, Abbott, Hologic, Inc., and QCMD. A.B.,
L.S.-M., and B.M.V.H. received travel reimbursements from Hologic, Inc. All other
authors declare no conflicts of interest.
The Cervista HPV HR test reagents for the intra-/interlaboratory reproducibility testing
were kindly provided by Hologic, Inc.; the funding sources did not have any influence
on the design of the study or on analysis of the results. The SHENCCASTII data
were kindly provided by J. Belinson (President, Preventive Oncology International
and Professor of Surgery, Cleveland Clinic Lerner College of Medicine, Cleveland
Clinic, OH, USA).
References:
1. Cuzick, J, Clavel, C, Petry, KU, Meijer, CJ, Hoyer, H, Ratnam, S, Szarewski, A, Birembaut, P,
Kulasingam, S, Sasieni, P, Iftner, T. 2006. Overview of the European and North American studies
on HPV testing in primary cervical cancer screening. Int. J. Cancer. 119:1095-1101.
70
2. Whitlock, EP, Vesco, KK, Eder, M, Lin, JS, Senger, CA, Burda, BU. 2011. Liquid-based cytology
and human papillomavirus testing to screen for cervical cancer: a systematic review for the U.S.
Preventive Services Task Force. Ann. Intern. Med. 155:687-97, W214-5.
3. Giorgi Rossi, P, Ronco, G. 2013. The present and future of cervical cancer screening programmes
in Europe. Curr. Pharm. Des. 19:1490-1497.
4. Meijer, CJ, Berkhof, J, Castle, PE, Hesselink, AT, Franco, EL, Ronco, G, Arbyn, M, Bosch, FX,
Cuzick, J, Dillner, J, Heideman, DA, Snijders, PJ. 2009. Guidelines for human papillomavirus DNA
test requirements for primary cervical cancer screening in women 30 years and older. Int. J. Cancer.
124:516-520.
5. Massad, LS, Einstein, MH, Huh, WK, Katki, HA, Kinney, WK, Schiffman, M, Solomon, D,
Wentzensen, N, Lawson, HW, 2012 ASCCP Consensus Guidelines Conference. 2013. 2012
Updated Consensus Guidelines for the Management of Abnormal Cervical Cancer Screening Tests
and Cancer Precursors. Obstet. Gynecol. 121:829-846.
6. Day, SP, Hudson, A, Mast, A, Sander, T, Curtis, M, Olson, S, Chehak, L, Quigley, N, Ledford, J,
Yen-Lieberman, B, Kohn, D, Quigley, DI, Olson, M. 2009. Analytical performance of the
Investigational Use Only Cervista HPV HR test as determined by a multi-center study. J. Clin. Virol.
45 Suppl 1:S63-72.
7. ThirdWave Technologies.2008. Cervista HPV HR package insert. Third Wave Technologies Inc.,
Madison, WI.
8. Du Chateau, BK, Schroeder, ER, Munson, E. 2013. Clinical laboratory experience with cervista
HPV HR as a function of cytological classification: comparison with retrospective digene HC2 high-
risk HPV DNA test data. J. Clin. Microbiol. 51:1057-1058.
9. Kurian, EM, Caporelli, ML, Baker, S, Woda, B, Cosar, EF, Hutchinson, L. 2011. Cervista HR and
HPV 16/18 assays vs hybrid capture 2 assay: outcome comparison in women with negative cervical
cytology. Am. J. Clin. Pathol. 136:808-816.
10. Belinson, JL, Wu, R, Belinson, SE, Qu, X, Yang, B, Du, H, Wu, R, Wang, C, Zhang, L, Zhou, Y, Liu,
Y, Pretorius, RG. 2011. A population-based clinical trial comparing endocervical high-risk HPV
testing using hybrid capture 2 and Cervista from the SHENCCAST II Study. Am. J. Clin. Pathol.
135:790-795.
11. Quigley, NB, Potter, NT, Chivukula, M, Knight, MZ, Welch, JR, Olson, MC. 2011. Rate of detection
of high-risk HPV with two assays in women >/= 30 years of age. J. Clin. Virol. 52:23-27.
71
Chapter 3
12. Ginocchio, CC, Barth, D, Zhang, F. 2008. Comparison of the Third Wave Invader human
papillomavirus (HPV) assay and the digene HPV hybrid capture 2 assay for detection of high-risk
HPV DNA. J. Clin. Microbiol. 46:1641-1646.
13 Boers A, Slagter-Menkema L, van Hemel BM, Belinson JL, Ruitenbeek T, Buikema HJ, Klip H,
Ghyssaert H, van der Zee AGJ, de Bock GH, Wisman GBA, Schuuring E. Comparing the Cervista
HPV HR test and Hybrid Capture 2 assay in a Dutch screening population: improved specificity of
the Cervista HPV HR test by changing the cut-off. Accepted for publication in PLos One.
14. EinsteinMH, Garcia FA, Mitchell AL, Day SP.2011. Age-stratified performance of the Cervista HPV
16/18 genotyping test in women with ASCUS cytology. Cancer Epidemiol. Biomarkers
Prev.20:1185–1189.
15. ZhaoJ, Zhang X, Ma J, Liu G, Yao D, Zhang W, Wang J, Wei L, Zhao Y, Zeng Y, Liao Q.2012.
Clinical performance characteristics of the Cervista HPV HR test kit in cervical cancer screening in
China. J. Low.Genit. Tract Dis.16:358–363.
16. GoldMA, Thomas MA, Huh WK, Sarto GE, Day SP.2013. High-risk human papillomavirus
detection in women with low-grade squamous intraepithelial lesions or higher-grade cytology using
the Cervista HPV HR test. J. Low. Genit. Tract Dis. 17:51–57.
17. YouensKE, Hosler GA, Washington PJ, Jenevein EP, Murphy KM. 2011. Clinical experience with
the Cervista HPV HR assay: correlation of cytology and HPV status from 56,501 specimens. J.
Mol. Diagn.13:160–166.
72
Discovery of new methylation markers to improve screening for CIN 2/3
Chapter 4
Discovery of new methylation markers to improve

screening for cervical intraepithelial neoplasia grade 2/3
Boers A.1, Wang R.1, van Leeuwen R.W.1, Klip H.G.1, de Bock G.H.2, Hollema H.3, van
Criekinge W.4, de Meyer T.4, Denil S.4, van der Zee A.G.J.1, Schuuring E.3, Wisman G.B.A.1
1 Department of Gynecologic Oncology, University of Groningen, University Medical Center

Groningen, the Netherlands
2 Department of Epidemiology, University of Groningen, University Medical Center Groningen,
the Netherlands
3 Department of Pathology, University of Groningen, University Medical Center Groningen,
the Netherlands
4 Department of Molecular Biotechnology, Ghent University, Ghent, Belgium
73
Chapter 4
Abstract
Aims: To identify new methylation markers for high-grade cervical intraepithelial neoplasia
(CIN2/3) using innovative genome-wide methylation analysis and to assess their diagnostic
performance in cervical scrapings.
Methods: Enrichment and capturing of methylated DNA from normal cervices and CIN2/3
lesions followed by next-generation sequencing (MethylCap-Seq) was performed to identify
differential methylation regions (DMRs). The top 15 highest ranking differentially methylated
genes were selected and validated by MSP in two steps: on the same DNA samples as used
for MethylCap-Seq and on DNA samples from an independent patient cohort with
(pre)malignant cervical neoplasia. For further diagnostic evaluation, the best differentiating
methylation markers were tested with quantitative MSP (QMSP) in cervical scrapings from 2
cohorts: 1) cervical carcinoma vs. healthy controls and 2) patients referred from population-
based screening with an abnormal Pap smear in whom HPV status was determined.
Results: With genome-wide MethylCap-Seq, 176 DMRs comprising 163 genes were
identified. After verification and validation of the top 15 genes with MSP, 9 genes showed
significant differential methylation in normal cervices versus CIN2/3 lesions (p<0.05).
Subsequently, methylation levels of 8/9 genes were significantly higher in carcinoma
compared to normal scrapings. For all 8 genes methylation levels increased with the severity
of the underlying histological lesion in scrapings from patients with an abnormal Pap smear.
In addition to the 8 new genes, also our previous four-gene panel (C13ORF18, JAM3,
EPB41L3 and TERT) was analyzed. The best combination of genes
(C13ORF18/JAM3/AL590705.4) revealed sensitivity (74%) for CIN2+ comparable to hrHPV
testing (79%), while specificity was significantly higher (76% vs 46%, p0.05) in a triage
setting after a positive Pap smear test in population-based screening.
Conclusion: We identified new CIN2/3 specific methylation markers using a genome-wide
DNA methylation analysis. The diagnostic performance of our new methylation panel shows
comparable sensitivity to hrHPV testing for CIN2+, but with higher specificity to prevent
referral for unnecessary colposcopy. The next step before implementation in primary
screening programs will be validation in population-based cohorts.
74
Introduction
Cervical cancer is characterized by a well-defined pre-malignant phase, cervical
intraepithelial neoplasia (CIN). Identification of these CIN lesions by population-
based screening programs and their subsequent treatment has led to a significant
1,2
reduction of the incidence and mortality of cervical cancer . Cytology-based testing
of cervical smears is the most widely used cervical cancer screening method, but is
3-5
not ideal, as the sensitivity for detection of CIN2 and higher (CIN2+) is only ~55% .
Cervical carcinogenesis is highly associated with high-risk human papillomavirus
(hrHPV) 6. Large randomized-controlled trials have shown that the sensitivity of
4,7-10
hrHPV testing is significantly higher than cytology testing . However, the
specificity of hrHPV testing, especially in a young screening population is relatively
low 3,11-13, which may lead to unnecessary referrals to the gynecologist, anxiety in the
false-positive women, and higher costs for the health-care system. Finally, in the
near future the prevalence of CIN and cervical cancer will probably decrease in
countries that have introduced primary prevention with hrHPV vaccination. With this
decrease in prevalence, the positive predictive value of the current screening tests
14
will by definition decrease . Therefore, other objective biomarkers with both high
sensitivity as well as high specificity are needed as new screening tools for cervical
cancer.
Different DNA methylation patterns in normal versus (pre)malignant lesions represent

excellent targets for diagnostic approaches based on methylation specific PCR
(MSP). Promoter hypermethylation of tumor suppressor genes is an early event in
cervical carcinogenesis and consequently hypermethylation analysis can be
15-17
especially relevant for the early detection of cervical neoplasia . Assessment of
methylation markers in cervical scrapings for the detection of CIN and cervical cancer
17-23
is feasible , but finding methylation markers with both high sensitivity as well as
high specificity remains a challenge. Through the years gradually more sophisticated
approaches have been developed to identify new methylation markers on a genome-
24
wide scale . Amidst comparable studies from other groups we have previously
reported our experience with pharmacological unmasking of the promoter region
combined with re-expression as analyzed by microarrays, high-throughput
quantitative methylation specific PCR (QMSP) on an OpenArray platform and methyl-
75
Chapter 4
DNA immunoprecipitation followed by microarray analysis (MeDIP), resulting in the

discovery and validation of the genes C13ORF18, JAM3, EPB41L3 and TERT 21,22,25.
The diagnostic performance of these genes showed sensitivities for detecting CIN2+
in a hrHPV positive population between 43%-71% and specificities between 89%-
21
100% . However, our strategies for discovering new methylation markers so far
were based on the difference between cancer and normal tissue resulting in markers
with high sensitivity for carcinoma, but with too low sensitivity for detecting CIN2/3
lesions. In our MeDIP study DNA methylomes of normal and CIN3 lesions were
25
analyzed . However, a disadvantage of this technique is that it primarily recognizes
bulk quantities of highly methylated repetitive DNA, resulting in less specificity. New
and more specific innovative genome-wide methylation analysis of DNA from CIN2/3
lesions versus normal cervical tissue should result in (new) CIN2/3 sensitive and
specific methylation markers. Methylated-CpG island recovery assay uses antibody-
coupled methyl-binding domain (MBD2) proteins to specifically purify methylated
26
DNA . The higher affinity of the MBD2 complex for double-stranded CpG-
methylated DNA results in a higher enrichment for methyl DNA sequences as
compared to MeDIP analysis. Next-generation-sequencing then shows the identified
novel methylated regions (MethylCap-Seq). After identification of novel methylation
markers for (pre)malignant cervical neoplasia through this approach, validation and
diagnostic evaluation of these newly found markers can be performed.
The aim of the present study was 1) to identify new methylation markers that can
differentiate between normal cervices and CIN2/3 lesions using MethylCap-Seq and
2) to validate the diagnostic performance of the newly found methylation markers in
cervical scrapings by QMSP.
Patients and Methods

General strategy
To characterize the DNA methylome of CIN2/3 lesions and to identify new CIN2 or
higher (CIN2+) methylation markers, we applied the following strategy (see figure 1):
76
MethylCap-Seq :
Identification
18 CIN2/3 lesions versus 20 normal cervices
MSP frozen tissue:

Verification
Correlating MSP bands with MethylCap-Seq data
MSP new patient cohort :

Validation
13 carcinoma, 19 CIN2/3, 17 normal cervices
QMSP scrapings: 1st Diagnostic evaluation

100 carcinoma versus 89 normal cervices
QMSP scrapings with abnormal cytology:

No CIN (n=27), CIN1 (n=38), CIN2 (n=49), CIN3 2nd Diagnostic evaluation
(n=57), miCa (n=44)
Figure 1: Flow scheme for the identification of new CIN2+ methylation markers
First, methylated DNA was enriched using MBD2 proteins with subsequent paired-
end sequencing (MethylCap-Seq) on DNA isolated from fresh-frozen macro-
dissected epithelial tissue of 18 CIN2/3 lesions (6 CIN2 and 12 CIN3), 20 normal
cervices and two pools of leukocyte DNA of healthy volunteers. In order to identify
differential methylated regions (DMRs), we retrieved the reads of promoter and exon
regions. We selected methylation markers that showed significant differences
between the normal and CIN2/3 cervices, while also the leukocyte count had to be
low, to prevent false-positive results. Markers were ranked on high specificity (no
methylation in the normal cervices) and high sensitviy (methylation in CIN2/3 lesions).
For the highest ranking top15 genes, methylation specific PCR (MSP) primers were
designed and methylation patterns were verified on the same DNA, which originally
was used for MethylCap-Seq. This first validation step enabled verification of
MethylCap-Seq data by correlating MSP band intensity with the number of reads
77
Chapter 4
from the MethylCap-Seq. In the second validation step high prevalence of

methylation in the CIN2/3 lesions and no methylation in the normal cervices was
analyzed by MSP analysis on DNA isolated from a completely independent cohort of
patients (cervical cancer (n=13), CIN2/3 lesions (n=19) and normal cervices (n=17)).
DNA was isolated from macro-dissected formalin fixed paraffin embedded (FFPE)
epithelial tissue.
Finally, diagnostic evaluation of the newly discovered methylation markers was

performed by QMSP on cervical scrapings. First, we tested the methylation ratios of
new biomarkers on a large series of randomly selected scrapings from cervical
cancer patients (n=100) and a similar age group of healthy controls (n=89). Secondly,
the potential of the new methylation markers as a diagnostic tool was evaluated in a
large series of scrapings (n=215) of randomly selected patients, referred with an
abnormal Pap smear at population-based screening. Histology was used as the
reference standard.
Patient samples
All patients referred to the outpatient clinic of the University Medical Center
Groningen (UMCG) with cervical cancer or an abnormal Pap smear at population-
based screening are routinely asked to participate in our ongoing ‘Methylation study’
which has been approved by the Institutional Review Board (IRB) of the UMCG.
Cervical tissue, scrapings and clinicopathologic data are prospectively collected and
stored in our tissue bank. Within our Methylation study tissue samples, scrapings and
clinicopathologic data from normal cervices are also collected from patients planned
to undergo a hysterectomy for non-malignant reasons. All cervical tissue that was
used for the normal control group was judged as histopathological normal. Patients
referred with cervical cancer are staged according to the FIGO criteria with pelvic
examination and biopsies under general anaesthesia. Cervical scrapings from both
groups (cervical cancer staging and benign gynecologic surgery) were collected
before surgery under general anaesthesia. All patients referred with an abnormal Pap
smear at population-based screening underwent an additional Pap smear prior to
colposcopy specifically for this study. At colposcopy, biopsies and/or Large Loop
Excision of the Transformation Zone (LLETZ) were performed. The tissue samples
78
were scored by an experienced gynecologic pathologist and the histological

classification was used as the reference standard. If no interference with routine
diagnostic evaluation was anticipated, specimens from the CIN lesions were retrieved
and stored at -80 °C. Clinicopathological data were retrieved from patient files and
stored in our large anonymous password-protected institutional Gynecologic
Oncology database. All patients gave written informed consent.
For the frozen tissue samples used in de MethylCap-Seq analysis, the median age of
the CIN2/3 patients was 35 years (IQR 30-39) and for the patients with normal
cervices 43 years (IQR 41-44). For the independent cohort of patients with FFPE
samples, the median age of the CIN2/3 patients was 37 years (IQR 34-41), for the
patients with normal cervices 43 years (IQR 40-44) and for the cervical cancer
patients 49 years (range 42-54). For the cervical scrapings the median age of cervical
cancer patients was 50 years (IQR 39-64) and for the patients with normal cervices
47 years (IQR 43-53). The stage of cervical cancer patients was: 1 (1%) FIGO stage
IA1, 31 (31%) FIGO stage IB1, 18 (18%) FIGO stage IB2, 21 (21%) FIGO stage IIA,
17 (17%) FIGO stage IIB, 1 (1%) FIGO stage IIIA, 8 (8%) FIGO stage IIIB and 3 (3%)
FIGO stage IV. Histological classification of the cervical cancer patients was: 70 (70%)
squamous cell carcinoma (SCC), 21 (21%) adenocarcinoma (ADC), 3 (3%)
adenosquamous (ASC) and 6 (6%) undifferentiated carcinoma. The median age of
the patients referred with an abnormal Pap smear was 37 years (IQR 32-43). The
histological classifications of these patients were: 27 without CIN, 38 CIN1, 49 CIN2,
57 CIN3 and 44 miCa (29 SCC, 12 ADC, 3 ASC). The Pap smears were classified
according to the Papanicolaou system.
From all frozen tissue samples used for MethylCap-Seq and the FFPE samples, 10
μm tissue sections were cut and macrodissection was performed to enrich for
epithelial cells. Before and after cutting a hematoxylin and eosin slide was made to
check presence of epithelial cells. Cervical scrapings were collected in 5 ml ice-cold
phosphate buffered saline (PBS: 6.4 mM NA2HPO4; 1.5 mM KH2PO4; 0.14 M NaCl;
2.7 mM KCl) and kept on ice until further processing. Of these 5 ml cell suspension, 1
ml was used for cytomorphological assessment. The remaining 4 ml was centrifuged
and the cell pellet was suspended in 1 ml TRAP wash buffer and divided in 4
79
Chapter 4
fractions. Two fractions were stored as dry pellet at -80°C for DNA isolation as
described previously 21.
DNA isolation
Tissue slides from FFPE tissue were deparaffinized using 100% xylene followed by
17
100% ethanol . Genomic DNA from fresh-frozen macro-dissected samples and
cervical scrapings was isolated by standard overnight 1% SDS and Proteinase K
treatment, salt-chloroform extraction and isopropanol precipitation as described
previously 21. DNA pellets were washed with 70% ethanol and dissolved in 150 μl TE-
4
(10 mM Tris/HCL; 0.1 mM EDTA, pH 8.0). Genomic DNA was amplified in a
multiplex PCR according to the BIOMED-2 protocol, to check the DNA’s structural
27
integrity . For the MethylCap-Seq samples, DNA quantity was measured using
Quant-iT™ PicoGreen® dsDNA Assay Kit according to manufacturer’s protocol
(Invitrogen, Carlsbad, CA, USA). For cervical scrapings DNA concentrations and
260/280 ratios were measured using the Nanodrop ND-1000 Spectrophotometer
(Thermo Scientific, Waltham, MA, USA). A 260/280 ratio of >1.8 was required for all
samples.
Methylated-CpG island DNA capturing followed by next-generation sequencing

(MethylCap-Seq)
Methylated DNA fragments were captured with methyl-binding domains using the
MethylCap kit according to manufacturers instructions (Diagenode, Liège, Belgium).
The kit consists of the methyl binding domain (MBD) of human MeCP2, as a C-
terminal fusion with Glutathione-S-transferase (GST) containing an N-terminal His6-
tag. Before capturing, DNA samples (500 ng) were sheared to a size range of 300-
1000 bps using a Bioruptor™ UCD-200 (Diagenode, Liège, Belgium) and fragments
of ~300 bp were isolated. Leukocyte DNA of 4 healthy controls were included in 2
sets of 2 samples. Captured DNA was paired-end-sequenced on the Illumina
Genome Analyzer II platform according to protocol (Illumina, San Diego, CA, USA).
28
Results were mapped on the nucleotide sequence using Bowtie software ,
visualized using BioBix' H2G2 browser (http://h2g2.ugent.be/) and processed using
the human reference genome (NCBI build 37). The paired-end fragments were
unique and located within 400 bp of each other 29.
80
MethylCap-sequencing analysis
For statistical analysis, reads of promoter (-2000 bp – to + 500 bp of transcription
start site) and exon regions were retrieved. In order to identify differences between
normal cervices and CIN2/3 lesions, we dichotomised the read data into methylation
positive or negative. Samples were considered negative if a sample showed either 0
or 1 read. Samples were considered methylation positive if a sample showed 3
reads. Subsequently, regions were ranked based on highest specificity and highest
sensitivity for CIN2/3. The candidate markers should fulfil the following criteria: 1)
Low/negative reads in the leukocytes to prevent false positive results. The region
was excluded if both leukocyte samples showed >1 read or if 1 leukocyte sample
showed >2 reads. 2) Unmethylated (0 or 1 read) in at least 75% (15/20) of the normal
cervix group. 3) Methylated (3 reads) in at least 28% (5/18) of the CIN2/3 lesion
group.
Verification and validation of MethylCap-sequencing data by methylation specific

PCR (MSP)
MSP primers were designed for the highest ranking top 15 genes (16 DMRs).
Sodium bisulfite treatment of isolated genomic DNA (1 μg/sample) was performed
according to the recommendations of the EZ DNA methylation kit (Zymo, BaseClear,
Leiden, the Netherlands). MSP design and analysis was performed using sequences
derived from the H2G2 browser. Each reaction was performed in 30 μl total reaction
volume, containing: 600 nM of each MSP primer, 1.5 μl of bisulfite treated DNA
(approximately 15 ng), standard PCR components (Applied Biosystems) and 0.5 U
AmpliTaq Gold DNA polymerase (Applied Biosystems). Condition of the MSP was:
10 min hot-start at 95°C; 95°C for 60 sec, 60°C for 60 sec, 72°C 60 sec for a total of
40 cycles, with a final elongation step of 7 min at 72°C. Leukocyte DNA from healthy
women was used as negative control and in vitro methylated (by SssI enzyme)
leukocyte DNA was used as positive control for each MSP.
Quantitative Methylation Specific PCR (QMSP)

QMSP was performed as described previously by our group with an internal (FAM-
21
ZEN/IBFQ)-labelled hybridisation probe for quantitative analyses . Primer and
probe sequences are summarized in Table 1. ȕ–actin was used as a methylation
81
Chapter 4
independent internal reference gene. QMSP reactions were performed in 10 μl final

volume, containing: 300 nM of forward and reverse primers, 250 nM of hybridisation
probe, 5 μl of 2* QuantiTech Probe PCR Master Mix (Qiagen Hilden, Germany) and
2.5 μl bisulfite modified DNA (approximately 25 ng). Each sample was analyzed in
triplicate by ABI PRISM® 7900HT Sequence Detection System (Applied Biosystems).
Negative and positive controls were the same as used for MSP. Standard curve
analysis was performed on each plate and by each primers-probe set on serial
dilutions of in vitro methylated leukocyte DNA. A DNA sample was considered
methylated if at least 2 out of the 3 wells were methylation positive with a Ct-value
below 50 and DNA input of at least 225 pg ȕ-actin. The relative level of methylation of
the region of interest was determined by the following calculation: the average
quantity of the methylated region of interest divided by the average quantity of the
reference ȕ–Actin gene and multiplied by 10000 30. In our analysis we also included 4
genes previously described by our group (C13ORF18, JAM3, EPB41L3 and TERT)
to compare sensitivity and specificity of these known genes with the newly identified
methylation markers. QMSP for these markers was performed as previously
described 21.
HPV testing
HrHPV testing was performed using general primer-mediated PCR (GP5+/6+) as
30
reported previously . For HPV-typing as well as detection of the clinical relevant
HPV infections, GP5+/6+ positive cases were tested by COBAS® 4800 HPV test.
The COBAS HPV test individually detects HPV 16 and 18, while at the same time
31
identifying 12 additional hrHPV types . The COBAS HPV test is routinely used in
our iso-15189-certified laboratory of molecular pathology on scrapings from the
national population-based screening program. For the COBAS® HPV testing in
this study, the PCR only workflow was used, since no liquid-based scrapings in
Preservcyt® were available but only already isolated DNA. This workflow was first
validated with DNA isolated from clinical samples that were tested previously in
the diagnostic routine and this showed comparable results to the liquid-based
samples.
82
83
Chapter 4
Statistical analysis
Statistical analysis was performed using SPSS software package (SPSS 20, Chicago,
IL, USA). Spearman's rank correlation coefficient was used to compare the
MethylCap-Seq reads with the MSP band intensity. Categorical methylation data
were analyzed using the Pearson F² test. Receiver operating characteristic (ROC)
curves were generated and the area under the ROC curve (AUC) was used as a
measure of test performance. The Mann-Whitney U test and Kruskall-Wallis test was
used to determine differences in methylation ratio in 2 groups or more, respectively.
The student T test was used to compare positive methylation and age. To compare
sensitivity and specificity of the patient group referred with abnormal cytology by DNA
methylation markers versus hrHPV, the extended McNemar test, described by
32
Hawass was executed . P-values lower than 0.05 were considered statistically
significant.
Results
Identification of differential methylated genes by MethylCap sequencing
Genome-wide MethylCap-Seq was used to compare the DNA methylation profiles of
CIN2/3 dysplastic cervical cells with normal cervical cells to identify CIN2/3 specific
DMRs. After applying our criteria, 176 DMRs comprising 163 genes remained. The
list of DMRs is shown in supplement table 1, ranked on the sum of unmethylated
normal samples and methylated CIN2/3 samples.
Verification and validation of the top15 differentially methylated genes

To verify the MethylCap-Seq data, the top 15, out of the 163 identified genes were
selected. MSP primers were designed and could be optimized for 14 out of the 15
genes. Verification of the selected 14 genes showed for 11 genes a significant
correlation between the MSP band intensity and the amount of reads from the
MethylCap-Seq data. One gene (PCDH17) showed high methylation levels in
leukocytes and was therefore excluded for further validation. The remaining 10
genes passed verification and continued to the subsequent validation step. Table
2 shows an overview of which genes continued through the different stages of
validation.
84
Table 2: Validation of the top15 genes (16 regions)
1st 2nd
Region
Rank Gene Optimized Verification Validation diagnostic diagnostic
ID
evaluation evaluation
1 ZSCAN1 7534754 Yes Yes Yes Yes Yes
2 PCDH17 7180524 Yes No*
3 ST6GALNAC5 7595284 Yes Yes Yes Yes Yes
4 CLIC6 7721996 Yes No
5 AC012354.1 7802984 Yes No
6 AL590705.4 8386972 Yes Yes Yes Yes Yes
7** PAX2 6998395 Yes Yes Yes No
8 CDH6 8060049 Yes Yes Yes Yes Yes
9 GFRA1 7006375 Yes Yes Yes Yes Yes
10 IRX1 8050044 Yes No
11 POU4F3 8099299 Yes Yes No*
12 GATA4 8293851 Yes Yes Yes Yes Yes
13 MKX 6962285 No
7** PAX2 6998393 Yes No
14 KCNIP4 7992660 Yes Yes Yes Yes Yes
15 LHX8 7594866 Yes Yes Yes Yes Yes
* Excluded due to high methylation in leukocytes

** Same gene, different region
The second validation step was performed by MSP on DNA from FFPE tissue of an
independent, randomly selected new patient cohort that consisted of 13 cervical
cancers, 19 HSIL lesions (8 CIN2, 8 CIN3 and 3 adCIS) and 17 normal cervices. Out
of the 10 genes analyzed, 9 showed low methylation levels in the normal samples,
significant differential methylation between normal versus HSIL lesions and again
little to no methylation in the leukocytes (p < 0.05) (Table 3). These 9 genes
(ZSCAN1, ST6GALNAC5, AL590705.4, PAX2, CDH6, GFRA1, GATA4, KCNIP4 and
LHX8) were selected for further diagnostic evaluation in cervical scrapings (Table 3).
85
Chapter 4
Table 3: Methylation positivity in an external cohort of FFPE samples to validate results of

high methylation in CIN2+ lesions and no methylation in normal cervices of the newly found
methylation markers.
Rank Gene Normal CIN 2 CIN3 adCIS carcinoma
1 ZSCAN1 4/16 8/8 7/8 3/3 12/13
3 ST6GALNAC5 0/16 1/6 4/8 2/3 9/12
6 AL590705.4 0/16 1/8 1/7 2/3 6/12
7 PAX2 1/14 6/8 7/8 3/3 5/13
8 CDH6 1/15 3/8 4/8 3/3 7/13
9 GFRA1 0/12 2/8 3/8 2/3 10/12
11* POU4F3* 2/14 6/7 3/7 3/3 11/12
12 GATA4 0/17 3/8 2/7 3/3 10/13
14 KCNIP4 0/17 6/8 5/8 3/3 10/12
15 LHX8 1/16 3/8 4/8 3/3 7/13
* Excluded due to high methylation in leukocytes
Diagnostic evaluation by QMSP for normal versus cancer scrapings

To evaluate the diagnostic value of the new methylation markers, cervical scrapings
from two cohorts of patients were used: 1) normal versus carcinoma scrapings and 2)
scrapings from patients referred from population-based screening with an abnormal
Pap smear (Pap2). In cohort 1, scrapings of 100 randomly selected cervical
carcinoma patients and 89 patients with histologically confirmed normal cervices
were used. QMSP analysis showed that the relative levels of DNA methylation were
higher in the carcinoma scrapings compared to the normal scrapings for 8 out of the
9 selected genes (p<0.001) (Figure 2). The area under the curve (AUC) for
methylation ratio in cervical carcinoma showed for 8 genes an AUC>0.91, and for
one gene (PAX2) an AUC of 0.59. Therefore PAX2 was excluded from further
analysis (Figure 3). In women with a normal cervix, methylation positivity for all 9
genes was not related to age (data not shown).
Diagnostic evaluation by QMSP for normal/LSIL versus HSIL scrapings

In cohort 2, scrapings of 215 consecutive patients referred from population-based
screening with an abnormal Pap smear were used. The 8 genes that showed
differential methylation in the normal versus the cancer scrapings were subsequently
tested in cohort 2. Methylation levels and frequencies for all 8 genes analyzed
(ZSCAN1, ST6GALNAC5, AL590705.4, CDH6, GFRA1, GATA4, KCNIP4 and LHX8),
increased with the severity of the underlying histological lesion (p<0.001) (Figure 4
and Table 4).
86
Figure 2: Methylation ratio of 9 genes tested with QMSP in scrapings from normal (Nl) and
cancer (Ca) patients. Methylation levels are significantly higher in the cancer scrapings
87
Chapter 4
Figure 3 ROC curve analyses of methylation ratios per gene
Without setting a cut-off value for achieving higher/lower sensitivity and/or specificity,
genes ZSCAN1, ST6GALNAC5 and KCNIP4 reached high sensitivity (90%) for
detection of CIN2+ lesions, while for CDH6, GATA4 and LHX8 sensitivity for CIN2+
was between 73-84% (Table 5a). For AL590705.4 and GFRA1 sensitivity for CIN2+
was between 46-61%, and these genes showed especially high specificity (82%-
92%). In our analysis, we also included a marker panel of 4 genes, previously
described by our group (C13ORF18, JAM3, EPB41L3 and TERT) to compare
sensitivity and specificity of these known genes with the newly identified methylation
markers. The gene C13ORF18 showed reproducible results as described
previously21 with high specificity (95%) and relatively low sensitivity for CIN2+ of 40%.
JAM3 and EPB41L3 showed sensitivities for CIN2+ between 63-69% and
specificities between 79-91%. The gene TERT was previously described with high
specificity, but this result could not be reproduced since specificity was only 46% in
our analysis, while sensitivity for CIN2+ lesions was 82%.
88
Table 4. Cytology according to the Papanicolaou system (Bethesda system) per histological
subgroup. Methylation and HPV positivity of the 8 new methylation markers and 4 known
markers tested with QMSP in cervical scrapings from patients with no CIN, CIN1, CIN2, CIN3
and (mi)Ca (n=215).
Cytology No CIN CIN1 CIN2 CIN3 miCA
Pap2 (ASCUS) 9/27 (33%) 9/38 (24%) 2/49 (4%) 0 0
Pap3A (LSIL) 18/27 (66%) 27/38 (71%) 38/49 (78%) 16/57 (28%) 5/44 (11%)
Pap3B (HSIL) 0 2/38 (5%) 8/49 (16%) 31/57 (54%) 27/44 (61%)
Pap4 (HSIL) 0 0 1/49 (2%) 10/57 (18%) 8/44 (18%)
Pap5 (MiCa) 0 0 0 0 3/44 (7%)
Unknown 0 0 0 0 1/44 (2%)
New genes
ZSCAN1 20/27 (74%) 28/38 (74%) 45/49 (92%) 51/57 (90%) 44/44
(100%)
ST6GALNAC6 22/27 (82%) 33/38 (87%) 41/49 (84%) 53/57 (93%) 41/44 (93%)
AL590705.4 3/26 (12%) 8/36 (22%) 23/49 (47%) 30/55 (55%) 37/43 (86%)
CDH6 10/26 (39%) 15/36 (42%) 28/49 (57%) 37/55 (67%) 42/43 (98%)
GFRA1 1/26 (4%) 4/36 (11%) 11/49 (22%) 21/55 (38%) 35/43 (81%)
GATA4 14/26 (54%) 21/36 (58%) 38/49 (78%) 44/55 (80%) 41/43 (95%)
KCNIP4 24/27 (89%) 36/38 (95%) 48/49 (98%) 57/57 43/44 (98%)
(100%)
LHX8 12/26 (46%) 20/36 (56%) 33/49 (67%) 44/55 (80%) 41/43 (95%)
Known genes
C13ORF18 2/27 (7%) 1/38 (3%) 10/49 (20%) 23/57 (40%) 27/44 (61%)
JAM3 3/27 (11%) 3/38 (8%) 21/49 (43%) 36/57 (63%) 37/44 (84%)
EPB41L3 2/27 (7%) 12/38 (32%) 21/49 (43%) 41/57 (72%) 41/44 (93%)
TERT 13/27 (48%) 22/38 (58%) 32/49 (65%) 48/57 (84%) 43/44 (98%)
HPV test
HrHPV 12/26 (46%) 24/36 (67%) 40/49 (82%) 45/55 (82%) 31/43 (72%)
89
Chapter 4
Figure 4: Methylation ratio of the eight genes tested with QMSP in scrapings from patients
with No CIN lesion, CIN1, CIN2, CIN3 and (mi)Ca. Relative levels of methylation significantly
increases with more severe histological abnormality
90
Figure 4: Methylation ratio of the eight genes tested with QMSP in scrapings from patients
with No CIN lesion, CIN1, CIN2, CIN3 and (mi)Ca. Relative levels of methylation significantly
increases with more severe histological abnormality
hrHPV status and triage testing

HrHPV testing was performed on the patients group referred with abnormal cytology
at population-based screening. For 6 out of 215 patients insufficient material was
available to perform HPV testing. HrHPV was detected in 152/209 (73%) samples by
the GP5+/6+ PCR and COBAS HPV test. Table 4 shows HPV status in relation to
underlying histological diagnosis. HrHPV was present in 12/26 (46%) patients without
CIN lesion, 24/36 (67%) CIN1 patients, 40/49 (82%) CIN2 patients, 45/55 (82%)
91
Chapter 4
CIN3 patients and 31/43 (72%) patients with miCa. The sensitivity of hrHPV testing
for CIN2+ was 79% with a specificity of 42%. For the genes CDH6, GATA4, and
LHX8 sensitivity and specificity results were comparable to hrHPV testing with
sensitivity for CIN2+ between 73-84% and specificity between 40-60% (Table 5a).
Table 5b shows sensitivity and specificity for CIN2+ and CIN3+ in scrapings of hrHPV
positive women (n=152), which were comparable to the results for the whole group,
as shown in Table 5a. The genes ZSCAN1, ST6GALNAC5 and KCNIP4 again
showed high sensitivity (92%) for the detection of CIN2+, while for CDH6, GATA4,
EPB41L3, TERT and LHX8 sensitivity for CIN2+ was between 72-85%. For
AL590705.4, JAM3, C13ORF18 and GFRA1 sensitivity for CIN2+ was between 43-
68%, however these genes showed high specificity between 86-94%. In the current
Dutch population based screening program, women with pap2/pap3a (ASCUS/LSIL)
scrapings are retested after 6 months with triage testing by hrHPV. Therefore, we
also show the results of triage testing by hrHPV and methylation markers in this
group (Table 5b). Triage testing by hrHPV shows a sensitivity for CIN2+ of 82% with
a specificity of 41%; GATA4, LHX8 and TERT show comparable results.
Table 5a. Sensitivity and specificity results for CIN2+ and CIN3+ in cervical scrapings from
patients referred from population-based screening with an abnormal pap smear (n=215)
Gen Sensitivity Specificity Sensitivity Specificity
CIN2+ CIN2+ CIN3+ CIN3+
ZSCAN1 93% 26% 94% 18%
ST6GALNAC5 90% 15% 93% 16%
AL590705.4 61% 82% 68% 69%
CDH6 73% 60% 81% 52%
GFRA1 46% 92% 57% 86%
GATA4 84% 44% 87% 34%
KCNIP4 99% 8% 99% 5%
LHX8 80% 40% 87% 41%
C13ORF18 40% 95% 50% 87%
EPB41L3 69% 79% 81% 69%
JAM3 63% 91% 72% 76%
TERT 82% 46% 90% 41%
hrHPV 79% 42% 78% 32%
92
Table 5b. Sensitivity and specificity results for CIN2+ and CIN3+ in scrapings of hrHPV
positive women (n=152). And in scrapings of Pap2/Pap3a (ASCUS/LSIL) patients (n=124).
Only hrHPV positive patients (n=152)
Sensitivity Specificity Sensitivity
Specificity CIN3+
CIN2+ CIN2+ CIN3+
ZSCAN1 94% 36% 96% 22%
ST6GALNAC5 92% 19% 95% 16%
AL590705.4 65% 86% 74% 68%
CDH6 72% 64% 83% 55%
GFRA1 51% 92% 65% 83%
GATA4 85% 47% 88% 33%
KCNIP4 98% 11% 99% 7%
LHX8 81% 53% 91% 45%
C13ORF18 43% 94% 55% 87%
EPB41L3 72% 78% 86% 66%
JAM3 68% 94% 80% 74%
TERT 81% 47% 91% 42%
Only Pap2/3A patients (n=124)
ZSCAN1 90% 27% 86% 19%
ST6GALNAC5 90% 16% 100% 16%
AL590705.4 38% 82% 40% 74%
CDH6 58% 59% 75% 55%
GFRA1 22% 92% 30% 88%
GATA4 78% 44% 80% 36%
KCNIP4 98% 8% 100% 6%
LHX8 70% 49% 85% 45%
C13ORF18 23% 95% 29% 89%
EPB41L3 51% 78% 76% 72%
JAM3 44% 91% 57% 80%
TERT 74% 48% 91% 43%
hrHPV 82% 41% 80% 32%
Different combinations of genes were analyzed to find the best methylation marker
panel with the highest combined sensitivity and specificity. For this analysis a sample
was considered positive if either of the genes in the combination tested was positive.
By adding more than 3 genes in a combination specificity of the methylation test
decreased, with minimal increase in sensitivity. The combinations of genes with the
highest combined sensitivity and specificity for CIN2+ was AL590705.4/EPB41L3/JAM3
93
Chapter 4
and AL590705.4/C13ORF18/JAM3 with a sensitivity of 76% and 74%, which is

comparable to hrHPV testing (79%). Specificity of both combinations was 71% and 76%,
which is significantly higher than for hrHPV testing (42%) (P0.05). Table 6 shows that
for all other combinations sensitivities for detecting CIN2+ lesions are between 64-
80%, with a combined specificity between 58-88%.
Table 6. Combinations of different methylation markers to create a panel of genes most

suited as triage test in scrapings, ranked on highest sensitivity (n=215).
Gene combination Sensitivity Specificity Sensitivity Specificity

CIN2+ CIN2+ CIN3+ CIN3+
JAM3/CDH6 80% 58% 86% 47%

AL590705.4/CDH6/EPB41L3 80% 55% 88% 47%
GFRA1/EPB41L3/CDH6 78% 57% 86% 49%
CDH6/EPB41L3 78% 57% 86% 49%
GFRA1/AL590705.4/CDH6 77% 57% 84% 48%
AL590705.4/CDH6 77% 57% 84% 48%
JAM3/EPB41L3/AL590705.4 76% 71% 85% 59%
C13ORF18/JAM3/AL590705.4 74% 76% 81% 60%
GFRA1/EPB41L3/AL590705.4 74% 74% 85% 62%
AL590705.4/EPB41L3 74% 74% 84% 62%
C13ORF18/CDH6 74% 58% 81% 51%
JAM3/GFRA1/AL590705.4 73% 77% 81% 62%
C13ORF18/JAM3/EPB41L3 73% 72% 83% 62%
GFRA1/CDH6 73% 60% 81% 52%
JAM3/AL590705.4 72% 79% 80% 63%
JAM3/EPB41L3 72% 75% 83% 65%
JAM3/EPB41L3/GFRA1 72% 76% 84% 65%
GFRA1/EPB41L3 69% 79% 83% 69%
C13ORF18/EPB41L3 69% 75% 81% 67%
C13ORF18/JAM3/GFRA1 66% 82% 77% 70%
JAM3/GFRA1 65% 86% 76% 73%
C13ORF18/AL590705.4 65% 79% 71% 66%
C13ORF18/JAM3 64% 88% 73% 74%
GFRA1/AL590705.4 64% 81% 71% 68%
94
Discussion
Due to introduction of primary prevention of cervical cancer through prophylactic
vaccination against hrHPV types 16 and 18, involved in 70% of cervical cancer, the
incidence of cervical neoplasia will decrease14. Current implementation of HPV
vaccination programs in Europe will not have a real impact on the incidence of CIN2/3+
within the next 10-15 years. However, this decline in incidence will most probably
impair the diagnostic performance of HPV testing and cytology triage testing even
14
more, resulting in less efficient population-based screening programs . There is
therefore an urgent need to further improve current methodology for cervical cancer
screening.
In this study we report new CIN2/3 specific methylation markers identified by a

genome-wide DNA methylation screening strategy comparing CIN2/3 and normal
cervical cells. Diagnostic evaluation in cervical scrapings shows that for 8 newly
identified genes the relative level of methylation increases with the severity of the
underlying histological lesion. Combining our newly identified genes with our
previously reported panel (C13ORF18, JAM3, EPB41L3 and TERT) reveals that for
the combinations AL590705.4/EPB41L3/JAM3 and AL590705.4/C13ORF18/JAM3
sensitivities for CIN2+ are between 74-76%, which is comparable to the sensitivity for
CIN2+ of hrHPV testing (79%). Specificities of our gene panel was between 71-76%,
which is significantly higher (p0.05) than the specificity for hrHPV testing (42%) in a
triage setting after a positive Pap smear test result in population-based screening.
Our strategy identified 163 genes, of which the highest ranking top 15 genes were
validated in different steps. From the 163 identified genes, 12 were described
previously in literature (POU4F3, PAX2, WT1, TBX3, SOX1, COL6A2, ALK, SOX17,
PCDH10, CTNND2, APOBEC2, hsa-mir-124-1) as being more frequently methylated
in CIN2/3 lesions and/or cervical cancer compared to normal cervices, indicating the
validity of our approach. More CIN2+ specific markers are necessary since literature
shows that methylation markers were often tested on CIN3 scrapings only 19,20,33.
Cervical cancer screening in the Netherlands in 2016 will change to primary hrHPV
screening and because the relatively low specificity of the hrHPV test, a triage test is
95
Chapter 4
necessary for hrHPV positive women to prevent unnecessary referral to the

gynecologist. Although triage testing with cytology is now mostly advocated, this test
has some disadvantages because of its subjectivity and unreliability to test on self-
34
sampler material . Therefore, we analyzed the performance of our methylation
markers in the hrHPV positive scrapings. The combination AL590705.4/JAM3 and
AL590705.4/C13ORF18/JAM3 showed in the hrHPV positive scrapings sensitivities
for CIN2+ between 76-77% and specificities between 81-83% (data not shown).
These results are better than for other reported triage strategies in literature, such as
immunohistochemical staining with p16INK4a that reports a sensitivity for detecting
35,36
CIN2/3 around 77% with a specificity of 61% or for HPV 16/18 genotyping which
37
reports sensitivity for CIN2/3 around 65% with a specificity of 73% . Since our
methylation panel was tested in a selected patient group that was referred from
population-based screening with abnormal cytology, further validation in hrHPV
positive scrapings collected from a large cohort of women from population-based
screening should be performed. These kinds of scrapings from real life cohorts will
become available in the Netherlands after 2016 when primary screening has
changed to hrHPV testing.
The advantage of methylation analysis is that is an objective test and that it can be
performed on the same material used for hrHPV testing, which makes it also
interesting for self-sampled material. Different methylation markers already have been
tested as a triage test in hrHPV positive women 19,21,33,38-41. However, for most markers
a cut-off value was set in order to obtain high specificity. The advantage of the newly
found and previously described markers by our group is that no cut-off value is needed.
If the PCR product was negative (i.e. no amplification of specific product), the
samples were called negative and any ratio above zero was called positive. This
unique feature of the selected genes allows an objective and easy to interpret test.
It is also interesting to speculate about the role of the newly found genes in relation to
carcinogenesis. The gene AL590705.4 is located on chromosome 9 and was not
described before as being methylated during carcinogenesis. The gene CDH6 belongs
to the family of Cadherins. Cadherins are membrane glycoproteins that mediate
homophilic cell-cell adhesion and play critical roles in cell differentiation and
96
morphogenesis. Decreased expression of this gene may be associated with tumor

growth and metastasis. It has recently been described as a new transforming growth
42
factor-ȕ (TGF-ȕ) target in thyroid tumor patients . The gene GFRA1 plays a key role
in the control of neuron survival and differentiation. It has been described as
differentally methylated between cancerous and non-cancerous tissue obtained from
lung cancer patients based on DNA methylation profiles 43.
The strengths of our current study are: 1) the genome-wide approach with
MethylCap-Seq for specific identification of differential methylation regions between
normal cervices and CIN2/3 lesions, 2) the systematic verification and validation of
the markers found, and 3) the selection of the best performing markers for diagnostic
evaluation in cervical scrapings. The limitation of the current study was that
diagnostic evaluation of the markers was performed on a selected group of patients
that were referred to our hospital for abnormal cytological screening and not on HPV
positivity since that is not allowed by law in the Netherlands (unless within clinical
trials). The sensitivity of the cytological screening test is lower than for hrHPV
screening, therefore some CIN2+ cases that are hrHPV positive, will be cytological
normal. We cannot exclude that inclusion of such cases might affect the results. This
is a target for future studies.
Population-based screening is in transition and methylation markers might be an

important component in future screening settings. It is important to validate the most
interesting markers described in literature in a population-based screening trial.
Verification of the results by different groups is important to assure the reproducibility
of the methylation analysis. The combination of genes with the highest possible
sensitivity and specificity should be evaluated.
In conclusion, we identified new CIN2/3 specific methylation markers for detection of

cervical neoplasia in cervical scrapings. These newly found markers might be applied
as a triage test in hrHPV positive women from population-based screening. However,
further validation in population-based cohorts is needed.
97
Chapter 4
References
1. Peto J, Gilham C, Fletcher O, Matthews FE. The cervical cancer epidemic that screening has
prevented in the UK. Lancet. 2004;364(9430):249-256.
2. Arbyn M, Raifu AO, Weiderpass E, Bray F, Anttila A. Trends of cervical cancer mortality in the
member states of the european union. Eur J Cancer. 2009;45(15):2640-2648.
3. Cuzick J, Clavel C, Petry KU, et al. Overview of the european and north american studies on HPV
testing in primary cervical cancer screening. Int J Cancer. 2006;119(5):1095-1101.
4. Mayrand MH, Duarte-Franco E, Rodrigues I, et al. Human papillomavirus DNA versus papanicolaou
screening tests for cervical cancer. N Engl J Med. 2007;357(16):1579-1588.
5. Cox JT, Castle PE, Behrens CM, et al. Comparison of cervical cancer screening strategies
incorporating different combinations of cytology, HPV testing, and genotyping for HPV 16/18:
Results from the ATHENA HPV study. Am J Obstet Gynecol. 2013;208(3):184.e1-184.e11.
6. Walboomers JM, Jacobs MV, Manos MM, et al. Human papillomavirus is a necessary cause of
invasive cervical cancer worldwide. J Pathol. 1999;189(1):12-19.
7. Ronco G, Dillner J, Elfstrom KM, et al. Efficacy of HPV-based screening for prevention of invasive
cervical cancer: Follow-up of four european randomised controlled trials. Lancet. 2013.
8. Ronco G, Giorgi-Rossi P, Carozzi F, et al. Efficacy of human papillomavirus testing for the detection
of invasive cervical cancers and cervical intraepithelial neoplasia: A randomised controlled trial.
Lancet Oncol. 2010;11(3):249-257.
9. Bulkmans NW, Berkhof J, Rozendaal L, et al. Human papillomavirus DNA testing for the detection
of cervical intraepithelial neoplasia grade 3 and cancer: 5-year follow-up of a randomised
controlled implementation trial. Lancet. 2007;370(9601):1764-1772.
10. Rijkaart DC, Berkhof J, Rozendaal L, et al. Human papillomavirus testing for the detection of high-
grade cervical intraepithelial neoplasia and cancer: Final results of the POBASCAM randomised
controlled trial. Lancet Oncol. 2012;13(1):78-88.
11. Arbyn M, Ronco G, Anttila A, et al. Evidence regarding human papillomavirus testing in secondary
prevention of cervical cancer. Vaccine. 2012;30 Suppl 5:F88-99.
12. Cuzick J, Arbyn M, Sankaranarayanan R, et al. Overview of human papillomavirus-based and

other novel options for cervical cancer screening in developed and developing countries. Vaccine.
2008;26 Suppl 10:K29-41.
13. Kulasingam SL, Hughes JP, Kiviat NB, et al. Evaluation of human papillomavirus testing in primary
screening for cervical abnormalities: Comparison of sensitivity, specificity, and frequency of referral.
JAMA. 2002;288(14):1749-1757.
98
14. Franco EL, Mahmud SM, Tota J, Ferenczy A, Coutlee F. The expected impact of HPV vaccination
on the accuracy of cervical cancer screening: The need for a paradigm change. Arch Med Res.
2009;40(6):478-485.
15. Baylin SB, Ohm JE. Epigenetic gene silencing in cancer - a mechanism for early oncogenic
pathway addiction? Nat Rev Cancer. 2006;6(2):107-116.
16. Steenbergen RD, Snijders PJ, Heideman DA, Meijer CJ. Clinical implications of (epi)genetic
changes in HPV-induced cervical precancerous lesions. Nat Rev Cancer. 2014;14(6):395-405.
17. Yang N, Nijhuis ER, Volders HH, et al. Gene promoter methylation patterns throughout the
process of cervical carcinogenesis. Cell Oncol. 2010;32(1-2):131-143.
18. Bierkens M, Hesselink AT, Meijer CJ, et al. CADM1 and MAL promoter methylation levels in
hrHPV-positive cervical scrapes increase proportional to degree and duration of underlying cervical
disease. Int J Cancer. 2013;133(6):1293-1299.
19. Hesselink AT, Heideman DA, Steenbergen RD, et al. Combined promoter methylation analysis of
CADM1 and MAL: An objective triage tool for high-risk human papillomavirus DNA-positive women.
Clin Cancer Res. 2011;17(8):2459-2465.
20. Lai HC, Lin YW, Huang RL, et al. Quantitative DNA methylation analysis detects cervical
intraepithelial neoplasms type 3 and worse. Cancer. 2010;116(18):4266-4274.
21. Eijsink JJ, Lendvai A, Deregowski V, et al. A four-gene methylation marker panel as triage test in
high-risk human papillomavirus positive patients. Int J Cancer. 2012;130(8):1861-1869.
22. Yang N, Eijsink JJ, Lendvai A, et al. Methylation markers for CCNA1 and C13ORF18 are strongly
associated with high-grade cervical intraepithelial neoplasia and cervical cancer in cervical
scrapings. Cancer Epidemiol Biomarkers Prev. 2009;18(11):3000-3007.
23. Reesink-Peters N, Wisman GB, Jeronimo C, et al. Detecting cervical cancer by quantitative
promoter hypermethylation assay on cervical scrapings: A feasibility study. Mol Cancer Res.
2004;2(5):289-295.
24. Laird PW. Principles and challenges of genomewide DNA methylation analysis. Nat Rev Genet.
2010;11(3):191-203.
25. Lendvai A, Johannes F, Grimm C, et al. Genome-wide methylation profiling identifies

hypermethylated biomarkers in high-grade cervical intraepithelial neoplasia. Epigenetics.
2012;7(11):1268-1278.
26. Rauch TA, Pfeifer GP. DNA methylation profiling using the methylated-CpG island recovery assay
(MIRA). Methods. 2010;52(3):213-217.
27. van Dongen JJ, Langerak AW, Bruggemann M, et al. Design and standardization of PCR primers
and protocols for detection of clonal immunoglobulin and T-cell receptor gene recombinations in
suspect lymphoproliferations: Report of the BIOMED-2 concerted action BMH4-CT98-3936.
Leukemia. 2003;17(12):2257-2317.
99
Chapter 4
28. Langmead B, Trapnell C, Pop M, Salzberg SL. Ultrafast and memory-efficient alignment of short
DNA sequences to the human genome. Genome Biol. 2009;10(3):R25-2009-10-3-r25. Epub 2009
Mar 4.
29. De Meyer T, Mampaey E, Vlemmix M, et al. Quality evaluation of methyl binding domain based
kits for enrichment DNA-methylation sequencing. PLoS One. 2013;8(3):e59068.
30. Wisman GB, Nijhuis ER, Hoque MO, et al. Assessment of gene promoter hypermethylation for
detection of cervical neoplasia. Int J Cancer. 2006;119(8):1908-1914.
31. Cui M, Chan N, Liu M, et al. Clinical performance of roche cobas 4800 HPV test. J Clin Microbiol.
2014;52(6):2210-2211.
32. Hawass NE. Comparing the sensitivities and specificities of two diagnostic procedures performed
on the same group of patients. Br J Radiol. 1997;70(832):360-366.
33. Overmeer RM, Louwers JA, Meijer CJ, et al. Combined CADM1 and MAL promoter methylation
analysis to detect (pre-)malignant cervical lesions in high-risk HPV-positive women. Int J Cancer.
2011;7(6).
34. Gok M, van Kemenade FJ, Heideman DA, et al. Experience with high-risk human papillomavirus
testing on vaginal brush-based self-samples of non-attendees of the cervical screening program.
Int J Cancer. 2012;130(5):1128-1135.
35. Carozzi F, Confortini M, Dalla Palma P, et al. Use of p16-INK4A overexpression to increase the
specificity of human papillomavirus testing: A nested substudy of the NTCC randomised controlled
trial. Lancet Oncol. 2008;9(10):937-945.
36. Carozzi F, Gillio-Tos A, Confortini M, et al. Risk of high-grade cervical intraepithelial neoplasia
during follow-up in HPV-positive women according to baseline p16-INK4A results: A prospective
analysis of a nested substudy of the NTCC randomised controlled trial. Lancet Oncol.
2013;14(2):168-176.
37. Rijkaart DC, Berkhof J, van Kemenade FJ, et al. Evaluation of 14 triage strategies for HPV DNA-
positive women in population-based cervical screening. Int J Cancer. 2012;130(3):602-610.
38. Brentnall AR, Vasiljevic N, Scibior-Bentkowska D, et al. A DNA methylation classifier of cervical
precancer based on human papillomavirus and human genes. Int J Cancer. 2014.
39. Vasiljevic N, Scibior-Bentkowska D, Brentnall AR, Cuzick J, Lorincz AT. Credentialing of DNA
methylation assays for human genes as diagnostic biomarkers of cervical intraepithelial neoplasia
in high-risk HPV positive women. Gynecol Oncol. 2014;132(3):709-714.
40. Hansel A, Steinbach D, Greinke C, et al. A promising DNA methylation signature for the triage of
high-risk human papillomavirus DNA-positive women. PLoS One. 2014;9(3):e91905.
41. Verhoef VM, Bosgraaf RP, van Kemenade FJ, et al. Triage by methylation-marker testing versus
cytology in women who test HPV-positive on self-collected cervicovaginal specimens
(PROHTECT-3): A randomised controlled non-inferiority trial. Lancet Oncol. 2014;15(3):315-322.
100
42. Sancisi V, Gandolfi G, Ragazzi M, et al. Cadherin 6 is a new RUNX2 target in TGF-beta signalling
pathway. PLoS One. 2013;8(9):e75489.
43. Selamat SA, Chung BS, Girard L, et al. Genome-scale analysis of DNA methylation in lung
adenocarcinoma and integration with mRNA expression. Genome Res. 2012;22(7):1197-1211.
101
Chapter 4
Supplement table 1: Identified DMRs with sensitivity and specificity ranked on the sum
of both
Sum unmeth
Size Nr normals Nr CIN2/3 normals and
Rank Region ID Gene region unmethylated methylated meth CIN2/3
1 7534754 ZSCAN1 195 19 9 28
2 7180524 PCDH17 133 19 8 27
3 7595284 ST6GALNAC5 180 19 8 27
4 7721996 CLIC6 123 20 6 26
5 7802984 AC012354.1 262 17 9 26
6 8386972 AL590705.4 412 17 9 26
7 6998395 PAX2 142 20 5 25
8 8060049 CDH6 139 20 5 25
9 7006375 GFRA1 313 19 6 25
10 8050044 IRX1 171 19 6 25
11 8099299 POU4F3 270 18 7 25
12 8293851 GATA4 264 18 7 25
13 6962285 MKX 436 17 8 25
14 6998393 PAX2 427 17 8 25
15 7992660 KCNIP4 305 17 8 25
16 7594866 LHX8 443 16 9 25
17 7610478 RP11-439A17.1 245 15 10 25
18 6998398 PAX2 197 19 5 24
19 7019754 DPYSL4 246 19 5 24
20 7037878 WT1 198 19 5 24
21 8000998 GABRA2 151 19 5 24
22 7319687 CACNG3 70 18 6 24
23 8021385 TNIP3 126 18 6 24
24 8323242 ZFHX4 282 18 6 24
25 6998396 PAX2 386 17 7 24
26 7256161 GJD2 527 17 7 24
27 7633581 AL136370.1 1 17 7 24
28 8070077 RGS7BP 339 17 7 24
29 8173181 RLBP1L2 223 17 7 24
102
30 8371278 C9orf71 41 17 7 24
31 7186432 SLITRK1 402 15 9 24
32 8036518 GALNTL6 325 15 9 24
33 8444752 FAM133A 264 15 9 24
34 7144043 TBX3 120 16 8 24
35 7198746 SOX1 432 16 8 24
36 7650220 EP11-10E13.1.1 138 16 8 24
37 7530232 CACNG7 343 15 8 23
38 7531604 EPS8L1 364 15 8 23
39 7533332 ZNF542 400 15 8 23
40 8389207 GRIN3A 372 15 8 23
41 7035376 SLC6A5 422 16 7 23
42 7262409 SHF 272 16 7 23
43 7614837 TCHH 160 16 7 23
44 7733864 COL6A2 317 16 7 23
45 7939583 AC092691.1 273 16 7 23
46 8006035 EPHA5 260 16 7 23
47 8028522 POU4F2 443 16 7 23
48 8153679 TFAP2B 213 16 7 23
49 8299826 PEBP4 55 16 7 23
50 7037882 WT1 195 18 5 23
51 7319688 CACNG3 206 18 5 23
52 8420025 MXRA5 166 18 5 23
53 7015183 FOXI2 371 17 6 23
54 7198756 SOX1 108 17 6 23
55 7500952 ZNF676 299 17 6 23
56 7534399 ZNF154 168 17 6 23
57 7795180 ALK 451 17 6 23
58 7853759 KCNH7 240 17 6 23
59 7932344 CADM2 159 17 6 23
60 7954401 ZIC4 331 17 6 23
61 8220882 ELMO1 229 17 6 23
62 8300584 NKX2-6 198 17 6 23
103
Chapter 4
63 8308835 ZMAT4 239 17 6 23
64 8354231 FOXH1 42 17 6 23
65 6997382 NKX2-3 468 15 7 22
66 7209496 SSTR1 224 15 7 22
67 7300000 BAIAP3 425 15 7 22
68 7488800 ZNF69 181 15 7 22
69 7949963 PPP2R3A 1 15 7 22
70 7950795 AC007159.2 554 15 7 22
71 8031815 NPY2R 512 15 7 22
72 8095341 5S_rRNA 192 15 7 22
73 8144463 AL033519.2 1 15 7 22
74 8161319 SNAP91 466 15 7 22
75 8161598 TBX18 569 15 7 22
76 8315432 SOX17 137 15 7 22
77 8369935 BX248098.1 160 15 7 22
78 8444558 PCDH11X 1 15 7 22
79 6973216 DRGX 516 16 6 22
80 7168291 PDX1 259 16 6 22
81 7216677 C14orf39 288 16 6 22
82 7558688 VWA5B1 16 16 6 22
83 7721995 CLIC6 143 16 6 22
84 7739980 PI1KAP2 144 16 6 22
85 8136141 NRSN1 147 16 6 22
86 8166500 C6orf220 206 16 6 22
87 8209902 THSD7A 229 16 6 22
88 8355044 FOXD4 384 16 6 22
89 7051745 KCNK4 388 17 5 22
90 7096188 KCNA6 207 17 5 22
91 7135359 C12orf42 257 17 5 22
92 7172465 C13orf36 548 17 5 22
93 7216685 SIX6 226 17 5 22
94 7255970 NOP10 1 17 5 22
95 7399713 AC091132.3 1 17 5 22
104
96 7429544 CBX2 219 17 5 22
97 7528453 ZNF578 132 17 5 22
98 7582721 MKNK1 1 17 5 22
99 7584213 DMRTA2 203 17 5 22
100 7769721 FAM19A5 307 17 5 22
101 7810978 OTX1 198 17 5 22
102 7848584 AC064865.1 1 17 5 22
103 7858997 HOXD13 135 17 5 22
104 8017441 DKK2 272 17 5 22
105 8136976 SCGN 312 17 5 22
106 8169072 MICAL1 1 17 5 22
107 8318803 BHLHE22 461 17 5 22
108 7031638 PARVA 138 15 6 21
109 7150264 AC084018.1 111 15 6 21
110 7251344 GABRG3 423 15 6 21
111 7255084 GREM1 426 15 6 21
112 7255394 RYR3 378 15 6 21
113 7489068 ZNF625 1 15 6 21
114 7589242 DAB1 426 15 6 21
115 7602550 AL136147.1 227 15 6 21
116 7613795 FAM63A 36 15 6 21
117 7689874 SLC32A1 191 15 6 21
118 7720644 OLIG1 308 15 6 21
119 7778034 SOX11 121 15 6 21
120 7843418 AC140481.1 354 15 6 21
121 7851097 GALNT13 179 15 6 21
122 7870465 PTH2R 337 15 6 21
123 7946628 KIAA1257 101 15 6 21
124 8024629 PCDH10 227 15 6 21
125 8218592 AVL9 63 15 6 21
126 8334444 RSPO2 402 15 6 21
127 6973214 DRGX 253 16 5 21
128 6998512 SEMA4G 1 16 5 21
105
Chapter 4
129 7035410 NELL1 115 16 5 21
130 7216686 SIX6 218 16 5 21
131 7473720 ONECUT3 506 16 5 21
132 7584206 DMRTA2 208 16 5 21
133 7619450 ETV3L 55 16 5 21
134 7645452 AC011700.2 291 16 5 21
135 7681392 GINS1 80 16 5 21
136 7893422 CNTN4 406 16 5 21
137 7946826 RAB43 7 16 5 21
138 8054394 CTNND2 321 16 5 21
139 8096965 PCDHGC5 233 16 5 21
140 8165363 AL137784.3 160 16 5 21
141 8237177 WBSCR17 323 16 5 21
142 8252950 AC006329.2 494 16 5 21
143 8300582 NKX2-6 217 16 5 21
144 8411390 GBGT1 201 16 5 21
145 6999108 FBXW4 100 15 5 20
146 6999851 PSD 325 15 5 20
147 7028932 TUB 120 15 5 20
148 7096257 KCNA1 338 15 5 20
149 7101957 GRIN2B 405 15 5 20
150 7209031 NKX2-1 137 15 5 20
151 7223639 ZFYVE1 63 15 5 20
152 7327883 MYLK3 99 15 5 20
153 7394459 KRT27 1 15 5 20
154 7398204 RUNDC3A 314 15 5 20
155 7528533 ZNF808 419 15 5 20
156 7552927 RP1-21O18.1 34 15 5 20
157 7614845 TCHH 330 15 5 20
158 7621232 TOMM40L 140 15 5 20
159 7692820 GDAP1L1 374 15 5 20
160 7762360 TTLL12 134 15 5 20
161 7825980 CNGA3 170 15 5 20
106
162 8000034 PHOX2B 221 15 5 20
163 8018500 ELOVL6 27 15 5 20
164 8022574 FAT4 369 15 5 20
165 8031643 DCHS2 62 15 5 20
166 8096823 PCDHB15 233 15 5 20
167 8147247 DNAH8 224 15 5 20
168 8148415 APOBEC2 50 15 5 20
169 8199655 UNC84A 190 15 5 20
170 8247593 CASD1 1 15 5 20
171 8287434 DLGAP2 391 15 5 20
172 8292161 hsa-mir-124-1 135 15 5 20
173 8293849 GATA4 179 15 5 20
174 8301460 EBF2 186 15 5 20
175 8413181 AL354796.1 379 15 5 20
176 8430345 BX842568.1 7 15 5 20
107
Chapter 4
108
Methylome analysis to discover hypermethylation biomarkers for both ADC and SCC
Chapter 5
Genome-wide methylome analysis to discover (novel)

hypermethylation biomarkers for both adeno- and
squamous cell cervical carcinoma
Wang R.1,3, Schuuring E.2, Boers A.1, van Leeuwen R.W.1,

van der Zee A.G.J.1, Wisman G.B.A. 1
1
Department of Gynecologic Oncology, University of Groningen, University Medical
Center Groningen, Cancer Research Center Groningen, Groningen, the Netherlands
2
Department of Pathology, University of Groningen, University Medical Center
Groningen, Groningen, the Netherlands
3
Department of Laboratory Medicine, Tianjin Medical University, Tianjin, China
109
Chapter 5
Abstract
Aims: Cervical adenocarcinomas (ADC) are mainly diagnosed at an advanced stage of
disease. In the last decade, the incidence of ADC increased in most developed countries and
represents about 20% of cervical cancers. One explanation for the increase of ADC is the
less effective cytomorphological detection of ADC and its precursors in population-based
screening programs. Analysis of DNA methylation markers might improve the detection of
ADC in earlier stages. However, no specific methylation markers have been described for the
detection of ADC. The aim of this study was to discover novel methylation markers for
cervical cancer detecting both ADC and SCC.
Methods: To generate a global methylation-profile of DNA from 20 normal cervices, 6 ADC
and 6 SCC, methylated DNA fragments were captured using the Methyl Binding Domain
(MBD) of human MeCP2 followed by next-generation-sequencing (MethylCap-seq).
Differential methylated markers were selected for verification by bisulfite pyrosequencing or
methylation specific PCR (MSP) on the same samples used for MethylCap-seq and validated
on an independent series of FFPE specimens from normal cervices and cervical cancers.
Further clinical validation was performed by quantitative methylation specific PCR (QMSP)
on cervical scrapings from an independent cohort of 89 women with a normal cervix and 125
cervical cancer patients.
Results: Validation of the highest ranking 15 differentially methylated candidate markers
resulted in 5 markers exhibiting different methylation between normal and cancer tissues
(p<0.05). Using QMSP analysis on cervical scrapings, the sensitivity of these 5 markers
varied from 80.5% to 91.9% to detect both ADC and SCC with almost all normal scrapings
negative (specificity: 94% -98.9%).
Conclusion: Using MethylCap-seq analysis, we identified 5 new methylation markers with a
high sensitivity for both ADC as well as SCC in cervical scrapings.
110
Introduction:
Cervical cancer is one of the common female cancers in the world. Each year, more
than 500,000 new cases and around 275,000 deaths occur globally1. Cervical
squamous cell carcinoma (SCC) and cervical adenocarcinoma (ADC) are two main
histological subtypes of invasive cervical cancer, which account for 75–90% and 10–
25%, respectively2-4. Compared to SCC, ADC is more common in European
countries with an incidence ranging from 5.5% to 30.0%5. Currently, the incidence of
SCC is declining in most developed countries. In contrast, there is a rise in the
absolute and relative incidence of ADC6. In Europe, ADC is increasing rapidly,
especially in younger women. In the Netherlands, the absolute incidence rate of ADC
increased with 15.8% in women aged 15-29 years and 2.5% in women aged 30-44
years7. Moreover, compared to SCC, ADC is mainly diagnosed in more advanced
stages, appears to be less sensitive to radiation therapy and chemotherapy and is
associated with a worse prognosis than SCC8-11.
Both the upward trend and postponed detection of ADC are probably due to the
present population-based screening programs, which are more effective in detecting
the precursors of SCC than those of ADC. Because of its localization higher up in the
cervix it is more difficult to obtain representative cytology samples and to observe
ADC or its precursors by colposcopy12. High risk human papillomavirus (hrHPV)
associated with cervical carcinogenesis is widely known and its detection is more
sensitive for the detection of cervical adenocarcinoma in situ (AdCIS) and ADC than
cytology. However, in population-based screening programs hrHPV-positive women
with normal cytology will require additional biomarkers, again more specific for SCC
and its precursors than for ADC to enable the gynecologist to decide on whether
performing endocervical curettage or not13. Therefore, novel biomarkers for cervical
cancer are required that ideally will identify and discriminate between ADC and SCC
as well as their precursors with high sensitivity.
Aberrant gene expression caused by epigenetic mechanisms are prominent features

of many types of cancer14, and promoter DNA methylation of tumor suppressor genes
(TSG) has been reported to be an early event in carcinogenesis15. DNA methylation
markers might be exploited in cancer diagnosis as variations in DNA methylation are
111
Chapter 5
observed more frequently than other genetic variations16. Although we17,18 and
others19 have reported many methylation markers associated with cervical cancer,
many of these markers are more frequently methylated in SCC compared to ADC17.
As to cervical cancer diagnostics, an important advantage of DNA methylation
markers is that they can be detected in the same scrapings as used for HPV
analyses20,21. However, so far only a limited number of methylated genes have been
identified that are specifically associated with ADC. Most of these markers have
22,23
lower sensitivity for ADC and SCC both or either one . Recently, 4 genes (PAX1,
PTPRR, SOX1, and ZNF582) were reported that are frequently methylated in ADC as
well as SCC24. However, data on screening using these markers is missing for larger
cohorts.
In the past ten years, advances in whole genome methylation profiling technologies
have revolutionized the field of cancer research. In order to identify cervical cancer
specific methylation markers, the pharmacological unmasking expression microarray
approach25 and chromatin immunoprecipitation (ChIP) combined with methylation-
specific oligonucleotide microarray have been performed26,27. Nevertheless,
microarray-based screening has drawbacks such as their design and production,
while also the inaccurate hybridization signals and antibody-based MeDIP are rather
variable, which leaves space for further improvement. Reduction in costs have
spurred the adoption of next generation sequencing (NGS) platforms with higher
sensitivity and accuracy compared to traditional microarray profiling28. Recently,
affinity-based methylation capture assay coupled using methyl binding domain (MBD)
complexes with NGS (MethylCap-seq) has been reported to be an effective
technique to comprehensively analyze the methylome in lung cancer, ovarian cancer,
and head and neck cancer29-31, and panels of hypermethylated loci have shown to
represent possible methylation markers for early detection. These technologies have
facilitated the discovery of potential biomarkers for disease development and
progression as well as our understanding of the complex, underlying molecular
mechanisms that lead to cancer.
Until now, no DNA methylome analysis has been performed using patient material
including cervical ADC. In this study, MethylCap-seq was applied to perform a
112
genome-wide DNA methylation screening of cervical cancer, including both ADC and
SCC, and normal cervix tissues. With this approach, we sought to identify genome-
wide aberrant methylation patterns of cervical cancer-specific markers with high
sensitivity to detect both ADC and SCC in cervical scrapings.
Materials and Methods

General Strategy
In order to identify and validate cervical cancer-specific methylation markers both for
ADC and SCC, the following strategy was applied (Fig1). Step1, DNA from frozen
tissue of cervical cancer (ADC=6, SCC=6) and normal cervices (n=20) was analyzed
using MethylCap-seq. Subsequently, the differentially methylated regions (DMRs)
were identified between normal cervices and cancers by statistical analysis. Step2,
among the methylation candidates, the top 15 were selected for verification by
methylation specific PCR (MSP) or pyrosequencing on the same frozen tissue as
used for step1. Step3, using the selected methylation candidates (in step2), MSP
was performed to validate on DNA of FFPE tissue from an independent series
(normal n=17, cancer n=13 composed of ADC n=6 and SCC n=7). Step4,
methylation candidates showing a significant difference methylation frequency in
normal and cancer were selected for further clinical validation on cervical scrapings
from a large series of cervical cancer patients (n=125 comprising n=57 ADC and
n=68 SCC) and healthy age-matched controls (n=89) by real time quantitative
methylation specific PCR (QMSP).
Patients:
Patients with cervical cancer referred to the outpatient clinic of the University Medical
Center Groningen (UMCG) are asked to participate in our on-going ‘Methylation
study’ that has been approved by the Institutional Review Board (IRB) of UMCG, the
Netherlands. All patients from whom material was obtained gave written informed
consent. Frozen tissue, paraffin embedded tissue and scrapings for this study were
prospectively collected and stored in our tissue bank.
Normal tissue samples and normal scrapings are collected from patients with non-
malignant disease. All cervical tissue that was used for the normal control group was
113
Chapter 5
judged as normal by histopathological examination. Patients referred with cervical
Fig1. Flow scheme for the identification of new cervical cancer markers
cancer are staged according to the FIGO criteria with pelvic examination and
biopsies under general anaesthesia. Cervical scrapings from both groups (cervical
cancer staging and benign gynecologic surgery) were collected before surgery under
general anaesthesia. The tissue samples were scored by an experienced
gynecologic pathologist and the histological classification was used as the reference
standard. All clinicopathological data were retrieved from patient files and stored in
our large anonymous password-protected institutional Gynecologic Oncology
database.
For MethylCap-seq and pyrosequencing, frozen tissue specimens were collected

from 20 normal cervices (IQR 33-45, median age: 43 years) and 12 cancers
composed of 6 SCC and 6 ADC (IQR 27- 69, median age: 44 years). Stage of
cervical cancer patients was for ADC: 3 FIGO stage IB1, 1 FIGO stage IB2 and 2
FIGO stage IIA. For SCC: 3 FIGO stage IB1, 1 FIGO stage IB2 and 2 FIGO stage IIA.
For MSP analysis, formalin fixed paraffin embedded (FFPE) tissue was collected from
17 normal cervices (IQR 40-44, median age: 43 years), 13 cervical cancers including
114
6 ADC and 7 SCC (IQR 42-54, median age: 49 years). Stage of cervical cancer
patients was for ADC: 2 FIGO stage IB1, 1 FIGO stage IB2, 2 FIGO stage IIA and 1
FIGO stage IIIA, for SCC: 2 FIGO stage IB1, 2 FIGO stage IB2, 1 FIGO stage IIA, 1
FIGO stage IIB and 1 FIGO stage IIIB.
For QMSP, scrapings were collected from 89 normal cervices (IQR 43-53, median
age: 47 years), and from 125 cervical cancer patients (IQR 23 to 84, median age: 50
years) compromising 68 SCC and 57 ADC. Stage of cervical cancer patients was for
ADC: 2 FIGO stage IA1, 1 FIGO stage IA2, 25 FIGO stage IB1, 8FIGO stage IB2, 8
FIGO stage IIA, 5 FIGO stage IIB, 1 FIGO stage IIIA, 6 FIGO stage IIIB and 1 FIGO
stage IV, for SCC: 1 FIGO stage IA1, 19 FIGO stage IB1, 14 FIGO stage IB2, 13
FIGO stage IIA, 14 FIGO stage IIB, 1 FIGO stage IIIA, 5 FIGO stage IIIB and 1 FIGO
stage IV.
Sample collection procedure and DNA isolation:

From the frozen tissue and FFPE samples, 10 μm tissue sections were cut and
macrodissection was performed to enrich for epithelial cells. Before and after cutting
hematoxylin and eosin slides were made to check presence of epithelial cells.
Cervical scrapings were collected in 5 ml ice-cold phosphate buffered saline (PBS:
6.4 mM Na2HPO4; 1.5 mM KH2PO4; 0.14 M NaCl; 2.7 mM KCl) and kept on ice until
further processing. Of these 5 ml cell suspension, 1 ml was used for
cytomorphological assessment. The remaining part (4 ml) was centrifuged and the
cell pellet was suspended in 1 ml TRAP wash buffer and divided in 4 fractions. Two
fractions were stored as dry pellet at -80°C for DNA isolation.
Tissue slices from FFPE were deparaffinized using 100% xylene followed by 100%
ethanol17. Genomic DNA from fresh-frozen macro-dissected samples and cervical
scrapings was isolated by standard overnight 1% SDS and Proteinase K treatment,
salt-chloroform extraction and isopropanol precipitation as described previously18.
DNA pellets were washed with 70% ethanol and dissolved in 150 μl TE-4 (10 mM
Tris/HCL; 0.1 mM EDTA, pH=8.0). Genomic DNA was amplified in a multiplex PCR
32
according to the BIOMED-2 protocol, to check the DNA’s structural integrity . For
the MethyCap-seq samples, DNA quantity was measured using Quant-i T™
115
Chapter 5
PicoGreen ® dsDNA Assay Kit according to manufacturer’s protocol (Invitrogen,

Carlsbad, CA, USA). For cervical scrapings DNA concentrations and 260/280 ratios
were measured using the Nanodrop ND-1000 Spectrophotometer (Thermo Scientific,
Waltham, MA, USA). A 260/280 ratio of >1.8 was required for all samples.
MethylCap-seq:
Genomic DNA samples (500 ng) were randomly sheared to a size range of 300-1000
bps using a Bioruptor™ UCD-200 (Diagenode, Liège, Belgium) and fragments of
~300 bp were isolated. Methylated DNA fragments were captured with methyl-
binding domains using the MethylCap kit according to manufacturer’s
instructions (Diagenode, Liège, Belgium). The kit consists of the MBD of human
MeCP2, as a C-terminal fusion with Glutathione-Stransferase (GST) containing
an N-terminal His6-tag. Leukocyte DNA of 4 healthy controls was included in 2
sets of 2 samples.
Captured DNA was paired-end-sequenced on the Illumina Genome Analyzer II

platform according to protocol (Illumina, San Diego, CA, USA). Results were mapped
on the nucleotide sequence using Bowtie software33, visualized using BioBix' H2G2
browser (http://h2g2.ugent.be/) and processed using the human reference genome
(NCBI build 37). The paired-end fragments were unique and located within 400 bp of
each other34.
For statistical analysis, only reads of the promoter (-2000 bp to + 500 bp of

transcription start site) were retrieved as these are mainly related with transcriptional
silencing of genes. In order to identify differences between normal cervices and
cervical cancer tissues, we dichotomised the read data into methylation positive or
negative. Samples from normal cervices were considered methylation negative if a
sample showed either 0 or 1 read. Cancer samples were considered methylation
positive if a sample showed 3 reads. Subsequently, Fisher exact test was performed
to determine the significant DMRs between ADC and normal or SCC and normal. To
downsize the number of DMRs and to pinpoint candidate methylation markers in
cervical cancer the following criteria were applied (see fig 2): 1) the methylation
frequency is significantly different between normal and cancer. 2) Unmethylated (0 or
116
1 read) in at least 75% (15/20) of the normal cervix group. 3) Methylated (3 reads) in
at least 50% (3/6) of ADC and SCC, respectively. 4) Low/negative reads in the
leukocytes to prevent false positive results. The region was excluded if both
leukocyte samples showed >1 read or if 1 leukocyte sample showed >2 reads. 5)
DNA region length should 30bp. 6) Comparable regions in both identified
histological subtype candidates group.
Bisulfite treatment of DNA:

Bisulfite treatment on denatured genomic DNA was performed as previously reported
35
. One microgram of genomic DNA per sample was modified with sodium bisulfite
using the EZ DNA methylation kit according to manufacturer’s protocol (Zymo
Research, Corp, Irvine, CA). Leukocyte DNA from healthy women and whole genome
amplified DNA (WGA) were used as negative controls, in vitro methylated (by SssI
enzyme) leukocyte DNA was used as positive control.
Pyrosequencing:
Bisulfite treated DNA was amplified using PyroMark PCR kit (Qiagen, Hilden,
Germany). PCR reaction and cycling conditions were according to the kit manual.
Subsequently, sample preparation and pyrosequencing was performed by PyroMark
Q24 instrument (Qiagen, Hilden, Germany) using the Pyro Gold Q24 Reagents
(Qiagen, Hilden, Germany). Data was analyzed and quantified with the PyroMark
Q24 software version 2.0.6 (Qiagen, Hilden, Germany). Non-template control (water),
positive and negative controls were used in each reaction. PCR and sequence
primers were available in Supple. TableS1.
MSP (Methylation specific PCR)

For MSP, each reaction was performed in 30 μl total reaction volume, containing: 600
nM of each MSP primer, 1.5 μl of bisulfite treated DNA (approximately 15 ng),
standard PCR components (Applied Biosystems, Carlsbad, CA, USA) and 0.5 U
AmpliTaq Gold DNA polymerase (Applied Biosystems, Carlsbad, CA, USA).
Condition of the MSP was: 10 min hot-start at 95°C; 95°C for 60 sec, 60°C for 60 sec,
72°C 60 sec for a total of 40 cycles, with a final elongation step of 7 min at 72°C.
PCR products were separated on a 2.5% agarose gel, stained with ethidium bromide
117
Chapter 5
and visualized by UV illumination. Non-template control (water), positive/negative

control and normal control were used in each reaction.
QMSP: Real time quantitative methylation specific PCR

QMSP was performed as we described previously with an internal (FAM-ZEN/IBFQ)-
labeled hybridization probe for quantitative analyses18. Primer and probe sequences
are available in Supple. TableS2. ȕ-actin was used as a methylation independent
internal reference gene. QMSP reactions were performed in 10 μl final volume in 384
well plates, containing: 300 nM of forward and reverse primers, 250 nM of
hybridization probe, 5 μl of 2* QuantiTech Probe PCR Master Mix (Qiagen, Hilden,
Germany) and 2.5 μl bisulfite modified DNA (approximately 25 ng). Each sample was
analyzed in triplicate by ABI PRISM ® 7900HT Sequence Detection System (Applied
Biosystems, Carlsbad, CA, USA). Negative and positive controls were included in
each QMSP. Standard curve analysis was performed on each plate and by each
primers-probe set on serial dilutions of in vitro methylated leukocyte DNA. A DNA
sample was considered methylated if at least 2 out of the 3 wells were methylation
positive with a Ct-value below 50 and DNA input of at least 225 pg ȕ-actin. The
relative level of methylation of the region of interest was determined by the following
calculation: (average DNA quantity of methylated gene of interest / average DNA
quantity for internal reference gene ȕ-actin) x 1000035.
Statistical analysis:
Statistical analysis was performed using SPSS Statistics 20.0 (SPSS 20, Chicago, IL,
USA). Chi-square test and Fisher’s exact test for small numbers were used to
analyze the different methylation frequency between normal and cancer. The
correlation between the average methylation level of each frozen tissue sample and
MelthyCap-seq reads were analyzed using Spearman’s rank test. The Mann-Whitney
U test was used to determine differences in methylation ratio between 2 groups. P-
value<0.05 was considered to be statistically significant. The sensitivity, specificity,
receiver operating characteristic curves (ROC) and area under ROC curve (AUC)
were calculated for the clinical validation36. The optimal threshold was calculated
based on the largest Youden’s index37,38.
118
Results:
Identification of methylated candidates by MethylCap-seq:
To identify DMRs in ADC and SCC compared to normal cervices, genome-wide
MethylCap-seq was performed. After we applied our criteria based-on methylation
frequency, 6,231 DMRs showed differential methylation in ADC compared to normal
cervices and 10,724 DMRs were identified in SCC compared to normal cervices (Fig
2). In ADC as well as in SCC also hypomethylation was more frequently observed
compared to normal cervices (Fig 3).
ADC group SCC group

ADC (n=6) vs. Normal (n=20) SCC (n=6) vs. Normal (n=20)
Statistical analysis P<0.05
6,231 DMRs 10,724 DMRs
Additional criteria:
Specificity 75% & Sensitivity 50%
Region length 30bp
Low methylation level in leucocyte pools
Fig2. Identification of methylated candidates by MethylCap-seq
119
Chapter 5
Fig3. Distribution of DMRs in both ADC and SCC groups
Additionally, we strengthened our criteria by focusing on the hypermethylated DMRs,

as these are more easily to translate to MSP assays, which can be implemented as
clinical diagnostic tests. Overall 446 DMRs, comprising 357 genes, were identified in
ADC and 93 DMRs, comprising 89 genes, in SCC. Gene ontology (GO) functional
analysis for these DMRs was performed to determine if similar pathways are affected
in both histological types of cancer. There were in total 328 and 49 GO terms
enriched in ADC and SCC, respectively. Most GO terms enriched in SCC were also
enriched in ADC, as 37/38 of the biological processes, 4/5 of the cellular components
and 5/6 of the molecular functions, respectively, were also shown in ADC. This
underlines that similar pathways are disrupted in the carcinogenesis of both
histological types of cervical carcinomas.
Figure 4 shows the P-values of the top5 GOs enriched in ADC together with the
associated P-values in SCC. The most significant common pathways were
sequence-specific DNA binding (GO:0043565), transcription factor activity
(GO:0003700), transcription regulator activity (GO:0030528), plasma membrane
(GO:0005886), neuron differentiation (GO:0030182), which indicates that some
hyperDMRs are associated with transcription regulation.
120
GO: Biological Process
GO: Cellular Component
GO: Molecular Function
Fig4. The top5 GOs enriched in ADC including the associated P-value of SCC.
121
Chapter 5
Verification and validation of the top15 candidates:

In order to identify methylation markers common in both histological groups, the
DMRs identified from the two cancer groups were cross-compared, which generated
53 DMRs comprising 50 genes (Fig 2). Of these 53 candidates, the top 15 candidate
markers (Supple Table S3) were selected to verify the MethylCap-seq data by either
MSP or pyrosequencing analysis. Using the same DNA as was used for MethylCap-
seq, it was shown that for 10 genes (SOX1, GFRA1, SLC6A5, TBX5, OLIG2,
AC004963.1, TBX20, AC096537.2(219), CR753863.1, SOX14) a significant
correlation remained between the PCR band intensity determined by MSP or the
percentage of pyrosequencing and the number of reads from the MethylCap-seq.
Table 1 shows an overview of which genes survived the different stages of
validation.
Table1: Verification, validation and diagnostic evaluation of the top15 candidates

Diagnostic
Rank Gene Verification Validation
evaluation
Significant
Primer Significant Primer Significant
difference
optimization association optimization difference
(Yes/No)
(Yes/No) (Yes/No) (Yes/No) (Yes/No)
1 OLIG2 Yes Yes No
2 CR753863.1 Yes Yes No
3 GFRA1 Yes* Yes* Yes Yes Yes
4 EVX2 Yes No
5 AL356961.2 Yes No
6 TBX5 Yes Yes Yes No
7# AC096537.2(218) Yes No
8 SOX1 Yes* Yes* Yes Yes Yes
9 SYT6 Yes No
10 # AC096537.2(219) Yes Yes No
11 FREM3 Yes No
12 TBX20 Yes Yes Yes Yes Yes
13 AC004963.1 Yes Yes No
14 SLC6A5 Yes Yes Yes Yes Yes
15 SOX14 Yes Yes Yes Yes Yes
# Same gene, different region
* These genes were verified by MSP, the remaining genes were verified by pyrosequencing
122
For these 10 markers (SOX1, GFRA1, SLC6A5, TBX5, OLIG2, AC004963.1, TBX20,
AC096537.2(219), CR753863.1, SOX14), MSP primers were designed. Four genes
showed high methylation levels in leukocytes and WGA and were therefore excluded
from further validation. Methylation patters of the remaining 6 genes were analyzed
on DNA from an independent series comprising normal cervix and cervical cancers
(ADC as well as SCC). Except for TBX5, 5 genes (SOX1, SOX14, GFRA1, SLC6A5,
TBX20) showed a significant difference of methylation positivity between normal and
cancer (P<0.05), with a methylation frequency in both ADC and SCC >50% (Table 2).
Table 2: Methylation positivity in external validation cohort of FFPF samples
Positive rate
Genes Positive rate
Cancer total ADC SCC
SOX1* 0% (0/15) 91.67% (11/12) 100% ( 5/5 ) 85.7% (6/7)
GFRA1* 0% (0/11) 83%(10/12) 83.33% (5/6) 83.33% (5/6)
SOX14* 25% (4/16) 83%(11/13) 83.33% (5/6) 85.7% (6/7)
SLC6A5* 6.67% (1/15) 83.33% (10/12) 100% (5/5) 71.4% (5/7)
TBX20* 5.88% (1/17) 83.33% (10/12) 100% (5/5) 71.4% (5/7)
TBX5 56.25% (9/16) 67% (8/12) 60% (3/5) 71.4% (5/7)

* Comparison of positive rate in normal cervices vs. cancer by Fisher exact test (P<0.05)
Diagnostic evaluation by QMSP for normal versus cancer scrapings

To determine their diagnostic performance, QMSP was set up for 5 genes (SOX1,
SOX14, GFRA1, SLC6A5, TBX20) and evaluated on scrapings from a large series of
cervical cancer patients (n=125, (ADC: n=57) and SCC: n=68) and healthy, age-
matched controls (n=89). QMSP analysis indicated that the level of DNA methylation
for all five genes was significantly higher in cancer scrapings compared to normal
scrapings (P<0.001), but as expected similar between ADC and SCC (Fig5).
123
Chapter 5
Fig5˖Scatter plot in cervical scrapings of women with normal cervix or cervical cancer
patients.
Mann-Whitney U test shows significant difference between normal and cancer
(P<0.001) for all 5 genes ;
Mann-Whitney U test shows no difference between SCC and ADC (P>0.05) for all 5
genes.
124
Because many markers were also methylated in normal scrapings, albeit at lower
levels as observed in cancer scrapings, a threshold was calculated at the highest
Youden´s index based on the ROC analyses of the individual genes. Subsequently,
sensitivity and specificity to detect ADC and/or SCC for all individual genes were
determined (Table 4). The sensitivity for the 5 genes ranged from 80.5%-91.9%, while
the specificity ranged from 94% -98.9% (Table 3a). For four genes (SOX1, SOX14,
SLC6A5, TBX20), the sensitivity to detect either ADC or SCC was comparable, albeit
slightly lower for ADC than for SCC.
In order to discover an optimal methylation marker panel with the highest sensitivity
and specificity different gene combinations were evaluated. For a combination, a
sample was considered positive if either one of the genes was positive. As expected,
the combination of 2-3 genes decreased the specificity with the highest specificity
calculated at 97.6% (Table 3b). However, simultaneously the sensitivity for cervical
cancer increased by most of the combinations. The combination of SOX1, GFRA1
and SLC6A5, showed the highest AUC (0.959) with a sensitivity of 94.3% and a
specificity of 97.6%. The methylation positivity for ADC and SCC was 89.5% and
98.5%, respectively (Table 3b). Addition of other genes did increase neither
sensitivity nor specificity.
Discussion
Currently, an effective early detection method for ADC and its precursors is lacking
and therefore ADC is mainly diagnosed in advanced stages. Screening for cervical
cancer by Pap smear analysis is associated with significant false positive and false
negative rates39 and especially ADC are easily missed.40,41 Compared to cytology,
hrHPV screening will detect more (pre)malignant cervical cancers irrespective of
histology. However, in population-based screening programs hrHPV-positive women
will require triage analysis, which are until now still more specific for SCC and its
precursors than for ADC42,13. Therefore, it is important to find better methods to
improve the detection of the ADC. Recently, many studies have shown that
inactivation of tumor suppressor genes by promoter hypermethylation is an early
event in (cervical) carcinogenesis,43,44 which may simultaneously also serve as
suitable markers to allow early diagnosis45.
125
Chapter 5
Table3a: Sensitivity and specificity of individual genes in cervical scrapings of women

with normal cervix or cervical cancer patients (with cut-off)
Gene AUC# Threshold Specificity Sensitivity
Name Cancer ADC SCC for
total ADC vs SCC
SOX1 0.962 25.085 98.9% 89.6%*** 87.7% 91.2% P=0.528
SOX14 0.960 70.309 96.4% 91.9%*** 87.7 % 95.5% P=0.118
GFRA1 0.919 40.925 97.6% 80.5%*** 70.2% 89.4% P=0.007
TBX20 0.935 140.383 94.0% 86.2%*** 82.5% 89.4% P=0.266
SLC6A5 0.928 315.308 97.6% 81.3%*** 82.5% 80.3% P=0.760
# AUC without threshold
*** Comparison of positive rate in women with a normal cervix or cervical cancer by chi-square
(P<0.001)
Table3b: Sensitivity and specificity of combination of genes1 in cervical scrapings of

women with normal cervix or cervical cancer patients
Gene names AUC# Specificity Sensitivity
Cancer ADC SCC
total
Combination 2 genes
SOX1/GFRA1 0.955 97.6% 93.5% 87.7% 98.5%
SOX1/SOX14 0.949 96.4 % 93.5% 89.5% 97.0%
GFRA1/ SOX14 0.945 96.4% 92.7% 87.7 % 97.0%
GFRA1/ SLC6A5 0.935 97.6% 89.4% 84.2% 93.9%
SOX1/ TBX20 0.931 92.8 % 93.5% 87.7 % 98.5%
SOX14/ SLC6A5 0.921 96.4% 92.7% 89.5% 95.5%
SOX1/SLC6A5 0.947 97.6% 91.9% 89.5% 93.9%
SOX14/ TBX20 0.921 91.6 % 92.7% 87.7% 97.0%
GFRA1/ TBX20 0.911 92.8 % 89.4% 82.5% 95.5%
SLC6A5/ TBX20 0.915 92.8 % 90.2% 86.0% 93.9%
Combination 3 genes
SOX1/GFRA1/ SOX14 0.953 96.4% 94.3% 89.5 % 98.5%
SOX1/GFRA1/ SLC6A5 0.959 97.6% 94.3% 89.5 % 98.5%
GFRA1/ SOX14/ SLC6A5 0.949 96.4% 93.5% 89.5% 97.0%
SOX1/SOX14/ SLC6A5 0.949 96.4% 93.5% 89.5 % 97%
SOX1/SOX14/ TBX20 0.929 91.6 % 94.3% 89.5% 98.5%
SOX1/GFRA1/ TBX20 0.931 92.8 % 93.5% 87.7 % 98.5%
SOX1/SLC6A5/ TBX20 0.935 92.8 % 94.3% 89.5% 98.5%
GFRA1/ SOX14/ TBX20 0.921 91.6 % 92.7% 87.7% 97.0%
GFRA1/ SLC6A5/ TBX20 0.923 92.8 % 91.9% 86.0% 97.0%
SOX14/ SLC6A5/ TBX20 0.925 91.6% 93.5% 89.5% 97.0%
1: Combination of genes is made after a threshold was set (threshold is shown in table 3a). For a
combination, a sample was considered positive if either one of the genes was positive.
# AUC based on combination of genes.
126
In this study, we performed an unbiased genome-wide DNA methylation analysis,

comparing cervical cancer, both SCC as well as ADC with and healthy cervical
epithelium. This study is one of the first providing an overview of the DNA methylome
for ADC with simultaneous identification of new methylation markers for ADC.
Finally, 5 new methylation markers were systematically validated for the early
detection of both ADC and SCC in cervical scrapings. A DNA methylation marker
panel with a specificity of 97.6% and sensitivity of 94.9% for cervical cancer was
identified with methylation positivity for ADC of 85.7% and SCC of 98.5%. In
carcinogenesis, cancer cells often exhibit imbalanced expression of oncogenes and
tumor-suppressor genes, thus acquiring preferential growth ability46. It is well
established that aberrant DNA methylation may lead to overexpression of oncogenes
and/or repression of tumor-suppressor genes. Analyseis of those aberrant
methylation patterns eg, by (Q)MSP indicate that alterations in DNA methylation
patterns may be used as cancer biomarkers eg, for early diagnosis.44,47-49 In our
study, many differentially methylated markers, either hypo- or hypermethylated, were
observed when normal cervices were compared with ADC and SCC, respectively.
ADC develops from mucus-producing glandular cells, while SCC most often occurs at
the squamous cells4,12. There is some evidence to suggest that ADC and SCC may
also be associated with different epidemiologic co-factors. Smoking and high parity
are risk factors for SCC50, while obesity is a risk factor for ADC51. Since epigenetic
marks reflect both an individual’s genetic background and exposure to different
environmental factors52, we performed gene ontology (GO) analysis on the
differentially hypermethylated markers in normal vs ADC and normal vs SCC to
identify possible common disrupted pathways. Most of the pathways disrupted by
hypermethylation in SCC were also disrupted in ADC, indicating highly similar
disruption of pathways by hypermethylation during carcinogenesis independent of
histological cancer subtype. Pathways identified were among others sequence-
specific DNA binding, transcription factor activity and transcription regulator activity,
53-55
all known to be involved in carcinogenesis . Of the 53 differentially methylated
candidates that were found in both ADC and SCC, 18 genes (Supplemental table 3)
were described previously in literature as being more frequently methylated in cancer,
and 4 genes were previously related to (squamous) cervical cancer (SOX156, SOX14,
127
Chapter 5
ONECUT156, WT1).
In this study, we combined MethylCap-seq to draw detailed methylome maps. Bock

et al57 compared MeDIP, MethylCap and bisulfite-based methods. MethylCap and
MeDIP provide broad coverage of the genome with higher coverage for MethylCap..
Both methods allow comparable distinction between methylated and unmethylated
regions as bisulfite-based methods, but are less accurate for quantifying the DNA
methylation levels in partially methylated genomic regions. However, as to DMRs,
MethylCap was able to detect roughly twice as many in comparison to MeDIP at
comparable sequencing depths. Nevertheless, every analytical technique has its
limitations. Current methodology cannot differentiate between the two alleles of a
gene. Presence of methylation does therefore not exclude that an active copy of the
gene might be present.
In order to identify methylation markers for a diagnostic setting, we focused on the

hypermethylated candidate genes. Using genome-wide MethylCap-seq it is difficult to
draw an authentic map for the methylation of single CG sites and therefore a more
sensitive and quantitative technique, i.e. pyrosequencing was applied for verification.
A significant correlation was observed between MethylCap-seq and pyrosequencing.
Subsequently, primers for (Q)MSP were designed as these assays are more suitable
in a diagnostic setting. Common clinical indexes such as sensitivity, specificity, ROC
and AUC were determined to allow a first impression of the diagnostic efficiency of
each candidate marker. All five identified candidate markers appeared to be quite
capable to discriminate between normal epithelium versus cancer. However, if no
threshold was set a relatively low specificity was observed for SOX14, SLC6A5 and
TBX20. Subsequently, an optimal threshold was set to improve the specificity.
Youden’s index is an easy method to set up an effective cut-off 38. In our dataset, with
the optimal cut-off, the specificity increased, while not losing too much sensitivity. In
addition, marker combinations are also common choices to enhance the accuracy of
clinical diagnosis42. From our analysis, the best result came from a 3 marker panel
with a specificity of 97.6% and a sensitivity of 94.3%. So far only a limited number of
methylated genes have been examined in ADC, especially using cervical scrapings in
a large series. These studies revealed no adequate markers for clinical application,
128
as most of them had lower sensitivity for both ADC and SCC or either one.23,24 Two
genes of the Wnt pathway, DKK3 and SFRP2, showed more methylation in ADC
tissue compared to SCC tissue (82% vs. 18% and 84% vs. 39%) and combined they
allowed detection of all AdCIS and ADC. However, this analysis was limited to a very
small number of scrapings (n=8)23. Recently, PAX1, PTPRR, SOX1 and ZNF582,
previously reported to be frequently methylated in SCC scrapings, were also
analyzed in ADC scrapings and showed a sensitivity of the single genes of
81.8%~93.3% with a specificity of 81.0%~95.2% in a Taiwanese population24.
However, the sample size of scrapings again is small and a threshold was used to
score samples methylation negative or positive. Combined this might easily affect the
specificity when a larger number of normal scrapings is analyzed. 24.
Some of the 5 genes, that we found to be potential biomarkers, have previously been
reported to be methylated in cancer. SOX1 and SOX14 belong to the SOX family.
SOX proteins are the best examples of transcription factors having similar DNA
58
binding specificities yet with divergent functions . SOX1 encodes a transcription
factor implicated in the regulation of embryonic development and in the determination
of cell fate. DNA methylation of SOX1 in cervical cancer has been reported by Lai et
al24,56. Furthermore, SOX1 was identified as a tumor suppressor gene, because it
interfered with Wnt/ȕ-catenin signaling in cervical cancer59 and hepatocellular cancer60.
Hypermethylated SOX1 was also found in ovarian cancer cells that are chronically
exposed to cisplatin61. SOX1 methylation, at least in part, is responsible for cisplatin
resistance in human non-small cell lung cancer (NSCLC)62. SOX14, in contrast to our
data, has been reported to be a good marker to differentiate between ADC and SCC,
with more methylation in SCC as determined by an array-based technique63.
GFRA1, GDNF (glial cell line-derived neurotrophic factor) family receptor alpha 1, is a
member of the GDNF receptor family, and mediates activation of the RET tyrosine
kinase receptor. This gene is a candidate gene for Hirschsprung disease64 and in lung
adenocarcinoma its methylation status determines tumor aggressiveness and outcome65.
Furthermore, in a recent study of our group GFRA1 methylation was also identified as a
methylation marker for cervical intraepithelial neoplasia CIN2/3 (see Chapter 4).
129
Chapter 5
T-box (TBX) transcription factors belong to an ancient gene family with critical roles in
embryogenesis, in early cell fate decisions and in control of differentiation and
organogenesis. TBX20 and TBX5 act synergistically to control vertebrate heart
morphogenesis. Currently, TBX20 and TBX5 are TBX genes defined to have multiple
protein isoforms created by alternative splicing and characterized by expression and
functional studies. These proteins are important for development as mutations lead to
severe developmental disorders in humans. Regulation of TBX transcription factor
activity has been characterized through protein interactions and DNA binding
affinities66 . Only TBX20 methylation has previously been related to the recurrence of
lung adenocarcinoma65.
SLC6A5 (GLYT2) encodes a sodium- and chloride-dependent glycine

neurotransmitter transporter important for the clearance of extracellular glycine during
glycine-mediated neurotransmission. Mutations in this gene cause hyperekplexia, a
heterogeneous neurological disorder. However, so far, there is no report showing the
relationship between SLC6A5 and cancer.
Conclusions:
Overall, our approach resulted in new cervical cancer methylation markers with high
specificity and high sensitivity for ADC as well as for SCC. Preliminary results
showed that especially SOX1, GFRA1 and SLC6A5 combined might be promising
markers. However, further research is needed to validate the clinical significance and
reproducibility for these markers. For instance, validation of these markers in a
population-based screening setting, particular for ADC precursor lesions such as
cervical AdCIS has not been tested yet. Knowledge of the pathogenesis-associated
epigenetic alterations based on the methylome analysis of ADC and SCC may result
in new targets for both therapeutic as well as diagnostic approaches.
Reference
1. Ferlay J, Shin HR, Bray F, et al: Estimates of worldwide burden of cancer in 2008: GLOBOCAN
2008. Int J Cancer 127:2893-917, 2010
2. Vizcaino AP, Moreno V, Bosch FX, et al: International trends in incidence of cervical cancer: II.
Squamous-cell carcinoma. Int J Cancer 86:429-35, 2000
130
3. Vizcaino AP, Moreno V, Bosch FX, et al: International trends in the incidence of cervical cancer: I.
Adenocarcinoma and adenosquamous cell carcinomas. Int J Cancer 75:536-45, 1998
4. Seoud M, Tjalma WA, Ronsse V: Cervical adenocarcinoma: moving towards better prevention.
Vaccine 29:9148-58, 2011
5. Organization. WH: WHO/ICO information on centre on human papillomavirus virus and cervical
cancer. aviailable at:http://apps.who.int/hpvcentre/statistics/dynamic/ico/DataQuery Select.cfm,
2010
6. Vinh-Hung V, Bourgain C, Vlastos G, et al: Prognostic value of histopathology and trends in cervical
cancer: a SEER population study. BMC Cancer 7:164, 2007
7. Bulk S, Visser O, Rozendaal L, et al: Cervical cancer in the Netherlands 1989-1998: Decrease of
squamous cell carcinoma in older women, increase of adenocarcinoma in younger women. Int J
Cancer 113:1005-9, 2005
8. Lai CH, Hsueh S, Hong JH, et al: Are adenocarcinomas and adenosquamous carcinomas different
from squamous carcinomas in stage IB and II cervical cancer patients undergoing primary radical
surgery? Int J Gynecol Cancer 9:28-36, 1999
9. Eifel PJ, Burke TW, Morris M, et al: Adenocarcinoma as an independent risk factor for
disease recurrence in patients with stage IB cervical carcinoma. Gynecol Oncol 59:38-44, 1995
10. Hopkins MP, Morley GW: A comparison of adenocarcinoma and squamous cell carcinoma of the
cervix. Obstet Gynecol 77:912-7, 1991
11. Chen RJ, Lin YH, Chen CA, et al: Influence of histologic type and age on survival rates for
invasive cervical carcinoma in Taiwan. Gynecol Oncol 73:184-90, 1999
12. Gien LT, Beauchemin MC, Thomas G: Adenocarcinoma: a unique cervical cancer. Gynecol Oncol
116:140-6, 2010
13. Kinney W, Fetterman B, Cox JT, et al: Characteristics of 44 cervical cancers diagnosed following
Pap-negative, high risk HPV-positive screening in routine clinical practice. Gynecol Oncol
121:309-13, 2011
14. Hoque MO, Kim MS, Ostrow KL, et al: Genome-wide promoter analysis uncovers portions of the
cancer methylome. Cancer Res 68:2661-70, 2008
15. Baylin SB, Ohm JE: Epigenetic gene silencing in cancer - a mechanism for early oncogenic
pathway addiction? Nat Rev Cancer 6:107-16, 2006
16. Esteller M, Corn PG, Baylin SB, et al: A gene hypermethylation profile of human cancer. Cancer
Res 61:3225-9, 2001
17. Yang N, Nijhuis ER, Volders HH, et al: Gene promoter methylation patterns throughout the process
of cervical carcinogenesis. Cell Oncol 32:131-43, 2010
18.Eijsink JJ, Lendvai A, Deregowski V, et al: A four-gene methylation marker panel as triage test in
high-risk human papillomavirus positive patients. Int J Cancer 130:1861-9, 2012
131
Chapter 5
19.Wentzensen N, Sherman ME, Schiffman M, et al: Utility of methylation markers in cervical cancer
early detection: appraisal of the state-of-the-science. Gynecol Oncol 112:293-9, 2009
20.Reesink-Peters N, Wisman GB, Jeronimo C, et al: Detecting cervical cancer by quantitative

promoter hypermethylation assay on cervical scrapings: a feasibility study. Mol Cancer Res 2:289-
95, 2004
21.Nijhuis ER, Reesink-Peters N, Wisman GB, et al: An overview of innovative techniques to improve
cervical cancer screening. Cell Oncol 28:233-46, 2006
22.Kang S, Kim JW, Kang GH, et al: Comparison of DNA hypermethylation patterns in different types
of uterine cancer: cervical squamous cell carcinoma, cervical adenocarcinoma and endometrial
adenocarcinoma. Int J Cancer 118:2168-71, 2006
23.van der Meide WF, Snellenberg S, Meijer CJ, et al: Promoter methylation analysis of WNT/beta-
catenin signaling pathway regulators to detect adenocarcinoma or its precursor lesion of the cervix.
Gynecol Oncol 123:116-22, 2011
24.Chang CC, Huang RL, Wang HC, et al: High methylation rate of LMX1A, NKX6-1, PAX1, PTPRR,
SOX1, and ZNF582 genes in cervical adenocarcinoma. Int J Gynecol Cancer 24:201-9, 2014
25.Ongenaert M, Wisman GB, Volders HH, et al: Discovery of DNA methylation markers in cervical
cancer using relaxation ranking. BMC Med Genomics 1:57, 2008
26.Weber M, Davies JJ, Wittig D, et al: Chromosome-wide and promoter-specific analyses identify
sites of differential DNA methylation in normal and transformed human cells. Nat Genet 37:853-62,
2005
27.Wilson IM, Davies JJ, Weber M, et al: Epigenomics: mapping the methylome. Cell Cycle 5:155-8,
2006
28.Hurd PJ, Nelson CJ: Advantages of next-generation sequencing versus the microarray in
epigenetic research. Brief Funct Genomic Proteomic 8:174-83, 2009
29.Huang RL, Gu F, Kirma NB, et al: Comprehensive methylome analysis of ovarian tumors reveals
hedgehog signaling pathway regulators as prognostic DNA methylation biomarkers. Epigenetics
8:624-34, 2013
30.Huang TT, Gonzales CB, Gu F, et al: Epigenetic deregulation of the anaplastic lymphoma kinase
gene modulates mesenchymal characteristics of oral squamous cell carcinomas. Carcinogenesis
34:1717-27, 2013
31.Carvalho RH, Haberle V, Hou J, et al: Genome-wide DNA methylation profiling of non-small cell
lung carcinomas. Epigenetics Chromatin 5:9, 2012
32.van Dongen JJ, Langerak AW, Bruggemann M, et al: Design and standardization of PCR primers
and protocols for detection of clonal immunoglobulin and T-cell receptor gene recombinations in
suspect lymphoproliferations: report of the BIOMED-2 Concerted Action BMH4-CT98-3936.
Leukemia 17:2257-317, 2003
33.Langmead B, Trapnell C, Pop M, et al: Ultrafast and memory-efficient alignment of short DNA
132
sequences to the human genome. Genome Biol 10:R25, 2009

34.De Meyer T, Mampaey E, Vlemmix M, et al: Quality evaluation of methyl binding domain based kits
for enrichment DNA-methylation sequencing. PLoS One 8:e59068, 2013
35.Wisman GB, Nijhuis ER, Hoque MO, et al: Assessment of gene promoter hypermethylation for
detection of cervical neoplasia. Int J Cancer 119:1908-14, 2006
36.Wang R, Wang G, Zhang N, et al: Clinical evaluation and cost-effectiveness analysis of serum
tumor markers in lung cancer. Biomed Res Int 2013:195692, 2013
37.Youden WJ: Index for rating diagnostic tests. Cancer 3:32-5, 1950
38.Fluss R, Faraggi D, Reiser B: Estimation of the Youden Index and its associated cutoff point. Biom
J 47:458-72, 2005
39.Koss LG: Cervical (Pap) smear. New directions. Cancer 71:1406-12, 1993
40.Zappa M, Visioli CB, Ciatto S, et al: Lower protection of cytological screening for adenocarcinomas
and shorter protection for younger women: the results of a case-control study in Florence. Br J
Cancer 90:1784-6, 2004
41.Smith HO, Tiffany MF, Qualls CR, et al: The rising incidence of adenocarcinoma relative to
squamous cell carcinoma of the uterine cervix in the United States--a 24-year population-based
study. Gynecol Oncol 78:97-105, 2000
42.Steenbergen RD, Snijders PJ, Heideman DA, et al: Clinical implications of (epi)genetic changes in
HPV-induced cervical precancerous lesions. Nat Rev Cancer 14:395-405, 2014
43.Jones PA, Laird PW: Cancer epigenetics comes of age. Nat Genet 21:163-7, 1999
44.Feinberg AP, Tycko B: The history of cancer epigenetics. Nat Rev Cancer 4:143-53, 2004
45.Baylin SB, Esteller M, Rountree MR, et al: Aberrant patterns of DNA methylation, chromatin
formation and gene expression in cancer. Hum Mol Genet 10:687-92, 2001
46.Hanahan D, Weinberg RA: The hallmarks of cancer. Cell 100:57-70, 2000
47.Ehrlich M: DNA methylation in cancer: too much, but also too little. Oncogene 21:5400-13, 2002
48.Esteller M: Epigenetics in cancer. N Engl J Med 358:1148-59, 2008
49.Tsai HC, Baylin SB: Cancer epigenetics: linking basic biology to clinical medicine. Cell Res 21:502-
17, 2011
50.Lacey JV, Jr., Frisch M, Brinton LA, et al: Associations between smoking and adenocarcinomas and
squamous cell carcinomas of the uterine cervix (United States). Cancer Causes
Control 12:153-61, 2001
51.Lacey JV, Jr., Swanson CA, Brinton LA, et al: Obesity as a potential risk factor for
adenocarcinomas and squamous cell carcinomas of the uterine cervix. Cancer 98:814-21, 2003
133
Chapter 5
52.Wild CP, Scalbert A, Herceg Z: Measuring the exposome: a powerful basis for evaluating
environmental exposures and cancer risk. Environ Mol Mutagen 54:480-99, 2013
53.Robertson KD, Jones PA: DNA methylation: past, present and future directions. Carcinogenesis
21:461-7, 2000
54.Herman JG, Baylin SB: Gene silencing in cancer in association with promoter hypermethylation. N
Engl J Med 349:2042-54, 2003
55.Baylin SB, Herman JG: DNA hypermethylation in tumorigenesis: epigenetics joins genetics. Trends
Genet 16:168-74, 2000
56.Lai HC, Lin YW, Huang TH, et al: Identification of novel DNA methylation markers in cervical cancer.
Int J Cancer 123:161-7, 2008
57.Bock C, Tomazou EM, Brinkman AB, et al: Quantitative comparison of genome-wide DNA
methylation mapping technologies. Nat Biotechnol 28:1106-14, 2010
58.Malas S, Duthie S, Deloukas P, et al: The isolation and high-resolution chromosomal mapping of
human SOX14 and SOX21; two members of the SOX gene family related to SOX1, SOX2, and
SOX3. Mamm Genome 10:934-7, 1999
59.Lin YW, Tsao CM, Yu PN, et al: SOX1 suppresses cell growth and invasion in cervical cancer.
Gynecol Oncol 131:174-81, 2013
60.Tsao CM, Yan MD, Shih YL, et al: SOX1 functions as a tumor suppressor by antagonizing the
WNT/beta-catenin signaling pathway in hepatocellular carcinoma. Hepatology 56:2277-87, 2012
61.Yu W, Jin C, Lou X, et al: Global analysis of DNA methylation by Methyl-Capture sequencing
reveals epigenetic control of cisplatin resistance in ovarian cancer cell. PLoS One 6:e29450, 2011
62.Li N, Li X, Li S, et al: Cisplatin-induced downregulation of SOX1 increases drug resistance by

activating autophagy in non-small cell lung cancer cell. Biochem Biophys Res Commun 439:187-
90, 2013
63.Senchenko VN, Kisseljova NP, Ivanova TA, et al: Novel tumor suppressor candidates on
chromosome 3 revealed by NotI-microarrays in cervical cancer. Epigenetics 8, 2013
64.Lui VC, Samy ET, Sham MH, et al: Glial cell line-derived neurotrophic factor family receptors are
abnormally expressed in aganglionic bowel of a subpopulation of patients with Hirschsprung's
disease. Lab Invest 82:703-12, 2002
65.Sato T, Arai E, Kohno T, et al: DNA methylation profiles at precancerous stages associated with
recurrence of lung adenocarcinoma. PLoS One 8:e59444, 2013
66.Yu J, Ma X, Cheung KF, et al: Epigenetic inactivation of T-box transcription factor 5, a novel tumor
suppressor gene, is associated with colon cancer. Oncogene 29:6464-74, 2010
134
135
Chapter 5
136
137
Chapter 5
138
139
Chapter 5
140
Summary and future perspectives
Chapter 6
141
Chapter 6
Summary
Although cytomorphology-based cervical screening has reduced the incidence of
cervical cancer in the Western world, it is predicted that by 2030 the number of
deaths from cervical cancer will increase to 410,000 worldwide1,2,3. The progress in
prevention, early detection and treatment of cervical cancer has been hampered by
the fact that the majority of cases occur in the developing world where infrastructure,
human resources and medical facilities with sufficient quality assurance are limited or
absent4. As one of the developing countries, China plays an important role in the
global fight against cervical cancer. However, several challenges play a major role in
this fight such as the huge population and area, the rapid, unbalanced economic
growth and the mass migration of the last years.
Due to its low sensitivity and high labor-intensity, cytomorphology-based examination

of cervical smears may not be an ideal screening method in China5. HPV testing is
another well-known approach for early detection of cervical cancer. Introduction in
China will generate a huge and chaotic market for numerous HPV detection
platforms with different specifications and therefore different (dis)advantages. It will
be necessary to choose an appropriate and validated platform before HPV detection
is introduced as a primary screening tool. Simultaneously, another important question
to be answered first is how to manage the HPV positive patient, particular for the
densely populated, but less-developed regions. Therefore, to explore feasible
cervical screening tools which may cover and benefit the large cohorts effectively,
new insights in cervical cancer biomarkers including HPV testing and DNA
methylation markers are investigated and discussed in this thesis.
Globally HPV DNA screening has been performed for more than 15 years and
demonstrated that the prevalence of HPV infection varies geographically and
racially6,7. In China, although several projects regarding the hrHPV infection rate and
cervical cancer screening were conducted8 (Table 4 in Chapter 1), the data are still
far from a nationwide standard and this information should be updated with time in
order to provide reference for effective screening. In Chapter 2 we described a
nationwide cross-sectional investigation to investigate HPV prevalence in most
regions in China, to follow up the potential changes of HPV infection in the regions
142
previously studied and to clarify the genotypic distribution of HPV in different regions
and age-grouped populations. In this study 120,772 liquid-based cytological samples
from women enrolled for population- or employee-based cervical screening programs
in 37 Chinese cities in 2012, were obtained and sent to the laboratory of molecular
infectious diseases of Guangzhou KingMed. Overall 111,131 samples were tested by
Hybrid Capture 2 and 9,641 samples were genotyped by TellgenplexTM HPV DNA
Assay. The total positive rate of high-risk HPV (hrHPV) was 21.07%, ranging from
31.94% to 18.42% in different regions. Age-specific prevalence showed a “two peak”
pattern. The youngest age-group (15-19 years) presented the highest hrHPV
infection rate (30.55%) and the second peak was found in the age group of 50-60
years. HPV16 (4.82% of all samples) and HPV52 (4.52% of all samples) were the
most prevalent hrHPV types, followed by HPV58 (2.74% of all samples); while HPV6
(4.01% of all samples) and HPV11 (2.29% of all samples) were predominant in the
low risk HPV (lrHPV) types. HPV16+HPV52 and HPV52+HPV58 were the most
common coexistence genotypes in case of multiple infections. In conclusion, HPV
infection rate in China has increased into the levels of HPV-heavy-burden country
zones and the rates have changed depending on regions along with time. Therefore,
this investigation provides an updated and valuable reference for an effective
nationwide cervical cancer screening program. Particularly, the analysis of genotypic
distribution of hrHPV has again demonstrated that HPV52 and HPV58 are two of the
dominant genotypes in China, which underlines that the currently available
commercial vaccines, Cervarix (HPV16/18) and Gardasil (HPV6/11/16/18) may not
be fully effective within the Chinese population. These epidemiological characteristics
should be taken into account when considering implementation of screening and
vaccination programs in China.
Currently, many new screening technologies are developed and “marketed” in China
with no validation of their effectiveness for population-based screening including HPV
assays9. A few investigators are trying to properly evaluate algorithms that can be
applied to the highest-risk areas in China as well as the more urban areas. Today
many commercial HPV DNA detection assays are available but only few have been
formally validated in primary cervical screening including the Hybrid Capture 2 (HC2)
hrHPV DNA test, the first hrHPV test approved by FDA in the US11. The Cervista
143
Chapter 6
HR HPV test (Hologic Inc), the second hrHPV assay approved by the FDA is widely
used in the US and in some Chinese studies. Cervista has a few advantages over
HC2: It requires less sample volume, includes an internal control, and has a shorter
processing time. However, Cervista has not previously been formally validated in
primary cervical screening. In Chapter 3, we demonstrated that Cervista fulfilled
the cross-sectional clinical performance and reproducibility criteria of international
guidelines for HPV test requirements for cervical screening. By non-inferiority analysis
following international guidelines the clinical sensitivity and specificity of Cervista were
compared to that of HC2 for detection of high-grade cervical lesions (CIN2+) in Chinese
women aged 30 years in >7,000 cervical screening samples selected from a multisite,
population-based, cross-sectional study. Cervista showed a clinical sensitivity for
CIN2+ of 89% (95%CI: 81.7-93.6) and a clinical specificity for CIN2+ of 91% (95%CI:
90.5-91.8). Both the relative clinical sensitivity and specificity were non-inferior to that of
HC2 (non-inferiority score tests, p=0.043 and p<0.0001, respectively). In addition, we
also determined the intra- and inter-laboratory reproducibility of Cervista in 510
scrapings, again following international guidelines. Intra- and inter-laboratory
agreements were 92% (lower bound 95%CI: 89.7%; kappa = 0.83; p<0.001) and 90.4%
(lower bound 95%CI: 88.4%; kappa = 0.81; p<0.001), respectively. In conclusion, the
Cervista HR HPV test met the cross-sectional clinical and reproducibility criteria of the
international guidelines for HPV test requirements and can be considered as clinically
validated for primary cervical screening purposes.
Screening identifies not only early stage invasive cancers that are expected to be
more amenable to cure, but also pre-malignancies that easily can be removed and
thus prevented from progressing to invasive lesions10. However, the relatively lower
specificity of hrHPV testing, especially in a young screening population, may lead to
unnecessary referrals to the gynecologist, anxiety in the false-positive women, and
higher costs for the health-care system and indeed is a challenge for cervical
screening programs. In addition, in the near future the prevalence of CIN and cervical
cancer will probably decrease in countries, that have introduced primary prevention
with hrHPV vaccination. With this decrease in prevalence, the positive predictive
value of the current screening tests will by definition decrease.
144
Therefore, other objective biomarkers with both high sensitivity as well as high
specificity are needed as new screening or triage tools for cervical cancer.
Differential DNA methylation patterns in normal cervical epithelium versus
(pre)malignant cervical lesions represent excellent targets for diagnostic approaches
based on methylation specific PCR (MSP). In our previous studies, the
pharmacological unmasking of the promoter region combined with re-expression as
analyzed by microarrays, QMSP on an OpenArray platform and methyl-DNA
immunoprecipitation followed by microarray analysis (MeDIP), resulted in the
discovery of 4 genes (C13ORF18, JAM3, EPB41L3 and TERT)11,12,13. The diagnostic
performance of DNA methylation analysis of these genes showed sensitivities for
detecting CIN2+ in a hrHPV positive population between 43-71% and specificities
between 89-100%11. However, our strategies for discovering new methylation
markers so far were based on the difference between cancer and normal tissue,
resulting in markers with high sensitivity for carcinomas, but with too low sensitivity
for detecting CIN2/3 lesions. In Chapter 4, in order to identify new methylation
markers for high-grade cervical intraepithelial neoplasia (CIN2/3), a novel technique
(MethylCap-seq) was applied to identify differential methylation regions (DMRs) and
generate a CIN2/3 lesion-specific methylome. Comparing 20 normal cervices with 18
CIN2/3 lesions, 176 DMRs comprising 163 genes were identified. The top 15 highest
ranking differentially methylated genes were selected and validated by methylation
specific PCR (MSP) in two steps: on the same DNA samples as used for MethylCap-
seq and on DNA samples from an independent patient cohort with (pre)malignant
cervical neoplasia. For further diagnostic evaluation, the best differentiating
methylation markers were tested with quantitative MSP (QMSP) on cervical
scrapings from 2 cohorts: 1) cervical carcinomas versus healthy controls and 2)
patients referred from population-based screening with an abnormal Pap smear in
whom HPV status was determined.
After verification and validation of the top 15 genes with MSP, 9 genes showed
significant differential methylation in normal cervices versus CIN2/3 lesions (p<0.05).
Subsequently, methylation levels of 8/9 genes were significantly higher in carcinomas
compared to normal scrapings. For all 8 genes methylation levels increased with the
145
Chapter 6
severity of the underlying histological lesion in scrapings from patients with an

abnormal Pap smear. In addition to the 8 new genes, also methylation levels of our
previous four-gene panel (C13ORF18, JAM3, EPB41L3 and TERT) were analyzed
on the same samples. The combination C13ORF18/JAM3/AL590705.4 revealed a
similar high sensitivity for CIN2+ (74%) compared to hrHPV testing (79%), while
specificity was significantly higher (76%) compared to hrHPV testing (42%) (p0.05).
With this genome-wide DNA methylation analysis we identified new CIN2/3 specific
methylation markers. The diagnostic performance of our new methylation panel
showed comparable sensitivity to hrHPV testing for CIN2+, but with higher specificity,
therefore possibly preventing unnecessary referral for colposcopy. The next step
before implementation in primary screening programs will be validation of our new
methylation panel in population-based cohorts.
Cervical cancer encompasses several histological types, of which adenocarcinomas

(ADC) and cervical squamous cell carcinoma (SCC) are the two main subtypes,
accounting for 10–25% and 75–90%, respectively14,15,16. Currently, the incidence of
SCC is declining in most developed countries, while there is a rise in the absolute
and relative incidence of ADC15,17. Although ADC is mainly increasing rapidly in
Europe, and relatively slower in Asia (2.8-22.6%)16, the absolute cases of ADC still
cannot be ignored, because of its large population. Moreover, in comparison with
SCC, ADC is mainly diagnosed at a more advanced stage of disease16,17. Current
screening programs are more effective in detection of the precursors of SCC than
those of ADC. Major reason for the inefficacy of current screening methodology for
detection of ADC is its location higher up in the cervical canal, which hampers
acquisition of adequate samples for cytology and also direct observation by
colposcopic examination17. Next to hrHPV testing, analysis of DNA methylation
markers might improve the detection of ADC in earlier stages. However, no specific
methylation markers have been described for the detection of ADC as well as for
SCC. The aim of the study described in Chapter 5 was to discover novel methylation
biomarkers for cervical cancer detecting both ADC and SCC. Novel methylation
markers were identified using MethylCap-seq by comparison of 20 normal cervices
with 12 cancers (6 ADC, 6 SCC). The top 15 markers were selected for verification
by bisulfite pyrosequencing or MSP on the same samples used for MethylCap-seq.
146
Methylation frequencies were validated on a series of paraffin-embedded specimens

from normal cervices and cervical cancers by MSP. Finally, QMSP was performed on
cervical scrapings from an independent cohort of women with a normal cervix and
cervical cancer. Validation of the highest ranking 15 differentially methylated
candidate genes resulted in 5 markers exhibiting different methylation frequencies
between normal and cancer tissues (p<0.05). Using QMSP on DNA of cervical
scrapings, the sensitivity of these 5 markers varied from 80.5% to 91.9% to detect
both ADC and SCC with almost all normal scrapings negative (specificity: 94% -
98.9%). In conclusion, using the MethylCap-seq analysis, we identified 5 new
methylation markers for early detection of both ADC and SCC in cervical scrapings.
Further validation in a large series of population-based scrapings is needed, in
particular in scrapings from ADC and its precursors, to confirm our findings.
Future perspectives
A successful and effective national cervical cancer screening program depends on
an optimal coverage of the screening population, high-level quality assurance of the
screening methodology and proper management of positive findings18,19. Recently,
many observational studies20,21,22 and a systematic meta- analysis23 have
demonstrated that a cervical screening program effectively will reduce the morbidity
and mortality of cervical cancer in most high-income settings. However, cytology-
based screening has reached a plateau owing to the low sensitivity, low
reproducibility, highly variable results among laboratories and poor performance in
detection of cervical adenocarcinoma24,25,26. Liquid-based-cytology (LBC) is
considered to be superior over conventional Pap cytology by its standardization,
resulting in easier handling and higher reproducibility with a similar sensitivity for the
detection of CIN2/3 and adenocarcinomas27.
The strong link between HPV and cervical carcinogenesis paves new ways for
cervical cancer prevention. Immunization for hrHPV as a means of primary
prevention and HPV DNA testing in cervical screening for the early detection of
(pre)malignant lesions are now widely applied. In 2006, Gardasil, a quadrivalent
147
Chapter 6
vaccine and Cervarix, a bivalent vaccine were fully licensed28. Until the beginning of
2012, around 40 countries have introduced HPV vaccination into their national
immunization program28.
However, there are still some obstacles with respect to the currently approved
vaccines that limit their efficiency: 1) Immunization may fail to protect a significant
proportion of women, because with the current vaccines women will only be
effectively protected against HPV16 and HPV18. On the other hand, many other
hrHPV genotypes have been detected and the prevalence of the different other
hrHPV genotypes varies in different countries and regions within countries6. For
example, we found (Chapter 2) that HPV52 and HPV58 have the highest prevalence
after HPV16 in most of the regions in China. Therefore, using the more recently
developed polyvalent HPV vaccine against HPV6, HPV11, HPV16, HPV18, HPV31,
HPV33, HPV45, HPV52 and HPV58 (Merck’s Investigational 9-valent HPV Vaccine
V503) might be more effective in immunization programs in China29
ˈ30
. 2) The costs
of the approved HPV vaccines has also been a hurdle in the introduction of the
vaccines, especially in developing countries which carry the greatest burden of HPV-
associated diseases31. 3) Concerns about implications for young women’s future
sexual, physical and reproductive health, parental refusal or religious beliefs are all
factors that influence participation rates in the vaccination program32.
Despite HPV vaccination, cervical screening will remain an important aspect of

cervical cancer prevention, both for the vaccinated and non-vaccinated women. In
the Netherlands, the prophylactic vaccine Cervarix® against HPV16 and 18 was
introduced in the national immunization program in 200933. However, as the current
vaccines are being given to adolescent girls only, it will take 10-15 years before the
effect of vaccination on cervical cancer prevalence becomes clear. Furthermore, not
all adolescent girls are vaccinated, since only 50-60% of the girls show up for the
vaccination program in the Netherlands. Moreover, even vaccinated girls should be
screened when they reach the screening age since the current vaccine will not
provide complete protection against all oncogenic HPV types34. Therefore,
continuation of cervical cancer screening will remain necessary.
148
Apart from the vaccines against HPV16 and 18, hrHPV DNA is an important
molecular biomarker for the detection of cervical cancer and its precursors.
Compared to cytology-based screening methods, clinical trials have shown that HPV-
based screening results in greater sensitivity (range 94.1– 95.4%) than cytology-
based screening (range 55.4 – 71.3%) with some loss in specificity for CIN2+ (CIN2+;
cytology specificity range 96.8–98.6% compared with HPV 94.1 – 94.2%)35,36,37,38.
However, HPV-based testing has a far better negative predictive ability than
cytology-based testing. A negative HPV test immediately gives a reassurance close
to 100% for absence of disease (cytology detection less than 60%) and almost
guarantees protection for the absence of HSIL over a prolonged period, safely
allowing lengthening of the screening interval for at least 5 years39,40. So, the results
from a HPV test provide significantly more valuable information from a disease
screening perspective. In addition, compared to cytomorphology screening,
molecular markers are convenient, cheap, user-friendly and less labor intensive,
especially when using high-throughput automated platforms. These advantages
make them also feasible for the low-resource settings, avoiding logistical complexity
and expertise required by cytology.
In China prophylactic vaccines have not officially been approved yet. Therefore, at
this moment the most feasible strategy for cervical cancer prevention and control is
organization of a national population-based screening program, thereby promoting
diagnosis and treatment of cervical neoplasia at an earlier stage41. Since nearly 70%
of Chinese rural women live in poor regions lacking financial resources and more
than 80% women diagnosed with cervical cancer have never been screened, the
Chinese government has executed a pilot program for free cervical cancer screening
for rural women between 2009-201142. This program has been continued for another
3 years (2012-2015), because only 10 out of 500 million Chinese rural women could
access this service. Meanwhile, in accordance with the feasibility and cost-
effectiveness evaluation, the Ministry of Health of China has launched a guideline for
region-driven screening protocols by using the cytology test plus HPV test in the
more developed regions and visual inspection with acetic acid (VIA) in low-resource
settings43. As China is a large population country, cytology-based technique is not an
ideal screening method because of the requirement of a large number of experienced
149
Chapter 6
cytologists/pathologists and maintaining the large numbers of clinics required for

mass population screening. As to VIA, a “screen-and-treat” method was
recommended for a large number of high-risk cases from patients living in low-
resource regions. However, due to the low sensitivity of VIA, cancer patients may be
missed, while on the other hand, because of no quality control, there is also a risk for
overtreatment of the women due to false positive results44. Therefore, it would be
more effective to introduce objective, reproducible and high-throughput molecular
biomarkers assays instead of VIA and/or morphology testing as the screening
method in China.
In the Netherlands, one of the leading countries in cervical cancer prevention,

primary hrHPV-based screening will start in 201645. However, because of the much
higher sensitivity of HPV testing, but relatively lower specificity, more patients with
false-positive results will be referred to gynecologists, which may result in
unnecessary higher costs and more unwanted anxiety of involved women. So, the
question arises how to manage women who test hrHPV positive. Analysis of DNA
methylation markers is objective, can be implemented in high-throughput systems
and analyzed on the same DNA sample as used for the HPV assay, which presents
DNA methylation analysis as an attractive triage testing method for HPV-positive
patients. So far, several DNA methylation markers have been successfully identified
for cervical neoplasia by our group11,12,46 including the markers as described in this
thesis. However, before our markers will be ready for clinical practice, a gradual
validation process should be performed. Pepe et al.47 recommended a five-phase
framework: After preclinical exploratory studies, assessment in non-invasive samples,
retrospective longitudinal studies, prospective screening studies and prospective
intervention studies should be performed. Our methylation markers (Chapter 4) to
detect CIN2/3 lesions were already tested in samples obtained from the present
cytology-based cervical screening program. Therefore, as a next step for future HPV
DNA based screening programs further validation of our markers needs to be
performed in hrHPV positive scrapings for which histological data is also available,
collected from a prospective population-based screening program. In the Netherlands
this type of scrapings will become available after 2016 when primary cervical
screening has changed to hrHPV testing. Likewise, our marker panel (Chapter 5) for
150
the early detection of both ADC and SCC should be validated in a screening setting,
particular for the precursor stage of ADC (AdCIS).
In China, so far there is one population-based study48 that reported the methylation
frequency of 3 genes (DAPK1, RAR-ȕ2 and MGMT) using liquid-based cytology
samples from a large cohort. Another study identified several hypermethylated genes
using captured methylated DNA combined with CpG-microarray analysis49. Most
other Chinese studies50
ˈ51ˈ52ˈ53ˈ54
on methylation markers are based on cell lines
and tissue samples and therefore cannot be translated to cervical cancer screening
directly. Because the distribution of HPV genotypes differs between the Chinese
population (Chapter 2) and Western world, it might well be that the specificity and
sensitivity for detection of CIN2+ lesions using our methylation panel will be different
in a Dutch and Chinese population. Therefore, clinical validation on Chinese HPV-
positive scrapings is required.
Last but not least, transient HPV infections are common, especially among younger
women55. In chapter 2 we reported that the youngest age-group (15-19 years)
presented the highest hrHPV infection rate (30.55%) in a Chinese screening cohort.
As a result, in young women hrHPV testing to discriminate between HPV-positive
scrapings with CIN2+ lesions and HPV-positive scrapings of those women with
transient HPV infections only is very challenging. For that reason, some international
guidelines stated that hrHPV DNA testing for screening purposes should not be
performed below the age of 30 years56,57,58. Therefore, in the Netherlands the cervical
screening program starts at the age of 30 years but in many other countries (e.g. in
UK) screening programs start at the age of 25. According to the current Chinese
cervical cancer screening guideline, the screening program in China even starts at
the age of 21 years. Hence, a DNA methylation biomarker panel could be very
relevant to identify women aged 21-29 years with HPV-positive scrapings and with
CIN2+ lesions, thereby excluding those women with transient HPV infections that are
not associated with CIN2+ lesions. The validation of the relevance of methylation
markers to identify Chinese women younger than 30 years with CIN2+ lesions
therefore needs to be performed in future studies.
151
Chapter 6
Reference˖
1 Kamangar F, Dores GM, Anderson WF. Patterns of cancer incidence, mortality, and
prevalence across five continents: defining priorities to reduce cancer disparities in different
geographic regions of the world. J Clin Oncol 2006; 24: 2137–50.
2 Mathers C, Boerma T MFD. The global burden of disease: 2004 update (Geneva, Switzerland:
World Health Organization).
http://www.who.int/healthinfo/global_burden_disease/GBD_report_2004update_full.pdf.
3 Thomas G. Are we making progress in curing advanced cervical cancer? J Clin Oncol 2011; 29:
1654–6.
4 Nogueira-Rodrigues A, Ferreira CG, Bergmann A, de Aguiar SS, Thuler LCS. Comparison of

adenocarcinoma (ACA) and squamous cell carcinoma (SCC) of the uterine cervix in a sub-
optimally screened cohort: A population-based epidemiologic study of 51,842 women in Brazil.
Gynecol Oncol 2014; published online Aug 14. DOI:10.1016/j.ygyno.2014.08.014.
6 De Sanjosé S, Diaz M, Castellsagué X, et al. Worldwide prevalence and genotype distribution

of cervical human papillomavirus DNA in women with normal cytology: a meta-analysis. Lancet
Infect Dis 2007; 7: 453–9.
7 Bruni L, Diaz M, Castellsagué X, Ferrer E, Bosch FX, de Sanjosé S. Cervical human

papillomavirus prevalence in 5 continents: meta-analysis of 1 million women with normal
cytological findings. J Infect Dis 2010; 202: 1789–99.
23.
9 Castle PE, de Sanjose S, Qiao Y-LL, et al. Introduction of human papillomavirus DNA
screening in the world: 15 years of experience. Vaccine 2012; 30 Suppl 5: F117–22.
10 Umar A, Dunn BK, Greenwald P. Future directions in cancer prevention. Nat Rev Cancer 2012;
12: 835–48.
12 Yang N, Eijsink JJH, Lendvai A, et al. Methylation markers for CCNA1 and C13ORF18 are
strongly associated with high-grade cervical intraepithelial neoplasia and cervical cancer in
cervical scrapings. Cancer Epidemiol Biomarkers Prev 2009; 18: 3000–7.
13 Lendvai Á, Johannes F, Grimm C, et al. Genome-wide methylation profiling identifies

hypermethylated biomarkers in high-grade cervical intraepithelial neoplasia. Epigenetics 2012;
7: 1268–78.
152
Int J cancer 1998; 75: 536–45.

Vaccine 2011; 29: 9148–58.

2010; 116: 140–6.
18 Attila. New strategies for HPV-based cervical screening. Womens Heal (l Engl) 2014; 9: 1–16.
19 Van Bogaert L. Are the currently existing anti-human papillomavirus vaccines appropriate for
the developing world? Ann Med Health Sci Res 2013; 3: 306–12.
20 Quinn M, Babb P, Jones J, Allen E. Effect of screening on incidence of and mortality from
cancer of cervix in England: evaluation based on routinely collected statistics. BMJ 1999; 318:
904–8.
21 Peto J, Gilham C, Fletcher O, Matthews FE. The cervical cancer epidemic that screening has
prevented in the UK. Lancet; 364: 249–56.
22 Smith-Bindman R, Chu PW, Miglioretti DL, et al. Comparison of screening mammography in

the United States and the United kingdom. JAMA 2003; 290: 2129–37.
23 Peirson L, Fitzpatrick-Lewis D, Ciliska D, Warren R. Screening for cervical cancer: a systematic

review and meta-analysis. Syst Rev 2013; 2: 35.
24 ACOG Practice Bulletin no. 109: Cervical cytology screening. Obstet Gynecol 2009; 114:
1409–20.
25 Herzog TJ, Monk BJ. Reducing the burden of glandular carcinomas of the uterine cervix. Am J
Obstet Gynecol 2007; 197: 566–71.
26 Baker E. HPV and Pap: shifting roles in cervical cancer screening. Med Lab Obs 2013; 45: 24,
26.
27 Arbyn M, Bergeron C, Klinkhamer P, Martin-Hirsch P, Siebers AG, Bulten J. Liquid compared

with conventional cervical cytology: a systematic review and meta-analysis. Obstet Gynecol
2008; 111: 167–77.
28 Markowitz LE, Tsu V, Deeks SL, et al. Human papillomavirus vaccine introduction--the first five
years. Vaccine 2012; 30 Suppl 5: F139–48.
29 Serrano B, Alemany L, Alonso P, et al. Potential impact of a 9-valent HPV vaccine in HPV-
related cervical disease in 4 emerging countries ( Brazil , Mexico , India and China ). Cancer
Epidemiol 2014; 38: 748–56.
153
Chapter 6
30 Joura EA, Ault KA, Bosch FX, Brown D, Cuzick J, Ferris D, Garland SM, Giuliano AR,
Hernandez-Avila M, Huh W, Iversen OE, Kjaer SK, Luna J, Miller D, Monsonego J, Munoz N,
Myers E, Paavonen J, Pitisuttithum P, Steben M, Wheeler CM, Perez G, Saah A, Luxembo VC.
Attribution of 12 high-risk human papillomavirus genotypes to infection and cervical disease.
Cancer Epidemiol Biomarkers Prev 2014; 23: 1997–2008.
31 Financing HPV vaccination in developing countries. Lancet 2011; 377: 1544.
32 Ferrer HB, Trotter C, Hickman M, Audrey S. Barriers and facilitators to HPV vaccination of
young women in high-income countries: a qualitative systematic review and evidence synthesis.
BMC Public Health 2014; 14: 700.
33 Westra TA, Rozenbaum MH, Rogoza RM, et al. Until which age should women be vaccinated
against HPV infection? Recommendation based on cost-effectiveness analyses. J Infect Dis
2011; 204: 377–84.
34 Harper DM, Franco EL, Wheeler CM, et al. Sustained efficacy up to 4.5 years of a bivalent L1
virus-like particle vaccine against human papillomavirus types 16 and 18: follow-up from a
randomised control trial. Lancet 2006; 367: 1247–55.
35 Naucler P, Ryd W, Törnberg S, et al. Efficacy of HPV DNA testing with cytology triage and/or
repeat HPV DNA testing in primary cervical cancer screening. J Natl Cancer Inst 2009; 101:
88–99.
36 Mayrand M-H, Duarte-Franco E, Rodrigues I, et al. Human papillomavirus DNA versus

Papanicolaou screening tests for cervical cancer. N Engl J Med 2007; 357: 1579–88.
37 Ronco G, Giorgi-Rossi P, Carozzi F, et al. Results at recruitment from a randomized controlled

trial comparing human papillomavirus testing alone with conventional cytology as the primary
cervical cancer screening test. J Natl Cancer Inst 2008; 100: 492–501.
38 Elfström KM, Herweijer E, Sundström K, Arnheim-Dahlström L. Current cervical cancer

prevention strategies including cervical screening and prophylactic human papillomavirus
vaccination: a review. Curr Opin Oncol 2014; 26: 120–9.
39 Monsonego J. HPV testing in prevention of cervical cancer: practices and current trends. Ann
Biol Clin (Paris) 2013; 71: 27–32.
40 Giorgi Rossi P, Baldacchini F, Ronco G. The Possible Effects on Socio-Economic Inequalities

of Introducing HPV Testing as Primary Test in Cervical Cancer Screening Programs. Front
Oncol 2014; 4: 20.
41 Liu XY, Feng AH, Cui YM, Tobe RG. Prevention of human papillomavirus (HPV) infection and
cervical cancer in China: How does HPV vaccination bring about benefits to Chinese women?
Biosci Trends 2013; 7: 159–67.
42 Women’s health in rural China. Lancet 2009; 374: 358.
154
43 Zhao FH, Lewkowitz AK, Hu SY, et al. Prevalence of human papillomavirus and cervical
intraepithelial neoplasia in China: a pooled analysis of 17 population-based studies. Int J
Cancer 2012; 131: 2929–38.
44 Cervical Cancer Screening in Developing Countries: report of a WHO consultation.

http://www.who.int/cancer/media/en/cancer_cervical_37321.pdf.
45 Boers A, Slagter-Menkema L, van Hemel BM, et al. Comparing the Cervista HPV HR Test and
Hybrid Capture 2 Assay in a Dutch Screening Population: Improved Specificity of the Cervista
HPV HR Test by Changing the Cut-Off. PLoS One 2014; 9: e101930.
46 Boers A, Bosgraaf RP, van Leeuwen RW, et al. DNA methylation analysis in self-sampled
brush material as a triage test in hrHPV-positive women. Br J Cancer 2014; 111: 1095–101.
47 Pepe MS, Etzioni R, Feng Z, et al. Phases of biomarker development for early detection of
cancer. J Natl Cancer Inst 2001; 93: 1054–61.
48 Sun LL, Cao DY, Yang JX, et al. Population-based case-control study on DAPK1, RAR-beta2
and MGMT methylation in liquid-based cytology. Arch Gynecol Obs 2012; 285: 1433–9.
49 Wu JH, Liang XA, Wu YM, Li FS, Dai YM. Identification of DNA methylation of SOX9 in cervical
cancer using methylated-CpG island recovery assay. Oncol Rep 2013; 29: 125–32.
50 Wu Y, Meng L, Wang H, et al. Regulation of DNA methylation on the expression of the FHIT
gene contributes to cervical carcinoma cell tumorigenesis. Oncol Rep 2006; 16: 625–9.
51 Song Y, Zhang C. Hydralazine inhibits human cervical cancer cell growth in vitro in association
with APC demethylation and re-expression. Cancer Chemother Pharmacol 2009; 63: 605–13.
52 WANG, SS., WANG N., YU X., YANG CX., YAN LP. WY. Study on methylation status of
RASSFIA gene promoter and exon 1 in cervical cancer cell lines. China Oncol 2013; 23: 777–
83.
53 Chen CL, Liu SS, Ip SM, Wong LC, Ng TY, Ngan HY. E-cadherin expression is silenced by
DNA methylation in cervical cancer cell lines and tumours. Eur J Cancer 2003; 39: 517–23.
54 Yu MY, Tong JH, Chan PK, et al. Hypermethylation of the tumor suppressor gene RASSFIA
and frequent concomitant loss of heterozygosity at 3p21 in cervical cancers. Int J Cancer 2003;
105: 204–9.
55 Kitchener HC, Almonte M, Thomson C, et al. HPV testing in combination with liquid-based
cytology in primary cervical screening (ARTISTIC): a randomised controlled trial. Lancet Oncol
2009; 10: 672–82.
56 Arbyn M, Anttila A, Jordan J, et al. European Guidelines for Quality Assurance in Cervical
Cancer Screening. Second edition--summary document. Ann Oncol 2010; 21: 448–58.
57 Wright TC, Massad LS, Dunton CJ, Spitzer M, Wilkinson EJ, Solomon D. 2006 consensus
guidelines for the management of women with abnormal cervical cancer screening tests. Am J
Obstet Gynecol 2007; 197: 346–55.
155
Chapter 6
58 Kitchener H, Denton K, Soldan K, Crosbie E. Developing role of HPV in cervical cancer

prevention. BMJ 2013; 4781: 1–7.
156
Nederlandse samenvatting
Chapter 7
157
Chapter 7
Samenvatting
Alhoewel door cytomorfologie-gebaseerde screening voor baarmoederhalskanker de
incidentie van baarmoederhalskanker in de westerse wereld heeft verminderd, wordt
voorspeld dat in 2030 het aantal sterfgevallen wereldwijd als gevolg van
1,2,3
baarmoederhalskanker toe zal nemen tot ongeveer 410.000 . De vooruitgang in
preventie, vroegtijdige opsporing en behandeling van baarmoederhalskanker wordt
belemmerd door het feit dat de meerderheid van de gevallen voorkomen in de derde
wereld, waar voor de vroegdetectie en behandeling van baarmoederhalskanker de
infrastructuur, het personeel en de medische voorzieningen met voldoende
4
kwaliteitsborging beperkt of zelfs afwezig zijn . China, als één van de
ontwikkelingslanden, speelt een belangrijke rol in de wereldwijde strijd tegen
baarmoederhalskanker. Toch zijn er verschillende uitdagingen die belangrijk zijn in
deze strijd, zoals de enorme bevolking en oppervlakte, de snelle, onevenwichtige
economische groei en de massale migratie van de laatste jaren.
Door zijn lage gevoeligheid en hoge arbeidsintensiteit, is cytomorfologie-gebaseerd

onderzoek van uitstrijkjes geen ideale screeningsmethode in China5. Het testen van
HPV is een andere bekende benadering voor de vroege detectie van
baarmoederhalskanker. De introductie van de HPV test voor de vroegdetectie van
baarmoederhalskanker in China is vanwege de talrijke beschikbare HPV detectie
platformen met uiteenlopende specificaties een grote uitdaging. Hiervoor is het
noodzakelijk om een optimaal en gevalideerd platform te kiezen voordat de HPV-test
wordt ingevoerd als primaire screeningsmethode. Tegelijkertijd, moet een andere
belangrijke vraag eerst worden beantwoord, namelijk hoe de HPV-positieve
patiënten, met name die in dichtbevolkte, maar minder ontwikkelde regio’s leven, te
vervolgen. Om de verschillende screeningsmethoden voor de vroegdetectie van
baarmoederhalskanker effectief op grote cohorten te onderzoeken, worden in dit
proefschrift de nieuwe inzichten in het gebruik van biomarkers specifiek voor de
screening voor baarmoederhalskanker, waaronder de HPV-testen en DNA-
methylatie markers, onderzocht en bediscussieerd.
HPV-DNA-screeningen worden al meer dan 15 jaar uitgevoerd en toont aan dat de

prevalentie van HPV-infectie geografisch en per ras varieert6,7. Hoewel verscheidene
158
projecten met betrekking tot de hrHPV-infectie frequentie en baarmoederhalskanker

8
screening al zijn uitgevoerd (Tabel 4 hoofdstuk 1), zijn de beschikbare gegevens in
China nog zo variabel dat conclusies om tot een landelijke standaard te komen nog
niet gemaakt kunnen worden. Deze informatie is nodig om de juiste
referentiewaarden te kunnen bepalen voor een effectieve screening. In hoofdstuk 2
beschrijven we een landelijk cross-sectionele studie naar de HPV-DNA-prevalentie in
verschillende regio’s binnen China, Tevens werd de verdeling van de verschillende
HPV genotypes tussen de verschillende regio's en verschillende leeftijdsgroepen
geanalyseerd. Voor deze studie werden uitstrijkjes verkregen via dunne-laag
cytologie van 120.772 vrouwen afkomstig uit bevolkingsonderzoek of werknemer-
gebaseerde baarmoederhalskanker screening programma’s in 2012 uit 37 Chinese
steden en opgestuurd naar het laboratorium van moleculaire infectieziekten van
Guangzhou KingMed. In totaal, werden 111.131 uitstrijkjes getest met behulp van de
Hybrid Capture 2 (HC2) test en van 9641 monsters werd het HPV-genotype bepaald
door middel van de TellgenplexTM HPV-DNA-test. De totale frequentie van hoog-
risico HPV (hrHPV) positiviteit was 21,07%, variërend van 31,94% tot 18,42% over
de verschillende regio's. Leeftijdsspecifieke prevalentie toonde een "twee pieken"
patroon. De jongste leeftijdsgroep (15-19 jaar) had de hoogste hrHPV-infectie
frequentie (30,55%) en de tweede piek werd gevonden in de leeftijdsgroep van 50-60
jaar. HPV16 (4.82% van alle monsters) en HPV52 (4,52% van alle monsters) waren
de meest voorkomende type hrHPV, gevolgd door HPV58 (2,74% van alle monsters);
terwijl HPV6 (4.01% van alle monsters) en HPV11 (2,29% van alle monsters) het
meest voorkomend waren in de laag risico HPV (lrHPV) types. HPV16/HPV52 en
HPV52/HPV58 dubbelinfecties waren de meest voorkomende co-existentie
genotypen in het geval van meerdere infecties. Deze data tonen dat de HPV-infectie
frequentie in China is toegenomen naar het niveau van landen met de hoogste HPV
infectiegraad. Daarom biedt dit onderzoek een waardevolle referentie voor een
effectief landelijk screeningsprogramma voor baarmoederhalskanker. Verder toont
de analyse van de genotypische verdeling van hrHPV aan dat HPV52 en HPV58 de
twee meest-voorkomende genotypen in China zijn. Dat betekent dat de momenteel
beschikbare commerciële vaccins, Cervarix (tegen HPV16/18) en Gardasil (tegen
HPV6/11/16/18 ) mogelijk niet volledig effectief zijn binnen de Chinese bevolking.
159
Chapter 7
Voor de keuze van de uitvoering van de screening en vaccinatie-programma’s moet

in China dus rekening gehouden worden met deze epidemiologische kenmerken.
Momenteel worden veel nieuwe screeningstechnologieën (inclusief de HPV-testen)
ontwikkeld en op de markt gebracht in China zonder validatie van hun effectiviteit
voor het bevolkingsonderzoek9. Onderzoekers proberen algoritmes te ontwikkelen
die in China kunnen worden toegepast voor zowel de hoogste risicogebieden als de
stedelijke gebieden. Momenteel zijn er vele commerciële HPV DNA testen
beschikbaar, maar slechts enkele testen zijn formeel gevalideerd voor primaire HPV
screening voor baarmoederhalskanker zoals de HC2 hrHPV DNA-test, de eerste
hrHPV-test goedgekeurd door de FDA in de VS11. De Cervista HR HPV test
(Hologic Inc), de tweede door de FDA-goedgekeurde hrHPV-test, wordt op grote
schaal gebruikt in de VS en in sommige Chinese studies. Cervista heeft enkele
voordelen ten opzichte van de HC2: Het vereist minder monstervolume, bevat een
interne controle en heeft een kortere verwerkingstijd. De Cervista is echter nog niet
formeel gevalideerd voor primaire screening voor baarmoederhalskanker volgens de
internationale richtlijn. In hoofdstuk 3 hebben we aangetoond dat Cervista voldeed
aan de internationale prestatie- en reproduceerbaarheidseisen die worden gesteld
aan HPV-testen voor screening voor baarmoederhalskanker. Door middel van non-
inferioriteitsanalyse volgens de internationale richtlijn, werd de klinische gevoeligheid
en specificiteit van de Cervista test vergeleken met die van HC2 voor de detectie
van hooggradige cervicale laesies (CIN2 +) in Chinese vrouwen die 30 jaar of ouder
waren. De Cervista test had een klinische gevoeligheid voor CIN2+ van 89% (95%
CI: 81,7-93,6) en een klinische specificiteit voor CIN2+ van 91% (95% CI: 90,5-91,8).
Zowel de relatieve klinische gevoeligheid als de specificiteit waren niet-inferieur aan
die van HC2 (non-inferioriteit scoren testen, p=0,043 en p<0,0001, respectievelijk).
Daarnaast hebben we ook de intra- en inter-laboratorium reproduceerbaarheid van
de Cervista test bepaald in 510 uitstrijkjes, opnieuw volgens de internationale
richtlijnen. De intra- en inter-laboratorium overeenkomsten waren 92% (ondergrens
95% CI: 89,7%; kappa = 0,83; p<0,001) en 90,4% (ondergrens 95% CI: 88,4%;
kappa = 0,81; p<0,001), respectievelijk. Samenvattend betekent dit dat de Cervista
HR-HPV-test voldoet aan de internationale richtlijnen gesteld voor nieuwe HPV-
160
testen en kan worden beschouwd als een klinisch gevalideerde test voor primaire
screening voor baarmoederhalskanker.
Door de screening van uitstrijkjes worden niet alleen vroege stadium carcinomen
geïdentificeerd, maar ook de voorstadia die cervicale laesies zoals CIN (cervicale
intra-epitheliale neoplasie) worden genoemd, representeren zich. Deze kunnen
gemakkelijk verwijderd worden en ontwikkelen niet tot een invasief carcinoom10.
Echter, de relatief lagere specificiteit van hrHPV testen, vooral bij een jonge
screeningspopulatie, kan leiden tot onnodige verwijzingen naar de gynaecoloog,
onnodige ongerustheid bij vals-positief geteste vrouwen, en hogere kosten voor de
gezondheidszorg, en zijn dus belangrijke redenen om de huidige
screeningsprogramma’s voor de vroegdetectie van baarmoederhalskanker te
verbeteren. Daarnaast zal in de nabije toekomst de prevalentie van CIN en
baarmoederhalskanker naar alle waarschijnlijkheid afnemen in de landen, die
primaire preventie door middel van hrHPV-vaccinatie hebben ingevoerd. Met deze
afname in de prevalentie zal de positieve voorspellende waarde van de huidige
testen afnemen.
Daarom zijn andere objectieve biomarkers nodig met zowel een hoge gevoeligheid
als een hoge specificiteit als nieuwe screeningsmethode voor triage test. Differentiële
DNA methylatie patronen van geselecteerde genen tussen normaal weefsel uit de
baarmoederhals en weefsel van (voorstadia van) baarmoederhalskanker kunnen
uitstekend gebruikt worden als diagnostische test op basis van methylatie specifieke
PCR (MSP). Eerder onderzoek van onze groep, had al geresulteerd in de
identificatie van 4 genen (C13ORF18, JAM3, EPB41L3 en TERT)11,12,13. De
diagnostische prestaties van de DNA-methylatie analyse van deze genen liet
gevoeligheden voor het opsporen van CIN2+ in een hrHPV-positieve populatie zien
variërend van 43% tot 71% en een specificiteit tussen de 89-100%11. Onze
strategieën voor het ontdekken van nieuwe methylatiemarkers waren dusver
gebaseerd op het verschil tussen kanker en normaal weefsel, resulterend in markers
met een hele hoge specifiteit en een hoge gevoeligheid voor kanker, maar met een
te lage gevoeligheid voor het detecteren van CIN2/3 laesies in uitstrijkjes. Om
nieuwe methylatiemarkers specifiek voor CIN2/3 te identificeren hebben we in
161
Chapter 7
hoofdstuk 4 een nieuwe techniek (MethylCap-seq) gebruikt om differentiële

methylatie regio's (DMR's) te identificeren. Door DNA van 20 normale weefsels uit de
baarmoederhals te vergelijken met DNA van 18 weefsels van CIN2/3 laesies, werden
176 DMR's van 163 genen geïdentificeerd. De top 15 van de meest differentieel
gemethyleerde genen werden geselecteerd en gevalideerd middels MSP in twee
stappen: Op hetzelfde DNA dat was gebruikt voor MethylCap-seq en op DNA van
een onafhankelijke patiënten cohort met (voorstadia van) baarmoederhalskanker.
Voor verdere diagnostische evaluatie, werden de best differentiële methylatiemarkers
getest met kwantitatieve MSP (QMSP) op baarmoederhalsuitstrijkjes van 2 cohorten:
1) baarmoederhalskanker versus gezonde controles en 2) patiënten verwezen vanuit
het bevolkingsonderzoek met een afwijkend uitstrijkje.
Na verificatie en validatie van de top 15 genen met MSP, waren 9 genen significant
differentieel gemethyleerd in normale baarmoederhalsweefsel versus CIN2/ 3 laesies
(p<0,05). Vervolgens waren de methyleringswaardes van 8/9 genen significant hoger
in uitstrijkjes van kankerpatiënten in vergelijking met uitstrijkjes van normale vrouwen.
Voor alle 8 genen namen de methylatieswaardes toe met de ernst van de
onderliggende histologische laesie in uitstrijkjes van patiënten verwezen met een
abnormaal uitstrijkje. Naast de 8 nieuwe genen, werd ook de methylering van ons
vorige vier-gen panel (C13ORF18, JAM3, EPB41L3 en TERT) geanalyseerd op
dezelfde monsters. De combinatie C13ORF18 /JAM3/AL590705.4 liet een
vergelijkbare hoge gevoeligheid voor CIN2+ (74-76%) zien als de hrHPV-test (79%),
terwijl de specificiteit significant hoger was (71-76%) als de hrHPV-test (42%)
(p0,05). Met deze genoom-brede DNA methylatie analyse werden nieuwe CIN2/3
methylatiemarkers geïdentificeerd. De diagnostische prestaties van ons nieuwe
methylatiepanel gaf vergelijkbare gevoeligheid voor de hrHPV-test voor het aantonen
van CIN2+, maar met een hogere specificiteit, waardoor mogelijk meer onnodige
verwijzingen voor colposcopie kunnen worden voorkomen. De volgende stap voor
implementatie in primaire screeningsprogramma’s is de validatie van ons nieuwe
methylatiepanel in een groot screeningspopulatie.
Baarmoederhalskanker omvat verschillende histologische types, waarvan

adenocarcinomen (ADC) en plaveiselcelcarcinoma (SCC) de twee belangrijkste
162
subtypen zijn, goed voor respectievelijk 10-25% en 75-90%14,15,16. Momenteel neemt

de incidentie van SCC af in de westerse landen, terwijl er een toename in de
15,17
absolute en relatieve incidentie van ADC is . Hoewel ADC voornamelijk snel
toeneemt in Europa, en relatief langzamer in Azië (2.8-22.6%)16, kunnen de absolute
gevallen met ADC niet worden genegeerd, vanwege zijn grote aantallen. Bovendien,
in vergelijking met SCC, wordt ADC hoofdzakelijk gediagnosticeerd in een verder
16,17
gevorderd stadium . Screeningsprogramma’s zijn effectiever in het detecteren
van de voorlopers van SCC dan die van ADC. Belangrijke reden voor de
ineffectiviteit van de huidige screeningsmethodologie voor de detectie van ADC is de
hoger gelegen locatie in het baarmoederhalskanaal, dat het verkrijgen van de nodige
monsters voor cytologie bemoeilijkt, maar ook de directe observatie door
colposcopische examinatie17.
Naast de hrHPV testen, zou analyse van DNA methylatiemarkers de detectie van
ADC in eerdere fasen kunnen verbeteren. Er zijn echter geen specifieke
methylatiemarkers beschreven voor de detectie van ADC of SCC. Het doel van de in
hoofdstuk 5 beschreven studie was om nieuwe methylatiemarkers voor
baarmoederhalskanker te identificeren voor zowel ADC als SCC. Nieuwe
methylatiemarkers werden geïdentificeerd met behulp van MethylCap-seq door DNA
van 20 normale weefsels uit de baarmoederhals te vergelijken met DNA uit 12
carcinomen (6 ADC en 6 SCC). De top 15 markers werden geselecteerd voor
verificatie middels bisulfiet-pyrosequencing of MSP op hetzelfde DNA als gebruikt
voor MethylCap-seq. Vervolgens werden de methylatiemarkers gevalideerd op een
reeks DNA monsters geïsoleerd van in paraffine ingebedde monsters van patiënten
met normaal weefsel of carcinoom uit de baarmoederhals met behulp van MSP.
Tenslotte werd met QMSP op uitstrijkjes van een onafhankelijk cohort van vrouwen
met normaal weefsel of carcinoom uit de baarmoederhals. Validatie van de top 15
differentieel gemethyleerde kandidaatgenen resulteerde in 5 markers die een hogere
methylatie frequentie hadden in kankerweefsels in vergelijking met normaal weefsel
(p<0,05). QMSP op DNA van uitstrijkjes toonde dat de gevoeligheid van deze 5
markers voor de detectie van baarmoederhalskanker varieerde van 80.5% tot 91.9%
voor zowel ADC als SCC, terwijl bijna alle normale uitstrijkjes negatief
waren(specificiteit: 94% -98,9%). Concluderend, identificeerden we met de
163
Chapter 7
MethylCap-seq analyse 5 nieuwe methylatiemarkers voor de vroegdetectie van ADC

en SCC. Verdere validatie in een grote reeks uitstrijkjes afkomstig vanuit het
bevolkingsonderzoek is nodig, in het bijzonder in uitstrijkjes van ADC en haar
voorstadia, om onze bevindingen te bevestigen.
Referenties
1 Kamangar F, Dores GM, Anderson WF. Patterns of cancer incidence, mortality, and
prevalence across five continents: defining priorities to reduce cancer disparities in different
geographic regions of the world. J Clin Oncol 2006; 24: 2137–50.
2 Mathers C, Boerma T MFD. The global burden of disease: 2004 update (Geneva, Switzerland:
World Health Organization).
http://www.who.int/healthinfo/global_burden_disease/GBD_report_2004update_full.pdf.
3 Thomas G. Are we making progress in curing advanced cervical cancer? J Clin Oncol 2011; 29:
1654–6.
4 Nogueira-Rodrigues A, Ferreira CG, Bergmann A, de Aguiar SS, Thuler LCS. Comparison of

adenocarcinoma (ACA) and squamous cell carcinoma (SCC) of the uterine cervix in a sub-
optimally screened cohort: A population-based epidemiologic study of 51,842 women in Brazil.
Gynecol Oncol 2014; published online Aug 14. DOI:10.1016/j.ygyno.2014.08.014.
6 De Sanjosé S, Diaz M, Castellsagué X, et al. Worldwide prevalence and genotype distribution

of cervical human papillomavirus DNA in women with normal cytology: a meta-analysis. Lancet
Infect Dis 2007; 7: 453–9.
7 Bruni L, Diaz M, Castellsagué X, Ferrer E, Bosch FX, de Sanjosé S. Cervical human

papillomavirus prevalence in 5 continents: meta-analysis of 1 million women with normal
cytological findings. J Infect Dis 2010; 202: 1789–99.
23.
9 Castle PE, de Sanjose S, Qiao Y-LL, et al. Introduction of human papillomavirus DNA
screening in the world: 15 years of experience. Vaccine 2012; 30 Suppl 5: F117–22.
10 Umar A, Dunn BK, Greenwald P. Future directions in cancer prevention. Nat Rev Cancer 2012;
12: 835–48.
12 Yang N, Eijsink JJH, Lendvai A, et al. Methylation markers for CCNA1 and C13ORF18 are
strongly associated with high-grade cervical intraepithelial neoplasia and cervical cancer in
cervical scrapings. Cancer Epidemiol Biomarkers Prev 2009; 18: 3000–7.
164
13 Lendvai Á, Johannes F, Grimm C, et al. Genome-wide methylation profiling identifies

hypermethylated biomarkers in high-grade cervical intraepithelial neoplasia. Epigenetics 2012;
7: 1268–78.
Int J cancer 1998; 75: 536–45.

Vaccine 2011; 29: 9148–58.

2010; 116: 140–6.
165
Chapter 7
166
Curriculum Vitae
Curriculum Vitae
167
Curriculum Vitae
Curriculum Vitae
Rong Wang was born on August 25th, 1977 in Tianjin, a

coastal city in northern China. After graduation from the
Tianjin No.14 high school, she was registered in Tianjin
Agriculture College in 1996, where she obtained her
bachelor degree in 2000. By a cooperation project on
Rare earth (Lanthanum, Cerium) inhibiting cancer cells
growth, invasion and metastasis related to matrix
metalloproteinases, she started her job at the Department
of Laboratory Medicine in Tianjin Medical University. In
2003, she pursued her master degree study supervised by Professor Yunde Liu on
the same project. During her study, she became a formal teacher in that department,
who mainly engaged in the teaching of Clinical Microbiology and Parasitology. After
she obtained her master degree in Clinical Laboratory diagnosis from Tianjin Medical
University in 2006, her research transferred to the microbe in the urinary and human
disease, especially on the clinical detection of Mycoplasma and Ureaplasma and their
epidemiology in China with the guidance of Professor Shangwei Wu.
In September 2010, she started her PhD project on Medical Microbiology under the
supervision of Professor Wu. In June 2011, she was selected as a PhD student of
Groningen University. In April 2012, she started at the Department of Gynecology and
Oncology, focusing on the identification of DNA methylation biomarkers to improve
cervical cancer screening. The results of the PhD research are presented in this
thesis. Hereafter, she intends to pursue her academic research on the translational
research of clinical diagnostic biomarkers in the coming future.
168
Acknowledgements
Acknowledgements
169
Acknowledgements
Acknowledgements
Just as the DNA has double strands, without the well-functioning of each base pair,
there will never be a good quality of gene. So everyone knows how important a good
team is! Science is a process to look for questions and answers. Therefore, if there
are some achievements in the thesis, that’s because of our strong Questions and
Answers Team (Q&A Team). I would like to express my deep and sincere gratitude to
all of those who contributed in different ways.
First and foremost, I would like to express my sincerely gratitude to my promotor Prof.
dr. A.G.J. van der Zee. Dear Ate, you guided me with a leading question which is,
“What should I do?” in a project. Thank you for your gentle guidance, continuous
encouragement, creative ideas and invaluable suggestions in each monthly thesis
meeting and the writing of this thesis as well, which have inspired me throughout the
whole course of this project. Thanks for help me to realize my original goal
“Collaboration”. Although it isn’t initiated by me currently, I was very happy when I
heard that the first place you came in China is Tianjin, my hometown. I am so
impressed by your good surgeon technique and presentation skills after I met you in
Tianjin. I wish you could spend more time in Tianjin next time so that I would have an
opportunity to learn more from you.
Second, my sincere gratitude also goes to my other promotor, Prof. dr. E. Shuuring.
Dear Ed, you pushed me forward with your rationality and logic for science. Thank you
for always reminding me the question which is, “Why do you do that?” which stimulate
my thoughts, and make myself understood my research more deeply and clearly. With
your guidance, my research became more professional. Moreover, thank you for your
great efforts for the papers and thesis. Your detailed comments and careful
consideration of my writings has contributed towards completion of this thesis.
170
Acknowledgements
Third, I would like to express my appreciation from the bottom of my heart to my

co-promotor and daily supervisor dr. G.B.A. Wisman. Dear Bea, you are indeed my
“SOS”, thank you for being available in the time of need and guided me “how to do”
things when they became beyond my capacity. Starting from the opening of my
research on DNA methylation, you have always been patient even to answer my
simple and stupid questions sometimes, providing detailed instructions on the
experiments and data analysis of my research and revising my presentations and
reports and continuously to improve my thesis work.
Next, thanks to my teammate Roland van Leeuwen, a very passionate and energetic
young man, who always instructed me “where to do” my experiments. Thanks for your
great effort on my tests and experiments. Your highly efficient, quick feedback, serious
and active communications and answers helped me to overcome lots of difficulties in
my work. I wish you great success on your doctoral study. Thank you Aniek Boers, a
girl who always told me “which samples can be used”. Thanks for your contribution in
the thesis, especially for the clinical part.
Looking back, I really enjoyed the small group meetings once in every two weeks with
you all, especially during the stage of marker identification and development of the
papers and thesis. Although some heating disputation, this enables me to understand
the research and project more deeply and completely.
I am also grateful to my Chinese supervisor, Prof. dr. Shangwei Wu. I appreciated you
for giving me a chance to enter your research. You broadened my horizon, inspired,
supported and encouraged me to realize my long awaited dream to study abroad.
Thanks for your strict and kind guidance in the early days of my research, those are
always in my mind. Thank you for your rich input editing the HPV paper. But I know, if I
want to follow your way to achieve more success, I still have a long way to go and
171
Acknowledgements
work hard in the future.
My sincere gratitude also goes to the members of the reading committee: Prof. dr.
P.J.F. Snijders, Prof. dr. L.F.A.G. Massuger and Prof. dr. H.W. Nijman for the intensive
reading, evaluating and assessing of my thesis.
I owe my sincere thanks to all kinds of our Q&A groups. Thanks for our methylation
group members, Tushar Tomar, Martijn Clausen and Gert Jan Meersma. All the
discussions, advices, ideas and experiences from you helped me a lot. It is
unforgettable for the happy time we “seized” the PyroMark Q24. Thanks for
Gynecology and Oncology group, Prof. dr. Steven de Jong, Harry Klip, Nicolette
Alkema, Jolanda Visser, Roelien Meijering, Neeltje Kooi, Joost Caumanns, Marco de
Bruyn, Phuong Le, Ximena Rosas Plaza, Annechien Plat, Gerke Ariaans and Gerda
de Vries, I enjoyed staying with you all not only in the discussion but also in the time
we go out for dinner, car racing and relaxation. Additionally, I would like to thanks to
Steven for giving me a constructive ideas always on my presentation. Thanks to Harry
for the sample collection.
Like people not only have skeleton, but also have more well-developed muscle and fat
to make us look beautiful. In addition to the Q&A team, my warmest thanks also go to
my teachers, classmates, friends, roommates, colleagues and those who gave me the
opportunity, walk alongside with me, everlasting support me since my beginning.
First of all, I’d like thank the people who help me to achieve the chance to go to UMCG,
Groningen, the Netherlands. I would like to thank Prof. dr. S. Poppema for all of your
efforts to Chinese students’ study in the Netherlands, it is a beautiful memory when I
conducted my first interview with you in China. Thank Prof. dr. A.J. Moshage for
helping me to both come to Groningen and seek for the research group suitable for
172
Acknowledgements
development. Thank Prof. dr. J.C. Wilschut and Prof. dr. C.A.H.H. Daemen for giving
me the first offer and initial assistance.
I greatly thank for all my heart to my family, for the invaluable care and attention.
Thank you my parent for gave me the freedom to choose my career and my life.
Thanks my young brother for the full responsibility to care the family while I was
studying. Thanks my uncles for your all too much hard work and endless support to
my study as well as care to my daily life all the time. Thanks my aunts for your
kindness and toleration. Thanks my dear cousins for your accompanies with me and
shared joy and sorrow.
My warmest thanks also goes to my lovely roommates, Urszula Domanska, Anton

Terwisscha van Scheltinga, Birgit Weyhenmeyer, Roelien Meijering and Pepijn
Schoonen. I will never forget the memorable time we had together and the precious
birthday dinner and gifts. Thanks Birgit, you always try to keep me not feeling alone
and encourage me to participate in various activities. Thanks Roelien and Pepijn, for
providing me a quiet space during the writing of my thesis. It was with a great pleasure
the moment when I invited you all to experience the Chinese food which also
beneficial for me to improve my cooking technics. Hope I will see you all again in
China in the future.
Thanks all my colleagues of M.O.L lab, the cherish and unforgettable memories in my
lifetime are those joy share with you all in all kinds of traditional activities, parties,
Lab-day, sushi and dumpling dinner, Sinterklaas, Christmas breakfast, birthday
cake…In addition, thanks to Hetty Timmer-Bosscha and Coby Meijer, for your
guidance and help for the lab manage, instrument usage, and reagent order. Thanks
to my neighbors: Hilde Nijhuis, Milind Pore, Anne Margriet Heijink, who always lend
your keys to me.
173
Acknowledgements
Thank colleagues from the Department of Pathology. Thanks to all colleagues of DNA
lab for your various instructions and help as well as all kinds of interesting chats in my
experiments. Thanks Lorian Slagter-Menkema for your technical guidance for the
LightCycler 480 experiments. Thanks Lydia Visser for your delicious evening
receptions and chocolate cakes elaborately prepared ever. I will remember that sweet
taste forever. Thanks colleagues in the office of Zheng and Rui for your tolerance and
understanding upon my “daily report”.
Special thanks to my dear friends in the Netherlands: Liangzheng, Wurui Zhaoyingru,

Wanghao and Pieter-Jan, Xie Xueqian, Yanxiaomei, Zhoukai, Caojunjun, Songjuan,
Wangcuifeng, Lijun, Chencheng and Libenhui, Hanbing, Liranran᧨Liyisi᧨Liuyuxuan ᧨
Hanjing, and Luanzhilin, with all your help, lots of difficulties became easy, tears shift
to laugh, especially for moving each time. I appreciate Yingru, Xiaomei's generously
for my daily necessities making my stuffs always rich enough for parties. Thanks to
Hao and Peter's enthusiasm and help. It seems that we always have endless words
with Hao. Looking forward to meet all your family in China. Thanks to Xueqian both for
the help of my paper writing and lots of valuable suggestions. In addition, thank you dr.
Qu Ning and Ren Yijin, both of you try your best to set up the collaboration between
the Netherlands and China. 2015 is the 30-year anniversary of the establishment of
friendship of Groningen and Tianjin. Best wishes to your cooperative projects proceed
smoothly this year and in the future. Looking forward to our future collaboration.
Thanks for the help of my long-lasting friends in China, you assisted me to deal with a
lot of issues during the past three years. Thanks for the everlasting supporting from
Department of Laboratory Medicine in Tianjin medical university.
My sincerest gratitude goes to my paranymphs again, Roland van Leeuwen, Pepijn

Schoonen and Rea Wu, many thanks for all your generous support and contribution
174
Acknowledgements
on my PhD defense. I couldn’t imagine without your help, how can I come to the
satisfactory end. Dear Rea, I indeed depend on you, I enjoy all the time I share with
you, shopping, weekend dinner, travel…all these will be cherished in my mind and my
heart.
The last but not the least, thanks the person who hurts me careless. Without the pain
of tear, how us can know the true happiness. I will try my best to find and cherish a
better tomorrow.
175
Acknowledgements
崤љ࠴аࣣো଻пোе۩‫▲ͫ׷ڐ‬૨暉Ѵ۩ͫ‫ۉ‬С‫۩ܴݵ‬ङ‫؟‬ыͫ৯٤ͫգ؆ͫգзͫ
߅ՄͫՠҁѪѴ澞
ߪ ȕ ‫ࡒם‬
176

New Insights in Methodology of Screening For Cervical Cancer

Hochgeladen von

Dokumentinformationen

Originaltitel

Copyright

Verfügbare Formate

Dieses Dokument teilen

Dokument teilen oder einbetten

Freigabeoptionen

Stufen Sie dieses Dokument als nützlich ein?

Sind diese Inhalte unangemessen?

Copyright:

Verfügbare Formate

New Insights in Methodology of Screening For Cervical Cancer

Hochgeladen von

Copyright:

Verfügbare Formate

University of Groningen

New insights in methodology of screening for cervical cancer

Link to publication in University of Groningen/UMCG research database

Citation for published version (APA):

Download date: 20-09-2019

Ph.D. thesis, University of Groningen, the Netherlands

ISBN: 978-90-367-7704-9 (printed version)

Copyright © 2015 RONG WANG, Groningen, the Netherlands

to obtain the degree of PhD at the

This thesis will be defended in public on

Wednesday 15 April 2015 at 11.00 hours

born on 25 August 1977

Chapter 1 General Introductionಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ11

Chapter 2 Nationwide prevalence of human papillomavirus infection

in 37 cities in China ಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ37

Submitted for publication

Chapter 3 Clinical validation of the Cervista HPV HR test according to the

international guidelines for human papillomavirus test requirements

for cervical cancer screening ಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ65

J. Clin. Microbiol. 2014,52(12):4391.

Chapter 4 Discovery of new methylation markers to improve screening for

cervical intraepithelial neoplasia grade 2/3 ಹಹಹಹಹಹಹಹಹಹ73

Submitted for publication

Chapter 5 Genome-wide methylome analysis to discovery of (novel)

hypermethylation biomarkers for both cervical adenocarcinoma and

squamous cell carcinomaಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ109

Submitted for publication

Chapter 6 Summary and future perspectives ಹಹಹಹಹಹಹಹಹಹಹಹಹಹ141

Chapter 7 Nederlandse samenvatting ಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ157

About the author ಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹಹ167

1.2. Clinicopathology of cervical cancer

Table 1. The database of the incidence and mortality of cervical cancer

in China (data adapted from Shi et al5)

NCCR 2004-2006  6% 6.0-9.1

national retrospective survey 14.6 11 sites (overestimated)

national retrospective survey ~10% 4.3 11 sites (underestimated)

national retrospective survey ~10% 2.5 11 sites (underestimated)

and is therefore detected easier by cytologic screening. In contrast, ADC develops

1.3. Treatment of cervical cancer

1.4. Cervical precursor lesions

Table 2: Overview of the different nomenclatures and classifications for

Dysplasia CIN Bethesda Papanicolaou

Normal Normal Within normal limits

Atypical cells Squamous atypia ASCUS Pap 2

Mild Dysplasia CIN1 Low-grade SIL (LSIL) Pap 3A1

Moderate Dysplasia CIN 2 Pap 3A2

Severe Dysplasia High-grade SIL (HSIL) Pap 3B

(Microinvasive) cancer (Microinvasive) cancer (Microinvasive) cancer Pap 5

Figure 1: The development of cervical carcinogenesis(AdaptedfromSnijdersetal

1.4.2 Treatment of precursor lesions

II. Cervical cancer screening program

In the Netherlands, pilot-studies on population- based screening by cytomorphologic

In China, cervical cancer screening remains opportunistic screening. Some women

be followed by development of more appropriate health-care policies and services for

III. Cervical cancer screening approach

3.2 Morphology-based testing

3.3 HPV as a marker for cervical cancer screening

Table 3. The main diseases associated with different HPV types

HPV Disease (% attributed

Cervical squamous cell

Cervical squamous cell

Benign genital lesions

Cutaneous HPV2,3,27,57 Skin warts

NCCR 2004-2006 6% 6.0-9.1