Dissertation M Sauer 2014 Druckversion

INTERACTION OF THE HOST
IMMUNE SYSTEM WITH
TUMOR CELLS IN HUMAN
PAPILLOMAVIRUS
ASSOCIATED DISEASES
MADELEINE SAUER
- 2014 -
Dissertation
submitted to the
Combined Faculties for the Natural
Sciences and for Mathematics
of the Ruperto-Carola University of
Heidelberg, Germany
for the degree of
Doctor of Natural Sciences
presented by
Madeleine Sauer
born in Sinsheim
oral examination: 29.01.2015
INTERACTION OF THE HOST
IMMUNE SYSTEM WITH
TUMOR CELLS IN HUMAN
PAPILLOMAVIRUS
ASSOCIATED DISEASES
Referees:
Prof. Dr. rer. nat. Martin Müller
Prof. Dr. med. Magnus von Knebel Doeberitz
The PhD project described in this thesis was started in March 2011 and completed in
November 2014 under the supervision of Prof. Dr. Magnus von Knebel Doeberitz in the
Department of Applied Tumor Biology at the Institute of Pathology, University of Heidelberg
in cooperation with the German Cancer Research Center (DKFZ), Heidelberg.
.
Meiner Familie
ACKNOWLEDGEMENT
First and foremost, I owe my deepest gratefulness to Prof. Dr. Magnus von Knebel Doeberitz, head of
the Department of Applied Tumor Biology at the Institute of Pathology, Heidelberg. He constantly
supported and encouraged me throughout my thesis and offered me many possibilities that promoted
my scientific skills.
I would like to thank Prof. Dr. Martin Müller (DKFZ, Heidelberg) for agreeing to be my first referee,
I am deeply grateful to my supervisor Dr. Miriam Reuschenbach who guided me with her experience,
patience and suggestion through the ups and downs of this thesis. Thank you for all the fruitful
scientific discussions - and beyond - and continuous support.
I would like to thank also PD Dr. Matthias Kloor (Department of Applied Tumor Biology) whose
suggestions and assistance guided me through some of the trickiest experiments of this thesis.
I gratefully acknowledge the great technical assistance of Lena Ehret-Maßholder, Nina Nelius,
Jonathan Dörre and Heike Sartor who helped me with bigger and smaller laboratory issues and
contributed to the successful accomplishment of the numerous projects of this thesis. Importantly, I
would like to thank all those people of the lab who are not mentioned here individually for all kinds of
help provided – often it is the small things that count.
I do not want to miss the opportunity to thank also Prof. Jürgen Kopitz, Dr. Johannes Gebert and
Dr. Svetlana Vinokurova for their advice, assistance and lively discussions and for sharing their
knowledge and experience with me.
I am grateful to Prof. Dr. Dirk Jäger (NCT, Heidelberg) who was not only one of my TAC members,
but also one of our cooperation partners and together with Dr. Inka Zörnig offered me to participate
in an interesting project.
Many thanks are likely extended to PD Dr. Niels Grabe and Dr. Bernd Lahrmann (TIGA Center,
Bioquant, Heidelberg) for the collaboration in the establishment of the automated cell quantification
tool.
I would like to thank the medical collaboration partners for their continuous support and provision of
fresh tumor samples and also fixed tissue specimens: Prof. Dr. Jens Peter Klußmann, Dr. Claus
Wittekindt and Dr. Steffen Wagner with their team (University Hospital Giessen, Department of
Otorhinolaryngology, Head and Neck Surgery), Dr. Daniel Weiß (University Hospital Muenster,
Department of Otorhinolaryngology), PD Dr. Stephan Polterauer and Prof. Dr. Reinhard Horvat
(Medical University of Vienna, Department of General Gynecology and Gynecological Oncology),
Prof. Dr. Dietmar Schmidt (Institute of Pathology, Mannheim), Prof. Dr. Peter Sinn (University
Hospital of Heidelberg, Institute of Pathology), PD Dr. Joachim Rom and Dr. Kristina Schäfer
(University Hospital of Heidelberg, Department of Gynecology and Obstetrics).
I will not forget the innumerable patients who kindly agreed to participate in our studies. Without
their trust in science and support many projects could not have been realized.
I also want to thank Prof. Dr. Braunbeck and Prof Dr. Buselmaier for taking their time to be a part of
my examining committee.
I would like to express my heartfelt thanks to my friends– in the lab and outside – for their wonderful
understanding, support and motivation. Without you life and work would only be half as nice…
I wish to extend my thanks especially to Katharina Essig, Dr. Elena-Sophie Prigge and Dr. Martina
Remus for proof-reading and valuable tips that surely improved the manuscript. Many thanks go also
to Hannelore and Klaus for the “printing services”.
I would like to express my heart-felt gratitude to my family to whom this thesis is dedicated and who
has always believed in me and supports me in all thinkable ways. Thank you for your never-ending
help and tireless encouragements.
Last but not least I would like to thank Sebastian for his calm when I got impatient, for his motivation
when nothing did work and for encouraging me to never forget about the aims and the way I chose.
Without his tireless support that even surmounts the English Channel this work would have never
become what it is now.
ABSTRACT
Human papillomaviruses (HPV) are very common in the sexually active population and contribute to
610,000 cancers per year occurring at different locations. The initial step of HPV-related
carcinogenesis is the induction of transforming processes in the host cells mediated by the viral
oncoproteins E6 and E7 that interfere with critical host cell pathways. The transforming infection is
highlighted by overexpression of the tumor suppressor protein p16 INK4a. Only a small number of
precancerous lesions progress while the majority can be controlled by the host’s immune system and
undergo regression. Progressing lesions under the immunoselective pressure seem to acquire
characteristics that enable them to circumvent the host’s immune attack and promote disease
progression. Immune evasion might be mediated by the immune microenvironment of the tumor as
well as by tumor cell intrinsic features.
The here presented thesis addressed different questions and strategies with regard to the role of the
immune system in HPV-associated diseases and can be subdivided in two main parts: In the first part
immunologic characteristics of precancerous lesions and cancers are investigated to gain insight into
possible immune evasion mechanisms developed during disease progression. In the second part
treatment options to positively influence the balance between immune evasion and anti-tumoral
immune responses are evaluated.
In the first part a) the immunohistochemical characterization of cervical precancers and cancers for
infiltration with different T cell phenotypes revealed that generally increasing T cell densities occur
late in carcinogenesis – and not yet with the onset of early transforming infection - and are
accompanied by immunosuppressive regulatory T cells (Tregs). Mean cell densities for Tregs in the
stroma significantly increased from 121.6 cells/mm2 (range: 24-286.8 cells/mm2) in low-grade lesions
to 308.8 cells/mm2 (24-724.8 cells/mm2) in high-grade lesions and 673.6 cells/mm2 (52.8-1564.8
cells/mm2) in cancer which points to their immunosuppressive role during carcinogenesis. The
demonstrated large variances in T cell densities within one diagnostic category, however, point to a
remarkable heterogeneity of the immune control with potential interesting prognostic implications. On
keratinocytes themselves b) a selective loss for human leukocyte antigen (HLA) class I heavy chain A
expression was observed in about 55% high-grade cervical intra-epithelial neoplasia (CIN) and 65% of
cervical cancers. HLA class II de novo expression was found in 50% of low-grade CIN and in about
85% of high-grade CIN and cervical cancers. These alterations could represent another fundamental
mechanism contributing to immune evasion. A c) longitudinal analysis of immune infiltrates in
patients treated with imiquimod, an immuno-modulatory Toll-like receptor (TLR) agonist, revealed
that the patient’s local immune constitution might be decisive for a possible response to immuneenhancing treatment strategies. Importantly, in patients responding to imiquimod immune cell
densities increased during the treatment as epithelial CD3+ T cell counts (from 160.8 to 371.1
cells/mm2) and CD8+ T cell counts (from 113.8 to 174.1 cells/mm2) demonstrated. The d)
development and establishment of an automated cell quantification tool for high-throughput analysis
allows the search for immune evasion markers and strategies to be continued in an objective,
standardized and faster way.
In consideration of the clinical efficacy of imiquimod and the observed stimulatory effects on the
immune infiltrate density in part one of this thesis e) a new second generation TLR-agonist (TMX202) potentially having less side-effects than imiquimod was tested for the first time in an in vitro T
cell stimulation model in part two of this thesis. Its potential to stimulate innate and adaptive immunity
was demonstrated by an enhanced killing capacity of T cells that were stimulated with HPV-related
antigens loaded on dendritic cells and then co-incubated with HPV16-positive CaSki cells. Based on
the dense infiltration with Tregs observed in part one of the presented thesis the f) immune stimulating
effects of Treg depletion was tested in an autologous in vitro model. In this regard, one major aim of
the thesis was the generation of a new HPV-positive tumor cell line derived from an oropharyngeal
squamous cell carcinoma that serves as model system for HPV-associated tumors. In combination with
peripheral blood lymphocytes obtained from the same patient this autologous system allowed to
address Treg depletion as an immunotherapeutic approach. The results demonstrated that this strategy
might enhance the cell-mediated immune response against tumor cells and emphasize the role that this
particular T cell phenotype is obviously playing in the carcinogenesis of HPV-associated tumors.
Based on the results obtained in the first part of the thesis it is well conceivable that the combination
of different immunologic markers contributes to the definition of a prognostic biomarker tool for
progression and regression of precancerous lesions. Such a prognostic “immune score” has a high
clinical relevance and allows risk-adapted treatment decisions minimizing the costs and long-term
sequelae of surgical interventions. In particular the newly developed microscopy based method in this
work allowing for the automated histological high-throughput quantification of infiltrating immune
cells in cervical intraepithelial neoplasia provides an important methodical tool to realize this long
term goal. The immuno-stimulating effects of the novel TLR7-agonist TMX-202 and Treg depletion
demonstrated in the second part of this thesis by in vitro models indicate that immunomodulatory
approaches could play an important role for the treatment of HPV-associated cancers in the future. In
this regard, the established novel tumor cell line in combination with autologous immune cells
provides a valuable in vitro model system for HPV-associated cancers that can be used to investigate
further immunotherapeutic intervention and treatment strategies.
KURZFASSUNG
Infektionen mit humanen Papillomviren (HPV) sind in der sexuell aktiven Bevölkerung weit verbreitet
und führen zu bis zu 610,000 teilweise unterschiedlich lokalisierter Krebserkrankung pro Jahr. Der
Beginn der HPV-assoziierten Karzinogenese stellt dabei die Induktion des Transformationsprozesses
durch die viralen Onkoproteine E6 und E7 dar, die mit essentiellen Signaltransduktionswegen der
Wirtszelle interagieren. Das transformierende Infektionsstadium korreliert hierbei mit der
Überexpression des Tumorsuppressorproteins p16INK4a. Nur eine geringe Anzahl der daraus
resultierenden präkanzerogenen Läsionen progrediert allerdings weiter zu einem Tumor während die
Mehrheit solcher Läsionen unter Kontrolle des Immunsystems wieder regrediert. Progredierende
Läsionen, die unter dem Selektionsdruck des Immunsystems stehen, scheinen dabei Charakteristika
erworben zu haben um einen Angriff des Immunsystems zu umgehen und ermöglichen dadurch die
Progression der Erkrankung. Entsprechende Immunevasionsstrategien könnten sowohl vom
Immunmikromilieu um den entstehenden Tumor herum ausgehen als auch auf zellinherenten
Tumoreigenschaften beruhen.
Die vorliegende Dissertation beschäftigte sich mit verschiedenen Fragestellungen und verfolgt
verschiedene Ansätze, die die Rolle des Immunsystems im Zusammenhang mit HPV-assoziierten
Erkrankungen näher beleuchten sollen und ist dabei in zwei Hauptteile untergliedert: Teil eins
beschäftigt sich mit der immunologischen Charakterisierung präkanzerogener Läsionen und invasiver
Tumore um einen tieferen Einblick in mögliche Immunevasionsmechanismen bei voranschreitender
Progredienz der Erkrankung zu gewinnen. In Teil zwei werden dagegen verschiedene
Therapiemöglichkeiten evaluiert mit dem Ziel das Gleichgewicht zwischen Immunevasion und antitumoraler Immunantwort positiv zu beeinflussen.
Im ersten Teil dieser Arbeit konnte durch a) die immunhistochemische Charakterisierung von
Infiltrationsraten verschiedener T-Zellphänotypen in Zervixkarzinomvorstufen und -tumoren gezeigt
werden, dass ein Anstieg der T-Zelldichte relativ spät in der Tumorentstehung erfolgt - und nicht mit
der Induktion des frühen Transformationsstadium korreliert - und dabei stets von einem Anstieg an
immunsupprimierenden regulatorischen T-Zellen (Tregs) begleitet wird. Die Mittelwerte der
gemessenen Zelldichten für Tregs im Stroma steigen dabei von 121.6 Zellen/mm2 (Varianz: 24-286.8
Zellen/mm2) in niedriggradigen Läsionen über 308.8 Zellen/mm2 (24-724.8 Zellen/mm2) in
hochgradigen Läsionen auf 673.6 Zellen/mm2 (52.8-1564.8 Zellen/mm2) in Tumoren an was auf ihre
immunsupprimierende Rolle während der Karzinogenese hinweist. Die beobachteten großen
Varianzen in den T-Zelldichten innerhalb einer diagnostischen Kategorie weisen dabei jedoch auf eine
bemerkenswerte Heterogenität der Immunsystemkontrolle mit möglicherweise vielversprechenden
prognostischen Implikationen hin. Auf Seite der sich verändernden Keratinozyten konnte weiterhin b)
ein selektiver Ausfall der Expression der schweren Kette A des humanen Leukozytenantigens (HLA)
Klasse I in 55% aller hochgradigen zervikalen intraepithelialen Neoplasien (CIN) und in 65% aller
Zervixkarzinome festgestellt werden. HLA Klasse II de novo Expression konnte dagegen in 50% aller
niedriggradigen CIN und in 85% aller hochgradigen CIN und Zervixkarzinomen beobachtet werden.
Die gefundenen Veränderungen könnten dabei einen anderen grundlegenden Mechanismus darstellen,
der zur Immunevasion der Tumorzelle beiträgt. Eine c) longitudinal ausgerichtete Analyse von
Immuninfiltraten von Patienten die mit Imiquimod behandelt wurden - einem immunmodulatorischen
Toll-Like-Rezeptor-(TLR)-Agonisten - ergab, dass die lokale Immunkonstitution des jeweiligen
Patienten
entscheidend
für
das
mögliche
Ansprechen
auf
immunstimulatorische
Behandlungsstrategien ist. In Biopsien von Patientinnen, die auf eine Imiquimodbehandlung
ansprachen konnten bemerkenswerterweise hohe Immunzelldichten im Behandlungszeitraum
beobachtet werden, wie die gemittelten epithelialen CD3+ T-Zellzahlen (Anstieg von 160.8 auf 371.1
Zellen/mm2) und CD8+ T-Zellzahlen (Anstieg von 160.8 auf 371.1 Zellen/mm2) belegen. Die d)
Entwicklung und Etablierung eines automatisierten Zellquantifizierungssystems, das speziell zur
Durchführung von Hochdurchsatzanalysen geeignet ist, ermöglicht die Suche nach
Immunevasionsmarkern und -strategien in objektiverer, standardisierter und auf schnellere Art und
Weise fortzusetzen.
Unter Berücksichtigung der klinischen Wirksamkeit von Imiquimod und den ermittelten
immunstimulatorischen Einfluss auf die Immuninfiltrationsraten in Teil eins wurde in Teil zwei dieser
Arbeit e) ein TLR-Agonist der zweiten Generation (TMX-202), der potentiell weniger Nebenwirkung
als Imiquimod aufweist, zum ersten Mal an einem in vitro T-Zellstimulationsmodell getestet. Das
Potential von TMX-202 das angeborene und adaptive Immunsystem zu stimulieren wurde in diesem
Zusammenhang an Hand gesteigerter zytotoxischer Aktivität von T-Zellen nachgewiesen. Diese
wurden mit dendritischen Zellen stimuliert, die mit HPV-assoziierten Antigenen beladen waren und
schließlich mit HPV16-positiven CaSki-Zellen koinkubiert. Aufbauend auf die in Teil eins der
vorliegenden Arbeit nachgewiesenen hohen Infiltrationsdichten an Tregs wurde zusätzlich f) der
immunstimulatorische Effekt einer Verminderung der Treg-Zellzahlen an einem autologen in vitro
Modell getestet. In diesem Zusammenhang war ein Hauptziel dieser Dissertation die Generierung
einer neuen HPV-positiven Tumorzelllinie aus einem Oropharynxkarzinom, die als Modell für HPVassoziierte Tumore dienen soll. In Kombination mit peripheren Lymphozyten, die aus dem Blut des
gleichen Patienten gewonnen werden, sollte das so gewonnene autologe System die Untersuchung
einer Treg-Depletion als einen Ansatz zur Immuntherapie ermöglichen. Die Ergebnisse zeigen, dass
diese Strategie die zellvermittelte Immunantwort gegenüber Tumorzellen verbessern kann und hebt
erneut die Rolle hervor, die dieser spezielle T-Zellphänotyp offensichtlich bei der Entstehung HPVassoziierter Tumor zukommt. Aufbauend auf den Ergebnissen der vorliegenden Dissertation ist es
denkbar, dass eine Kombination verschiedener immunologischer Marker zur Definition eines
prognostischen Biomarkersystems führt, das die Vorhersage der Progression und Regression
präkanzerogener Läsionen ermöglicht. Ein solcher „Immunindex“ ist von hoher klinischer Relevanz
und soll risikoangepasste Behandlungsentscheidungen ermöglichen, um somit die Kosten und
Spätkomplikationen von chirurgischen Eingriffen zu minimieren. Insbesondere das in dieser Arbeit
neu
entwickelte
Mikroskopie-basierte
Verfahren
zur
automatischen
histologischen
Hochdurchsatzquantifizierung von Immuninfiltraten in zervikalen intraepithelialen Neoplasien sollte
ein entscheidendes methodisches Werkzeug darstellen, um dieses Langzeitziel zu erreichen. Die in
dieser Arbeit an in vitro Modellen für HPV-assoziierte Krebsarten gezeigten immunstimulierenden
Effekte des neuartigen TLR7-Agonisten TMX-202 und der Treg-Depletion zeigen, dass
immunmodulatorische Ansätze bei der Behandlung solcher Erkrankungen in Zukunft eine wichtige
Rolle einnehmen könnten. Eine Kombination der neu generierten Tumorzelllinie mit autologen
Immunzellen sollte in diesem Zusammenhang ein verlässliches in vitro Modellsystem für HPVassoziierte Krebsarten darstellen, das weiterführende Studien zu immuntherapeutischen Interventionsund Behandlungsstrategien ermöglicht.
Table of contents
TABLE OF CONTENTS
1.
Introduction
1
1.1
The discovery of human papillomaviruses in the causation of cancer
1
1.2
1.3
Characteristics and life cycle of human papillomaviruses
HPV-associated cancers
2
5
1.3.1
1.3.2
Prevalence, incidence and mortality of HPV-associated diseases
Viral oncogene overexpression and the transforming infection stage
5
7
1.3.3
1.3.4
1.3.5
A biomarker for transforming infections: p16INK4a overexpression
Histomorphological classification of cervical precancers
HPV infection stages interpreted as a progression model of cervical cancer
10
12
12
1.4
1.5
The immunobiology of HPV infections
1.4.1 The role of the host’s immune system in the defense against HPV
1.4.2 Immune evasion strategies developed by human papillomaviruses
Immunologic intervention strategies in HPV-associated diseases
1.5.1 The urgent need for therapeutic interventions in HPV-associated
precancers and cancers
1.5.2 Toll-like receptors are key players in linking the innate and
adaptive immune responses
1.5.3 Toll-like receptor ligands have immune-stimulatory properties
16
16
19
23
23
24
25
2.
Motivation and Rationale
27
3.
Materials and Methods
31
3.1
Materials
3.1.1 Technical equipment, instruments
3.1.2 Chemicals and Reagents
31
31
32
3.1.3
3.1.4
3.1.5
3.1.6
Consumables
Commercially available kits
Antibodies
Enzymes
34
35
36
37
3.1.7
3.1.8
3.1.9
3.1.10
3.1.11
3.1.12
Peptides
Primers
Buffers and Solutions
Cell culture media
Cell lines
Patients’ material
37
38
38
40
42
42
3.1.13 Software
43
Table of contents
3.2
4.
Methods
44
3.2.1
3.2.2
Immunohistochemistry for archived tissue samples
Molecular Biology Methods
44
47
3.2.3
3.2.4
3.2.5
Biochemical Methods
Cell culture methods
Statistical Methods
51
52
63
Immune cell infiltrates and possible immune evasion mechanisms in
cervical lesions
4.1
Development of an automated quantification system for the computational
profiling of cervical intraepithelial neoplasia and its microenvironment
4.1.1
4.1.2
4.2
4.3
4.4
65
65
intraepithelial neoplasia
4.1.3 Calculation of cell densities from the output data
The local immune cell infiltration in cervical intraepithelial neoplasia in
relation to p16INK4a expression
4.2.1 p16INK4a-expression status of the lesions
4.2.2 Comparison of T cell infiltrates in p16INK4a-positive and
66
71
p16INK4a-negative low-grade CIN
4.2.3 T cell infiltrates in p16INK4a-positive high-grade CIN
4.2.4 T cell infiltrates in cervical carcinomas
Alterations of human leukocyte antigen expression in cervical intraepithelial
neoplasia and cancers
4.3.1 Altered HLA class I antigen expression in cervical intraepithelial
neoplasia and cervical carcinoma
4.3.2 Human leucocyte antigen class II expression in cervical intraepithelial
neoplasia and cervical cancer
72
76
78
Immune cell infiltrates under immune-stimulatory treatment
4.4.1 Characterization of the study cohort
87
88
4.4.2
4.4.3
5.
Scanning and digitalization of stained tissue sections
Development of an image processing tool adapted to cervical
64
T cell infiltrates in non-responders and responders to imiquimod
before treatment
T cell infiltrates in non-responders and responders to imiquimod
after treatment
71
72
79
80
84
89
91
Treatment options for HPV-associated precancers and cancers
94
5.1
94
Effects of TLR agonist treatment on immune cells
5.1.1
5.1.2
The effect of TLR7 agonist treatment on the TLR7 mRNA
expression levels in PBMCs
95
The effect of TLR7 agonist treatment on the TLR7 protein expression
in PBMCs
97
Table of contents
5.1.3
Release of the pro-inflammatory cytokine IL-6 of PBMCs upon
treatment with TLR7 agonists
5.2
Effects of TMX-202 treatment on the in vitro priming of naïve T lymphocytes
with HPV-associated and host cell antigens and the generation of
antigen-specific T cells
5.2.1 Determination of L1 peptides bound to HLA class I antigens with high
affinity for stimulation assays
5.2.2 The effect of TMX treatment on dendritic cell maturation
5.2.3 The effect of TMX treatment on stimulation of naïve T cells with
5.2.4
5.3
5.4
100
100
101
105
T cells against CaSki cells
Establishment of an autologous system for the development and evaluation
106
of therapeutic intervention strategies in HPV-associated diseases
5.3.1 The cell line HN038M: general features and patient’s characteristics
5.3.2 Determination of the HPV-status and oncogene activity
5.3.3 Cell line validation via short-tandem-repeat profiling
Effect of regulatory T cell depletion on the cellular immune response against
autologous tumor cells
5.4.1 T cell purity and Treg depletion
5.4.2 Characterization of the effect of Treg depletion on the killing
108
108
111
113
potency of autologous T cells against the tumor cell line HN038M
The killing capacities of T cells co-incubated with autologous tumor
115
cells can also be monitored in real-time
117
5.4.3
6.
HPV-associated antigenic peptides
The effect of TMX treatment on the killing potency of stimulated
98
114
114
Discussion and Conclusion
120
6.1
Overview of the results obtained during the thesis
120
6.2
An automated cell quantification tool allows the analysis of the immune cell
contexture of cervical precancerous lesions in high-throughput approaches
Immune cell densities and composition are different in high-grade lesions
and cancers compared with low-grade lesions
6.3
6.4
6.5
6.6
6.7
6.8
HLA class I and class II antigen expression is altered in cervical
intraepithelial neoplasia and cancers
The density and composition of immune cell infiltrates can be influenced
122
125
127
by immuno-modulatory drugs
The search for the prognostic markers characterizing the immune evasion
phenotype has to be continued
A new immune modulatory drug, TMX-202, shows promising effects the
132
priming of naïve T cells to HPV-associated antigens
The generation of a HPV-associated head and neck squamous cell carcinoma
cell line for immunological studies based on an autologous system
139
136
143
Table of contents
6.9
Regulatory T cells seem to have an inhibitory effect on anti-tumoral immune
responses against autologous tumor cells of a HPV-positive HNSCC patient
6.10 Future prospects
145
148
7.
References
152
8.
Publications, Posters and Oral Presentations
174
9.
Supplementary Material
175
9.1
Supplementary figures
176
9.2
Supplementary tables
185
Abbreviations
ABBREVIATIONS
°C
µ
AB
ad inj
AP
APC
aqua bidest.
ARF
ATCC
ATP
B
Β2M
BC
bp
Ca
CD
CDK
cDNA
CIN
CTL
CTLA
CUL
DAB
DAC
DC
dest
DMEM
DMSO
DNA
Dnmt1
DTT
E
E2F
E6
E7
EDTA
ELISA
EtOH
F
FACS
Fas
FCS
FDA
FFPE
degree Celsius
micro
antibody
ad injectabilia
activator protein
antigen-presenting cell
bidistilled water
alternate reading frame
american type culture collection
adenosine triphosphate
bone marrow
beta-2-microglobulin
B cell, B lymphocyte
base pairs
carcinoma
cluster of differentiation
cyclin dependent kinase
complementary DNA
cervical intraepithelial neoplasia
cytotoxic T lymphocytes
cytotoxic T lymphocyte-associated protein
cullin
3,3'- Di-amino-benzidine
decitabine
dendritic cell
destillata
Dulbecco’s Modified Eagle Medium
dimethyl sulfoxide
deoxyribonucleic acid
DNA methyltransferase 1
dithiothreitol
early gene/protein
transcription factor
early HPV protein 6
early HPV protein 7
ethylenediaminetetraacetic acid
enzyme linked immunosorbent assay
ethanol
fragment
fluorescence-activated cell sorting
apoptosis stimulating fragment
fetal calf serum
Food and Drug Administration
formalin-fixed paraffin-embedded
Abbreviations
FI
FITC
FOXP3
g
G
gDNA
GM-CSF
GranB
Gy
h
H
H2O
HCA
HEPES
HLA
HNSCC
HPLC
HPV
HR
HRP
HSV
Hz
ICD-O
IF
IFN
Ig
IG
IHC
IHS
IL
IMDM
iNOS
Jak
JPEG
KDM
KMT
L
l
LC
LCR
LPS
LR
M
m
MACS
MCP-1
MeOH
MFI
MHC
fluorescence intensity
fluorescein isothiocyanate
forkhead box P3
gram
growth
genomic DNA
granulocyte macrophage colony-stimulating factor
granzyme B
Gray
hour/s
histone
water
human carcinoma antigen
4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid
human leukocyte antigen
head and neck squamous cell carcinoma
high pressure liquid chromatography
human papillomavirus
high risk
horse raddish peroxidase
herpes simplex virus
Hertz
international classification of diseases for oncology
immunofluorescence
interferon
immune globulin
immunoglobulin
immunohistochemistry
Intensity-Hue-Saturation
interleukin
Iscove's modified Dulbecco's medium
increased inducible nitric oxide synthase
Januskinase
joint photographic expert group
histone lysine demethylases
histone lysine methyltransferases
late gene/protein
liter
Langerhans cell
long control region
lipopolysaccharide
low risk
Molar, Mol / liter
meter
magnetic cell sorting
monocyte-chemoattractant-protein-1
methanol
mean fluorescence intensity
major histocompatibility complex
Abbreviations
MICA
min
miRNA
MMP
mol
mRNA
MSI
MT
MyD88
n
n
NFkB
NK
NKT
NMSC
NO
OIS
OPC
ORF
OSCC
p
p
p.a.
p16INK4a
p53
PAGE
PAMP
Pap test
PBL
PBMC
PBS
PBS-T
PCR
PDZ
PE
PFA
pRB
PRR
PVDF
qPCR
Rb
RB
RGB
RNA
ROI
rpm
RT
RT
MHC class I chain-related molecule
minute/s
micro ribonucleic acid
matrix-metalloproteinase
molar
messenger ribonucleic acid
microsatellite instability
methyl transferase
myeloid differentiation primary-response protein 88
sample size
nano
nuclear factor-kappa B
natural killer cell
natural killer T cell
non-melanoma skin cancer
nitric oxide
oncogene induced stress
oropharynx cancer
open reading frame
oropharyngeal squamous cell carcinoma
probability
protein, peptide
per analysis
tumor suppressor kinase inhibitor of CDK4
tumor suppressor protein 53
polyacrylamide gel electrophoresis
pathogen-associated molecular pattern
Papanicolau test
peripheral blood lymphocytes
peripheral blood mononuclear cell
phosphate-buffered saline
PBS containing 0.05% Tween 20
polymerase chain reaction
post synaptic density protein (PSD95), Drosophila disc large tumor suppressor
(Dlg1), zonula occludens-1 protein (zo-1)
phycoerythrin
paraformaldehyde
phosphorylated retinoblastoma protein
pattern recognition receptor
polyvinylidene difluoride
quantitative polymerase chain reaction
retinoblastoma
retinoblastoma protein
red green blue
ribonucleic acid
region of interest
rounds per minute
room temperature
reverse transcription
Abbreviations
S
sec
SCC
SDS
SSC
STAT
STR
T
t
TAP
TAP-1
Taq
TBE
TBS
TCR
TEMED
TGF
Th
TIL
TLRs
TMB
TMX
TNF
TNM
Treg cell
Tris-HCl
V
v/v
VAIN
VIN
VLP
w/v
synthesis
second/s
squamous cell carcinoma
sodiumdodecylsulfate
squamous cell carcinoma
signal transducers and activators of transcription
short-tandem-repeat
thymus
time
transporter associated with antigen presentation
antigen-processing-protein-1
thermus aquaticus
tris-borate EDTA
tris-buffered saline
T cell receptor
tetramethylethylenediamine
transforming growth factor
T-helper cell
tumor infiltrating lymphocyte
Toll-like receptor
Tetramethyl benzidine
Telormedix
tumor necrosis factor
tumor nodes metastasis
regulatory T cell
Tris-(hydroxymethyl)aminomethane hydrochloride
vacuolar
volume per volume
vaginal intraepithelial lesion
vulvar intraepithelial neoplasia
virus like particle
weight per volume
1. Introduction
1
1. I
NTRODUCTION
“Do not follow where the path may lead. Go instead where there is no
path and leave a trail.” (George Bernard Shaw)
1.1
The discovery of human papillomaviruses in the
causation of cancer
The knowledge of the relationship between the carcinogenesis of cervical tumors and sexually
transmittable agents has already dated back to the 19th century when a link between the sexual
behavior and the risk for development of cervical cancers could be established. The Italian physician
Domenico Antonio Rigoni-Stern analyzed the causes of death of Veronese women who had died
between 1760 and 1839 and observed a significantly higher frequency of cervical cancers occurring in
sexually active women, compared to virgins and nuns who were affected by cervical tumors only very
rarely (GASPARINI and PANATTO, 2009).
The awareness that cancers can be attributed to infectious agents raised in the beginning 20 th century.
Peyton Rous, in 1911, demonstrated that cell-free extracts of chicken sarcoma can be transferred from
one individual to another (ZUR HAUSEN, 2011).
The history of the discovery of the human papillomavirus (HPV) as causing agent in the development
of cervical cancer however started considerably later in the 1970s. At that time, Harald zur Hausen
subverted the widespread opinion among scientists and clinicians that the sexually transmittable
herpes simples virus type 2 (HSV-2) may be causally linked to carcinomas of the anogenital tract and
proposed papillomaviruses to contribute to cervical carcinogenesis (ZUR HAUSEN et al., 1974). He
proved that not HSV-2 but HPV DNA is detectable in the tumor tissue and hypothesized that the viral
genome is persistently present and transcriptionally active in HPV-infected cancer cells (ZUR
HAUSEN et al., 1975).
His findings considerably contributed to a detailed description of the phylogenetic heterogeneity of the
human papillomaviruses, to the identification of the major HPV types associated with cancer and the
characterization of the oncogenic potential of the HPV proteins E6 and E7. His work awoke the
general interest of the scientific community on viruses as cancer causing agents in general and on HPV
in particular leading to the opening of a completely new research area which aims at understanding of
the molecular mechanisms and, in a second step, combating HPV-related diseases. Having the courage
to leave the path chosen by the research community of his time and thus paving the way for the
following generations of researchers to participate in the battle against cancer caused by one of the
most common sexually transmittable agents he was rewarded the Nobel Prize in 2008.
2
1.2
1. Introduction
Characteristics and life cycle of human papillomaviruses
Human papillomaviruses are non-enveloped DNA viruses containing one single-stranded circular
DNA molecule with a size of about 8000 base pairs. They belong to the family of papillomaviridae
which can further be subdivided into five genera (Alpha-, Beta-, Gamma-, Mu- and
Nupapillomaviruses) (DOORBAR, 2006). The viral genome (Figure 1.1) comprises eight open
reading frames (ORF) covering three distinct functional parts, namely the early genes’ region (E1-E7),
the late genes (L1 and L2) and a non-coding part called long control region (LCR) containing cisregulatory elements (ZHENG and BAKER, 2006). HPV is characterized by an icosahedral capsid of a
diameter of about 55 nm and consisting of 72 pentameric subunits, the capsomers. The capsomers are
made of two structural proteins, the late proteins L1 and L2.
FIGURE 1.1
THE HPV16 GENOME. The genome comprising 7904 bp is represented by a black circle with the early
(p97) and the late (p670) promoters marked by arrows. The six early ORFs (E1-E7) are expressed from
either p97 or p670. The late ORFs (L1 and L2) are represented as yellow structures and the long control
region (enlarged and visualized as black line) contains the E2-binding sites, the binding sites for E1 and
the TATA element of the p97 promoter. Adapted from (DOORBAR, 2006).
Human papillomaviruses are characterized by a strong tropism for epithelial cells and the mucosa. The
virus’ life cycle is initiated with infection of keratinocytes of the basal squamous epithelium which
requires either microlesions as they can be found for example in the cervical epithelium or other
anatomical structures like the characteristic tonsillar crypt epithelium. HPV infection therefore
exclusively affects undifferentiated epithelial stem cells as cell division and keratinocyte
differentiation are the prerequisite for the completion of the virus’ life cycle and replication (CHOW et
al., 2010). The infectivity is predominantly mediated by the viral proteins L1, but also L2 capsid
proteins (BUCK et al., 2013) enabling the virus to bind to the host cells which is followed by uptake
of the virus via clathrin-based endocytosis. The late proteins also seem to be involved in transferring
the viral DNA to the nucleus following disassembling in the late endosomes and lysosomes
(DOORBAR, 2006). Initially, following infection the keratinocytes undergo lateral cell divisions and
thus build a reservoir of stem cells harboring the virus (NGUYEN et al., 2014). Viral DNA replication
1. Introduction
3
in this phase is tightly synchronized with the amplification of the host cell DNA during S-phase.
During this latent phase of the infection the early proteins E1 and E2 are expressed which fulfill
different functions during early infection (see also Table 1.1 for HPV protein functions). In addition to
the role they play in the replication of the viral genome, they assure that the viral DNA is maintained
as an episome at low copy numbers of about 10-200 copies. The viral early proteins E1 and E2 seem
to prevent the integration of the viral DNA into the host’s genome as well as they assure the correct
viral genome segregation during stem cell divisions (MCBRIDE, 2013). E2 initiates the amplification
of the viral DNA by binding to the HPV upstream regulatory region and by forming together with E1
that is recruited to this non-coding region the E1/E2 initiation complex. The E1 protein functions as a
DNA helicase, and recruits several other, cellular, proteins to the viral origin of replication such as
RPA (replication protein A) and DNA polymerase α primase (CONGER et al., 1999; DOORBAR,
2006). Furthermore E2 controls the early promoter of high-risk HPV types, called p97 in HPV16, and
thus strictly regulates the early proteins, which is of special importance regarding the expression rates
of the viral oncogenes E6 and E7 expressed at low levels only (DOORBAR, 2005).
With the migration of virally infected basal stem cells into the suprabasal cell layer of the epithelium
the cells quit the cell cycle in order to undergo terminal differentiation. This initiates the productive
phase of the viral infection characterized by activation of the viral genes in parallel to the
differentiation program of the keratinocytes.
FIGURE 1.2
THE HPV LIFE CYCLE. The cell layers of the mucosal epithelium are indicated on the left. Cells
expressing cell cycle markers (red) that occur in the suprabasal cell layer are characterized by viral
oncogene expression (E6 and E7) (green cells). The activation of p670 in E6 and E7 expressing cells of
the upper epithelium leads to expression of viral proteins required for viral genome replication. The
successive viral protein expression stages indicated by arrows represent distinct steps of the viral life
cycle directly influencing also the host cell: low-levels of E1, E2, E4 and E5 (light green) accompanied
by viral oncogene expression (E6 and E7) leads to induction of cell proliferation. Elevation of the proteins
involved in replication (dark green) allows increased viral genome amplification. L1 and L2 (yellow) are
expressed in the upper epithelium, where viral genome is packaged into infectious particles. Here, E4 also
is expressed and probably contributes to the viral release. Adapted from (DOORBAR, 2006).
As suprabasal keratinocytes that are terminally differentiated undergo cell cycle arrest, the virus has to
reactivate S-phase of the host cells to complete its life cycle. The viral oncogenes E6 and E7 are
required for the reactivation of the host replication cycle. However, their expression is still under
control and locally restricted to a few cells in the lower part of the epithelium. These cells assure the
4
1. Introduction
viral replication to be maintained and infectious virions to be produced once the keratinocytes quit the
basal cell layer and migrate towards suprabasal layers (DOORBAR, 2006). The mode of action of the
viral oncoproteins acting in the host cells are explained in more detail in section 1.3.2. In the mid and
upper layers of the epithelium, the viral DNA is replicated, the amplified genome is packed into viral
capsids and infectious virions are finally released. These processes are controlled by the late promoter
which is dependent on the differentiation program and activated with the migration of keratinocytes
through the epithelium. Its activation leads to increased expression rates of viral early proteins E1, E2,
E4 and E5 which are involved in HPV DNA replication (Figure 1.2). However, E6 and E7 expression
levels still are tightly controlled by the repressive functions of E2 on the early viral promoter (HAMID
et al., 2009).
viral
protein
function in the viral life cycle
activities in the host cell
Early proteins:
E1
viral genome replication
DNA-binding activity, helicase activity, ATPase
E2
viral gene transcription, viral genome
replication, viral genome maintenance
transactivation/transrepression, DNA-binding
activity, DNA segregation in mitotic cell
E4
viral genome replication (enhanced
amplification)
destruction of keratin network, induction of G2M
arrest of cell cycle
E5
possibly involved in proliferation and/or
inhibition of apoptosis
interference with cellular signaling pathway
E6
reactivation of cellular replication
mechanisms, proliferation, immortalization,
inhibition of apoptosis, viral genome
maintenance
interaction with various cellular proteins, e.g. p53,
c-Myc, Bak, Bax, PDZ domain
E7
reactivation of cellular replication
mechanisms, proliferation, genomic
instability, inhibition of apoptosis, viral
interaction with various cellular proteins, e.g. pRB,
HDAC, E2F6, p21, p27, CDK/cyclin
genome maintenance
Late proteins:
L1
major capsid protein
L2
minor capsid protein
TABLE 1.1
THE HPV16 PROTEIN FUNCTIONS. Depicted are the early and late proteins, their main functions in
the viral life cycle and their activities in the host cell. Adapted from (KAJITANI et al., 2012).
The viral proteins E4 and E5 also seem to be involved in viral DNA replication. E5 is involved in
EGF-mediated signaling in order to maintain an environment that is favorable for replicative
processes. The viral E4 protein expression levels increase during genome amplification and induce G2
cell cycle arrest of host keratinocytes and thus prevents cell proliferation by counteracting E7 effects;
it enhances viral genome amplification and thus is likely to contribute to an increased viral synthesis
rate. Its interaction with and destabilizing effects on the keratin network of host cells implies that E4
could also be involved in viral release (reviewed in DOORBAR, 2013).
Finally, the structural L1 and L2 proteins necessary for capsid formation and packaging of the viral
DNA are expressed and accumulate during viral replication. Following assembly – in which E2 is also
involved by binding viral DNA for loading - of capsids containing one copy of the HPV DNA the
1. Introduction
5
infectious particles are released from the fully differentiated cells that reach the outer surface of the
squamous cell epithelium (DOORBAR, 2006).
The viral gene expression pattern is tightly associated with the biological infection stage. The above
described life cycle represents the productive or permissive infection stage characterized by viral DNA
replication and release of newly assembled virions. In case of persistence of the HPV infection, the
productive live cycle may be quit and the proliferation of terminally differentiated epithelial cells in
the lower third of the epithelium can be induced (DOORBAR, 2005).
This shift occurs under the influence of deregulated expression of E6 and E7 oncoproteins leading to
abolishment of cell cycle arrest while DNA damage responses are inhibited. These processes induce
the transforming infection stage which will be described in more detail in chapter 1.3.2.
1.3
HPV-associated cancers
More than 100 different HPV genotypes have been described so far (BERNARD, 2005). Those that
are the clinically most relevant ones belong to the genera of alpha-papillomaviruses and cause not only
cancerous precursor lesions and cancer but also genital warts (DE VILLIERS et al., 2004).
Furthermore, papillomaviruses are classified into two groups depending on their potential to cause
cancer: low-risk types leading to ano-genital warts or common skin warts whereas the so called highrisk types - representing only a handful of all genotypes described so far - are involved in cancer
development. Among those the high-risk types HPV16 and HPV18 are the most prevalent ones with
their DNA being detectable in around 80% of cervical cancers (SCHIFFMAN et al., 2007). The most
prevalent low-risk HPV types are HPV11 and HPV6 that cause the vast majority of genital warts
(STEBEN and GARLAND, 2014).
This chapter deals with the contribution of HPV to tumor development and the mechanisms involved
in the establishment of precancerous lesions and their progression towards invasive disease.
1.3.1
Prevalence, incidence and mortality of HPV-associated diseases
HPV infections are common within the sexual active population and thus considerably contribute to
the global health burden: the prevalence in the younger population aged 15 to 25 years is very high
and an individual’s life time risk to enter in contact with these viruses and get infected is about 80%
(DUNNE et al., 2007). Although the majority of HPV infections are cleared within two years – the
mean clearance time even is 5 months - remaining without any consequences, more than 5% of
cancers appearing worldwide are related to HPV infections that persisted (DE MARTEL et al., 2012).
Incidences for different HPV-associated diseases vary between women and men (Figure 1.3).
Cervical cancer is the fourth most common cancer in women following breast cancer, colorectal
cancer and lung cancer in the latest GLOBOCAN statistics. Nearly 530,000 new cases of cervical
cancer are diagnosed each year (2012), the main burden however occurs in developing countries that
show a 10-fold higher incidence compared with industrial nations. Here, cervical cancer represents
almost 12% of all female cancers and it still remains the most common cancer in women in Eastern
and Middle Africa (FERLAY et al., 2010). This difference is explained by lacking screening programs
and early detection of precancerous stages and cervical cancer. The pap-test, developed in the 1930s
6
1. Introduction
by George Papanicolaou, allows the early detection and treatment of precancerous stages and since its
introduction has led to a significant decline of the cervical cancer incidence in industry nations (GIBB
and MARTENS, 2011). In 2012, an estimated number of around 270 000 women died from cervical
cancer worldwide. This makes a percentage of 7.5% of all female cancer deaths. Here again, the
majority, more than 85% of all deaths related to cervical cancer, occur in low- or middle-income
countries (FERLAY et al., 2010).
Even though cervical cancer is the best characterized among the HPV-associated diseases HPV
infection can also occur on other epithelial or mucosal sites and can cause several other cancer types.
While virtually all cancers of the cervix uteri are attributable to precedent infection with human
papillomaviruses, persistent HPV infection also causes precancerous lesions and cancers at other anogenital sites and contributes to a proportion of vulvar, vaginal, penile and anal cancers (PARKIN and
BRAY, 2006).
FIGURE 1.3
CONTRIBUTION OF HPV TO THE HEALTH BURDEN. Estimated incidence rates of HPV-associated
diseases (related to HPV6, 11, 16, 18) in women and men in Europe. Adapted from (STANLEY, 2012b).
The contribution of HPV in the carcinogenesis of a proportion of head and neck squamous cell
carcinoma (HNSCC) today is widely accepted (GILLISON et al., 2000). Especially, the oropharyngeal
tract can be concerned by HPV infection and recent studies revealed an increasing incidence of HPVassociated cancers of the oropharynx (OPC), the pharyngeal region located at the back of the throat,
and here primarily the tonsils, the base of the tongue and the soft palate are affected (CHATURVEDI
et al., 2011). Also here HPV16 is the most prevalent type with about 90% of HPV-positive carcinomas
located in the oropharynx being positive for HPV16 DNA. Also in the other ano-genital sites, except
the cervix uteri where HPV18 plays a non-negligible role, the vast majority of cancers are associated
with HPV16 infections (BOSCOLO-RIZZO et al., 2013). The incidence rate of head and neck
squamous cell carcinomas equals that of cervical cancer with about 550 000 new HNSCC cases per
year worldwide making it the 7th most common cancer in men. Thereof around 85,000 cases represent
oropharyngeal cancers. The mortality rate is higher than in cervical cancer with around 305,000 deaths
per year related to HNSCC. However, the etiological heterogeneity – with tobacco and alcohol being
the major risk factors for HNSCC – makes it difficult to estimate to which extent HPV contributes to
oropharyngeal cancers (GILLISON et al., 2014). Estimates for HPV-association among oropharyngeal
cancers are higher for industrial nations (60-70% for the United States) which have experienced an
increase in incidence for oropharyngeal cancer in the last two decades, whereas less than 10% of OPCs
are believed to be caused by HPV in developing regions (CHATURVEDI et al., 2013). Worldwide,
1. Introduction
7
these data lead to an estimated proportion of about 25% of OPCs attributable to HPV infection
following an IARC review published in 2012 (CHATURVEDI et al., 2013).
Distinct HPV types were found to play a role in the autosomal recessive hereditary skin disorder
epidermodysplasia verruciformis which is characterized by a higher susceptibility for persistent
infections and development of benign lesions and also malignancies of the skin (HARWOOD et al.,
2004). HPV also contributes to non-melanoma skin cancer (NMSCC) (reviewed in MOLHOPESSACH and LOTEM, 2007 and SMOLA, 2014) and might be of higher relevance in
immunosuppressed individuals (REUSCHENBACH et al., 2011).
The mechanisms of HPV-induced carcinogenesis in the following chapters will be explained by means
of the well-studied cervical cancer and its precursor lesions, the so called cervical intraepithelial
neoplasia (CIN). The underlying tumorigenic mechanisms, however, are the same in other HPVassociated cancers.
1.3.2
Viral oncogene overexpression and the transforming infection stage
A small percentage of HPV infections – those that are not cleared and persist over months – finally
shift from permissive/productive to transforming infection which is accompanied by a massively
deregulated expression of the E6 and E7 oncoproteins. The affected cells that had undergone cell cycle
arrest are driven continuously into S-phase and finally shift from the production of viral infectious
agents to an intensively supported proliferation (DOORBAR, 2005). The deregulated expression of the
E6 and E7 oncoproteins therefore can be considered as the prerequisite for the establishment of
precancerous lesions and the development of invasive cancer. The transformation zone of the cervix is
preferentially affected by HPV-induced neoplasia. Here, the cells of the stratified squamous
epithelium and the columnar endocervical cells converge, for which reason this area is also called the
squamocolumnar junction. Under hormonal influence the transition zone during the female life cycle
is subject to substantial anatomical changes and rebuilding processes with changing proportions of
columnar and stratified epithelium (BURD, 2003).
FIGURE 1.4
TRANSFORMING INFECTION IS INDUCED BY FUNDAMENTAL CHANGES IN THE VIRAL
GENE EXPRESSION PATTERN. While the majority of low-grade lesions (CIN1) represent permissive
infections with the underlying viral gene expression patterns described above, high-grade lesions (CIN 2
and CIN3) are characterized by an increasing proportion of cells with deregulated E7 expression which is
accompanied by decreased expression of the early proteins involved in viral replication. This may be
accompanied by the integration of the viral DNA into the host cell genome. In CIN3 and cervical cancers
are characterized by more and more decreased or absent viral replication and increased E7 expression.
Adapted from (DOORBAR, 2006).
8
1. Introduction
Recently, a cell population retaining embryonic characteristics within the junction has been identified
that is speculated to be particularly susceptible to HPV-induced carcinogenesis (HERFS et al., 2012).
The occurrence of HPV-associated HNSCC at oropharyngeal sites, specifically the tonsils, could also
be explained by the histologic characteristics of the lymphoepithelium of the tonsillar crypts. Here, the
so called reticulated epithelium seems to be the preferred site for HPV infections (WESTRA, 2012).
The loss of the repressive effects of viral E2 on the early promoter p97 is considered to be the reason
for the deregulated oncogene expression and leads to a massive up-regulation of E6 and E7 expression
rates in higher lesion grade (Figure 1.4). Different events are discussed as underlying mechanisms for
the loss of E2 function. One is the integration of the viral episome into the host’s genome which
disrupts the gene locus for E2 (VERNON et al., 1997). However, also cells with unintegrated,
episomal viral DNA show E6 and E7 overexpression, a finding for which other, epigenetic
mechanisms such as methylation of the E2-binding sites changing the binding affinities and thus the
transcription rates might be responsible (CHAIWONGKOT et al., 2013).
The oncoproteins E6 and E7, once they are expressed, perfectly act together, complementing the
functions of each other, to re-induce proliferation in terminally differentiated cells and to circumvent
apoptosis at the same time.
FIGURE 1.5
HPV16 E7 CIRCUMVENTS ONCOGENIC STRESS INDUCED CELL CYLCE ARREST BY
TARGETING pRB FOR DEGRADATION. A) In normal cells CDK4/6 is negatively regulated by
p16INK4a. If absent CDK4/6 is activated by binding of cyclin D and phosphorylates pRB which is
degraded. Loss of the suppressive subunit pRB activates the E2F transcription factor which mediates Sphase entry. B) Oncogenic stress induces KDM6B expression and increased p16 INK4a levels which
inhibits CDK4/6 activity and phosphorylation of pRB, resulting in G1 cell cycle arrest and senescence. C)
HPV E7 targets pRB for degradation and circumvents growth arrest. Adapted from (MCLAUGHLINDRUBIN and MUNGER, 2013).
Cell cycle progression, the transition from G1-phase to S-phase in dividing cells, is tightly regulated
by the complex built of the cyclin-dependent kinase 4 (CDK4) and cyclin D. Under normal
circumstances binding of the regulatory subunit cyclin D to CDK4 activates the complex and allows
CDK4 to phosphorylate pRB which then dissociates from the E2F transcription factor. E2F migrates
to the nucleus where it induces the transcription of genes necessary for cell cycle progression, such as
cyclins A and E. In normal, non-dysplastic cells, the cyclin-dependent kinase inhibitor p16INK4a
negatively regulates the kinase activity of CDK4 and CDK6 by binding to them and inhibiting the
1. Introduction
9
formation of active complexes with cyclin D. This prevents hyperphosphorylation and inactivation of
pRB and the release of E2F transcription factors. Consequently, cells expressing p16 INK4a under
normal conditions are retained in the G1 phase and do not enter S-phase (Figure 1.5)
(MCLAUGHLIN-DRUBIN and MUNGER, 2013).
Due to the potency to drive host cells into S-phase which is crucial for the induction of the
transforming infection stage, E7 directly contributes to carcinogenesis (reviewed in MCLAUGHLINDRUBIN and MUNGER, 2009). E7 is assumed to bind to the retinoblastoma (pRB) tumor suppressor
thus interfering with the pRB pathway and abrogating the host’s capacities to control cell cycle
progression in the way similar to how other viruses achieve the same goal (JONES and WELLS,
2006). Binding of the E7 oncoprotein to pRB leads to the disruption of complexes built of pRB and
transcription factors belonging to the E2F family (Figure 1.5). Although no external growth stimuli are
present, E2F is released from pRB and activates other host cell proteins involved in DNA replication
such as the cyclins A and E (DOORBAR, 2006). Furthermore E7 was demonstrated to interact with
other proteins of the host’s cell cycle regulation machinery, among others the activator protein 1(AP1)
transcription complex (ANTINORE et al., 1996), histone deacetylases (LONGWORTH et al., 2005)
and also the cyclin-dependent kinase inhibitors p21 and p27 (NOYA et al., 2001) are concerned.
Interestingly, during natural infection E7 does not always induce cell cycle progression in
differentiated keratinocytes. In some cells which express high levels of p21 and p27 the CDKs seem to
be resistant to oncoprotein effects as here E7 builds inactive complexes with cyclin E. Consequently,
mitosis is induced only in cells that are characterized by low p21 and p27 levels or by E7 levels that
are high enough to overcome the cycle arrest (DOORBAR, 2006).
FIGURE 1.6
HPV16 E6 INTERFERES WITH THE P53 PATHWAY. Binding of E6 to p53 promotes its degradation
and thus prevents cell cycle arrest and apoptosis. Adapted from (YIM and PARK, 2005).
The interference of E7 with the pRB pathway is complemented by E6 which abolishes the p53mediated apoptotic signaling. Modulation of apoptotic pathways is a common and effective
mechanism known from a number of oncogenic viruses and contributes to malignant progression
(reviewed in FUENTES-GONZALEZ et al., 2013)). During transforming HPV infections, the cellcycle entry of upper epithelial cells normally should lead to apoptosis mediated by ARF (ADP
10
1. Introduction
ribosylation factor). However, under the influence of the viral oncoprotein E6 binding to E6AP (E6associated protein) p53 is ubiquitinylated and degraded (DOORBAR, 2006). The interplay of E7
together with E6 impairing different host cell pathways leads to deregulated cell proliferation while
the central apoptotic pathway is impaired. Apart from the interaction with p53 the viral E6 protein is
also reported to target telomerase and different PDZ proteins involved in cell signaling and other
cellular processes and thereby further supports transforming processes within the host cells
(GANGULY and PARIHAR, 2009; WISE-DRAPER and WELLS, 2008).
The inhibition of central DNA repair mechanisms at the same time when host cells undergo
deregulated DNA synthesis has massive further consequences on genome integrity and provokes
additional genomic alterations (DUENSING and MUNGER, 2002).
In addition to the well-established mode of action of the viral oncoproteins E6 and E7 also E5 is
discussed as another protein being involved in the development of cancers (reviewed in MIGHTY and
LAIMINS, 2014)). One possible mechanism might be its ability to also inhibit ubiquitination and
subsequent degradation of Bax and thereby preventing hydrogen-peroxide induced apoptosis (OH et
al., 2010).
1.3.3
A biomarker for transforming infections: p16 INK4a overexpression
The expression of cyclin-dependent kinase inhibitor p16INK4a is induced in aging cells and therefore a
sign of senescence accompanied by cell cycle arrest and chromatin condensation. Its ability to prevent
cells from further proliferation is of special interest in premalignant and malignant cells that have
acquired genomic damages. In these cells p16INK4a acts as a tumor suppressor preventing cell cycle
progression and further accumulation of DNA damages. Due to the biological importance of the CDK
inhibitor there seems to be an evolutionary pressure for loss of p16INK4a gene function in neoplastic
context. Indeed, many cancers of different sites show evidence of a functional loss of p16INK4a by
epigenetic modifications, deletions or point mutations. (LIGGETT and SIDRANSKY, 1998; ROCCO
and SIDRANSKY, 2001).
In cervical intraepithelial neoplasia and carcinomas however, p16INK4a is overexpressed. This can be
directly linked to the deregulated oncogene expression and therefore p16INK4a represents a wellestablished and recognized marker for transforming HPV infections (VON KNEBEL DOEBERITZ,
2002). In this context its biological function as growth arrest inducing protein is abolished by the viral
oncoprotein E7 that inactivates the down-stream inhibitory signals of p16INK4a. The overexpression of
the viral protein E7 in dysplastic cells with underlying transforming HPV infection causes oncogenic
stress to the host cell. This leads to an epigenetic remodeling particularly of the CDKN2a
(p16INK4a/ARF) gene locus and a substantially increased p16INK4a expression.
Histone lysine methylation is one epigenetic mechanism involved in transcriptional activation and
repression and for this reason plays a non-negligible role in cell cycle regulations. The enzymes
involved in this epigenetic chromatin remodeling are histone lysine methyltransferases (KMTs) and
demethylases (KDMs) that influence lysine methylation pattern of histones. The trimethylation mark
at lysine 27 of histone H3 (H3K27me3) results in epigenetic silencing of the gene. The histone
demethylases KDM6A (UTX) and KDM6B (JMJD3) however are able to remove the repressive
methylation pattern and therefore are involved in transcriptional activation. The expression of the viral
1. Introduction
11
E7 leads to oncogene induced stress (OIS) and transcriptional induction of histone demethylases
KDM6A and KDM6B which remove the H3K27me3 mark. This results in epigenetic reprogramming
by changing the levels of histone methylation of the p16IN4a/ARF gene locus and enhanced p16 INK4a
expression (MCLAUGHLIN-DRUBIN et al., 2011).
Cells in the transforming infection stage however are not subject to p16 INK4a mediated cell cycle arrest
and the cells continue to proliferate in presence of the overexpressed cyclin-dependent kinase
inhibitor. This is explained by the interference of E7 with p16INK4a downstream targets, namely pRB
which is degraded under the influence of E7 activating the CUL2 pathway (HUH et al., 2007).
In this context, p16INK4a therefore cannot be considered as a senescence marker anymore and remains
without mechanistic relevance. It is rather a marker for the transforming processes ongoing within the
virally infected host cells. The immunohistochemical staining pattern in these cases shifts from patchy
in case of real senescence to strong and diffuse in dysplastic lesions (KLAES et al., 2001).
FIGURE 1.7
EXAMPLES OF p16INK4a IMMUNOHISTOCHEMISTRY ON CERVICAL TISSUE SAMPLES:
shown are representative examples of A) normal epithelium negative for p16 INK4a, B) cervical
intraepithelial neoplasia (CIN3) with p16INK4a-positive epithelium and C) cervical carcinoma with strong
p16INK4a staining. Details of the epithelium are shown on right side.
Recently published data (MCLAUGHLIN-DRUBIN et al., 2013) on p16INK4a functions demonstrated
that p16INK4a overexpression is not only a bystander effect of oncogenic stress induced by HPV E7
protein and not a consequence of pRB inactivation by the viral oncogene, but is elementary to
maintain the neoplastic phenotype and the continuous growth of cells with underlying transforming
HPV infections. Dysplastic cells under the influence of viral E7 become dependent on KDM6B and
p16INK4a expression for survival.
12
1.3.4
1. Introduction
Histomorphological classification of cervical precancers
In histomorphology, HPV-induced lesions are subdivided in three distinct progression steps and
described by successive grade of cervical intraepithelial neoplasia (CIN) depending on the extent of
morphological aberrations (MARTIN and O'LEARY, 2011). With increasing dysplastic cellular
alterations beginning in the basal and suprabasal layers and eventually reaching throughout the
complete epithelium the lesions are termed with increasing lesion grades from 1 to 3. CIN1 usually is
described by the occurrence of so called koilocytes, indicating that viral replication is ongoing in the
suprabasal layers of the epithelium. The altered cellular morphology concerns less than one third of
the thickness of the epithelium. Lesions characterized as CIN1 are not yet considered as
premalignancy in the narrower sense and therefore usually are not treated. With persistence and
progressing disease cells with more severe dysplastic cellular alteration expand and may grow beyond
the lower third of the thickness of the affected squamous epithelium and these lesions are named
CIN2. The lesions that grow further and even beyond two thirds of the epithelium are referred to as
CIN3 lesions (DARRAGH et al., 2013; RICHART, 1973).
The Histomorphological defined CIN grades cannot be translated unequivocally into the biological
infection stages. CIN1 and a part of CIN2 lesions retain the capacity to undergo the normal squamous
epithelial differentiation and thus for viral replication and represent the permissive (productive)
infections. However, the control of the viral oncogene expression in basal and suprabasal cells may
have already been lost in a part of CIN1. The vast majority of CIN2 lesions and virtually all CIN3
lesions are in the advanced transforming infection stage. Due to this discrepancy markers are needed
to highlight the biological infection stage in biopsies. As it is directly link to oncogene activity
p16INK4a overexpression represents a reliable biomarker to identify lesions that have quit productive
infection and entered the transforming infection stage independently of their histomorphological
appearance (BERGERON et al., 2014; VON KNEBEL DOEBERITZ et al., 2012).
1.3.5
HPV infection stages interpreted as a progression model
of cervical cancer
The cervical carcinogenesis can be subdivided in clearly defined successive steps (Figure 1.8). Latent
infections (step 1), during which the viral DNA has been replicated yet, usually remain clinically
innocuous and are characterized by basal infected keratinocytes that divide continuously to establish a
reservoir of cells harboring the episomal viral DNA. Permissive or productive infections (step 2) are
characterized by viral replication cycles which become induced in suprabasal differentiating epithelial
cells. This stage is accompanied by the occurrence of visible low-grade lesions. Here, the viral
proteins are expressed - with E6 and E7 under transcriptional control -, viral DNA is synthesized and
finally viral particles are released. In case of persistent HPV infection the lesions may progress
towards the transforming infection stage (step 3) which is accompanied by a fundamental shift in the
viral gene expression pattern with E6 and E7 oncogene overexpression as described in chapter 1.3.2.
1. Introduction
13
This shift is the key event for the development of high-grade precancerous lesions and cancers with
critical host cell pathways being reprogrammed to overcome the cell cycle arrest of fully differentiated
cells and promote the proliferation of these keratinocytes.
With regard to the histomorphological classification precancerous stages are graded from CIN1 lesions
to CIN2 and CIN3 depending on the severity and the extent of the affected epithelium as described in
chapter 1.3.5. Thereby CIN1 overlaps with both biological categories with the majority of them
representing permissive infections and a smaller proportion of them being in the transforming
infection stage.
As the shift to E6 and E7 oncogene overexpression is accompanied by the induction of p16 INK4a
overexpression (chapter 1.3.4) in the affected cells p16INK4a is a surrogate marker for transforming
processes ongoing under the influence of the viral oncoproteins that interfere with the host cell
replication machinery and tumor suppressors. A strong p16INK4a expression can be detected in around
40 % of low-grade lesions, the vast majority of high-grade lesions and virtually all cervical cancers
(DARRAGH et al., 2013; TSOUMPOU et al., 2009). As it is consistently expressed along with viral
proteins E6 and E7 it represents an interesting target for immune therapies.
Although virtually all cervical cancers can causally be linked to human papillomaviruses, a woman
who is infected does not inevitably develop a precancerous lesion and cancer. Most of the HPV
infections are cleared spontaneously within several months. Only a long-lasting persistent HPV
infection may lead to the development of precancers which is a rare event. The natural history of HPV
infections and the development of premalignancy and eventually cancer can be considered a dynamic
process in opposite directions, progression and regression. With more than 90% of HPV infections
being cleared within 2 years only a small percentage of women originally infected will develop a
precancerous lesion (SCHIFFMAN and WENTZENSEN, 2010). And also established CIN lesions
show a clinically heterogeneous behavior: Here again, only a part of them will further progress
towards higher lesion grades and the majority of them (60% of CIN 1 and 40% of CIN2) will regress
in dependence of the host’s immune surveillance capacities (MCCREDIE et al., 2008; OSTOR, 1993).
Even CIN3 lesions with extensive morphological abnormalities are reported to regress to a certain
extent (MUNK et al., 2007).
There are sporadic reports on a higher progression risk of p16 INK4a-positive CIN1 compared to the
p16INK4a-negative ones (WANG et al., 2004). Nevertheless high-grade lesions, although they are all in
the transforming infection stage and thus all show p16INK4a overexpression, do not all progress. This
demonstrates that p16INK4a which is a reliable surrogate for oncoprotein activity and the biological
infection stage cannot predict progression of a lesion and other markers are necessary for the
development of a prognostically relevant tool.
The prevalence of HPV-infection is highest in young women aged 15-25, and the mean age of women
diagnosed with high-grade cervical lesions is approximately 28 years, while invasive cervical cancer is
established much later in women aged approximately 50 years at time of diagnosis (DOORBAR,
2006). Only a part of the precancerous lesions persist, progress and grow further out over time to
bigger lesions and to higher precancerous stages and finally only the minority of individuals that have
acquired a HPV-infection during their life time develop cancer. This demonstrates that cervical
carcinogenesis as very slow process. Considering these clinical characteristics one might speculate that
14
1. Introduction
accumulation of genetic changes of the host in combination with predisposing factors might be
decisive for whether a tumor develops or not.
The progression towards cancer could biologically be explained by accumulation of specific cellular
and chromosomal changes and subsequent outgrowth of distinct cell clones. It has been demonstrated
that the viral oncoprotein overexpression affects the integrity of the host genome in different ways:
Both oncoproteins E6 and E7 are able to cause major numeral and structural chromosomal aberrations
and also DNA damages (DUENSING and MUNGER, 2002). These changes are caused by disruption
of the centrosome duplication control mechanisms and the simultaneous induction of multiple spindle
poles (KORZENIEWSKI et al., 2011). The resulting mis-segration of chromosomes and aneuploidy of
the host cells contribute to further genomic aberrations. Although deadly for most of the cells raising
during such process this might generate some cells with growth advantage and lead to the outgrowth
of these cell clones (DUENSING and MUNGER, 2002; KORZENIEWSKI et al., 2011).
FIGURE 1.8
CERVICAL CANCER PROGRESSION MODEL. Cervical carcinogenesis is a characterized by
successive biological infection. It is, however, a dynamic processes as lesions also can undergo
regression. Persistent transforming infection is accompanied by accumulation of secondary genomic
alterations that might provide distinct cells with growth advantage. Selection for these cell clones and
expansion leads to tumor growth and invasive disease. The transforming infection stage is highlighted by
p16INK4a overexpression. Histomorphological classification represents a two-tiered system that does not
match the biological infection stages totally as transforming infections can be observed in a proportion of
low-grade lesions already. Adapted from (DOEBERITZ and VINOKUROVA, 2009a).
Studies that are based on comparative genomic hybridization report on different genomic changes in
cervical squamous cell carcinoma such as gains at chromosome 3q, losses at 3p and losses at 11q with
the aberrations mainly located at the terminal chromosomal regions. In cervical precancerous stages
there are also chromosomal aberrations, with increasing frequency from low-grade CIN to high-grade
1. Introduction
15
lesions and finally cancer. The same gains and losses affecting 3p, 3q and 11q of cervical squamous
cell carcinoma are already present in high-grade lesions, however at a lower frequency. Genomic copy
number alterations have substantial effects on gene dosage that may involve overexpression of
oncogenes on the one side and decreased expression of tumor suppressor genes on the other side. As
the above described changes of different chromosomal regions are all present in SCC one might
speculate that these aberrations provide growth advantage for tumor cells and are selected during
cervical carcinogenesis (reviewed in THOMAS et al., 2013).
The chromosomal instability of keratinocytes is likely to induce secondary genomic alterations that
may give rise to cells having distinct features providing them with growth advantage (BECKMAN and
LOEB, 2005). During continued persistence these cells are further selected by evolutionary
mechanisms leading to clonal expansion of cells that are adapted best to the host’s immunologic
environment.
Still, the causal relationship between integration of the viral genome and induction of chromosomal
instability leading to further genomic alterations is discussed controversially. Different
hypotheses/concepts regarding the chronology of the events may be discussed: On the one hand
integration of the viral DNA is hypothesized to be the first event leading to genomic rearrangement
and for this reason is responsible for chromosomal instability and aneuploidisation (HOPMAN et al.,
2006; PETER et al., 2010; PETT et al., 2004). On the other hand genomic instability is considered to
be an early event and rather prepares the integration of the viral DNA by creating fragile chromosomal
sites (DUENSING and MUNGER, 2004; MELSHEIMER et al., 2004, reviewed in WENTZENSEN et
al., 2004). Nonetheless, cervical carcinogenesis can be seen as a multi-step process including
deregulation of viral protein expression and breakthrough of the host’s cell cycle machinery finally
leading to chromosomal instability, accumulation of DNA damages and secondary (epi)genetic
alterations that altogether favor the outgrowth of cancer cells (SNIJDERS et al., 2006).
Whatever the chronological succession is, chromosomal instability seems to be the decisive event for
the onset of malignant processes and the transition from precancerous lesions to invasive disease
(BIGNOLD, 2002, 2003). The resulting destabilizing effects of aneuploidy in terms of chromosome
synthesis, segregation and repair during mitosis lead to further secondary genomic changes giving rise
to a huge number of cells provided with different characteristics (reviewed in DUESBERG et al.,
2011). Those that are best adapted to the environment of their host, especially the immunologic
environment, will survive and undergo clonal expansion and thus promote carcinogenesis.
An effective immune response is considered to be crucial for the clearance of infections and for
regression of established lesions. Considering the complexity of the immunobiology of HPV
infections (described in section 1.4.1) and the multitude of mechanisms developed by the virus to
circumvent the host’s immune attack (see “immune evasion”, section 1.4.2) it appears that secondary
genomic alterations are likely to affect mechanisms that contribute to immune tolerance or
immunosuppression and that enables the virus to remain undetected. Another frequently observed
genomic loss (LOH) is that at 6p21.3 locus which harbors the genes for HLA class I antigens
(CHATTERJEE et al., 2001; KERSEMAEKERS et al., 1999). This results in MHC-class I downregulation which substantially contributes to impaired antigen presentation and recognition by immune
cells that could eliminate HPV infected or precancerous cells, such as cytotoxic T cells (CTLs). Other
secondary genomic changes that alter the immunological features also might contribute to survival and
growth advantages of cells which undergo clonal selection to finally grow out to precancerous lesions
16
1. Introduction
and invasive cancers. The HPV-transformed cells within these lesions and tumors are able to
circumvent the host’s immune attack by different mechanisms and therefore constitute the “immune
evasion phenotype”.
Considering the fact that all projects of this thesis are centered on questions of immunology, the
following chapter will address different immunologic aspects in general and in particular related to
HPV-infections.
1.4
1.4.1
The immunobiology of HPV infections
The role of the host’s immune system in the defense against HPV
The host’s immune system plays a crucial role in whether a HPV infection persists or is cleared and
whether a developing lesion regresses spontaneously or persists and finally leads to invasive cancer. It
has been demonstrated that immune deficiency and immunosuppression of allograft transplanted
patients increase the risk for persistence of the precancerous disease and development of cancer
(DENNY et al., 2012; PALEFSKY, 2009). In contrast, cytotoxic T lymphocytes (CTLs) are associated
tumor control and a decreased risk for cancer (MATSUI et al., 1999). The presence or absence of
distinct immune cell phenotypes in the lesion and the surrounding tissue is considered to be highly
important for the prediction of the clinical outcome of the patients and should be considered in the
treatment plans as a prognostic parameter.
The following sections will address the different arms – innate and adaptive - of the immune system
and their role in HPV-related diseases.
INNATE IMMUNITY AND HPV
The innate responses represent the first line defense against invading pathogens and comprise
mechanisms that, in contrast to adaptive immune responses, act independently from antigen
specificity. Activation of the innate immune system leads to an immediate reaction without
establishing however an immunologic memory of the encountered pathogens (MOGENSEN, 2009).
Cells of the innate immune system recognize highly conserved molecular patterns that are shared by
many different pathogens leading to a general activation of the immune system. Here, Toll-like
receptors (TLRs) play an important role (HEINE and LIEN, 2003) which will be explained in more
detail in section 1.4.x. Cells of the innate immune system comprise dendritic cells (DCs) and
Langerhans cells (LCs) which are professional antigen-presenting cells (APCs), and also
macrophages, natural killer (NK) cells and natural killer T (NKT) cells. They release proinflammatory cytokines and thereby substantially change the immune milieu of the infection site by
attracting further innate immune cells and, in a second step, by induction of the adaptive immune
response (JANEWAY and MEDZHITOV, 2002). Antigen-processing and cross-presentation by DCs
and LCs enable T lymphocytes to get activated and to perform their tasks as cells of the adaptive
immune system (reviewed in AMADOR-MOLINA et al., 2013).
NK cells are an important cell type mediating innate immunity as they are able to recognize abnormal
cells, for example by aberrant Human Leukocyte Antigen (HLA) class I molecule expression. NK cell-
1. Introduction
17
mediated cytotoxicity which is induced upon stimulation of activating NK cell receptors such as
NKp30, NKp46 and NKG2D then eliminate these virally infected or precancerous cells (reviewed in
AMADOR-MOLINA et al., 2013).
In the setting of established HPV-associated precancers and cancers however, many innate immunity
mechanisms are impaired, such as cytokine release, antigen-presentation by LCs and type I interferon
(IFN)-responses favoring the persistence and progression of the lesions and carcinogenesis, that will
be explained in more detail in section 1.4.2 (reviewed in STANLEY, 2008).
HUMORAL IMMUNE RESPONSE TO HPV
The viral protein that most potently induces antibody responses in patients with underlying HPVinfections is the late protein L1. However, titers of neutralizing antibodies remain relatively low in
naturally occurring HPV infections. This might be due to mechanisms developed by the virus to evade
recognition and elimination by the host’s immune system that will be discussed in chapter 1.4.2. These
immune evasion strategies prevent the induction of a strong immune response (STANLEY, 2008).
HPV infections studied in animal models revealed that even low antibody titers provided protection
against subsequent HPV infections. The observation that the protective characteristics of these sera
could be transferred to other individuals gave rise to the development of the currently used
prophylactic vaccines that use L1 of different HPV-types as an immunogen in a bivalent (Cervarix®,
HPV16 and 18) and a quadrivalent formulation (Gardasil®, HPV 6, 11, 16 and 18) (reviewed in
STANLEY, 2006).
T CELL MEDIATED IMMUNE RESPONSE TO HPV
T lymphocytes can - independently of antigen-specificity - be quantified as different T cell subtypes in
the epithelium where the lesion or the tumor is located and the adjacent stromal compartment which is
generally characterized by higher densities of immune cells (GUL et al., 2004; SHAH et al., 2011). T
cell phenotypes that could be relevant in the course of HPV infection and development of
precancerous lesions belong to different arms of the immune system and might contribute to either an
effectively mounted response against the infected keratinocytes or to immune suppression and T cell
anergy and thus to disease progression (GARCIA-CHACON et al., 2009; PATEL and
CHIPLUNKAR, 2009). Immune markers that provide information about the quality of the immune
cell composition in the locally confined regions around the lesions include different T cell markers:
CD3+ T cells generally are quantified in order to obtain information about the total T cell numbers
present in the affected tissue. To define the proportions of different specialized T cell subpopulations
the total infiltrate consists of, other markers might be interesting: CD4 is generally used to characterize
T helper cells whereas CD8 identifies the presence of cytotoxic T lymphocytes (CTLs) that can further
be characterized by Granzyme B (GranB) which is typically expressed in activated, granzymeproducing CTLs (BONTKES et al., 1997; NEDERGAARD et al., 2007). Forkhead box transcription
factor 3 (Foxp3) however is a marker for regulator T cells (Treg cells) and rather indicates the
activation of the opposite, immunosuppressive arm of the immune response (WU et al., 2011). CD3 ζchain is a dimeric signal transducing molecule of the T cell receptor (TCR) that is responsible for the
activation of T cells following binding to and recognition of HLA-bound antigens. It therefore is
18
1. Introduction
considered to be a parameter for the successful T lymphocyte activation and effectiveness of the T
cell-mediated immune response (WHITESIDE, 2004).
With regard to the first approach, previous cross-sectional studies evaluating the densities and
phenotypes of tissue infiltrating T cells in general report on elevated numbers of different T
lymphocyte subtypes, such as CD3+ T cells, cytotoxic CD8+ CTLs and also Treg infiltration, with
increasing histomorphologically defined CIN stages (EDWARDS et al., 1995; NEDERGAARD et al.,
2007) (ADURTHI et al., 2008; WU et al., 2011). However, only the minority of CTLs in CIN is
reported to be in the activated, GranB-expressing state (BONTKES et al., 1997). Single studies also
report the inverse correlation between global T cell infiltration and histomorphologically defined
stages (SILVA et al., 2010). Regulatory T cells are considered to exert immunosuppressive functions
in the microenvironment and are reported to be increased in high-grade lesions and invasive cancer
and might contribute to the progression of the lesions towards cancer (ADURTHI et al., 2008;
JAAFAR et al., 2009; NAKAMURA et al., 2007; WU et al., 2011) and (reviewed in PATEL and
CHIPLUNKAR, 2009). Down-regulation of CD3 ζ-chain expression has been demonstrated in
different tumor entities such as melanoma (DWORACKI et al., 2001), head and neck cancers (KUSS
et al., 1999) and also colorectal carcinomas (NAKAGOMI et al., 1993) which emphasizes
(NAKAGOMI et al., 1993)its relevance for disease progression. However, the data published for
cervical cancer and its precancerous lesions are way scarcer (ZEHBE et al., 2002). Interestingly, CD3
ζ-chain expression was found to be down-regulated in cervical carcinoma patients compared to healthy
controls but not in precancerous lesions.
These previous studies on T cell infiltration reported on varying cell densities and phenotypes in
correlation with histomorphological disease stages without taking however into consideration the
underlying biological infection status defined as non-transforming and transforming infections.
Therefore, it has been remained unclear until now whether these changes are induced with beginning
transforming infections in low-grade lesions or rather occur later in well-established high-grade
lesions that have accumulated secondary genomic alterations and clonal selection.
The density and phenotype of tissue infiltrating T cells has been reported to correlate with outcome in
various cancer types (GALON et al., 2006) and it is conceivable that the quantity and the quality of
immune cells in CIN is of prognostic importance.
With regard to the characterization of antigen-specific T cells in HPV-related cervical precancers and
cancers data are published for peripheral as well as for tumor-infiltrating T cells. T cells specific for
HPV-antigens are rare, however several studies demonstrated that at low frequencies they exist.
The immunogenicity of the viral Ll proteins that is able to induce a humoral immune response is also
reported to contribute to proliferative T cell responses. Both T cell subtypes, CD4+ and CD8+ T cells
were shown to contribute to the cellular immunity (PASSMORE et al., 2002). However, CD4+ T
lymphocytes were characterized by a higher IFN-γ release upon stimulation with L1-peptides
demonstrating that the antigen is able to induce a T helper cell type 1 (Th1) memory response that
enhances the mechanisms of the cell-mediated immunity (SHEPHERD et al., 1996). Interestingly, T
cell responses predominated in patients who had cleared the HPV infection or resolved precancerous
lesions and one could speculate that CD4+ memory T cell responses are established during the battle
against HPV infections and also provide long term protection (CHAN et al., 2011). Conversely,
patients with advanced persisting or progressing precancerous lesions lack pro-inflammatory cytokines
1. Introduction
19
(DE VOS VAN STEENWIJK et al., 2008) and although isolated from tumors and lymph nodes of
cervical cancer patients CD4+ and CD8+ T cells show only low levels of IFN-γ release upon
stimulation with HPV antigens (DE VOS VAN STEENWIJK et al., 2010).
This again is indicative for the inability of these patients to mount an effective T cell-mediated
immune response leading to functionally inactive or even immunosuppressive T cell fractions
invading the tumors and circulating in the peripheral blood. Factors that might contribute to the
impairment of the host’s cellular immune response and the inability of these patients to clear the
infection and thus prevent malignancies will be highlighted in the following section (1.4.2).
The published data imply that antigen-specific cellular immune responses is associated a successful
establishment of immune response against HPV and clearance of the infection and thus occurred in
patients whose immune system successfully raised an immune response against HPV. Also the lack of
IFN-γ secretion, the cytokine directing cellular cytotoxic responses, suggests that immunosuppressive
mechanisms inhibit the raise of an effective cell-mediated immune response and immune attack of the
lesions and tumors. This might be an interesting starting-point for therapeutic interventions.
1.4.2
Immune evasion strategies developed by human papillomaviruses
Human papillomaviruses have developed different immune escape mechanisms allowing them not
only to modulate the host’s immune response that might be raised during infection and establishment
of precancerous lesions but also to avoid to be recognized and to remain largely invisible for the host’s
defense mechanisms. The viral life cycle regarding the site of infection and also the viral replication
within the epithelium is perfectly adapted to assure immune ignorance (reviewed in KANODIA et al.,
2007 and STANLEY, 2012a).
In general, the infection site is located in the basal cell layer of the stratified squamous epithelium and
occurs via microlesions in the tissue giving the virus access to the basal stem cells. During permissive
infection the viral genome is maintained in that cell layer at low copy numbers which minimizes the
antigen exposure for immunocytes that could potentially invade the epithelium from the stromal tissue
beneath. The expression level of the viral proteins, especially the early proteins, remains very low
during replication (Figure 1.9). The protein considered to be most immunogenic, the viral capsid
protein L1, is not expressed in the basal, undifferentiated layers of the squamous cell epithelium. As
only in the upper layers of the squamous epithelium, which circulating immune cells hardly have
access to, the viral proteins become expressed at higher levels, and an effective immune response
cannot be initiated. The low expression level of HPV proteins in epithelial layers that are closer to
immune cells thereby represents an effective immune evasion strategy (FELLER et al., 2010).
Furthermore, the assembled virions are released during naturally occurring cell death of differentiated
keratinocytes arriving at the upper side of the epithelium and not due to cytolysis or necrosis of the
infected host cells. Again, this strategy prevents stromal or epithelial immune cells from having direct
contact with the infectious agent and also inhibits inflammation at the infection site. The non-lytic life
cycle of HPV does not provoke release of pro-inflammatory cytokines and supports the virus’
invisibility to the host’s immune defense (STANLEY, 2006). Altogether the viral replication cycle is
perfectly adapted to the differentiation of the host’s keratinocytes with the aim to reduce antigen
exposure and pro-inflammatory mechanisms and thereby to avoid an effective immune response to be
20
1. Introduction
raised. At a later time point, the interaction of the viral proteins E6 and E7 with host cell pathways
regulating apoptosis and cell cycle progression contribute to the survival of the virus within
keratinocytes by preventing apoptosis and delaying the differentiation program of keratinocytes that
consequently remain in the proliferative phase (chapter 1.2).
FIGURE 1.9
HPV EVADES THE HOST’S IMMUNE SYSTEM. The virus’ life cycle (left) is adapted to the host
cells’ differentiation program and thus prevents immune recognition: there is no blood-born phase, low
viral protein levels in the basal cell layers, no cell death for viral release. Furthermore, HPV actively
down-regulates inflammatory processes (right) preventing thus the activation of innate and adaptive
immune responses by different mechanisms. Adapted from (STANLEY, 2012a).
Also the fact that the viral genome is not optimized for the mammalian translation machinery and
shows a different codon usage that results in decreased translation rates of the viral proteins might
explain the low expression profile of viral antigens (reviewed in ZHAO and CHEN, 2011).
Furthermore, a mechanism that can be subsumed under the keyword “molecular mimicry” contributes
to immune evasion by hindering recognition by and reactivity of immune cells (OLDSTONE, 1998). It
has been demonstrated that human papillomavirus proteins display similar epitopes compared with the
host cell proteome that leads to less effective immune response due to self-tolerance mechanisms of
the host’s cells (NATALE et al., 2000).
The immune evasion strategy of HPV does not only include adaptions to remain invisible for the
host’s immune system but also includes strategies evolved to actively counteract immune responses
that are raised by the host.
Immune tolerance and immune suppression per se are important and helpful mechanisms in the
regulation of the different arms of the immune system and balance the activation and termination of
immune attacks. Mediators of immune suppression such as regulatory T cells are inherent to the
immune defense and have evolved to prevent damages by excessive cytotoxic responses especially if
they are directed against self-antigens. Adopted by tumors, however, immune tolerance or suppression
are mechanisms that enable cancer cells to evade the host’s immune attack. They become able to
modulate the tumor environment in order to create an immunotolerant micromilieu and thereby to
immobilize the host’s immune responses.
1. Introduction
21
Chemokines and cytokines are signaling molecules and the key players in the regulation of a complex
immunologic network of activating and inhibitory immune responses. They are decisive for the
immune cell types attracted to the lesion and the outcome of the immune response. Immune evasion
strategies developed by tumors that involve aberrant cytokine or chemokine secretion do not only have
singular but rather systemic effects on different immunological pathways. The most immediate effect
of HPV infections is the down-regulation of type I IFN responses inhibiting antiviral innate immune
defense mechanisms and also the induction of a secondary adaptive response (Figure 1.9). Type I IFN,
especially IFN-α, is normally produced by infected cells, has anti-viral effects and also recruits
neutrophils, macrophages, NK cells and DCs to the infection site (BASLER and GARCIA-SASTRE,
2002). Its activation is necessary to induce both the innate and the adaptive immune response. The
viral oncoproteins have been shown to interact with IFN signaling pathways that normally lead to
transcriptional induction of IFN downstream target genes necessary for anti-viral defense, induction of
immune response and cell growth regulation. Both E6 and E7 directly interfere with IFN downstream
targets (IFN response genes, nuclear factor-kappa B (NFkB)) and signaling pathways to prevent IFNmediated immune responses. They inhibit among others the transcription of transporter associated
with antigen-processing 1 (TAP-1), IFN-β and monocyte-chemoattractant-protein-1 (MCP-1) and
interfere with the Jak-STAT-pathway that upon activation regulates DNA transcription and is also
involved regulating the activity of immune cells (reviewed in STANLEY, 2008). Further chemokines,
such as interleukin (IL)-8, and cytokines (IL-18, IFN-γ) are suppressed in HPV-infected cells.
Normally involved in the onset of the inflammatory responses and attracting different sorts of immune
cells such as monocytes, memory T cells, NK cell or being involved in the priming of CD8+ T cells
the down-regulation of these molecules favors the persistence of the viral infection and the
development of precancerous lesions (reviewed in STANLEY, 2012a). Changes of the polarity of the
Th1/Th2 cytokine profile to a pronounced Th2 response has also been demonstrated to have
immunosuppressive effects and to result in impaired cytotoxic immune responses (BAIS et al., 2005).
Such a change is accompanied by reversal of the immune cell composition in precancerous stages and
cancers as demonstrated by immunohistochemical analyses and also flow cytometry data (ADURTHI
et al., 2012; SHAH et al., 2011).
HPV infection also has effects on antigen presentation via HLA class I molecules. It has been shown
several times independently that the viral early proteins E7 and E5 are associated with HLA class I
antigen expression on keratinocytes and this impairs recognition of the infected cells by CD8+ T
lymphocytes and the induction of a cytotoxic response (BOTTLEY et al., 2008; CAMPO et al., 2010).
Theoretically, down-regulation of HLA class I molecules on the cell surface increases the
susceptibility to be killed by NK cells. However, as the immunosuppressive cytokine IL-10 is also
associated with HLA class I down-regulation (RODRIGUEZ et al., 2012), it is likely that the
keratinocytes evade a possible NK cell-mediated immune attack due to the generally
immunosuppressive micromilieu probably disturbing the recruitment of immune cells to the lesion. It
could be demonstrated that E5 and the oncoprotein E7 both directly regulate HLA class I expression
levels (reviewed in KANODIA et al., 2007). The interaction of E7 with the promoter of the HLA class
I heavy chain gene has repressive effects on the transcription and leads to down-regulation of HLA
class I antigen levels (GEORGOPOULOS et al., 2000). Viral E5 in contrast affects the stability and
the transport of HLA class I complexes loaded with peptides that both depend on an acid pH: via
interaction with the H(+)-ATPases (V-ATPase) it inhibits the acidification of endosomes and the Golgi
22
1. Introduction
complex and thus massively disturbs peptide-HLA-complex trafficking to the cell membrane
(ASHRAFI et al., 2005; SCHAPIRO et al., 2000).
Down-regulation of type I IFN production is associated with lacking antiviral innate immune defense
mechanisms and consequently inhibition of a secondary adaptive response. It was shown that the
initiation of any of these responses by antigen-presenting cells (APCs) is defective because
Langerhans cells, the specialized epithelial APCs, are decreased in number and are not activated
during HPV-infections and uptake of L1 antigen, leading to inhibition of both innate and adaptive
immune response (STANLEY, 2008). Different mechanisms are discussed to contribute to this lacking
LC activation: (1) E-cadherin expression in keratinocytes which is required for APCs to migrate
through the epithelium is down-regulated under the influence of DNA methyltransferase 1 activity
(Dnmt1) (LAURSON et al., 2010). (2) The inhibition of a HPV-specific immune response in the
epithelium is caused by activation of phosphoinositide-3-kinase pathway (FAUSCH et al., 2005).
(3) Immunosuppressive cytokines (transforming growth factor (TGF)-β, Fas-ligand) may be released,
that among other effects are responsible for the recruitment of regulatory T cells (Treg cells) that again
change the cytokine milieu in the lesions by releasing TGF-β and IL-10. This release inhibits thereby
the functional activity of CTLs and in the long run favor a deficient recognition of and cytotoxicity
against HPV-infected and transformed cells and promote the outgrowth of the lesions and progression
towards cancer.
Altogether, these factors inhibit the influx of immune cells into the epithelial compartment, impair the
migration and thus lead to a decreased likelihood that immune cells detect HPV and initiate an
effective immune response.
Importantly, the rates of HPV clearance and lesion regression proof that in the majority of the patients
the immune system is able to combat the disease and that the quality of the immune system, either
humoral (mainly against L1) or cell-mediated (against late or early viral proteins), is decisive in the
natural course and also treatment of cervical intraepithelial neoplasia and cancers. Immunotherapeutic
inventions therefore should aim at the activation of a strong, T cell-based immune response that may
induce destruction of HPV infected keratinocytes by CTLs either by recognition of tumor antigens or
tumor-associated antigens.
Also, the composition of the T cell infiltrates and the cytokine profile in the microenvironment may be
indicative for the clinical course of the disease and the patient’s outcome and therefore represent
potential markers for progression or regression. They also contain information that could potentially
be considered for treatment decisions in precancers in order to minimize unnecessary surgical
interventions in patients that - from their immune status - are likely to overcome the disease.
1. Introduction
1.5
1.5.1
23
Immunologic intervention strategies in HPV-associated
diseases
The urgent need for therapeutic interventions in HPV-associated
precancers and cancers
The prophylactic L1 VLP based vaccines Cervarix and Gardasil aim at the induction of a systemic
immune response and are based on the production of neutralizing antibodies floating the whole body.
The represent, however, HPV type-specific approaches and are considered - aside from sporadic
reports on cross-protection - to offer protection against only two high-risk HPV types, HPV16 and 18,
(reviewed in KAWANA et al., 2009). Also, they do not have any known advantageous effects on
preexisting HPV infections or established lesions.
A recently developed mathematical model can be used to estimate the impact of the prophylactic
vaccines on the development of the incidence of HPV-associated cervical precancers and cancers and
also anogenital warts. This model is based on epidemiological information of the natural history of
HPV infections, the frequency and natural history of HPV-infections and resulting precancerous
lesions and considers also cervical cancer screening program implemented in Germany. This model
predicts that with a vaccination coverage of about 50% only, which reflects the actual situation in
Germany, over the next 100 years about 22% of cervical intraepithelial neoplasia and 37% of cervical
cancers will be prevented by the available prophylactic vaccines (HORN et al., 2013).
These data demonstrate that the situation within the next 20-30 years will not substantially change and
that cervical precancerous lesions and cancers as well as anogenital warts still are a major health
problem to resolve. Therefore, there is a non-negligible need for secondary vaccination strategies or
other approaches enabling the immune system to recognize and eliminate HPV-infected and
transformed cells. In contrast to primary vaccines, secondary therapeutic intervention strategies aim at
establishing an effective cellular immunity and especially enhancing the T cell responses to antigens
expressed by HPV-infected cells that additionally might undergo transforming processes
(BRINKMAN et al., 2007). A multitude of secondary vaccination approaches is under investigation in
preclinical trials and some are investigated in clinical trials (reviewed in ALBERS and KAUFMANN,
2009 and KANODIA et al., 2008).
It has been reported that the majority of the therapeutic vaccination trials only sporadically induce an
efficacious clinical response accompanied by cytotoxic cellular responses that might be able to
overcome the viral evasion mechanisms and subvert immune suppression. Some other approaches led
to promising results in patients with vulvar precancerous lesions but rarely in cervical intraepithelial
neoplasia. This might be due to special features of the mucosal immunity in comparison to the
epithelial immune reactions in VIN that hinders the induction of cellular response upon systemic
vaccine administration (reviewed in KAWANA et al., 2012).
For these reasons other approaches, alone or in combination with the above described therapeutic
vaccines represent an interesting option to enhance the host immune reaction, also at mucosal sites in
order to elicit a strong T cell mediated immune response. One of these treatment strategies involves
immuno-modulatory agents that modify the quality of the immune response and reverse the
immunosuppressive environment. The following chapter is dedicated to Toll-like receptors that can be
targeted by specific immuno-stimulatory compounds. Once they are activated by these compounds
24
1. Introduction
they are able to link innate and adaptive immune responses to finally induce a strong anti-viral and
anti-tumoral cellular response.
1.5.2
Toll-like receptors are key players in linking the innate and adaptive
immune responses
Discovered in the 1990s and since then subject to many functional studies Toll-like receptors were
identified as main molecules of the innate immune system and major players of the first wall
encountered by pathogens that enter the body. TLRs can mainly be found in immune cells of the
innate immune system including DCs, monocytes and mast cells. They can, however, occasionally
also be found in T and B lymphocytes as well as in NK cells (CRAIN et al., 2013). Their expression
also was demonstrated in cells of the endothelium and epithelium and in a subset of tumor cells
(HOLLDACK, 2014).
FIGURE 1.10
TLRs LINK INNATE WITH ADAPTIVE IMMUNE RESPONSES. Upon recognition of pathogenassociated molecule patterns (PAMPs) by TLRs dendritic cells release cytokines (IL-6, IL-12) that via
complex regulatory mechanisms stimulate further innate immune cells but also adaptive immune
responses. Adapted from (STEVENSON and RILEY, 2004)
TLRs fall into the category of the so called pattern recognition receptors (PRRs) that are able to detect
microbial infections. They bind to and recognize highly conserved microbial structures, so called
pathogen-associated molecular patterns (PAMPs) that are common to a broad variety of infectious
agents and comprise molecules such as lipopolysaccharides (LPS), bacterial DNA or double-stranded
RNA (CHAN et al., 2009). Binding of PAMPs results to interaction of the TLRs with adaptor
molecules, for example with myeloid differentiation primary-response protein 88 (MyD88) (AKIRA
and TAKEDA, 2004). This initiates complex intracellular signaling pathways mediating the signal to
the nucleus where NFκB becomes activated regulating the expression of downstream target genes
including primarily pro-inflammatory cytokines such as IL-1, IL-6, IL-8, IL-12, tumor necrosis factor
(TNF)-α, IFN-α and IFN-β. Their expression further enhances the innate immunity. Immediate
1. Introduction
25
protection against pathogens is provided, in an antigen non-specific manner, by activation of NK cells,
recruitment of macrophages and activation of the complement cascade (reviewed in MEDZHITOV,
2007). The induction of the expression of co-stimulatory molecules of antigen-presenting cells such as
CD40, CD80 and CD86 that contribute to T cell activation (ZHOU et al., 2013) as well as the created
pro-inflammatory milieu that recruits further immune cells to the infection site enables an antigenspecific, adaptive immune response to be raised against the tumor (Figure 1.10) (DE GIORGI et al.,
2009). Obviously, TLRs activated in locally confined regions can also induce NK cell mediated
killing, enhance MHC class I expression on tumor cells and interfere with apoptotic pathways of the
tumor cells leading to tissue destruction and a further release of pro-inflammatory cytokines enhancing
the immune response (HOLLDACK, 2014).
1.5.3
Toll-like receptor ligands have immuno-stimulatory properties
TLR agonists represent a promising approach for the activation of the innate and adaptive immune
response by binding to and stimulation of TLRs. In the context of this thesis TLR7 is of special
interest which is an intracellular non-catalytic receptor that is located within the endosomal
compartment of immune cells (CRAIN et al., 2013). The natural ligands for TLR7 are single-stranded
RNA molecules which preferentially are rich in guanine and uridine (DIEBOLD et al., 2004).
However, they also respond to synthetic small molecules such as imidazoquinolines and molecules
that structurally resemble purine bases (Figure 1.11) (HEMMI et al., 2002) making them an interesting
target for immuno-modulatory treatment strategies.
One of these synthetic TLR ligands is the imidazoquinoline compound imiquimod that acts as an
immune modifier by changing the immune milieu by binding to TLR7 and, to a lesser extent, to TLR8
(TERLOU et al., 2010). Imiquimod is approved by the Food and Drug Administration (FDA) as a 5%
cream – and as such called Aldara® - and clinically applied for the treatment of genital warts,
superficial basal cell carcinoma and actinic keratosis (GASPARI et al., 2009). Imiquimod appears as
important non-invasive treatment option also in patients with vulvar intraepithelial neoplasia allowing
the conservation of the vulvar anatomy and is of special interest in multifocal VINs that show high
rate of recurrence. Two imiquimod-treatment studies, a pilot study and the following placebocontrolled, randomized trial, were conducted that included patients with diagnosed vulvar
precancerous lesions (grade 2 or 3). The trials demonstrated the efficiency of locally applied (topical)
imiquimod treatment that induces (at least partial) clinical response in the majority of the patients
defined as reduction of the lesion size, histologic regression and HPV clearance (VAN SETERS et al.,
2002; VAN SETERS et al., 2008).
In general, Aldara® is well-tolerated by patients if locally applied. Nevertheless, clinical studies
demonstrated that systemic and local adverse effects cannot be avoided. They appear as fever,
arthralgia, headache, myalgia or lymphadenopathy and all these symptoms are caused by proinflammatory cytokine release in the blood stream having systemic effects (CRAIN et al., 2013).
These observations induced chemists and biologist to search for new derivatives with an at least as
high immune stimulatory potential as imiquimod but reduced side effects. In this context, a TLR7specific ligand called SM360320 (Figure 1.11) was synthesized on the basis of an adenine skeleton
and pharmacologically evaluated. In a mouse model, this substance demonstrated to have an adjuvant
effect in combination with DNA vaccination (DHARMAPURI et al., 2009). It could also be shown
26
1. Introduction
that SM360320 is up to 100-fold more potent in inducing interferons compared with imiquimod
(KURIMOTO et al., 2004).
In order to further improve the effects mediated by SM360320 further derivative molecules based on
this core molecules were synthesized by conjugating it to different macromolecules such as proteins,
lipids or polyethylene glycol. These attempts gave rise to TMX-202 which is the core TLR7 agonist
conjugated to a C-12 phospholipid (Figure 1.11).
FIGURE 1.11
CHEMICAL STRUCTURES OF TLR AGONISTS. Shown are the chemical structures of the
imidazoquinoline imiquimod and the purine-like TLR7 ligand SM360320 which represents the core
molecule of TMX-202 obtained by conjugation of a C-12 phospholipid.
In a cooperation project with a company specialized in TLR agonist, Telormedix S.A., Bioggio,
Switzerland, we could gain access to the new TLR agonist TMX-202 and characterize its immunostimulatory potential on different levels and in vitro experiments.
2. Motivation and Rationale
27
2. M
OTIVATION AND RATIONALE
Infections with human papillomaviruses are very common in the sexually active population and under
certain circumstances might give rise to dysplastic abnormalities of the squamous cell epithelium – for
example in the cervix uteri where the lesions then are called cervical intraepithelial neoplasia (CIN).
The initial step of cervical carcinogenesis is the transition from permissive infection to transformation
of the HPV-infected cells, induced by expression of the oncogenes E6 and E7 that interfere with
critical host cell pathways and which can be highlighted by p16INK4a overexpression. However, the
induction of transformation is not sufficient to drive a lesion into further progression and development
of invasive cancer, as proven by the proportion of transformed high-grade lesions that are reported to
undergo regression. The deregulated host cell pathways lead to chromosomal instability and
subsequent secondary genomic aberrations. These might give rise to cell clones having distinct
features providing them with growth advantage over normal cells promoting disease progression. The
higher progression rates in immunosuppressed or immunocompromised individuals clearly
demonstrate the importance of the host’s immune system in HPV clearance and prevention from HPVrelated cancer. Clonally selected and expanding cells might have developed mechanisms that promote
immune evasion and thus disease progression (see Figure 2.1 for conceptual background). These
mechanisms might affect the immune microenvironment by creating an immunosuppressive milieu as
well as tumor cell intrinsic features enabling tumor cells to directly evade the immune attack.
FIGURE 2.1
CONCEPTUAL MODEL ILLUSTRATING CENTRAL QUESTIONS ADDRESSED IN THIS THESIS.
Immune evasion mechanisms in cervical intraepithelial neoplasia might contribute to the induction of
transforming infection and progression of cervical intraepithelial neoplasia towards invasive cancer.
Clonal selection for tumor cells favoring immune evasion leads to the outgrowth of a so called immune
evasion phenotype which might be reversed by proper immuno-modulatory drug intervention. Adapted
from (DOEBERITZ and VINOKUROVA, 2009b).
28
2. Motivation and Rationale
The objective of this thesis was to gain a better understanding of the battle of the host’s immune
system against HPV-associated precancerous lesions of different grades and also cancer. This is
crucial to improve prognostic markers for guiding treatment decisions and also therapeutic
interventions. Therefore, in this thesis the following two central goals were pursued:
In the first part of this thesis, the immune evasion phenotype of cervical precancerous lesions was
characterized to improve the understanding of the quality of possibly initiated immune responses
against HPV-induced neoplasia and evaluate how successful the immune system battles against
dysplastic cells. Possible immune evasion or suppression mechanisms were investigated and their
occurrence was correlated with different time points of the disease progression. Thereby features
observed on the immune cells’ side as well as mechanisms inherent to HPV-infected keratinocytes
were investigated.
FIGURE 2.2
GRAPHICAL OVERVIEW OF THE AIMS AND THE WORKFLOW OF THE PRESENTED THESIS.
Shown are the main questions of each part (dark blue), the different aspects investigated in this context
(grey) and the central methodological approaches (light blue).
The second part of this work aimed at the development of non-invasive therapeutic strategies to
circumvent possible immune evasion mechanisms in HPV-associated diseases. Different approaches
aiming at immune-modulation of potentially suppressed immune responses were evaluated that
possibly influence the balance between immune evasion and anti-tumoral immune response and thus
might lead to an effect tumor attack.
All these aspects of the interaction of the host’s immune system with lesion or tumor cells in HPVassociated diseases were investigated by a broad spectrum of different approaches to address specific
problems (outlined in Figure 2.2).
2. Motivation and Rationale


29
Objective and standardized quantification methods for immune cell infiltrate in HPV-associated
precancerous lesions of the cervix are lacking.
Development of a computational tissue analysis platform combining histological analysis and
automated whole slide imaging allowing the objective quantification of tissue infiltrating
immunocytes. This method is required to establish an immune-based prognostic biomarker tool
for cervical intraepithelial neoplasia that would also allow the monitoring of the efficacy of
tumor therapies (chapter 4.1).


It is still unclear which mechanisms trigger the infiltration of immune cells into lesions and a
shift in immune cell densities and composition still cannot be related to any time point of the
natural history of CIN. It is of particular interest whether this correlates with the biological
infection status, e.g. with the shift from permissive to transforming infection.
Characterization of the immune cell infiltrates as CD3, CD8, GranB, Foxp3 and CD3 ζ-chain
expressing cells in an antigen-independent way in cervical precancerous lesions of different
grades and infection stages by immunohistochemistry. The main focus was lying on possible
differences between permissive and transforming infections as represented by p16 INK4a
overexpression (chapter 4.2).




HPV might interfere with antigen-presentation mechanisms and thus contribute to immune
evasion. Altered HLA expression is reported in cervical cancer, the time point when this occurs
during progression of precancerous lesions however remains to be elucidated.
Analysis of the expression of molecules involved in antigen-processing and -presentation (HLA
class I heavy chains and light chains, and HLA class II antigens) on keratinocytes by
immunohistochemistry in different progression grades of cervical precancerous lesions and
cancers (chapter 4.3).
Topical treatment with an immune-modulatory drug, imiquimod, might have effects on immune
cell densities and composition in cervical intraepithelial neoplasia. These changes might be
associated with the clinical outcome of patients.
Longitudinal characterization of the immune cell infiltrates as CD3 and CD8 expressing cells by
immunohistochemistry in cervical precancerous lesions that were topically treated with a Tolllike receptor agonist-based immune modulator (chapter 4.4).

Second generation Toll-like receptor agonists might have less side effects compared with
imiquimod, however their effects on immune cells and the potential to raise an anti-tumoral
immune response in the context of HPV-associated diseases have to be demonstrated.

Evaluation of the effects of a new immuno-modulatory drug (TMX-202) on TLR7 expression in
immune cells on the transcript and protein level by quantitative real-time PCR and western blot.
Evaluation of its efficiency to induce a pro-inflammatory cytokine milieu (as measured by IL-6
ELISA). Investigation of its potency to enhance the immune attack against HPV-associated
cancers by an in vitro priming experiment of naïve T cells and measurement of cytotoxic
responses against CaSki cells monitored by CD107a degranulation assay (chapter 5.1).
30
2. Motivation and Rationale

Models based on HPV-positive tumor cell lines and autologous immune cells for the
investigation of immune responses against HPV-associated tumors are lacking. These, however,
are of special importance to test immune-modulatory treatment options.

Generation of a HPV-associated cell line from HNSCC patients in order to establish an
autologous model for in vitro functional analyses of tumor cell and immune cell interaction.
Tumor samples could be obtained from HPV-positive head and neck squamous cell carcinomas
that develop through the same tumorigenic mechanisms as cervical cancers and therefore
represent a valuable model for HPV-related cancers (chapter 5.2)


The immunophenotypic characterization of CIN and cervical cancer revealed regulatory T
lymphocytes as one possible contributor to HPV-related carcinogenesis and possible target for
immuno-modulatory intervention strategies.
Analysis of the immunosuppressive effects mediated by regulatory T lymphocytes in HPVassociated diseases using the established autologous tumor model and evaluating Treg depletion
as one possible therapeutic intervention strategy to reverse immune evasion phenotype. The
antitumoral effects of Treg depleted T cells and the total T cell fraction was compared in
CD107a degranulation assay and impedance measurement (chapter 5.3).
3. Materials and Methods
31
3. M
ATERIALS AND METHODS
3.1
Materials
3.1.1
Technical equipment, instruments
Agarose gel carriage
Agarose gel running chamber SubCell GT
Tecnomara (Fernwald)
Biorad (Munich)
Analytical balance BP 210D
Balance BP 310S
Sartorius (Goettingen)
Sartorius (Goettingen)
Bond Autostainer
Camera Electrophoresis Docu System 120
Centrifuge 5810R
Digital camera Leica DFC480
Electrophoresis chamber (Sub Cell GT)
Leica Microsystems (Wetzlar)
Kodak (Stuttgart)
Eppendorf (Hamburg)
Leica Microsystems (Wetzlar)
Biorad (München)
Flow cytometer (FACSCalibur)
Fluorometer for microtiter plates Luminex 100
Gel Documentation GelDoc 2000
Incubator
Leica Bond Autostainer II
Magnetic stirrer MR 2002
Microscope for cell culture Olympus CK40
Microscope Leica DMRBE
Becton Dickinson (Franklin Lakes, USA)
Luminex (Austin, USA)
Biorad (München)
Memmert (Schwabach)
Leica Microsystems (Wetzlar)
Heidolph (Schwabach)
Olympus optical CO (Center Valley, USA)
Leica Microsystems (Wetzlar)
Microtiter plate reader Multiscan EX
Microtome Leica RM 2035
Microwave
Minishaker Vortex MS1
NanoZoomer 2.0-HT Scan System
PCR system (Mastercycler Gradient)
Thermo Electron Corporation (Karlsruhe)
Leica Microsystems (Wetzlar)
Panasonic (Hamburg)
IKA-Works (Wilmington, NC)
Hamamatsu (Herrsching)
Eppendorf (Hamburg)
PH meter (PB-11)
Photometer for microtiter plates GENios
Sartorius (Göttingen)
Tecan (Crailsheim)
Photometer Ultrospec 7000
Pipettes 2-1000µl Pipetman
GE Healthcare (Uppsala, Sweden)
Gilson (Bad Camberg)
Pipettor 8-5010
Power supply Power Pac 300
Real-Time PCR system (StepOnePlus)
Robocycler Gradient 96
Rolling mixer CAT RM 5
RTCA Analyzer W380
Neolab (Heidelberg)
Biorad (Munich)
Applied Biosystems (Foster City, USA)
Stratagene (Santa Clara, USA)
Neolab (Heidelberg)
ACEA Biosciences (San Diego, USA)
32
3. Materials and Methods
RTCA SP Station 1x96
Safety cabinet Class II, SL-130 Blue Series
Shaker for microtiter plates Titramax 100
ACEA Biosciences (San Diego, USA)
Kojair (Vilppula, Finland)
Heidolph (Schwalbach)
Speed Vac DNA Speed Vac 110
Table top centrifuge 5424
Savant (Holbrook, USA)
Eppendorf (Hamburg)
Thermomixer 5436
Vacuum manifold Vacusafe Comfort
Eppendorf (Hamburg)
Millipore (Billerica, USA)
Vortex (MS1 Minishaker)
Water bath Grant SUB14
IKA (Staufen)
Grant Instruments (Cambridge, UK)
3.1.2
Chemicals and Reagents
Acetic acid 100%
Merck (Darmstadt)
Acryl amide RotiphoreseGel 30
Agarose Ultra Pure
Albumine Bovine Fraction V pH 7.0
Amphotericin B
Aquatex
B2-microglobuline (human)
Roth (Karlsruhe)
Invitrogen (Carlsbad, CA, USA)
Serva (Heidelberg)
Invitrogen (Karlsruhe)
Merck (Darmstadt)
Sigma Aldrich (Steinheim)
Boric acid
Brefeldin A
Bromphenol blue
Caseine from bovine milk
Merck (Darmstadt)
Sigma-Aldrich (Steinheim)
Serva (Heidelberg)
Sigma Aldrich (Steinheim)
CellGro® DC
Citric acid
CellGenix Technologie Transfer (Freiburg)
Merck (Darmstadt)
DAB+ Substrate Chromogen System
Decitabine (5-Aza-2‘-deoxycytidine, DAC)
Developer for photographic processing
Dimethyl sulfoxide (DMSO) 99.5%
Dipotassium phosphate
Disodium hydrogenphosphate
Dako (Carpinteria, USA)
Sigma-Aldrich (Steinheim)
Adefo-Chemie (Dietzenbach)
Sigma Aldrich (Steinheim)
Gerbu (Gaiberg)
VWR (Darmstadt)
Disodium phosphate
DMEM / Ham’s-F12
DNA ladder 100 bp
Dulbeco’s PBS (1x)
VWR International (Leuven, Belgium)
Gibco (Paisley, UK)
Invitrogen (Karlsruhe)
Gibco (Paisley, UK)
Ethanol 96%
Ethanol 99%
Ethanol absolute
Ethanol absolute
Sigma Aldrich (Steinheim)
Sigma Aldrich (Steinheim)
Sigma Aldrich (Steinheim)
Sigma-Aldrich (Steinheim)
Ethylenediaminetetraacetic acid (EDTA)
Fetal bovine
Fixer for photographic processing
GelRed Nucleic Acid Stain
Merck (Darmstadt)
Gibco (Paisley, UK)
Adefo-Chemie (Dietzenbach)
Biotium (Hayward, CA)
3. Materials and Methods
Gentamycin
Glutamine
Glycerol (86%)
Invitrogen (Karlsruhe)
PAA Laboratories (Pasching, Austria)
Carl Roth (Karlsruhe)
Glycerol
Glycine
Roth (Karlsruhe)
AppliChem (Darmstadt)
GM-CSF
H2O HPLC-grade
PromoCell (Heidelberg)
VWR (Darmstadt)
Haematoxylin
Hemalaun
Sigma Aldrich (St. Louis, USA)
AppliChem (Darmstadt)
Heparin-Natrium 25000
Horse serum
Human epithelial growth factor (EGF)
Human serum
Hydrochloric acid (HCl 37%)
Hydrocortisone
Ratiopharm (Ulm)
Vector Laboratories (Burlingame, CA)
Sigma Aldrich (Steinheim)
PAA Laboratories (Pasching, Austria)
Carl Roth (Karlsruhe)
Sigma Aldrich (Steinheim)
Hydrogen peroxide 30%
IMDM
Merck (Darmstadt)
Gibco (Paisley, UK)
Imiquimod
Insuline
Isopropyl alcohol
Calbiochem (San Diego, USA)
Sigma Aldrich (Steinheim)
Sigma Aldrich (Steinheim)
Luminol Reagent for Western Blot sc-2048
Lymphocyte Separation Medium LSM 1077
Magnesium chloride (MgCl2)
Mercaptoethanol
Methanol
Paraformaldehyde (PFA)
PBS
Penicillin/Streptomycin (100x)
Ponceau S Solution 0.1%
Potassium chloride
SC Biotechnology (Santa Cruz, USA)
PAA Laboratories (Pasching, Austria)
Solis Biodyne (Tartu, Estonia)
Merck (Darmstadt)
Sigma Aldrich (Steinheim)
Carl Roth (Karlsruhe)
Gibco (Paisley, UK)
Gibco (Paisley, UK)
Sigma Aldrich (Steinheim)
Merck (Darmstadt)
Potassium dihydrogen phosphate
Precision Plus Protein Standard
Gerbu (Gaiberg)
Bio-Rad (Munich)
Protease Inhibitor Cocktail
Protein Assay Dye Reagent Concentrate
Sigma (Steinheim)
BioRad (München)
Quantum 263 Medium for Tumor Cells
Reaction buffer BD (10 x)
PAA Laboratories (Pasching, Austria)
Solis Biodyne (Tartu, Estonia)
Recombinant Human Interleukin-2
Recombinant Human Interleukin-4
Recombinant Human Interleukin-7
RIPA Buffer
PromoCell (Heidelberg)
PromoCell (Heidelberg)
PromoCell (Heidelberg)
Sigma Aldrich (Steinheim)
RNase Out
ROX size standard
RPMI 1640
Invitrogen (Karlsruhe)
Applied Biosystems (Darmstadt)
Gibco (Paisley, UK)
33
34
3. Materials and Methods
Sodium acetate
Sodium carbonate
Sodium chloride (NaCl)
Roth (Karlsruhe)
J.T. Baker (Deventer, Netherlands)
Sigma Aldrich (Steinheim)
Sodium dihydrogen phosphate dihydrate
Sodium dihydrogenphosphate monohydrate
J.T. Baker (Deventer, Netherlands)
J.T. Baker (Deventer, Netherlands)
Sodium heparin
Sodium hydrogene carbonate
Ratiopharm (Ulm)
Merck (Darmstadt)
Sodium hydroxide (NaOH)
Sodium thiosulfate
Sigma Aldrich (Steinheim)
Gerbu (Gaiberg)
Sodiumdodecylsulfate (SDS)
Supplement Insulin-Transferrin-Selenium
Tetramethylbenzidine (TMB)
Tetramethylethylenediamine (TEMED)
TMX-202
Tris-(hydroxymethyl)aminomethane
Serva (Heidelberg)
Gibco (Paisley, UK)
Sigma Aldrich (Steinheim)
Sigma Aldrich (Steinheim)
Telormedix (Bioggio, Switzerland)
Roth (Karlsruhe)
Tris-HCl
Trypan blue solution
Roth (Karlsruhe)
Sigma-Aldrich (Steinheim)
Trypsin-EDTA
Tween 20
Ultima Gold
Gibco (Paisley, UK)
BioRad (München)
PerkinElmer (Waltham, USA)
Xylene
Merck (Darmstadt)
3.1.3
Consumables
96-well plates, U-bottom
BD Falcon (Durham, NC, USA)
96-well plates, V-bottom
Cannulas Microlance 3
Cell culture plate Cellstar (6/12/96 wells)
Cell culture plate Nunclon Surface
Cell strainer (100 µm)
Columns (LS)
BD Falcon (Durham, NC, USA)
Becton Dickinson (Fraga, Spain)
Greiner Bio-One (Frickenhausen)
Nunc (Roskilde, Danmark)
BD Biosciences (Erembodegen, Belgium)
Miltenyi Biotec (Bergisch Gladbach)
Columns (MS)
E-Plate VIEW 96
Examination gloves, Sempercare® nitrile
Filter paper
Miltenyi Biotec (Bergisch Gladbach)
ACEA Biosciences (San Diego, USA)
Semperit (Vienna, Austria)
Whatman (Dassel)
Filter systems (0.22 µm)
Filter tips (10 µl, 20 µl, 200 µl, 1000 µl)
Filter tips Tip One
Freezing tubes Cryo.s with screw cap (2ml)
Corning incorporated (New York, USA)
Corning Incorporated (New York, USA)
StarLab (Ahrensburg)
Greiner Bio-One (Frickenhausen)
Laboratory film Parafilm „M“
Microscope cover glasses
Microscope slides Superfrost Plus
Microscopy Aquatex
American National Can (Greenwich)
Marienfeld (Lauda-Königshofen)
Menzel (Braunschweig)
Merck (Darmstadt)
3. Materials and Methods
35
Microtome blades R35
Microwell plate NUNC 96 flat bottom
PCR tubes
PFM Medica (Köln)
NUNC (Langenselbold)
Steinbrenner (Wiesenbach)
Petri dishes
Petri dishes (sterile)
Greiner Bio-One (Frickenhausen)
Nunc (Roskilde, Denmark)
Photographic film Kodak
Pipet tips
Sigma-Aldrich (St. Louis, USA)
Greiner Bio-One (Frickenhausen)
Polystyrene tubes
PVDF membrane Hybond N+
BD Biosciences (Erembodegen, Belgium)
Amersham (Buckinghamshire, UK)
Reaction tubes (1.5 and 2 ml)
Reaction tubes (15 ml and 50 ml)
Scalpel No. 11
Serological pipette (2-25 ml)
Serological pipette Costar Stripette
Sterile filter (0.2 µm)
Eppendorf (Hamburg)
Greiner Bio-One (Frickenhausen)
PFM Medical (Köln)
Sarstedt (Nümbrecht)
Corning (New York, USA)
Corning Incorporated (New York, USA)
Tissue culture flasks (sterile, pyrogen-free) Cellstar
Western Blotting Luminol Reagent
Greiner Bio-One (Frickenhausen)
Santa Cruz (Heidelberg)
3.1.4
Commercially available kits
CD4+CD25+ Regulatory T Cell Isolation Kit human
CINtec Plus (Cytology, Histology)
DNeasy Blood & Tissue Kit
Miltenyi Biotec (Bergisch Gladbach)
Roche (Mannheim)
Qiagen (Hilden)
dNTP Set 100 mM
Hexanucleotide mix (10 x)
Invitrogen (Karlsruhe)
Roche Diagnostics (Mannheim)
Multiplex HPV Genotyping Kit
Mycoplasma detection kit MycoAlert™
Pan T Cell isolation Kit II
Power SYBR Green PCR Master mix
QIAamp DNA FFPE Tissue Kit
Quantikine®ELISA Human IL-6
Diamex (Heidelberg)
Lonza (Köln)
Miltenyi Biotec (Bergisch Gladbach)
Applied Biosystems (Foster City, USA)
Qiagen (Hilden)
R&D Systems (Abingdon, UK)
RNeasy Mini Kit
SuperScript II Reverse Transcriptase
Vectastain Elite ABC Kit
Qiagen (Hilden)
Invitrogen (Karlsruhe)
Vector (Burlingame, USA)
36
3. Materials and Methods
3.1.5
Antibodies
Modi-
Application
fication
(dil./conc.)
E6H4
none
ready to use in
CINtec® Plus Kit
(Roche, Mannheim)
CD3
PS1
none
IHC
(1:50)
Acris antibodies (Herford)
Foxp3
236A/E7
none
IHC
(1:50)
eBioscience (Frankfurt a. M.)
GranzymeB
11F1
none
IHC
(1:50)
Novocastra (Newcastle upon
Tyne, UK)
CD8
4B11
none
IHC
(1:50)
Novocastra (Newcastle upon
Tyne, UK)
CD3-ζ
6.B10.2
none
IHC
(1:200)
Santa Cruz (Heidelberg)
HC-10
n.a.
none
IHC
(1:50)
kind gift of Soldano Ferrone
HCA-2
n.a.
none
IHC
(1:50)
kind gift of Soldano Ferrone
L368
n.a.
none
IHC
(1:50)
kind gift of Soldano Ferrone
LG-612.14
n.a.
none
IHC
1:300
kind gift of Soldano Ferrone
mouse IgG /
rabbit IgG
polyclonal
biotin
CD4
RPA-T4
FITC
CD8
RPA-T8
FITC
CD25
4E3
PE
CD107a
n.a.
PE
FACS
(1:50)
BD Pharmingen (Heidelberg)
HLA-A2
n.a.
FITC
FACS
(1:50)
AbD Serotec (Puchheim)
HLA-A/B/C
W6/32
FITC
FACS
(1:50)
eBioscience (Frankfurt a. M.)
epithelial
antigen
BerEP4
FITC
FACS
(1:50)
Dako (Eching)
isotype
control
IgG1
n.a.
FITC
Reactivity
Clone
p16INK4a
IHC
(1:50)
FACS
(1:50)
FACS
(1:50)
FACS
(1:50)
FACS
(1:50)
Supplier
Vector Laboratories
(Burlingame, USA)
BD Pharmingen (Heidelberg)
BD Pharmingen (Heidelberg)
Miltenyi Biotec (Bergisch
Gladbach)
BD Pharmingen (Heidelberg)
3. Materials and Methods
37
isotype
control
IgG1
n.a.
PE
isotype
control
IgG2a
n.a.
FITC
isotype
control
IgG2b
n.a.
PE
isotype
control
IgG2b
n.a.
PE
FACS
(1:50)
BD Pharmingen (Heidelberg)
TLR7
monoclonal
none
WB
(1:1000)
Abcam (Cambridge, UK)
HPV16 E7
NM2
none
WB
(1:500)
Santa Cruz (Heidelberg)
actin
n.a.
none
WB
(1:20000)
MP Biomedicals, Heidelberg.
rabbit IgG
n.a.
HRP
WB
(1:2000)
Promega (Mannheim)
mouse IgG
n.a.
HRP
WB
(1:4000)
GE Healthcare (Freiburg)
FACS
(1:50)
FACS
(1:50)
FACS
(1:50)
BD Pharmingen (Heidelberg)
BD Pharmingen (Heidelberg)
BD Pharmingen (Heidelberg)
n.a = not available; IHC = immunohistochemistry; WB = Western Blot
3.1.6
Enzymes
Collagenase Type IV
Sigma-Aldrich (Steinheim)
DNase I
Hyaluronidase
Super Script Reverse Transcriptase (200U/µl)
Tag DNA Polymerase (5U/µl)
Sigma-Aldrich (Steinheim)
Sigma-Aldrich (Steinheim)
Invitrogen (Karlsruhe)
Invitrogen (Karlsruhe)
3.1.7
Peptides
p16INK4a
peptide name
p16INK4a peptide R1
INK4a
Amino acid positions
51-59
amino acid sequence
VMMMGSARV
The p16
9mer peptide was synthesized by the core facility for peptide synthesis, German Cancer Research
Center, Heidelberg.
38
3. Materials and Methods
HPV16 L1
peptide name
Amino acid positions
amino acid sequence
HPV16 L1_2
2-11
SLWLPSEATV
HPV16 L1_12
12-21
YLPPVPVSKV
HPV16 L1_60
60-68
ILVPKVSGL
HPV16 L1_67
67-75
GLQYRVFRI
HPV16 L1_97
97-105
RLVWACVGV
HPV16 L1_249
249-257
YLRREQMFV
HPV16 L1_323
323-331
ICWGNQLFV
The HPV16 L1 9mer and 10mer peptides were synthesized by Genaxxon Bioscience, Ulm.
Influenza virus matrix protein
peptide name
Amino acid positions
amino acid sequence
viral MP
57-68
GILGFVFTL
The virus matrix protein was synthesized by the core facility for peptide synthesis, German Cancer Research
Center, Heidelberg.
3.1.8
Primers
primer name
sequence (5’-3’)
TLR7 forward
AAGCCCTTTCAGAAGTCCAAGTT
TLR7 reverse
GGTGAGCTTGCGGGTTTGT
β-actin forward
ATGTGGCCGAGGACTTTGATT
β-actin reverse
AGTGGGGTGGCTTTTAGGATG
The primers were obtained from Thermo Scientific, Ulm.
3.1.9
Buffers and Solutions
Agarose Gel (1.5%):
1.5 g Agarose
100 ml TBE buffer
1 µl Gel Red
10% APS:
10 % (w/v) Ammonium persulfate in aqua bidest
3. Materials and Methods
Blotting Buffer (10x):
30.37 g Tris
144.13 g Glycine
ad 1l Aqua bidest.
Blotting Buffer (1x) working solution: 100 ml 10x Blotting Buffer
200 ml methanol
700 ml Aqua bidest.
10x Citrate buffer:
100 mM Citric acid monohydrate (21 g)
ad 1l Aqua bidest.
adjust pH to 6.0 with NaOH
DNA loading buffer (6x):
25 ml Glycerol
125 mg Xylenecyanol
25 ml H2O dest.
Laemmli sample buffer (4x) :
2.5 ml 1M Tris pH 8.0 (125mM)
8 ml 10% SDS (4%)
2 ml glycerin (10%)
2 ml β-mercaptoethanol (10%)
4 mg bromphenol blue (0.02%)
ad 20 ml H2O
5 M NaOH:
100 g NaOH
ad 500 ml Aqua bidest
10x PBS:
84 g NaCl (= 0.8 % w/v)
2 g KCl (= 0.02 % w/v)
11.5 g Na2HPO4 (= 0.1 % w/v)
2 g KH2PO4 (= 0.02 % w/v)
ad 800 ml Aqua bidest, pH 7.4
4% PFA Stock Solution:
4g paraformaldehyde
100 ml PBS
10x SDS-PAGE running buffer:
30.3 g Tris-Base
144 g Glycine
100 ml 10% SDS
ad 1 l Aqua bidest.
39
40
3. Materials and Methods
10x TBE:
108 g Tris
55 g boric acid
40 ml 0.5 M EDTA, pH 8.0
ad 1 l Aqua bidest.
10x TBS:
60.55 g Tris
87.66 g NaCl
ad 1 l Aqua bidest.
adjust pH to 7.6 with 37% HCl
1x TBS:
dilute 10x TBS 1:10 in Aqua bidest
1x TBS-T for Western Blot:
dilute 10x TBS 1:10 in Aqua bidest
add 0.1% Tween
Tris 0.5 M, pH 6.8:
20.29 g Tris
20 ml 10% SDS
ad 500 ml Aqua bidest.
adjust pH to 6.8, autoclave
Tris 1.5 M, pH 8.8 :
3.1.10
90.9 g Tris
20 ml 10% SDS
ad 500 ml Aqua bidest
adjust pH to 8.8 with HCl, autoclave
Cell culture media
B cell/T cell basis medium
500 ml IMDM
50 ml human serum
6 ml L-glutamine
25 µg/ml gentamicin
Dendritic cell medium
CellGro medium
1% penicillin / streptomycin
3% human serum
Freezing medium for PBMCs
human serum
+ 10 % DMSO
3. Materials and Methods
MACS Buffer
PBS
5% human AB serum
1 mM EDTA
sterile filtrate
Peptide-Load medium
500 ml IMDM (serum free)
25 µg/ml gentamicin
Quantum Tumor Medium
Quantum 263 for Tumor cells
(used for cell line generation)
5 µg/ml insulin
0.5 µg/ml hydrocortisone
10 ng/ml hEGF
25 µg/ml gentamicin
T cell medium:
Bc/Tc basis medium
1 x Insulin Transferrin Selenium
10 U/ml IL-2
10 U/ml IL-7
Tumor digestion solution:
10 ml Tumor preparation solution
1 mg/ml Collagenase Type IV
0.1 mg/ml Hyaluronidase
20 µg/ml DNase I
Tumor preparation solution
200 ml RPMI 1640
25 mM HEPES
3.6 ml Penicillin/Streptomycin (100x)
5 µg/ml amphotericin B
2 mM Glutamin
Tumor transport solution
DMEM medium
10% FCS
100 µg/ml gentamycin
10 µg/ml amphotericin B
Tumor cell line Medium
RPMI 1640
10% FCS
25 µg/ml gentamicin
(used for standard cell lines)
41
42
3. Materials and Methods
3.1.11
Cell lines
cell line
origin, characteristics
experiment
CaSki
HPV16 positive cervical cancer cell line
killing assay (chapters 3.2.4, 5.2.4)
Raji
B cell lymphoma cell line, high TLR7 expression
Positive control for TLR7 expression
(chapters 3.2.4, 5.1.2)
T2
TAP-deficient T-B lymphoblastoid hybridoma
Peptide binding assay (3.2.4, 5.2.1)
3.1.12
Patients’ material
1) Cervical intraepithelial neoplasia and cervical cancer patients
study “immune cell infiltration in relation to p16INK4a expression” (chapter 4.2)
University Hospital, Heidelberg and Institute of
Pathology, Mannheim, Germany
period of recruitment
number of patients
resected tissue
stages
CIN1
CIN2
CIN3
invasive cervical carcinoma
2003-2004
(Mannheim)
(Heidelberg)
and
2007-2010
69
cervical cone biopsies
n = 22
n = 11
n = 19
n = 17
p16INK4a-positive [n, (%)]
13 (59.1%)
11 (100.0%)
19 (100.0%)
17 (100.0%)
2) Cervical intraepithelial neoplasia and cervical cancer patients
antigen-presentation HLA class I and HLA class II study (chapter 4.3)
Institute of Pathology, Mannheim, Germany
period of recruitment
2003-2004
number of patients
n = 41 (* n = 69)
resected tissue
cervical cone biopsies
disease stages
* CIN1
CIN2
CIN3
invasive cervical carcinoma
n = 19
p16INK4a-positive [n, (%)]
10 (52.6 %)
n = 9 (* n = 9)
n = 13
n = 19
9 (*9) 100.0%
13 (100.0%)
19 (100.0%)
* cohort enlarged by CIN1 (n=19) and an additional subset of CIN2 (n=9) for HLA class II analysis (deriving from cohort 1)
3. Materials and Methods
43
3) Cervical intraepithelial neoplasia patients:
imiquimod study (chapter 4.4)
University Hospital, Vienna, Austria
period of recruitment
2009-2010
number of patients
10
resected tissue
punch biopsies
disease stages
all patients had CIN2/3 at study entry
(inclusion criterion)
4) Oropharyngeal carcinoma patients
Generation of HPV-positive tumor cell line (chapter 5.2)
University Hospital, Giessen, Germany
period of recruitment
2011-2014
number of patients
58
resected tissue
primary tumors located in the oropharynx and
metastatic lymph nodes in the head and neck
region
HPV-status
HPV-positive
HPV-negative
n = 31
n = 27
5) healthy blood donors
TMX-202 treatment on PBMCs
Institute of Pathology, Heidelberg, Germany
period of recruitment
2013/2014
number of patients
4
Sex
male
50%
female
50%
3.1.13
Software
Adobe Acrobat Reader 6
BIMAS
CellQuest pro (5.2)
Diskus Bilddarstellung (4.30)
Endnote X6
Magellan Standard
MedCalc, version 11.5.1.0
Adobe (San Jose, CA, USA)
(PARKER et al., 1994)
Becton Dickinson (San Jose, USA)
Techn. Büro Hilgers (Königswinter)
Thomson Reuters (New York, NY, USA)
Tecan Group Ltd. (Männedorf, Switzerland)
MedCalc Software (Ostend, Belgium)
NDP.view Software
Hamamatsu (Herrsching)
44
3. Materials and Methods
RTCA software 2.0.0
SPSS Statistics 22
STATISTICA (7)
ACEA Biosiences (San Diego, USA)
IBM (Ehningen)
StatSoft (Europe) GmbH (Hamburg)
Statistica, version 8.0.3.6
StepOne (2.1)
Statsoft (Hamburg)
Applied Biosystems (Foster City, USA)
SYFPEITHI
TissuemorphDP™M
(RAMMENSEE et al., 1999)
Visiopharm, Hørsholm, Denmark
3.2
Methods
3.2.1
Immunohistochemistry for archived tissue samples
p16INK4a Immunohistochemistry
The identification and diagnosis of precancerous lesions of the cervix uteri is based on the biomarker
p16INK4a, a cyclin-dependent kinase inhibitor which is normally involved in tumor suppression. It is a
well-established biomarker for early oncogenic processes in HPV-related cancers, especially cervical
cancer as p16INK4a becomes markedly overexpressed in persistent HPV-infections in which the
oncogenic transformation is induced by HPV proteins which is the first step necessary for the
development of cervical cancer.
The CINtec® PLUS Kit was used for the qualitative detection of p16 INK4a and the
immunohistochemical staining procedure was carried out as proposed in the manufacturer’s protocol.
The provided reagents were used as described in the protocol with the following exceptions: the
substrate incubation with DAB was carried out in two consecutive steps comprising 8 minutes each.
Immunohistochemical staining for CD3-, Foxp3-, Granzyme B- and CD8-positive cells and the
antigen-presentation molecules HLA class I and HLA class II
Different T cell markers were qualitatively and quantitatively investigated by immunohistochemical
analyses. The global T cell infiltration in tissue specimens was analyzed by staining for CD3. T cell
markers representing different T cell subtypes were used to characterize T cell activation (CD8,
Granzyme B and CD3 ζ-chain) and T cell inhibition (Foxp3). Precancerous lesions and cancers were
characterized for HLA class I heavy chain (HLA-A/B/C) and light chain (beta2-microglobuline, β2m)
and HLA class II antigen expression.
Formalin-fixed paraffin-embedded tissue sections were mounted on aminopropylsilane-coated slides
and following deparaffinisation in xylene and rehydration in decreasing ethanol concentrations (100%70%) the slides were heated for 15 minutes in 10mM citrate buffer (pH=6) in order to retrieve antigen
epitopes to be analyzed. Blocking of endogenous peroxidase was performed by using 0.6% H 2O2 in
methanol. In order to reduce non-specific antibody binding and background staining the tissue sections
were then blocked with 10% horse serum in PBS before the various first antibodies were applied (for
dilutions see section “Antibodies”) and incubated at 4°C overnight. Slides were then incubated with
biotinylated anti-mouse/anti-rabbit IgG secondary antibodies for 30 minutes at room temperature.
3. Materials and Methods
45
Following the application of avidin-biotin reagent according to the manufacturer’s instructions, the
color reaction with 3,3-diaminobenzidine (DAB+ chromogen) allowed the detection of the antigens to
be characterized. Finally, the slides were counterstained with hematoxylin and mounted with Aquatex.
Automated immunohistochemical staining protocol (CD3 and CD8)
Tissue sections (2µm) were mounted on aminopropylsilane-coated slides and subject to automated
immunohistochemical staining with the Leica-Bond II Max autostainer by applying the following
staining protocol with reagents provided by Leica Biosystems:
Step (solution applied)
duration
temperature
BOND Dewax Solution
30 sec
72°C
BOND Dewax Solution
30 sec
72°C
BOND Dewax Solution
30 sec
RT
ethanol (99%) (3 repetitions)
RT
BOND wash solution (3 repetitions)
RT
BOND ER Solution 1 (citrate buffer, pH 6.0) (2x)
RT
BOND ER Solution 1 (citrate buffer, pH 6.0)
20 min
100°C
BOND ER Solution 1 (citrate buffer, pH 6.0)
12 min
RT
BOND wash solution (3 repetitions)
peroxide block
37°C
20 min
BOND wash solution (3 repetitions)
serum block (10% goat serum in PBS)
RT
30 min
BOND wash solution (3 repetitions)
primary antibody in TBS/10% FBS)
RT
RT
RT
30 min
BOND wash solution (3 repetitions)
RT
RT
post primary (polymer penetration enhancer
in TBS/10% FBS)
8 min
RT
BOND wash solution (3 repetitions)
2min
RT
polymer (secondary antibody, poly-HRP-anti-mouse/anti-rabbit IgG)
8 min
RT
BOND wash solution (2 repetitions)
2min
RT
deionized water
RT
mixed DAB Refine
???
RT
mixed DAB Refine
10 min
RT
deionized Water (3x)
RT
46
3. Materials and Methods
hematoxylin counterstaining
5 min
deionized Water (3x)
BOND wash solution
RT
RT
5 min
RT
deionized water
RT
embed slides in Aquatex
RT
Microscopic evaluation
p16INK4a staining
Sections were defined to be negative in cases where no p16 INK4a expression was detectable or where
p16INK4a-positive cells showed a focal staining pattern (patchy, restricted to single cells). Lesions with
a strong and diffuse p16INK4a staining were considered to p16INK4a-positive.
T cell markers
Immunohistochemically stained slides were analyzed independently in a blinded fashion during two
sessions and blinded to histopathological grade. For counting and evaluation of the tumor-infiltrating
lymphocytes, a Leica DMRBE microscope with a 10x10 ocular grid covering an area of 0.0625 mm2
at a 400-fold magnification was used. In total, seven grid areas were counted in the in lesion/tumor
and the adjacent stromal tissue, three located in the epithelium and four located in the stroma. In total,
an area of 0.4375 mm2 was considered for counting.
HLA class I and II
Lesions and tumors that showed a strong cytoplasmic or membranous staining in more than 75% of
cells were classified as positive for HLA class I or class II expression. Heterogeneous expression was
defined as faint and patchy staining (cytoplasmic or membranous) observable in 75% to 25% of the
cells of a lesion or tumor. Lesions were defined to be negative for HLA expression when the staining
was absent or restricted to single cells (representing invading APCs) or could be identified as locally
induced expression due to the presence of immune cells (faint staining, locally restricted in areas with
infiltrating immune cells) and concerned less than 25% of the lesion cells.
Automated evaluation of immunohistochemically stained slides
The establishment of an automated immune cell quantification platform and the adaption of the
underlying imagine analysis algorithms were a major goal of this thesis and are described in detail in
chapter 4.1. The major steps of the workflow are automated staining, whole-slide-imaging and
computational image analysis and were carried out under the supervision of PD Dr. Niels Grabe and
Dr. Bernd Lahrmann, TIGA Center, Heidelberg.
3. Materials and Methods
3.2.2
47
Molecular Biology Methods
Isolation of genomic DNA from cells or tissue
DNA was isolated from either FFPE tissue sections or from cultured cells deriving from fresh tumor
tissue following the manufacturer’s instructions.
Briefly, for the purification of genomic DNA (gDNA) from fresh or frozen cells, the pellet was
resuspended in 200µl PBS and 20µl proteinase K were added. Then 200µl Buffer AL were added and
the sample was incubated at 56°C for 10 min before it was resuspended in 200µl ethanol and loaded
on a spin column by centrifuging at 8000 rpm for 1 min. After having washed the column with bound
DNA two times with the provided wash buffers AW1 and AW2, DNA was eluted with Buffer AE in
two subsequent steps and by using 30µl buffer only for each step to increase the final DNA
concentration without losing to much of the maximum DNA yield.
For the isolation of gDNA from formalin fixed paraffin-embedded tissue sections, samples had to be
pretreated by xylene to remove paraffin. Following centrifugation and removal of the supernatant by
pipetting, ethanol was added to the pellet to remove residual xylene. Ethanol was removed by
pipetting following centrifugation. This step was repeated once before the pellet was dried in the
SpeedVac at 37°C for 15 min, resuspended in 180µl buffer ATL and completely lysed by adding 20µl
Proteinase K at 56°C (minimum 1 hour until overnight incubation). Then the samples were incubated
at 90°C for 1 hour before 200 µl Buffer AL and 200µl ethanol were added to the sample. The lysate
was transferred and the provided QIAamp MinElute column which – following binding of DNA to the
column – was washed twice with Buffers AW1 and AW2. Finally, following complete drying of the
membrane by centrifugation at full speed, DNA was eluted in two steps with 30µl buffer in each step.
The concentration of eluted gDNA was determined by measuring the absorbance at 260 nm with the
elution buffer used as blank for the zero adjustment.
GP5+/6+ PCR for Luminex® -based HPV-Genotyping
amount
reagent
26.75 µl
H2O
5.0 µl
10x PCR Buffer
7.0 µl
50 mM MgCl2
1.5 µl
10 mM dNTPs
2.0 µl
primer set 1
0.5 µl
primer set 2
0.25 µl
Taq polymerase
7.0 µl
template
48
3. Materials and Methods
Temperature profile:
94°C
10 min initial denaturation
94°C
30 sec denaturation
38 °C
30 sec primer annealing
72°C
80 sec primer extension
72°C
6 min final extension
4°C
forever
40 x
HPV genotyping based on Luminex® technology
Luminex Technology based on polystyrene beads with various but specifically identifiable absorption
spectra allow the multiplexed detection of different factors. Specifically amplified DNA from tumor
samples is bound to the beads that are coupled to HPV specific oligonucleotide probes. By this
approach 24 of the most common HPV types (15 high-risk and 6 low-risk and 3 putative high-risk
types) can be detected simultaneously in one sample by reporter fluorescence. For the assay procedure
the manufacturer’s protocol was followed:
First, 40µl/well of the Bead Mix were pipetted to each required sample well of a 96 well Hybridization
Plate. As a negative control 10 µl H 2O, 10µl Hybridization Control (1:10 diluted in H2O) ad 10µl PCR
product per sample well were pipetted. The PCR plate was covered tightly with a seal foil and
incubated at 95°C for 10 min in a preheated PCR machine. The plate was then incubated on ice for 1
min and then for hybridization subsequently transferred to the PCR machine and incubated at 41°C for
30 min. In the meantime a filter plate was equilibrated by pipetting 100 µl Assay Buffer in each well
and incubating the plate for 30min at room temperature. Wash Buffer was removed by vacuum
filtration and the Bead Mix PCR samples were transferred to the filter plate after having the samples
mixed vigorously by pipetting up and down and with the hybridization plate still being located in the
PCR machine. Liquid is removed from the filter plate by vacuum filtration and the plate was washed
twice with 100µl/well Assay Buffer. 70µl Staining solution were added to each well and incubated
protected from light for 30min at room temperature under slight agitation. Then the liquid was again
removed by vacuum filtration and the plate was washed trice with 100µl/well Assay Buffer
respectively. The beads were then resuspended in 100µl Assay Buffer and transferred to a 96 lockmicrotiter plate to measure samples then in the Luminex analyzer.
RNA extraction from cultured cells
RNA purification from human cells was performed with RNeasy Mini Kit from QIAgen with slight
modifications to the manufacturer’s protocol. All centrifugation steps were carried out at 10000 rpm if
not indicated otherwise. Cells were disrupted by adding Buffer RLT and β-mercaptoethanol (1:100) to
the cells and vortexed. The lysate was homogenized by adding 70% ethanol and vigorous vortexing or
pipetting. Then 700µl of the sample were transferred onto the membrane of an RNeasy spin column
and centrifuged for 90 s. The flow-through was discarded. For DNA elimination 350 µl Buffer RW1
3. Materials and Methods
49
were added to the spin column, centrifuged for 90 sec to wash the membrane and the flow-through
was discarded. DNase 1 incubation mix (consisting of 62µl H2O, 7µl 10xDNase Buffer + 1µl DNase 1
(Invitrogen) per sample) was added onto spin column membrane and incubated for 15 min at RT.
350µl Buffer RW1 were added to the membrane, centrifuged for 90 sec at 10000 rpm and the flowthrough was discarded. To wash the spin column membrane, 500µl Buffer RPE were then added to the
membrane, centrifuged for 90 sec and the flow-through was discarded. This washing step was repeated
once by centrifuging the spin column for 3 min. The spin column was then dried by centrifuging it at
14000 rpm for 2 min and was then placed in a new 1.5 ml reaction tube. Then, 50µl RNase-free water
were pipetted on the spin column membrane, incubated for 7 min on ice and then centrifuged at 10000
rpm for 2 min to elute the RNA. This step was repeated to increase the overall RNA yield
accompanied however by decreased RNA concentration.
DNA concentration was measured at 260nm wavelength via photometer and RNA purity was assessed
as the ratio of absorbance measured at 260nm to the absorbance measured at 280nm wavelength.
Reverse Transcription
Isolated RNA underwent reverse transcription for the generation of complementary DNA (cDNA).
Complete RNA samples or a negative control (H2O HPLC-grade) were used for reverse transcription
in addition with the following components:
amount
reagent
11.0 µl
RNA (1 µg, prediluted with H2O)
4.0 µl
5 x First-Strand Buffer
2.0 µl
0.1M DTT
0.5 µl
Oligo-dT-nucleotide
0.5 µl
Hexanucleotide Mix (1:10 pre-diluted)
1.0 µl
10 mM dNTPs
1.0 µl
Reverse Transcriptase (200U/µl)
The First-Strand Buffer, DTT and the Reverse Transcriptase (all contained in the SuperScript II
Reverse Transcription Kit) were mixed with remaining reagent and RNA as listed above. The mixture
was incubated at 70°C for 10 min, the briefly put on ice, incubated at 37°C for 15 min and finally at
42°C for 60 min. The reverse transcription was completed with a denaturation step carried out at 90°C
for 5 min. As the resulting cDNA concentration was assumed to equal the initial RNA concentration,
cDNA was diluted based on RNA concentrations to 20ng/µl in H2O (HPLC-grade). The samples were
either stored at -20°C until further usage or immediately used in quantitative real-time PCR.
50
3. Materials and Methods
Real-time quantitative PCR
Quantitative Real-time RT-PCR was performed with primers to detect human toll-like receptor 7
(TLR7) gene expression. The human β-actin gene expression was used as a normalization control
(primer sequences listed in section 3.1.8). Quantitative real-time PCR was performed in triplicates in a
96-well plate format. Power SYBR Green Master Mix (5µl), the corresponding forward and reverse
primers (final concentration 150 nM) and cDNA template (5µl) or water for the non-template controls
were mixed. The cycling conditions are shown in the table below.
Temperature
Duration
Enzyme activation
95°C
15 min
Denaturation
95°C
15 sec
Annealing
60°C
30 sec
Extension
72°C
30 sec
Number of cycles
1 cycle
40 cycles
Calculation of TLR7 mRNA levels
The threshold cycle PCR values (Ct) were obtained during exponential amplification. The calculation
of relative changes in TLR7 mRNA levels was based on the ∆∆Ct method, which means that TLR7
gene expression – in relative units – was calculated by comparing the Ct values of the target gene with
the normalization control gene. The Ct values for β-actin and TLR7 of technical triplicates of each
sample were averaged. Standard deviations (threshold cycle differences) between triplicate reactions
less than 0.5 cycles were considered to be acceptable and the Ct values were used for further
calculation. The relative expression level of TLR7 mRNA was calculated in comparison to β-actin
mRNA expression. In a first step, ∆Ct-values were calculated for TLR7:
∆Ct gene =Ct target – Ct control cDNA
Then, the ∆∆Ct value for each treated samples was calculated by subtracting the ∆Ct of the control
(untreated or DMSO-treated) from the ∆Ct of the sample.
∆∆Ct = ∆Ct gene (treated) - ∆Ct gene (untreated/control)
The fold exchange in TLR7 expression was then obtained by calculation 2 -∆∆Ct and visualized as bar
graphs in a log2 scale.
3. Materials and Methods
3.2.3
51
Biochemical Methods
Whole cell lysates
Whole cell lysates were prepared by resuspending cell pellets containing a defined numbers of cells in
4x Laemmli buffer and heated for 10 min at 95°C before subjected to gel electrophoresis.
Protein lysates and Bradford assay
Cell pellets were resuspended in 50µl RIPA Buffer containing Protease Inhibitor Cocktail and
incubated for 20 min on ice. Samples were centrifuged at 13000rpm for 15 min at 4°C and
supernatants were transferred in new 1.5 ml reaction tubes for further processing. Protein
concentrations were determined using Bradford protein assay, a photometric method based on the dye
Coomassie Brillant Blue changing its color from red to blue if complexes with proteins are formed.
For the quantification of protein concentrations a 10 mg/ml aqueous BSA solution was serially diluted
to produce a standard curve ranging from 0.0 mg/ml to 2.0 mg/ml BSA. The samples to be tested were
diluted 1:20 in water and Bradford Reagent which was filtered with a 0.22µm sterile filter was diluted
1:5 in water. 5 µl of the standards and the diluted samples were pipetted into a 96 well flat bottom
plate and 250 µl of the Bradford solution were added to each well and incubated for 5 min at room
temperature while shaking until measurement. The absorbance was measured at 595 nm without
wavelength correction and the protein concentrations of the samples were determined by using the
formula of the best-fit curve for the standard values. Samples were either directly used for SDS-PAGE
or stored at -20°C for further analysis.
Sodium-dodecyl-sulfate polyacrylamide gel electrophoresis (SDS-PAGE)
Protein separation was performed with polyacrylamide gel electrophoresis based on the method by
Laemmli. Proteins are denatured and negatively charged after binding of the detergent SDS. In an
electric field the negatively charged proteins migrate towards the anode and are thus separated
according to their molecular weight.
For TLR7 SDS-PAGE whole cell lysate were used, while for HPV16 E7 electrophoresis protein
lysates as described above were used
Dependent on the size of the protein of interest, the resolving gels were produced with different
percentages of acrylamide contained in the formulation to vary the pore size for protein separation. For
the early protein of HPV16, E7, a 20% resolving gel was used, whereas for the separation of TLR7
protein a 7% gel was required.
52
3. Materials and Methods
stacking gel
resolving gels
5% gel
7% gel
20% gel
Aqua dest.
2.9 ml
5.1 ml
0.73 ml
Tris Buffer
1.25 ml
(0.5 M Tris, pH 6.8)
2.5 ml
(1.5 M Tris, pH 8.8)
2.5 ml
(1.5 M Tris, pH 8.8)
Bis-Acrylamide
(30%)
0.85 ml
2.3 ml
6.67 ml
10% APS
50 µl
100 µl
100 µl
Temed
5 µl
10 µl
10 µl
Dependent on the required total protein amount, protein lysates were mixed with 4x Laemmli sample
buffer and water and adjusted to a total volume of 50 µl. The samples are cooked for 10 minutes at
95°C for protein denaturation and then loaded on the gel. Gels were run at 200V, 250 mA and 50W for
about 50 minutes.
Western Blot
Filters and sponges for the Western Blot chamber were treated with blotting buffer before use. The
PVDF membrane was prewetted with 100% methanol for 30 seconds before use and was then washed
in Blotting Buffer. The membrane and the acryl amide gel were stuck between three filter papers and
one sponge from both sides. The blot was performed in Blotting Buffer at 400mA, 50W for 60-90
minutes. After the blotting step the membrane was stained with Ponceau Red solution indicating the
success of the protein transfer. After removing of the color by applying distilled water, the membrane
was blocked with 5% casein solution (in TBS buffer) for one hour at room temperature on the rolling
mixer and was then incubated over night at 4°C with first antibody diluted in 5% casein solution (for
antibody concentrations see Table). The next day, the membrane was washed 3 times for 10 minutes
with TBS containing 0.1% Tween before incubated with the corresponding secondary antibody diluted
in 5% casein solution. After another washing step (3x10 minutes in TBS-T) the membrane was
incubated with the premixed ECL substrate (solution A and B) for 1 min before the antigens could be
detected by developing the photographic films exposed to luminescent light in the darkroom. Exposure
time was variable ranging from 10 seconds to 1 hour depending on the strength of the signal.
3.2.4
Cell culture methods
Flow cytometry analysis
In order to characterize the phenotypes of different primary cells as well as to monitor the generation
of tumor cell lines and purification of different cell types from the whole PBMC fraction, flow
cytometry analysis was performed. Therefore living cells were stained with directly fluorochrome-
3. Materials and Methods
53
labeled antibodies specific for extracellular antigens expressed by different cell types. The light
emitted by the fluorochromes following absorption is three-dimensionally scattered and registered as
forward scatter representing the cell size and as sideward scatter representing the cell granularity. Not
only the percentage of positive cells can be determined, but also the mean fluorescence intensity
providing information about the levels of antigen expression can be measured.
Depending on their availability an average of 2.5 x105 cells were used per staining. Cells were
harvested, washed twice in PBS by centrifugation for 10 min at 4°C and 1200 rpm. Incubation with
directly labeled primary antibodies (each diluted 1:50 in PBS) was carried out on ice for 30 min and
protected from light. If double staining was performed, the second antibody was applied in a second
incubation step following an additional washing step to remove the first antibody. Additionally, for
every staining an isotype control based on an antibody directed against the corresponding IgG subtype
(diluted 1:50 in PBS) was included. Following the incubation with the antibodies, cells were washed
twice again in PBS by centrifuging at 4°C and 1200 rpm for 8 min and finally fixed in 1% PFA
solution and stored in the dark until measurement. The samples were analyzed in a FACSCalibur and
fluoresce data obtained were evaluated using CellQuest Pro Software.
Density gradient centrifugation for the isolation of mononuclear cells from peripheral blood
For immunological studies peripheral blood mononuclear cells (PBMCs) were isolated by density
gradient centrifugation. PBMCs are comprise different cell types such as lymphocytes, monocytes and
macrophages that can be used directly or following further separation in different immunological
approaches. Density gradient centrifugation allows isolating PBMCs from whole blood and is carried
out with a ficoll-based separation medium that contains a hydrophilic polysaccharide of a distinct
density. Centrifugation of overlaid blood leads to the separation of the whole blood sample into
plasma on the top of tube, PBMCs (middle fraction), and a fraction mainly consisting of erythrocytes
on the bottom of the tube.
For extraction of PBMCs heparinized blood samples were diluted with equal amounts of RPMI 1640
medium and the mixture was carefully layered on 15ml lymphocyte separation medium. Density
gradient centrifugation was carried out at 2500 rpm for 15 minutes with inactivated brake. The
supernatant containing plasma was taken off and stored at –20°C. Then the interphase containing
peripheral blood mononuclear cells (PBMCs) was collected and cells were washed twice in RPMI
1640 medium by centrifugation at 1800 rpm and 1500 rpm (15 min each) to purify PBMCs from
eventually contaminating separation medium and also from thrombocytes. Following the washing
steps the pellet was resuspended in 20 ml RPMI 1640 medium and cells were counted with trypan blue
solution. PBMCs then were either directly used in experiments or stored at -80°C in human serum
containing 10% DMSO.
Tumor preparation and tumor tissue culture
Tumor samples of HNSCC patients were sent overnight within 16-24 hours after surgical resection
while lying in transport medium and on ice packs. All washing steps were performed with cooled
54
3. Materials and Methods
solutions and at 4°C. The tumor sample deriving from primary tumor of different localizations (base of
the tongue, edge of the tongue, tonsils) or lymph node metastasis were transferred into a sterile 50 ml
tube and washed twice in 10 ml tumor preparation solution by centrifuging for 10 min at 1200 rpm.
Then the tissue was transferred to a sterile petri dish and eventually necrotic tissue, fatty tissue
surrounding lymph nodes and larger blood vessels were removed mechanically. Following two
additional washing steps in a new 50 ml tube the samples were again transferred to a new petri dish
and cut into very small pieces (≥ 1 mm2) with a scalpel. The tissue pieces were again transferred into a
new 50 ml tube with the help of forceps and 10 ml pipettes washed twice in a new 50 ml tube with
tumor preparation solution and then digested overnight (16-20 hours depending on the size of tissue
pieces) by applying 5 ml tumor digestion solution. Due to the limited size of tumor tissue, isolation of
tumor infiltrating lymphocytes was not carried out and the complete amount of digested tumor tissue
was used for generation of HNSCC tumor cell lines. The next day the tissue was washed twice in
RPMI 1640 medium by centrifuging for 10 min at 1200 rpm and finally resuspended either in
Quantum medium or FAD medium and cultured in cell culture flasks or plates. Cultures with adherent
and outgrowing tumor cells were checked regularly via flow cytometry for the proportion of tumor
cells. Cultures with at least 10% tumor cells were further cultured, whereas cultures with less tumor
cells and those containing only fibroblasts were discarded. Sequential trypsinization of fibroblasts
from young tumor cell cultures was performed to remove fibroblasts and if necessary a second
trypsinization step of the remaining adherent tumor cells was carried out to detach single cells from
tumor cell clusters and allow these areas to expand.
T2- peptide binding assay
The human T2 cell line deriving from T-B lymphoblast hybrids was used to test the binding affinities
of different L1 peptides that were synthesized based on epitope prediction algorithms. T2 cells are
transporter associated with antigen processing (TAP1/TAP2) deficient and therefore defective in
loading human leukocyte antigen (HLA) class I molecules with endogenous peptides. However, HLA
class I molecules of T2 cells can be loaded exogenously with peptides present in the medium, whereby
different epitopes bind to HLA class I molecules with higher or lower affinities. Only peptide-HLAcomplexes are stable and can be detected by flow cytometry analysis following staining with a HLA
class I specific antibody whereas free HLA class I molecules not bound to any peptide are instable and
degraded and therefore cannot be detected. The higher the affinity of predicted epitopes to HLA class
molecules, the more stable is the complex formed and the higher is the fluorescence intensity
measured.
T2 cells were harvested and resuspended in T2 medium at a density of 0.5x106 cells/100 µl. 100
µl/well was pipetted into a 96-well round bottom plate. Then β-2-microglobulin at concentration of 5
µg/ml as well as the newly synthesized peptides to be analyzed for their binding affinity (at 50 µg/ml)
were added to each well. Peptides already known to have a high affinity to HLA class I were included
as positive controls and determined the cut-off (viral MP, p16_R1, L1_323). All antigens were tested
in quadruplicates. The plate was incubated over night at 27°C for 17 hours. Following the 17 hours
incubation period the plate had to be incubated another 2.5 hours at 37°C, 5% CO2. Then cells were
harvested and transferred into a 1.5 ml reaction tube. They were washed one with PBS and then
3. Materials and Methods
55
stained with directly labeled HLA-A2 antibody. Finally, samples were washed twice with PBS, fixed
with 1% PFA and stored at 4°C in the dark until measurement.
For analysis the MFI values for all antigens were recorded and compared with the positive controls
after background subtraction (obtained by measuring T2 cells incubated in absence of any antigen).
Peptides were considered to have sufficient binding capacity to HLA-A2 if the MFI was significantly
higher as negative controls and at the same time exceed the lowest MFI measured for the positive
controls. Peptides that fulfilled both criteria were considered to form stable peptide-HLA-complexes
and to be suitable for simulation experiments.
IL-6 ELISA
Enzyme-linked immunosorbent assay (ELISA) allows the detection and quantification of antigen by
specific antibodies. The ELISA used for measuring IL-6 levels is a classical “sandwich” ELISA with
the antigen contained in a sample being attached to the surface of wells coated with a first antigenspecific antibody. Bound antigens are detected by a second specific antibody linked to an enzyme
allowing the detection of antigen-antibody complexes via a color reaction after adding the
corresponding substrate for the enzyme. Antigen concentrations can be calculated by comparing the
measured optical density values with those of defined standard concentrations. For measuring the
interleukin-6 release following PBMC stimulation, a commercially available IL-6 ELISA was used
providing pre-coated and blocked plates. The assay was performed following the manufacturer’s
protocol and all standards, samples and controls were tested in duplicate. Briefly, after having
prepared an IL-6 standard dilution series, the assay diluent provided was added to the wells, followed
by 100µl/well of standard, sample or control. After 2 hours incubation, wells were washed, 200µl/well
of IL-6 conjugate was added, incubated for another 2 hours and washed again. Substrate solution
(200µl/well) was added, incubated for 20 minutes protected from light and then the color reaction was
stopped by adding 50µl/well of stop solution. The optical density was determined at a wavelength of
450nm and the reference wavelength for wavelength corrections was 540nm. A standard curve was
created by plotting the mean absorbance for each standard against the concentrations and drawing a
best fit curve through the data points which allowed the concentration of the samples to be calculated.
Generation of dendritic cells from monocytes
Dendritic cells are the most potent antigen-presenting cells and play an important role in the raise of an
antigen-specific immune response as the process and present antigens to T cells. Except from culturing
dendritic cells from hematopoietic progenitor cells they also can easily be generated using CD34positive monocytes circulating in the peripheral blood. Generation of dendritic cells requires external
granulocyte-macrophage colony stimulating factor (GM-CSF) and IL-4 contained in the media.
The standard protocol requires PBMCs freshly isolated from peripheral blood. They were washed in
serum free CellGro medium and resuspended in CellGro medium supplemented with 3% human
serum. After one hour of adherence in a 10cm cell culture plate at 37°C the non-adherent cells were
56
3. Materials and Methods
detached by tapping to avoid clumping and the cell culture flask was then incubated overnight at 37°C,
5 % CO2. The next day, the non-adherent cells were removed and the remaining adherent monocytes
were cultured to generate mature dendritic cells. To stimulate the differentiation of monocytes into
dendritic cells, the monocytes were cultured in CellGro medium containing GM-CSF (1000U/ml) and
IL-4 (667 U/ml) for 6 days at 37°C and 5% CO2 in a humified atmosphere. The standard maturation
cocktail was supplemented by addition of TMX-202 to the medium (treatment schedule see below) on
days 0, 2, 4 and 5 of the culture. Cells were harvested on day 6 by using 0.05% EDTA solution and a
cell scraper. Cells were immediately used for immunoassays. For the T cell in vitro priming DCs
generation with cells from the same donor was repeated weekly (every 6 days) and used as fresh cells
for T cell re-stimulation.
Isolation of T lymphocytes from PBMCs
In order to obtain total T cells from the whole PBMC fraction, isolated mononuclear cells were subject
to plastic adherence for 4 hours in RPMI containing 5% human serum as described for the generation
of dendritic cells. T cells then were isolated from the non-adherent cell fraction by Pan T cell isolation
based on magnetic depletion of non-T cells. Therefore floating cells were harvested, passed through a
40µm cell strainer to avoid clumping, counted and washed once in MACS Buffer. T cells were
purified from that fraction by using MACS Pan T cell Isolation Kit II from Miltenyi Biotec following
the manufacturer’s protocol. Briefly, cells were resuspended in 40 µl MACS buffer per 10 7 cells and
then incubated for 5 minutes with 10 µl per 107 cells of the biotin-antibody cocktail (containing
monoclonal antibodies against CD14, CD15, CD16, CD19, CD34, CD36, CD56, CD123, and
CD235a) targeting non-T cells. Following addition of 30 µl MACS buffer and 20 µl anti-biotin
microbeads per 107 cells to the sample and incubation for 10 minutes, T cells could be isolated by
magnetic separation. The specific binding of magnetic microbeads to labeled cells allows the depletion
of the non-target cells by retaining them in the magnetic field of the MACS column while unlabeled T
cells pass through. Eluted T cells were counted and immediately used in immunoassays. T cells were
cultured in T cell medium containing IL-2 and IL-7.
Isolation of regulatory T cells from PBMCs (Regulatory T cell depletion)
Treg depletion was performed in order to compare the effects of T cell mediated killing of tumor cells
between the total T cell fraction (including regulatory T cells) and T cells that are depleted from
regulatory T cells. Tregs were depleted from the total CD3+ T cell fraction by using MACS
technology based on magnetic labelling of CD25+ T cells with CD25 MicroBeads and isolation of the
labelled cells by positive selection over a MACS column in a magnetic field. Tregs were depleted in
two successive steps following the manufacturer’s protocol and comparable to the procedure describe
for T cell isolation. In the first step non-CD4 positive cells were magnetically labelled and separated
from CD4+ T cells to enrich the CD4+ T cell fraction. These cells were eluted from the column and
stored for the experiments. In the second step, CD4+CD25+ regulatory T cells were labelled and
separated from the remaining CD4+ T cell population over a column. Briefly, the cell pellet of CD4+
3. Materials and Methods
57
T cells was resuspended in 90 µl MACS buffer per 107 cells and 10 µl of anti-CD25 microbeads per
107 cells were added and incubated for 15 minutes on ice. Cells were then washed twice in 2 ml of
MACS buffer by centrifuging at 1200 rpm for 10 minutes and finally resuspended in 500 µl of MACS
buffer. Cells were then subject to magnetic separation for positive selection of labeled CD25+ cells
while unlabeled cells pass through the column. Non-CD4+ T cells separated in first step were
combined with the CD4+ enriched and Treg depleted fraction and used for further experiments.
In vitro priming of T cells with HPV16 L1 and p16INK4a peptides
The induction of a specific T cell response depends on the recognition of the antigen via MHC
complex and the activation by co-stimulatory molecules. The activation of naïve T cells after the
recognition of antigens presented by antigen presenting cells and their development into effector T
cells is called “priming” and can be simulated in vitro to monitor the ability of peptides to induce a
cell-mediated immune response in naïve individuals giving rise to T cells that are able to recognize
and target tumor cells that express the protein.
In order to induce a primary cell-mediated immune response against HPV16 L1-peptides and a
p16INK4a-peptide naïve T cells of a HLA-A*0201 positive healthy donor were stimulated with 9mer
and 10mer L1 and p16INK4a peptides predicted for HLA-A2 and validated in peptide binding assay.
Dendritic cells as potent antigen-presenting cells were used to prime naïve T cells to the peptides and
were generated in 4 cycles as described above. T cells were obtained by T cell isolation from PBMCs
as described above. The ratio between DC and T cell during stimulation was 1:10.
FIGURE 3.1
TMX-202 AND DMSO TREATMENT SCHEDULE FOR DENDRITIC CELLS AND T CELLS
DURING THE IN VITRO PRIMING APPROACH.
Dendritic cells were harvested and loaded with peptides by incubating them in peptide-load medium
with 20µg/ml of each peptide and in presence of Lipofectamine 2000 for 2.5 hours at 37°C, 5% CO2.
Cells were irradiated with 30 Gray after loading and washed twice. They were then added to T cells in
a 12-well plate in T cell medium and were co-cultured until the next restimulation 6 days later. For
restimulation T cells were harvested, washed and counted and the required amounts of DC were
loaded again with peptides by repeating the procedure described above. In total, T cells had four
stimulation cycles over 24 days (Figure 3.1). The experiment was based on two distinct T cell
fractions that were treated with either TMX-202 or DMSO during the complete duration of the
58
3. Materials and Methods
experiment (4 treatments per cycle) and also were stimulated with either TMX- or DMSO-treated
dendritic cells. The treatment schedule is shown in Figure 3.1.
PBMC treatment with TMX-202
PBMCs were obtained by density gradient centrifugation as described above and cultured in 24-well
plates in medium for T cells. PBMCs were cultured for 3 days and were treated daily with either 1 µM
TMX-202, 10µM TMX-202, 30µM imiquimod or the same amount of DMSO as added with the
substances as control. Cells were harvested after 72 hours. Due to the adherence capacities of
monocytes two distinct cell fractions had to be harvested: non-adherent peripheral blood lymphocytes
(PBLs) and adherent monocytes that were harvested by using a cell scraper. Cells were washed and
pellets were stored at -20°C until used for further analyses. Supernatants were also harvested,
centrifuged to remove cells and stored at -80°C for cytokine analysis.
Tumor cell line maintenance
Tumor cell lines were cultured in the corresponding tumor cell media listed above. Adherent tumor
cell lines were split when confluent. Prior to use in killing assays tumor cells were treated with 1 µM
DAC following a standard treatment protocol to increase antigenicity of the tumor cells which was
developed in the department: tumor cells were treated for 96 hours, with daily change of half of the
media and addition of 1 µM final concentration with the supplemented media. Cells were then
harvested and used in the corresponding experiments.
xCELLigence Impedance Measurement
The xCELLigence system is based on a microelectronic readout using electronic cell sensor array
technology and allows for real-time monitoring of cellular processes without requiring labeling of
cells with additional compounds and therefore being less invasive and allowing more physiological
conditions. The assay principle is based on changes of the electrode impedance by adherent cells
(Figure 3.3) As the measurement reflects the entire duration of the assay, the conditions can be
monitored in real-time allowing the characterization of the kinetic response of cells within an assay,
prior and following certain treatments. Thereby information regarding the biological status of the cell
(growth rate, growth arrest, morphology, apoptosis) can be obtained rendering the assay also suitable
for the quantification of compound-mediated or cell-mediated cytotoxicity. The assay principle is
based on the measurement of changes of the electrode impedance due to cell-electrode interactions, as
adherence of cells onto the electrodes affects the local ionic environment at the electrode/solution
interface. Impedance increase is dependent on the numbers of cells attached to the electrodes but also
on the quality of the interaction between cells and electrodes. The electrode impedance is represented
by a dimensionless value, termed cell index, which indicates the relative change in measured electrical
impedance and thus the cell status. It contains information about cell viability, cell growth or growth
arrest, apoptosis, morphology and adhesion degree (Figure 3.3).
3. Materials and Methods
electrode without cells
59
Z
Z = Z0
baseline
Z
electrode with attached cell
FIGURE 3.2
Z = Z cell
impedance
PRINCIPLE OF THE xCELLigence TECHNOLOGY. Adherence of cells to the electrodes affects the
electrode impedance (Z cell) compared with the baseline impedance (no cells, non-adhered cells) by
changing the local ionic environment at the electrode/solution interface). Adapted from
www.aceabio.com.
Without cells or cells not adhered to the electrodes the cell index is zero. Under the same conditions,
cell index values increases with adherence of cells to the electrodes, and even more increases if cells
spread over the electrodes or become more strongly attached to them. The values decrease with cells
detaching from the electrodes due to apoptosis or cytotoxicity.
FIGURE 3.3
CHANGES OF THE CELL INDEX REPRESENTATIVE OF THE ELECTRODE IMPEDANCE OVER
TIME UNDER DIFFERENT CONDITIONS. A) Electrode and cell index (CI) is zero if no cells or only
non-adherent cells are contained in the wells and B) increases with adherence of cells to the electrodes.
The CI positively correlates with C) the cell number and D) the strength of adherence. E) Detaching cells
due to apoptosis or cytotoxicity lead to decreasing CI values. Adapted from www.aceabio.com.
60
3. Materials and Methods
The xCELLigence system was used to compare and quantify the effects of depletion of regulatory T
lymphocytes from the total T cell fraction on the killing rate of tumor cells. It served as platform to
characterize the cell-mediated cytotoxicity in an autologous tumor model.
The 96 well E-plate was prepared by adding 100 µl PBS in all interspaces between the wells to reduce
evaporation of the medium and drying-out of the plate. Then, 75 µl/well Quantum tumor cell medium
were added in well (150 µl/well in the wells designated for medium control). The plate then was
incubated for 30 min at RT, put onto the SP station for measurement and impedance was measured for
determination of the background (sequence 1). Per well, 25000 tumor cells were seeded in a volume of
150 µl Quantum tumor cell medium and were grown for 96 hours while pre-treated with 1 µM DAC
following the standard treatment protocol developed in our laboratory (see above). Therefore the
measurement was interrupted every 24 hours until day 4 (96 hours) when effector cells were added.
This was done by changing half of the media and adding 25000 T cells per well in 75 µl T cell
medium. The plate was then measured for an additional period of 96 hours without interruption while
T cells and tumors were co-incubated. Throughout whole experiment the electrode impedance was
measured every 30 minutes leading to approx. 48 time points measured for each 24 hours-interval and
approx. 190 time points recorded during the tumor cell growing and the co-incubation phase.
CD107 degranulation assay
Cytotoxic T lymphocytes (CTLs) can get activated upon contact with and recognition of target cells.
In the activated state during the CTL-target interaction they start to release cytotoxic granules which is
accompanied by the mobilization of CD107a (lysosomal-associated membrane protein-1, LAMP-1) to
the cell surface which is normally present in vesicle membranes. The CD107a surface expression thus
correlates with the cytotoxic activity of T cells and the killing rate of target cells and can be
quantitated by flow cytometry analysis. This method allows also for the simultaneous staining with
other markers to gain further information about T cell phenotypes. The assay principle is displayed in
Figure 3.4.
CD107a mobilization assay was used in two different settings: for the analysis of the killing potential
of T cells after in vitro priming under treatment with immune modulators and of the killing effect of T
cells after Treg depletion in an autologous setting.
The specific conditions and setups for each of these assays are demonstrated in Tables x and x. In
general, 2.5x105 T cells (effectors) were co-incubated with tumor cells (targets) in a 1:1 ratio.
Following isolation of fresh cells or harvesting of cultured cells, T cells were washed once in RPMI
medium and adjusted to 2.5x106 cells per 100 µl in T cell medium. Effector cells and target cells were
co-incubated in a total volume of 200µl in a 96-well round bottom under sterile conditions. As the
experiment was demonstrated in previous approaches to provide reliable results with small standard
deviations between quadruplicates it was performed in duplicates as T cell numbers were restricted. As
controls for spontaneous CD107a release and background reactivity T cells were incubated without
tumor cells. To the corresponding wells (except those where only T cell markers were investigated or
served as isotype controls), 10 µl of fluorescent-labeled anti-CD107a antibody were added (see Table
“Antibodies” section 1.3.5).
3. Materials and Methods
FIGURE 3.4
61
PRINCIPLE OF THE CD107A DEGRANULATION ASSAY. Effector cells (T cells) are co-incubated
with target cells (tumor cells) and an antibody against CD107a is added to the culture (A). Upon T cell activation CD107a is
mobilized to the cell surface with the release of cytotoxic granules and can be bound by the antibody. CD107a surface
expression can then be analyzed by flow cytometry analysis (B).
The plate was then incubated at 37°C in a humidified atmosphere with 5% CO 2. After 1 hour of coincubation, brefeldin A at a final concentration of 5µg/ml was added to each well and the plate was
incubated for further 4 hours. The plate was then centrifuged (1200 rpm, 10 minutes, room
temperature) and the supernatant was removed. T cell/tumor cell conjugates were dissolved by
resuspending the pellets in 200 µl PBS/0.5mM EDTA buffer and cells were then transferred into a 1.5
ml reaction tube. The wells were washed a second time with 200 µl PBS/0.5mM EDTA buffer and
remaining cells were also transferred into the tubes. After centrifugation (1200 rpm/10 minutes/ 4°C),
cells were washed once with FACS-PBS and samples were either directly fixed with 1% PFA solution
(CD107a single staining) or subjected to FACS staining (isotype control and T cell markers as single
staining or in double staining with CD107a) by applying the FACS staining protocol described above.,
Finally, cells were fixed with 1% PFA solution and transferred into FACS tubes for measurement.
During FACS analysis, T cells could be distinguished from tumor cells in the FSC/SSC based on their
size and granularity. The corresponding FITC and PE isotype controls allowed the definition of
quadrant borders and single stains for CD107a and T cell markers were used to adjust the fluorescence
compensation for the measurement of double stains. Samples containing only tumor cells were used to
verify that with the instruments settings and gates chosen for analysis only T cells are included in the
analysis and tumor cells are excluded from the quantitation. Then the samples of co-incubated T cells
and tumor cells for the analysis of the killing rate were measured in duplicates by applying the same
settings and conditions.
62
3. Materials and Methods
In vitro priming of T cells with subsequent CaSki killing
sample no
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16-17
18-19
20-21
22-23
24-25
26-27
Tc untreated
yes
yes
yes
yes
yes
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
no
no
no
Tc treated
no
no
no
no
no
yes
yes
yes
yes
yes
no
no
no
no
no
no
no
no
yes
yes
yes
tumor cells
no
no
no
no
no
no
no
no
no
no
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
yes
anti-CD107a
none
none
yes
none
yes
none
none
yes
none
yes
none
none
yes
none
yes
no
yes
yes
no
yes
yes
FACS-stain
IgG1-FITC
IgG1-PE
none
CD8
CD8
IgG1-FITC
IgG1-PE
none
CD8
CD8
IgG1-FITC
IgG1-PE
none
CD8
CD8
CD8
none
CD8
CD8
none
CD8
Samples 1-5 represent controls for untreated T cells, samples 6-10 controls for treated cells. Controls for tumor
cells are represented by samples 11-15. The killing experiment (co-incubation of T cells with tumor cells) is
represented by samples 16-23.
Treg depletion with subsequent killing of the autologous cell line HN038M
sample no
1
2
3
4
5
6
7
8
9
10
11
12-13
14-15
Tc total
yes
yes
yes
yes
no
no
no
no
no
no
no
yes
no
Tc depleted
no
no
no
no
yes
yes
yes
yes
no
no
no
no
yes
tumor cells
no
no
no
no
no
no
no
no
yes
yes
yes
yes
yes
anti-CD107a
none
none
yes
none
none
none
yes
none
yes
none
yes
yes
yes
FACS-stain
IgG1-FITC
IgG1-PE
none
CD4
IgG1-FITC
IgG1-PE
none
CD4
none
CD4
yes
yes
yes
Samples 1-4 represent controls for total T cells, samples 5-8 controls for Treg depleted T cells. Controls for
tumor cells are represented by samples 9-11. The killing experiment (co-incubation of T cells with tumor cells)
is represented by samples 12-15.
3. Materials and Methods
63
Mycoplasma detection assay
Mycoplasma are the simplest prokaryotes and a major problem in cell culture as contamination is very
common and may influence the cell proliferation of cell lines but also their gene expression patterns
which might be a problem for different assays. To exclude infections with mycoplasma, cell lines,
especially those of primary cell culture, were regularly tested with the MycoAlert™ assay. This assay
is a rapid and easy method to detect mycoplasma contamination in cell cultures and is based on the
selective biochemical analysis of the activity of special mycoplasma enzymes. Mycoplasma contained
in the culture are lysed and set free enzymes that react with the MycoAlert substrate catalyzing the
conversion of ADP to ATP. The ATP levels measured in a sample before and after the substrate is
added allow the calculation of a ratio that indicates presence or absence of mycoplasma. The
underlying biochemical reaction is based on the oxidation of Luciferin in presence of ATP by
Luciferase and allows the quantification of emitted light.
The intensity of the emitted light is linearly related to the ATP concentration and can be measured by a
luminometer. The assay was performed following the manufacturer’s protocol. Briefly, 2 ml of cell
culture supernatant or cell culture were transferred into a reaction tube and any cells contained in the
sample were pelleted at 1500 rpm for 5 minutes. 100 µl of the cleared supernatant were transferred
into a luminometer cuvette and 100 µl of MycoAlert reagent were added to each sample and incubated
for 5 minutes. The cuvette was placed in the luminometer and read (with a program set to 1 minute
integrated reading) to obtain a value for Reading A. Then 100 µl of the MycoAlert substrate were
added to each sample and incubated for 10 minutes before the cuvette was measured again (Reading
B). The ratio “Reading B/Reading A” was calculated and interpreted as follows:
3.2.5
Ratio
Interpretation
< 0.9
negative for mycoplasma
0.9 - 1.2
borderline: quarantine cells and retest in 24 hours
> 1.2
mycoplasma contamination
Statistical Methods
For the comparison of continuous data between two groups either Student’s t-test or Mann-Whitney U
test for non-parametrical data were used.
For the estimation of differences in categorical in terms of between two groups chi-square test was
used.
For all tests, differences were considered to be significant if the calculated p-value was 0.05 or less.
64
4.
4. Results I
IMMUNE CELL INFILTRATES
AND POSSIBLE IMMUNE
EVASION MECHANISMS IN
CERVICAL LESIONS
The present chapter deals with the immunological characterization of cervical intraepithelial neoplasia
and cancers. Final goal of this part is to gain a better understanding of the clinically heterogeneous
behavior of the precancerous lesions in terms of regression and progression rates.
In the first part a central methodological approach of histological analyses of immune cell infiltrates in
CIN was established using a computer-based tool for the standardized quantification of immune cell
infiltrates in cooperation with the TIGA Centre Heidelberg (chapter 4.1).
In the following part immune cell infiltrates were investigated in CIN to find out if samples of
different infection stages are different in terms of immune cell phenotypes. Changes in immune cell
densities and composition might be a hint for either underlying immune-regulatory mechanisms or
effective anti-tumoral immune responses. In this context, the time point of when changes in the
immune cell infiltration become apparent during the natural history of CIN lesions and a possible
association with the initiation of the transforming infection stage as represented by p16INK4a
overexpression are of special interest. Therefore different T cell markers of which most are wellcharacterized, and which are representative of T cell activation and also of immuno-regulatory
mechanisms were investigated (chapter 4.2).
To better understand to which extent intrinsic features of the epithelial cells play a role in the
pathogenesis of cervical cancer the antigen-presentation capacity of the lesion cells was investigated.
Antigen presentation might be influenced by HPV infection and alterations regarding the expression of
the involved molecules probably promote disease progression by causing immune escape despite the
presence of infiltrating lymphocytes (chapter 4.3).
Finally, the immune infiltrates of patients with CIN who were topically treated with the clinically
approved immuno-modulatory substance imiquimod, were characterized in a longitudinal approach
(chapter 4.4). These analyses aim at a better understanding of how the immune cell composition could
be positively influenced and how these changes might correlate with the clinical course of the disease.
4. Results I
4.1
65
Development of an automated quantification system for the
computational profiling of cervical intraepithelial neoplasia
and its microenvironment
4.1.1
Scanning and digitalization of stained tissue sections
As the analysis of immune cell infiltrates in the lesions and the adjacent stroma was based on
digitalized images of the tissue sections, the slides were scanned after having been fully automated
stained with monoclonal antibodies against CD3 and CD8. The tissue sections were automatically
imaged with the Hamamatsu NanoZoomer 2.0-HT Scan System (Figure 4.1) at 20-fold magnification
resulting in a resolution of 0.46µm/pixel. The scan system is equipped with three 4096x64 pixel Time
Delay and Integration (TDI) CCD (charge-coupled device) sensors enabling imaging based on a threedimensional XYZ-zoom technology (ROJO et al., 2006). This type of sensor allows multilayer
scanning and is not restricted to one single, two-dimensional layer. The resulting virtual slides can
then be analyzed in a similar manner as using classical microscopy allowing focusing through
different layers of the tissue, dependent on the number of layers that were scanned and the distance
between them. With a total capacity of 210 slides per batch and a scanning speed of 60 seconds for a
tissue sample sized 15x15mm² the system allows for high-throughput scanning, digitalization and
archiving of tissue samples. While scanning a glass slide, the system automatically detects the region
of interest defined by presence of any tissue and also automatically chooses the correct and valid focal
plane for the scanning processing. The file size of the virtual slides resulting from the digitalization
process originally is up to 20 GB as uncompressed files and depends on the total size of the scanned
area, the number of scanned layers and the magnification used during scanning. The file size
retroactively can be reduced (approximately by the factor 25) by applying lossless JPEG compressing
algorithms reducing for example a 16.3 GB slide to a 636 MB JPEG file.
FIGURE 4.1
THE NANOZOOMER 2.0-HT SCAN SYSTEM USED FOR DIGITALIZATION OF STAINED
SLIDES AND THE USER INTERFACE OF THE NDP SLIDE SERVER. With the NDPView software
digitalized slides can be analyzed on a computer and allows the user to navigate through the slide in all
three dimensions. Slides can be annotated, screenshots can be made and parameters such as intensity of
the color channel, contrast and brightness can be adapted.
66
4. Results I
The scanned slides were made accessible on the TIGA’s Slide Server (http://tigacenter.bioquant.uniheidelberg.de/ndp-slide-server.html) for all cooperation partners and thus facilitated the exchange of
data and information. This tool was also used for the definition of the lesion based on the p16 INK4a
staining and in cases of unclear morphology in low-grade lesions served as a platform for the
pathologist’s review of the tissue (Figure 4.1).
4.1.2
Development of an image processing tool adapted to cervical
intraepithelial neoplasia
The algorithms used for image processing have been developed using TissuemorphDP™M from
Visiopharm, a company specialized in tissue analysis. The image processing software applied in this
project was based on different algorithms and developed in cooperation with the TIGA Center,
Heidelberg. An overview of the user interface of the Visiopharm image processing software with
exemplary tools developed in the frame of this project that are applicable to the analysis of immune
cells infiltrates in cervical intraepithelial neoplasia is given in Figure 4.2.
FIGURE 4.2
OVERVIEW OF THE USER INTERFACE OF THE VISIOPHARM IMAGE PROCESSING
SOFTWARE. Tools were developed for the annotation of the lesion and basal membrane, generation of
ROIs, clearance of non-ROIs before starting the processing for cell segmentation.
4. Results I
67
Image processing was developed and adapted to CIN lesions using the Visiopharm image processing
software before algorithms could finally be applied to the digitalized slides. Image processing is
performed in four distinct steps.
(i)
Automated tissue detection
The fully automated detection of all analyzable tissue contained on the glass is the first step towards
the cell quantification in the lesion and its microenvironment. The region of interest (ROI) is defined
as the tissue area that shall be subjected to further analysis. ROI detection was performed on the whole
slide after converting a color overview image (RGB) into a greyscale image. By applying simple
thresholding methods on the grey scale image as described previously (OTSU, 1979), ROIs could be
separated from the background regions of the slides. Thereby the background representing any nontissue regions is separated from the relevant tissue regions which can then be subject to further image
processing steps and analysis (Figure 4.3 A). As a post-processing step for the ROI detection, areas
that cannot be analyzed because they are too small or inappropriate such as small tissue fragments,
folded tissue, staining artifacts or dust particles were removed (Figure 4.3 B) by applying
morphological operations like opening or closing (GONZALES, 2009). These are standard imaging
operations to remove small disturbing objects from the image or to remove small holes contained in
the tissue.
FIGURE 4.3
(ii)
EXAMPLE OF THE ROI DETECTION PROCESS. A) Regions of interest are detected automatically
using thresholding methods to separate the tissue from the background (green line). B) Post-processing
steps remove artifacts and areas (indicated by arrows) that are too small for further analysis.
Manual annotation of the lesion and the basal membrane and automated generation of
different invasive margins in the stroma
In a second step the regions to be analyzed had to be defined which was done partially by manual
annotation and partially by automated generation of regions that were then subjected to further
analysis. Due to the high tissue heterogeneity in CIN and the resulting difficulties to separate normal
tissue from the lesion and also the presence of p16INK4a-positive and p16INK4a-negative lesion areas
among the low-grade lesions renders fully automated computational image processing challenging.
One major concern is that p16INK4a-negative low-grade lesions would not have been identified as such
by the established automated annotation algorithm and would have falsely been annotated as “normal”
tissue. Therefore in this first approach for the establishment of the basic method, automated
68
4. Results I
tumor/lesion-identification was replaced by manual annotation based on a comparison with the
p16INK4a-stained reference slide. Lesions positive for p16INK4a-overexpression were visually identified
and the corresponding region was annotated manually in the slides stained for the defined T cell
markers (white line, Figure 4.4). In unclear cases due to aberrations between the histological stage
given by the pathologist and the morphology of the lesion, tissue sections were reviewed by the
pathologist again. The manual approach described here also allowed for the separate investigation of
p16INK4a-positive and -negative lesion areas within the same sample. In the second step the basal
membrane underneath the annotated lesion was also manually marked. These annotation steps are the
prerequisite to proceed to the next step that divided the adjacent stromal tissue into several distinct
areas (invasive margins). The algorithms applied for region growing are used from the baseline (basal
membrane/lamina) and separate the tissue into specific regions by growing in fixed and determined
directions. Starting at the basal lamina, the first region grows with a distance of 100µm into the
surrounding tissue of the epithelial region (yellow line, Figure 4.4). Then the second defined region
grows with further 400µm into the tissue (border at 500µm, pink line) and is followed by the last
growing with 500µm (border at 1000µm) leading finally to the last margin with a maximal distance of
1000µm located from the basal membrane (green line, Figure 4.4). After processing the slides were
manually inspected and regions that did not represent typical stromal tissue (artifacts such as
disruptions, or glandular tissue, endothelial cells and cavities of large blood vessels) and that therefore
had to be excluded from further processing were removed manually from the regions by annotating
them as regions to be cleared (Figure 4.4).
FIGURE 4.4
EXAMPLES OF PROCESSED SLIDES WITH AND THE CLEANING POST-PROCESSING STEP.
The regions of interest (ROIs) are visualized by color-coded lines and represent the epithelium (white),
margin 100 (yellow), margin 500 (pink) and margin 1000 (green). Cleared regions are marked by an
asterisk.
4. Results I
(iii)
69
Cell segmentation
During the last image processing step positively stained and unstained cells were detected by cell
segmentation and subsequently the expression level (determined as brown (positive) or non-brown
(negative)) is determined. The segmentation of the cell nuclei was performed separately in all
determined stromal ROI generated in the previous step and also of all nuclei in the epithelial region.
The segmentation of all nuclei (brown and blue) is based on a watershed segmentation described
elsewhere (BEUCHER, 1992; JUNG and KIM, 2010) on the IHS (Intensity, Hue, Saturation)-S color
band. The basic principle of watershed segmentation is the transformation of an intensity image into a
three-dimensional topographic image. The intensity of each pixel of an image thereby is represented
by the altitude of the relief. Watershed algorithms then are applied, the relief is “floated” and the
watersheds around peaks can be interpreted as borders defining different components which can thus
be segmented from each other.
FIGURE 4.5
EXAMPLE OF THE CELL DETECTION STEP. Shown is the annotated tissue (A) before cell
segmentation and (B) after cell segmentation. DAB-negative cells are displayed in green, DAB-positive
cells in red.
Finally, the DAB-positive (brown stained) cells were detected within a HDAB-DAB color band,
provided by a color deconvolution algorithm (RUIFROK and JOHNSTON, 2001). In dependence on
the DAB staining signal of the surrounding membranes nuclei were categorized into two groups by
simple thresholding, namely blue nuclei with brown (DAB-positive) membranes and blue nuclei
without brown staining signal (Figure 4.5). In a post-processing step nuclei detected that were defined
as being too small where removed by an area-filter. An overview of all image processing steps is given
in Figure 4.6.
70
FIGURE 4.6
4. Results I
EXAMPLE OF THE SUCCESSIVE STEPS OF THE AUTOMATED QUANTIFICATION PROCESS.
Based on the p16INK4 reference slide (1) on which the lesion was marked after reviewed by a pathologist
(1*) the slides stained for CD3 (left side, (A)) and CD8 (right side, (B)) were annotated by demarking the
epithelium and the basal membrane (2). Then the different invasive margins with 100 µm, 500 µm and
1000 µm reaching into the stromal compartment were generated (3). Finally, cells stained for the
corresponding immune cell markers (red) and those that are negative (green) are detected and quantified.
4. Results I
4.1.3
71
Calculation of cell densities from the output data
The successive application of image processing steps described above resulted in different output
variables comprising number and area of the nuclei for every staining category (negative = blue,
positive = brown) and, in addition, the white areas surrounding the nuclei and representing cytoplasm.
All output variables are listed below in table 4.1.
ROI 001
RO002
ROI 003
ROI 004
epithelium
margin 100
margin 500
margin 1000
counts of negative nuclei inside the corresponding ROI (blue signal)
counts of positive nuclei inside the corresponding ROI (brown signal)
area covered by negative nuclei inside the corresponding ROI
area covered by positive nuclei inside the corresponding ROI
remaining (non-nuclei) area inside the corresponding ROI (white area)
TABLE 4.1
OVERVIEW OF THE OUTPUT VARIABLES OBTAINED FOR ALL DEFINED REGIONS OF
INTEREST (ROI) FOLLOWING APPLICATION OF IMAGE PROCESSING STEPS.
The output data comprise cell counts of positive and negative cells and the areas that are covered by
cell in a distinct ROI. Single values are given for the negative nuclei, positive nuclei and the white
surroundings representing cytoplasm. The total areas of all compartments, ROIs, could then be
calculated from these values. This was done for each compartment separately (margin 100, margin 500
and margin 1000), but also for the continuous regions that reach from the basal membrane up to the
500µm and the 1000µm borders. The ratios between cell counts in a distinct ROI and the
corresponding area of this compartment were calculated in order to obtain the cell densities as
“positive cells/mm2”) from the output data.
4.2
The local immune cell infiltration in cervical intraepithelial
neoplasia in relation to p16INK4a expression
The study presented in this chapter adresses the question whether changes in the composition and
densities of immune cell markers are correlated to p16INK4a overexpression in cervical dysplasia, as a
marker stratifying CIN into two infection states (permissive infections, p16 INK4a-negative, and
tansforming infections, p16INK4a-positive). The correlation of possible changes in immune cell density
and composition with p16INK4a-defined biologic stages may reveal import insights into the immune
control and changes of these mechanisms during cervical carcinogenesis.
72
4. Results I
For the investigation of a possible link between the infection stage and infiltrating immune cells in
CIN mainly well-characterized standard T cell markers were chosen (for details see Introduction
chapter 1.4.1).
A mixture of activation and inhibition markers should allow to investigate to which extent T cells
present in the lesion microenvironment are in an activated state and possibly able to combat the HPV
infection and transformed cells or are inhibited. The global T cell infiltration was characterized by
CD3-expressing cells while CD8 and Granzyme B were used as markers for cytotoxic T lymphocytes
(CTLs) and activated CTLs displaying lytic activity. Forkhead box transcription factor 3 (Foxp3), a
marker for regulatory T cells and thus representing the suppressive state of immune cells was also
included. CD3-ζ was included as a marker for the susceptibility of T cells for activation upon antigen
recognition.
4.2.1
p16INK4a-expression status of the lesions
As a marker highlighting transforming HPV infections (BERGERON et al., 2014; VON KNEBEL
DOEBERITZ et al., 2012) p16INK4a was used to biologically define the different lesion grades that
were available for this study. Immunohistochemical staining for 16INK4a (chapter 3.21) revealed that all
cervical carcinoma samples and all high-grade CIN (CIN2/3) were p16INK4a-positive. However, lowgrade CIN (CIN1) could be classified into two groups with 9 of 22 lesions being p16INK4a-negative
(permissive infection) and 13 of 22 lesions being p16INK4a-positive representing the early transforming
infection stage in CIN (Table 4.2).
number of patients
p16INK4a-positive
samples: n (%)
TABLE 4.2
4.2.2
CIN1
CIN2
CIN3
CxCa
Overall population
22
11
19
17
69
13 (59.1%)
11
(100.0%)
19
(100.0%)
17
(100.0%)
60
(86.96%)
SAMPLE CHARACTERISITCS REGARDING THE HISTOLOGICAL CLASSIFCIATON AND THE
TRANSFORMING INFECTION STAGE AS REPRESENTED BY THE p16INKa STATUS.
Comparison of T cell infiltrates in p16 INK4a-positive and p16INK4anegative low-grade CIN
In terms of histomorphological classification CIN1 lesion are regarded as a uniform group.
Biologically they are, however, more diverse with a proportion of these lesions being already in the
early transforming infection stage which is highlighted by beginning p16 INK4a-overexpression.
T cell infiltrates of all phenotypes were compared between p16INK4a-negative and p16INK4a-positive
low-grade lesions (representative examples for the immunohistochemical characterization of
infiltrating immune cells are given in Figure 4.7).
4. Results I
FIGURE 4.7
73
REPRESENTATIVE DETAILS OF IMMUNOHISTOCHEMICAL STAININGS (AT 200x
MAGNIFICATION) FOR p16INK4a, CD3, CD8, GRANB, CD3ζ AND FOXP3. Representative areas of
the epithelium (upper part of the tissue) and the adjacent stroma (lower part) are shown and examples of
positive cells are indicated by arrows.
Low-grade lesions (all CIN1 irrespective of the p16INK4a expression state) and the adjacent stromal
compartment had generally lower total numbers of infiltrating immune cells compared with the higher
grade CIN (mean cell numbers, ranges and standard deviations are summarized in Table S9.1).
Nevertheless, the comparison of p16INK4a-negative and p16INK4a-positive samples within the group of
low-grade CIN did not reveal significant differences regarding the infiltration densities of the five
investigated T lymphocyte phenotypes (4.8 and Table S9.1). The ratio between epithelial and stromal
cell numbers representing the percentage of T cells invading from the lesion-adjacent stroma into the
lesion neither did reveal significant differences between p16INK4a-negative and p16INK4a-positive CIN1
lesions (Figure 4.8 and Table S9.1). Furthermore, the ratios of all T cell subtypes to CD3+ cell counts
were calculated for both compartments as a measure for the proportion of distinct T lymphocyte
phenotypes among all present T cells. Here again, no significant differences between p16INK4a-negative
and p16INK4a-positive low-grade lesions were observed (Figure 4.9 and Table S9.2).
74
FIGURE 4.8
4. Results I
DISTRIBUTION OF T CELL SUBTYPES IN DIFFERENT COMPARTMENTS IN p16INK4aNEGATIVE LOW-GRADE LESIONS COMPARED WITH p16INK4a-POSITIVE LOW-GRADE
LESIONS. A) Absolute T cell counts per 0.0625mm² in the lesion and lesion-adjacent stroma. B) Ratio
between the lesion and lesion-adjacent stroma for all T cell phenotypes. The dot in the center of each box
represents the median value of the distribution; the borders of the box represent the upper and lower
quartiles (25%-75%). Significant levels are indicated by asterisks:
*
p<0.05 (significant)
**
p<0.01 (very significant)
***
p<0.001 (extremely significant)
4. Results I
FIGURE 4.9
75
RATIOS OF T CELL SUBTYPES TO CD3+ T CELLS PRESENT IN THE LESION AND LESIONADJACENT STROMA IN p16INK4a-NEGATIVE LOW-GRADE LESIONS COMPARED WITH
p16INK4a-POSITIVE LOW-GRADE LESIONS. The dot in the center of each box represents the median
value of the distribution; the borders of the box represent the upper and lower quartiles (25%-75%).
Significant levels are indicated by asterisks:
*
p<0.05 (significant)
**
p<0.01 (very significant)
***
p<0.001 (extremely significant)
76
4.2.3
4. Results I
T cell infiltrates in p16INK4a-positive high-grade CIN
The high-grade CIN (CIN2/3) were all p16INK4a-positive indicating true transforming HPV infection in
these lesions that furthermore probably have acquired secondary genomic alterations.
The comparison of T cell counts between high-grade CIN (CIN2/3, all p16INK4a-positive) and all lowgrade CIN (of which n=13 were p16INK4a-positive and n=9 were p16INK4a-negative) revealed that the
number of total T cells represented by CD3+ cells significantly was increased in high-grade lesions
compared with low-grade CIN in both the epithelium (p=0.0273) and the stromal compartment
(p<0.0001). This general increase is also reflected by the higher stromal infiltration of Foxp3+ T cells
(p=0.0076), the higher infiltration with GranB+ T cells in the epithelium (p=0.0028 for the epithelium
and p=0.0014 for the stromal), of CD8+ T cells (p=0.0012 for the epithelium and p<0.0001 for the
stroma) and also of CD3ζ+ T cells (p=0.0286 for the epithelium p=0.0022 for the stroma) (Figure
4.10, Table S9.1). With regard to the epithelial to stromal cell number ratios a trend for decreased
ratios was found for CD3+ cells (p=0.0799) and also CD3ζ+ cells (p=0.0672) in high-grade CIN
compared to low-grade lesions (Figure 4.10 and Table S9.1). Again, the ratios for all T cell
phenotypes to CD3+ T cell counts were calculated and were found to be significantly increased in
high-grade CIN for GranB+ T cells (p=0.0041) and for CD8+ T cells (p=0.0258) in the epithelium.
The ratio calculated for CD3ζ+ T lymphocytes showed the inverse correlation and tended to be
decreased in the stromal compartment of high-grade CIN (p=0.0700) (Figure 4.11 and Table S9.2).
Interestingly, the absolute T cell numbers but also the ratios calculated for different T cell subtypes to
CD3+ T cells are very heterogeneous in high-grade lesions (Figures 4.10 and 4.11) and within a
distinct histomorphological category (Table S9.1) and span wide ranges. Epithelial numbers for CD8+
T cell for example range from 3.7 to 32.3 cells per 0.0625mm² in CIN2 lesions. These enormous
variances can also be observed for epithelial Foxp3+ T cell numbers in CIN2 lesions ranging from 2.0
to 17.0 cells per 0.0625mm2 (Table S9.1).
4. Results I
FIGURE 4.10
77
DISTRIBUTION OF T CELL SUBTYPES IN DIFFERENT COMPARTMENTS IN LOW-GRADE
(LG) LESIONS COMPARED WITH HIGH-GRADE (HG) LESIONS. A) Absolute T cell counts per
0.0625mm² in the lesion and lesion-adjacent stroma. B) Ratio between the lesion and lesion-adjacent
stroma for all T cell phenotypes. The dot in the center of each box represents the median value of the
distribution; the borders of the box represent the upper and lower quartiles (25%-75%). Significant levels
are indicated by asterisks:
*
p<0.05 (significant)
**
p<0.01 (very significant)
***
p<0.001 (extremely significant)
78
FIGURE 4.11
4.2.4
4. Results I
RATIOS OF T CELL SUBTYPES TO CD3+ T CELLS PRESENT IN THE LESION AND LESIONADJACENT STROMA IN LOW-GRADE (LG) LESIONS COMPARED WITH HIGH-GRADE (HG)
LESIONS. The dot in the center of each box represents the median value of the distribution; the borders
of the box represent the upper and lower quartiles (25%-75%). Significant levels are indicated by
asterisks:
*
p<0.05 (significant)
**
p<0.01 (very significant)
***
p<0.001 (extremely significant)
T cell infiltrates in cervical carcinomas
With regard to the total T cell numbers the infiltration is even higher in cervical carcinoma samples in
comparison to high-grade CIN for most of the different T cell phenotypes (mean cell numbers, ranges
and standard deviations are shown in Table S9.1). Especially the stromal compartment showed an
enhanced T cell infiltration where significant differences compared to the high-grade lesions could be
found for the global T cell infiltration with CD3+ T lymphocytes (p=0.0414), GranB+ T cells
(p=0.0095) and also Foxp3+ T cells (p=0.0243). The higher total cell numbers were accompanied by
4. Results I
79
decreased epithelial to stromal ratio for GranB+ (p=0.0467) and Foxp3+ T lymphocytes (p=0.0464).
For the other cell types (CD3+, CD8+ and CD3ζ+ T lymphocytes) no significant differences in the
epithelial/stromal ratio could be observed (Table S13.x). Also most of the ratios calculated for all T
cell subtypes to CD3+ cell counts as a measure for the proportion of distinct cell phenotypes among all
T lymphocytes, were decreased in cervical carcinomas compared to high-grade CIN. The decrease was
significant for the intraepithelial CD8/CD3 ratio (p=0.0090) and the stromal CD3ζ/CD3 ratio
(p=0.0090), which represents the lowest CD3ζ/CD3 ratio of all stages. The only exception is the
significantly higher GranB/CD3 ratio (p=0.0418) in cervical carcinoma samples compared to highgrade lesions.
In summary, cervical precancerous lesions displayed generally increasing T lymphocyte densities with
worsening lesion grade from low-grade lesions to high-grade lesions and towards cancer. Thereby, T
cell densities in the transforming infection stage of low-grade CIN were not yet different form nontransforming CIN1 lesions. Although the increase of immune cell densities could be observed for
different T cell markers, the presence of regulatory T cells could be identified in all lesion stages and
is more pronounced in the stroma than in the epithelium. Based on the data shown in Table S9.1 an
increase from low-grade lesions (stromal mean cell density for both non-transforming and
transforming low-grade lesions together: 10.8 cells/0.0625 mm2) to CIN3 (mean 19.3 cells/0.0625
mm2) could be observed (p=0.0076). The Foxp3+ T cell density was further increased in invasive
cancer with a mean density of 42.1 cells/0.0625 mm2 compared with high-grade lesions (p=0.0243).
The ranges of densities were remarkable in all diseases stages with 0.0-20.0 cells/0.0625 mm2 in lowgrade lesions, 1.5-16.8 cells/0.0625 mm2 in high-grade lesions and 3.3-97.8 cells/0.0625 mm2 in
cervical cancers.
4.3
Alterations of human leukocyte antigen expression in
cervical intraepithelial neoplasia and cancers
As shown in section 4.2 there is a striking contradiction between high numbers of infiltrating
lymphocytes in high-grade cervical dysplasia and carcinomas indicating that immune cells are
attracted to the lesion site. However, these lesions obviously have progressed to finally become an
established and morphologically visible high-grade lesion demonstrating that despite the presence of T
cells in the microenvironment a certain number of already established high-grade lesions cannot be
completely eradicated and will further progress to become invasive tumors. High T lymphocyte
infiltration of both CD4+ and CD8+ T lymphocytes in association with cancer development has also
been observed in other tumor entities (HAN et al., 2014; MATKOWSKI et al., 2009).
These observations might imply that tumor cells under the immunoselective pressure evolve strategies
that provide protection from recognition and elimination by cytotoxic T cells (GARCIA-LORA et al.,
2001). Indeed, as adaption to the host’s immune system and in order to circumvent an immune attack
tumor cells are able to modulate the immune response by changing their own characteristics. One of
these changes represent the alteration of the expression and function of human leucocyte antigen
(HLA) class I and class II on the surface of tumor cells. In comparison with the modification of the
80
4. Results I
tumor microenvironment by changes in the cytokine milieu and immune cell composition, is a much
more immediate mechanism. In the context of HPV-associated diseases this might also be of
importance: transforming cervical lesions and carcinomas constitutively express the viral oncoproteins
E6 and E7 which could be degraded for antigen processing and subsequent presentation by HLA class
I molecules and might be recognized by effector cells such as cytotoxic T lymphocytes. As outlined in
chapter 1.4.2 alterations in antigen-presentation pathways might result in a less effective presentation
of viral and tumor-associated antigens and prevents the tumor from being recognized by the host T
cells.
HLA class I antigens are composed of a heavy chain (glycoprotein) which is encoded by genes within
the HLA regions of chromosome 6p (HLA-A, -B, -C) and a light chain (β2m) encoded by a gene
located on chromosome 15q. HLA class I antigens are normally expressed on all nucleated cells of the
body. HLA class II antigens are also heterodimeric molecules composed of an alpha and a beta chain.
This class of antigen-presenting molecules is usually expressed by professional antigen-presenting
cells of the immune system. Tumors of different origins have been reported to show altered human
leucocyte antigen expression which can be gradual and range from down-regulation to total loss of
classical HLA class I antigens but also gradual induction of de novo expression of HLA class II
antigens.
The project described in this chapter aims at the characterization of altered HLA class I antigen and
HLA class II antigen expression profiles in cervical intraepithelial neoplasia and cancers to answer the
question if these modifications might contribute to cervical carcinogenesis. Some reports on altered
HLA class I expression are conflicting and it remains still unclear whether HLA class I antigens are
completely lost during cervical carcinogenesis – suggesting a strong selection pressure for negative
cell clones, or whether their expression is only reduced – suggesting functional impairment, but
potentially enabling re-expression by drug intervention, vaccination or immune modulation.
Cervical lesions of different grade, CIN2 (n=9), CIN3 (n=13) and invasive squamous cell carcinoma
(SCC) samples (n=19) were analyzed by immunohistochemical staining for HLA class I antigen heavy
chains (HLA-A, HLA-B and HLA-C) and the light chain (beta-2-microglobuline, β2m) and also HLA
class II antigens in order to find out if these molecules are differentially expressed in increasing
histomorphological lesion grades.
4.3.1
Altered HLA class I antigen expression in cervical intraepithelial
neoplasia and cervical carcinoma
For the characterization of HLA class I antigen expression a panel of antibodies was used as described
previously (KLOOR et al., 2005) to determine the expression levels of the HLA class I heavy and light
chains separately. The monoclonal antibodies HC-10 and HCA-2 recognize different epitopes of the
HLA class I heavy chains: while HC-10 recognizes a determinant expressed on β2m-free HLA-A10,
HLA-A28, HLA-A29, HLA-A30, HLA-A31, HLA-A32 and HLA-A33 heavy chains and on β2m-free
4. Results I
81
HLA-B and HLA-C heavy chains the monoclonal antibody HCA-2 binds to a determinant expressed
on β2m-free HLA-A (excluding HLA-A24), HLA-B7301 and HLA-G heavy chains.
To determine the expression of the HLA class I light chain the monoclonal antibody L368 recognizing
β2m was used.
Importantly, HLA class I complexes are denatured by formalin fixation during the tissue processing
and dissociate into the heavy chain and the light chain. Therefore, it is not possible do detect intact and
functionally active HLA class I complexes. Thus only a combination of antibodies can allow the
distinction between free heavy chains or β2m molecules respectively and those assembled to HLAclass I heavy chains/β2m complexes transferred to and located on the cell surface. Membranous
localization of HLA heavy chains (A/B/C) indicated intact HLA class I complexes transferred to the
tumor cell surface. In contrast, altered expression or complete loss of membranous β2m staining and
disturbances in membranous HLA class I heavy chain staining is a sign for defects in the antigen
presentation pathway being either impaired or non-functional.
Lesions were classified as having normal, heterogeneous or negative HLA class I staining pattern
based on criteria summarized in Table 4.3.
staining pattern
% cells positive within
lesion/tumor
positive
strong, homogeneous overall expression
> 75%
heterogeneous
faint and patchy, weak overall expression
25-75%
absent or restricted to single cells (immune cells or locally
induced expression)
< 25%
score
negative
TABLE 4.3
SCORING SYSTEM FOR THE EVALUATION OF HLA CLASS I AND II STAINING PATTERNS.
Examples of staining pattern are shown in Figure 4.12 for the HCA-2 antibody.
FIGURE 4.12
REPRESENTATIVE HCA-2 STAINING PATTERNS OBSERVED IN CIN AND CERVICAL
CANCER SAMPLES (200x MAGNIFICATION). Shown are examples for A) positive staining (strong
and membranous) in normal, non-dysplastic epithelium, B) positive staining of invasive SCC, C)
heterogeneous expression pattern and D) invasive SCC with negative HCA-2 staining pattern.
82
4. Results I
Cytoplasmic and membranous staining of cells of the normal, non-dysplastic epithelium, precancerous
lesions and tumors was recorded separately and are summarized in Table 4.4. Representative staining
results for p16INKa and all HLA class I antigen markers are shown in Figure 4.13.
HLA class I heavy chain
HC-10 cytoplasm
HC-10 membrane*
HCA-2 cytoplasm
HLA class I light chain
β2m cytoplasm
HCA-2 membrane*
β2m membrane*
non-neoplastic epithelium
positive (%)
heterogeneous (%)
negative (%)
Samples analyzed
19
0
0
19
100.0%
0.0%
0.0%
19
0
0
19
100.0%
0.0%
0.0%
19
0
0
19
100.0%
0.0%
0.0%
15
4
0
19
78.9%
21.1%
0.0%
19
0
0
19
100.0%
0.0%
0.0%
19
0
0
19
100.0%
0.0%
0.0%
CIN 2
positive (%)
heterogeneous (%)
negative (%)
Samples analyzed
9
0
0
9
100.0%
0.0%
0.0%
9
0
0
9
100.0%
0.0%
0.0%
6
1
2
9
66.7%
11.1%
22.2%
1
3
5
9
11.1%
33.3%
55.6%
9
0
0
9
100.0%
0.0%
0.0%
9
0
0
9
100.0%
0.0%
0.0%
CIN 3
positive (%)
heterogeneous (%)
negative (%)
Samples analyzed
12
1
0
13
92.3%
7.7%
0.0%
9
3
1
13
69.2%
23.1%
7.7%
6
2
3
11
54.5%
18.2%
27.3%
2
3
6
11
18.2%
27.3%
54.5%
12
1
0
13
92.3%
7.7%
0.0%
8
4
1
13
61.5%
30.8%
7.7%
invasive SCC
positive (%)
heterogeneous (%)
negative (%)
Samples analyzed
16
2
1
19
84.2%
10.5%
5.3%
15
3
1
19
78.9%
15.8%
5.3%
8
4
5
17
47.1%
23.5%
29.4%
2
4
11
17
11.8%
23.5%
64.7%
16
3
0
19
84.2%
15.8%
0.0%
8
6
5
19
42.1%
31.6%
26.3%
TABLE 4.4
HLA CLASS I ANTIGEN EXPRESSION IN CIN2, CIN3 AND INVASIVE SCC. Data for HC-10 and
HCA-2 heavy chain antibodies and β2m are shown for the cytoplasmic and membranous separately.
Normal, non-dysplastic epithelium if present and analyzable on the same slide was characterized for
HC-10, HCA-2 and β2m staining patterns. In total, n=19 regions could be found that were adjacent to
CIN2 or CIN3 lesions. The normal epithelial regions showed positive staining in 100.0% of the cells
and also a clear membranous staining for all three antibodies.
In cervical precancerous lesions and cancers a high frequency of HLA class I alterations could be
observed.
All samples investigated for HLA class I antigen expression were p16 INK4a-positive. The staining
results for HC-10 showed that all CIN2 samples displayed normal expression in both cytoplasm and
membranous localization (100.0%). A heterogeneous membranous staining could be observed in 3 of
13 (23.1%) of CIN3 lesions and 3 out of 19 samples (15.8%) of invasive SCC. Lesions totally negative
for membranous HC-10 staining were rare and represented 1of 13 (7.7%) of CIN3 and 1 of 19 (5.3%)
of invasive SCC.
The HCA-2 staining demonstrated that heterogeneous or absent cytoplasmic staining occurred more
frequently in comparison with HC-10 antibody staining. Positive HCA-2 cytoplasmic staining could
only be observed in 6 of 9 (66.7%) of CIN2, in 6 of 11 (54.5%) of CIN3, and 8 of 17 (47.1%) of
invasive SCC samples. Conversely, heterogeneous expression and total losses were frequent:
regarding the membranous expression more than half of CIN2 (5 of 9, 55.6%) and CIN3 (6 of 11,
54.5%), and 11 of 17 (64.7%) of invasive SCC are negative for membranous HCA-2 staining.
4. Results I
FIGURE 4.13
83
EXEMPLARY STAINING RESULTS FOR ALL MARKERS IN A CERVICAL CANCER SAMPLE
(SCC) AT 200x MAGNIFICATION. Shown are the p16INK4a-staining and the slides stained for all three
HLA class I antigen markers (HC-10, HCA-2 and L368).
Regarding the staining for β2m the results demonstrated the vast majority of cervical precancers and
cancers are positive for cytoplasmic β2m (100.0% of CIN2, 92.3% of CIN3 and 84.2% of invasive
SCC). Heterogeneous expression is found in a small proportion of CIN3 and invasive SCC (7.7% and
15.8%) and none of the samples is negative for cytoplasmic β2m expression. Regarding the
membranous expression of β2m all CIN2 samples displayed normal expression (100.0%) while CIN3
and invasive SCC to a certain extent display altered membrane staining. However, still 61.5% of CIN3
and 42.1% of invasive SCC are positive for membranous β2m.
The correlation analyses between expression intensities (negative, homogenous and positive) and stage
of the disease showed that the HC-10 membranous staining was differently distributed between all
precancerous lesions (CIN2 and CIN3) and invasive cancers (SCC) with p<0.0001. CINs lesion more
often showed a positive staining (in 13/22 samples) while in SCC more often a heterogeneous staining
pattern could be observed (in 15/19 samples). Regarding the HCA-2 staining no differences between
the two groups could be shown for the membranous staining, but the overall cytoplasmic expression
was different between CIN and SCC: CIN lesions more frequently showed positive staining patterns
(in 9 out of 20 CINs), while 9 of 17 SCC samples were negative for HCA-2 staining (p=0.0005). For
membranous β2m-expression a strong trend towards more positive staining pattern in CIN (17 of 22
samples) in comparison to SCC samples (8 of 19 samples) could be observed. In contrast, SCC
samples showed a higher tendency to be negative for membranous β2m-expression (5 of 19 samples)
compared with CIN samples (1 out of 22).
84
4.3.2
4. Results I
Human leucocyte antigen class II expression in cervical intraepithelial
neoplasia and cervical cancer
HLA class II antigens are normally expressed on professional antigen-presenting cells (APCs), but
have also been reported to be expressed by distinct solid tumors (ALTOMONTE et al., 2003;
DENGJEL et al., 2006). The mechanisms involved in the expression of HLA class II antigens and
their role in the interaction of the tumor cells with the host’s immune system as well as the role of
immunoselection in HLA class II antigen loss are largely unknown. To investigate the role of HLA
class II antigen expression in the development of cervical intraepithelial neoplasia and progression
towards cancer, cervical lesions were stained with a monoclonal antibody against HLA class II chains
DR, -DQ, -DP (LGII-612.14).
The analysis was performed in the cohort used for the characterization of HLA class I antigen
expression. With CIN2 already displaying strong HLA class II antigen de novo expression the
question arose whether or not low-grade CIN (CIN1) also showed this expression pattern. To
explicitly address this question the cohort was enlarged by an additional set of CIN1 samples (n=19)
and a further subset of CIN2 samples (n=9). In parallel to the study of immune cell infiltrates in
different infection stages of low-grade CIN lesions (chapter 4.x) the HLA class II expression pattern
was correlated with the p16INK4a status of these additionally included lesions.
The same categories of staining patterns were applied as for HLA class I antigen staining (Table 4.3).
Examples of the different HLA class II antigen staining patterns are shown in Figure 4.14.
FIGURE 4.14
EXEMPLARY LGII.612-14 STAINING PATTERNS OBSERVED IN CIN AND CERVICAL CANCER
SAMPLES. Shown are A) normal, non-dysplastic epithelium which is negative for HLA class II antigen
expression B) the transition from adjacent normal epithelium to a CIN lesion with a strongly positive
staining pattern, C) positive invasive SCC and D) heterogeneous staining pattern in invasive SCC with
areas negative and positive for LGII.612-14 staining.
4. Results I
85
The results of HLA class II antigen expression were recorded separately for the cytoplasm and the
membranous localization and are summarized in Table 4.5.
Again, if normal non-dysplastic epithelium adjacent to the lesions was present, it was also analyzed
for HLA class II antigen expression (n=29). Positive staining was completely absent in the normal
stratified cervical epithelium or restricted to single cells in the epithelium only. However, HLA class II
antigen expression can frequently be detected in dysplastic epithelium as shown in Figure 4.14.
Interestingly, 15 of 18 investigated CIN2 samples showed cytoplasmic HLA class II antigen
expression (heterogeneous or positive staining) in the lesion and only 3 of 18 (16.7%) were negative
for staining with the LGII.612-14 antibody. Importantly, more than half of the CIN2 lesions (55.6%)
(10 out of 18 cases) displayed strong and positive HLA class II antigen expression. This suggests that
HLA class II antigen expression is a very common event during the initial steps of transforming HPV
infection.
HLA class II
LGII-612.14 cytoplasm
LGII-612.14 membrane*
0
0
29
29
0.0%
0.0%
100.0%
0
0
29
29
0.0%
0.0%
100.0%
positive (%)
heterogeneous (%)
negative (%)
Samples analyzed
6
4
9
19
31.6%
21.1%
47.3%
3
4
12
19
15.8%
21.1%
63.1%
positive (%)
heterogeneous (%)
negative (%)
Samples analyzed
10
5
3
18
55.6%
27.8%
16.6%
8
6
4
18
44.4%
33.3%
22.3%
positive (%)
heterogeneous (%)
negative (%)
Samples analyzed
7
4
2
13
53.8%
30.8%
15.4%
5
6
2
13
38.4%
46.2%
15.4%
invasive SCC
normal (%)
heterogeneous (%)
negative (%)
Samples analyzed
11
5
3
19
57.9%
26.3%
15.8%
10
6
3
19
52.6%
31.6%
15.8%
non-neoplastic epithelium
positive (%)
heterogeneous (%)
negative (%)
Samples analyzed
CIN 1
CIN 2
CIN 3
TABLE 4.5
HLA CLASS II ANTIGEN EXPRESSION IN NORMAL EPITHELIUM, CIN1, CIN2, CIN3 LESIONS
AND INVASIVE SCC. Data are shown for the cytoplasmic and membranous expression separately.
In CIN3 lesions 84.6% (11 out of 13 samples) of the lesions were found to be positive for HLA class
II antigens with more than half of them (53.8%) being strongly stained and considered positive. The
same trend could also be observed in invasive cancers. Here, a positive HLA class II staining pattern
could be observed in 11 of 19 cases (57.9%).
86
4. Results I
The observation that the majority of CIN2 lesions displayed HLA class II antigen expression prompted
the idea to characterize low-grade CIN1 lesions - included retrospectively - for the expression of HLA
class II antigens in order more precisely determine the time point of the induction of its expression.
Again, in a non-negligible proportion of samples (10 of 19, 52.7%) HLA class II expression could be
observed. In comparison with high-grade lesions (CIN2/3) and cancers, however, the percentage of
negative lesions was relatively high (47.35%).
As for the immune cell infiltrates (chapter 4.2) the low-grade lesions were stratified for their p16INK4astatus representing thus non-transforming (p16INK4a-negative) and transforming (p16INK4a-positive)
CIN1 in order to estimate a possible correlation of HLA class II with the biological infection stage.
FIGURE 4.15
REPRESENTATIVE STAININGS FOR THE CORRELATION OF (A) P16INK4A EXPRESSION AND
(B) HLA CLASS II ANTIGEN EXPRESSION IN LOW-GRADE CIN (CIN1). Shown are examples for
1) perfectly matching p16INK4a-positive areas with HLA class II positive regions 2) a p16 INK4a-positive
lesion that is HLA class II negative and C) a p16 INK4a-negative (focal p16INK4a-expression) that is HLA
class II positive.
Among CIN1 9 out of 19 (47.4%) were p16INK4a-negative and 10 out of 19 (52.6%) were p16INK4apositive. A possible association between p16INK4a expression reflecting the infection stage and HLA
class II antigen expression in CIN1 lesions could not be found when HLA class II antigen expression –
cytoplasmic or membranous – and the p16INK4a expression status in low-grade lesions were correlated.
4. Results I
87
This result confirmed the observations made during the microscopic evaluation with regard to the
occurrence of all possible combinations of p16INK4a expression with HLA class II antigen presence or
absence (Figure 4.15). The distribution HLA class II antigen expressing lesions among p16 INK4anegative and p16INK4a-positive CIN1 is shown in Table 4.6.
p16INK4a status
LGII-612.14 negative
LGII-612.14 positive
p16INK4a-negative
3/9 (33.3%)
6/10 (60.0%)
P16INK4a-positive
6/9 (66.6%)
4/10 (40.0%)
p=0.245
p-value
TABLE 4.6
DISTRIBUTION OF HLA CLASS II EXPRESSION IN p16INK4a-NEGATIVE AND p16INK4a-POSITIVE
CIN1 LESIONS.
The distribution of HLA class II antigen expression was also correlated to the grade of the disease by
comparing all precancerous lesions with the invasive cancer samples: no correlation could be observed
between the membranous HLA class II antigen expression and the disease stage represented by all
CIN lesions and invasive SCC samples (p=0.182). The comparison of single, unpooled CIN stages
(CIN1, CIN2 and CIN3 separately) and SCC samples with each other revealed that membranous HLA
class II antigen expression was significantly different lower in CIN1 lesions than all high-grade lesions
(CIN2, CIN3) and cancers (p=0.019).
In order to find out if there was a correlation between HLA class II expression and the alterations of
HLA class I antigen expression reported in the previous section (4.3.1) the samples that were initially
included in the study (CIN2, CIN3 and invasive SCC) before enlargement by CIN1 and further CIN2
samples and for which both staining data sets were available, were investigated. A significant
association between HLA class II and class I antigen expression was not observed. The presence and
absence of HLA class II expression was correlated with the HC-10 staining pattern (p=0.996) and
HCA-2 staining (p=0.532) and also β2m expression (p=0.361). A significant association between HLA
class II and class I antigen expression was not observed.
While the normal, non-dysplastic epithelium was negative for HLA class II staining, a strong and
uniform staining pattern was observed in glandular cells and the columnar epithelium of the
transformation zone of the cervix uteri (Figure 4.15 3B).
4.4
Immune
treatment
cell
infiltrates
under
immuno-stimulatory
It has been demonstrated that immune modulation by topical treatment with imiquimod, a TLRagonist, might positively influence the local immune response and lead to regression of dysplastic
lesions (TERLOU et al., 2010).
88
4. Results I
The efficacy of topical imiquimod treatment in patients with cervical intraepithelial neoplasia has been
tested for the first time in the frame of a phase I (double-blind randomized, placebo-controlled) trial
conducted in Austria (GRIMM et al., 2012). The treatment protocol and the clinical outcome of the
patients analyzed in the here presented study are summarized in Figure 4.16.
The patients included in the Austrian trial represent an exceedingly precious cohort. Although the
sample size is relatively small, the included biopsies represent a precious source of tissue of nonexcised lesions that were treated with an immuno-modulatory agent and observed for 20 weeks. This
cohort therefore provides highly important longitudinal information about the influence of immunomodulatory agents on the immune cell composition and the clinical behavior of these lesions.
FIGURE 4.16
4.4.1
SCHEME OF THE AUSTRIAN IMIQUIMOD TRIAL WITH TIMING OF THE OBTAINED PUNCH
BIOPSIES. Procedure is shown for the 10 patients of the imiquimod arm that were analyzed in the
presented study.
Characterization of the study cohort
In a cooperation project with the Medical University of Vienna, Austria samples of the above
described imiquimod trial could be obtained for immunological characterization. 10 patients with a
CIN2/3 diagnosis that had received a 16-week imiquimod treatment were included in the analysis each
providing cervical biopsies before (week 0), during (week 8) and after (week 20) treatment. Tissue
sections of the biopsies were stained for p16INK4a, CD3 and CD8. Image annotation and processing
were performed based on the method described in section 4.1 and blinded to the patient ID and the
clinical outcome. All patient related information at this stage of the analysis was subjected to
pseudonymisation except the histomorphological classification (lesion grades) as the lesion grade that
led to the diagnosis was needed for the definition of the region to be analyzed on the p16INK4a reference
4. Results I
89
slide as well as for the annotation of the slides stained with T cell markers. Once the immune cells
were quantified the clinical parameters were uncovered: 6 of the patients had regressing disease
(defined as CIN1 or less) and 4 of the patients had persistent disease or had even progressed (defined
as CIN2 or CIN3). The characteristics of all 10 patients are listed in Table 4.7.
patient
1
2
3
4
5
6
7
8
9
10
TABLE 4.7
4.4.2
week 0
(CIN grade)
CIN2
CIN2
CIN3
CIN2
CIN2
CIN2
CIN2
CIN 3
CIN2
not available
week 8
(CIN grade)
(no CIN)
CIN1
CIN2
no CIN
n.a
n.a
CIN1
CIN3
no CIN
no CIN
week 20
(CIN grade)
CIN3
no CIN
CIN1/no CIN
CIN1/no CIN
CIN2
CIN2
no CIN
CIN3
CIN1
no CIN
clinical outcome
progression
regression
regression
regression
persistence
persistence
regression
persistence
regression
regression
OVERVIEW OF THE CHARACTERISITICS OF THE PATIENTS SELECTED FOR THIS
APPROACH. All patients received a 16-week imiquimod treatment; n.a = not analyzable. Histologic CIN
grades were recorded to evaluate the treatment efficacy for CIN2/3 patients which was defined as
histologic regression of the initial high-grade lesions to histologically proven CIN1 or less (normal
epithelium).
T cell infiltrates in non-responders and responders to imiquimod
before treatment
Immune cell infiltrates, as CD3+ and CD8+ T cells, were quantified by application of the automated
quantification method presented in chapter 4.1. The lesions were annotated as described previously on
the basis of the p16INK4a reference slide. The areas for all regions of interest as well as the T cell
densities in these regions were calculated. Cell densities of both T cell phenotypes were compared
between patients that had persistent or progressing disease and did not respond to the imiquimod
therapy (“non-responders”) and patients whose lesions had regressed during the treatment
(“responders”). Densities for each phenotype separately as well as ratios of CD8+ T cells to all CD3+
cells in the different regions were compared in week 0 and week 20 biopsies.
Before treatment (week 0 biopsy) the infiltration with CD3+ T cells is very high in patients who did
not respond to the imiquimod therapy (progressing/persistent lesions) compared with patients whose
lesion had regressed after imiquimod therapy (Figure 4.17). The mean cell density of CD3+ T cells in
the epithelium of non-responders is much higher (537.0 cells/mm2) compared with responders (160.8
cells/mm2). However, these differences are not statistically significant (p=0.190). This trend can also
be observed in the stromal compartments where again non-responders had higher CD3+ T cell
90
4. Results I
numbers (1883.9 cell/mm2) compared with non-responders (945.9 cells/mm2) (p=0.190) (margin 500)
(supplementary Figures S9.1 and S9.2 and supplementary Table S9.3).
Interestingly, with regard to CD8+ T cell infiltrating the lesion and the stroma the densities are higher
in responders than in non-responders in week 0 before treatment is started. The mean cell densities for
CD8+ T cells in the epithelium of non-responders is 82.1 cells/mm2 compared with 113.8 cells/mm2 in
responders (p=0.730). The difference in the stromal compartment margin 100 is even more
pronounced (394.2 vs. 973.3 cells/mm2, p=0.286) (supplementary Figures S9.1 and S9.2). The same
trend could also be observed for the more distant stromal compartments and also for the CD8/CD3 cell
ratios in all regions of interest (supplementary Figures S9.1 and S9.2 and supplementary Table S9.4).
FIGURE 4.17
CD3+, CD8+ CELL COUNTS AND CD8/CD3 RATIO IN THE INITIAL BIOSPSY (WEEK 0) IN
NON-RESPONDERS AND RESPONDERS. Results are shown as Box-Whisker-Plots for A) the
epithelium and B) the first stromal compartment (margin 100). The line in the center of each box
represents the median value of the distribution; the borders of the box represent the upper and lower
quartiles (25-75%).
4. Results I
4.4.3
91
T cell infiltrates in non-responders and responders to imiquimod after
treatment
In the biopsies taken 4 weeks after the treatment (week 20 biopsy) CD3+ T cell densities are
comparably high in non-responders and in responders to imiquimod in the lesion and stromal
compartment (Figure 4.18 and supplementary Figures S9.3 and S9.4). For example, in the epithelium
the mean cell density is 287.8 cells/mm2 in non-responders and 371.1 cells/mm2 in responders
(p=0.429) (supplementary Table 9.3).
FIGURE 4.18
COMPARISON OF CD3+, CD8+ CELL COUNTS AND THE CD8/CD3 RATIO IN THE INITIAL
BIOSPSY (WEEK 0) AND THE LAST BIOPSY (WEEK 20) IN NON-RESPONDERS AND
RESPONDERS. Results are shown as box-whisker-plots for A) the epithelium and B) the first stromal
compartment (margin 100). The line in the center of each box represents the median value of the
distribution; the borders of the box represent the upper and lower quartiles (25-75%).
92
4. Results I
As the direct comparison between week 0 and week 20 shown in Figure 4.18 demonstrates the
assimilation of responders and non-responders in terms of T cell densities is caused by an increased
CD3+ T cells densities in responders compared to non-responders. This can be observed in the
epithelium and the stroma of responders is also reflected by the comparison of the mean cell densities
of week 0 and week 20 (see also supplementary Figures S9.6 and S9.7 and supplementary Table S9.3).
With regard to CD8+ T cells the direct comparison of both time points (week 0 and week 20) for nonresponders and responders revealed that CD8+ T cell densities also slightly increase over time in
patients responding to the treatment but not in non-responders (Figure 4.18 and supplementary Figures
S9.6 and S9.7). Non-responders in contrast show decreasing CD8+ T cell densities in week 20
compared with week 0 which results in a more pronounced difference between the groups at the end of
the treatment. In week 20 the CD8 mean cell density in the epithelium of non-responders is 58.2
cells/mm2 compared with 174.1 cells/mm2 in responders (p=0.643).
To get a better insight in how the T cell infiltrates develop during the treatment in the two groups, the
mean cell densities of every single patient at each time point is shown in a line chart and both groups
(non-responders vs. responders) were directly compared (Figure 4.19). This contrasting juxtaposition
revealed that the majority of the responders’ infiltrate densities is located above the highest value of
the non-responders’ T cell densities in week 20. However, the groups are different in the middle of
treatment were non-responders show an increase and responders a decrease in T cell densities.
Interestingly, these differences are completely reversed in the last weeks of the treatment until week
20. T cell densities in non-responders show a massive decrease while those of responders continuously
increase. The majority of responders therefore quit the treatment with clearly higher T cell densities
compared with non-responders
4. Results I
FIGURE 4.19
93
DEVELOPMENT OF CD8+ T CELL DENSITIES OVER TIME IN NON-RESPONDERS COMPARED
WITH RESPONDERS. Results for non-responders (red) and responders (green) are shown as line chart
for A) the epithelium and B) the first stromal compartment (margin 100). The dashed line represents the
highest count of non-responders in week 20.
94
5. Results II
5.
TREATMENT OPTIONS FOR
HPV-ASSOCIATED
PRECANCERS AND CANCERS
Despite important advances in the prevention of HPV infections and screening programs the worldwide incidence rates for cervical and other HPV-associated ano-genital precancerous lesions and
cancers are not expected to decrease significantly within the next 15 to 20 years. On the contrary, the
incidence is expected to increase in developing countries. The introduction of the prophylactic
vaccines was demonstrated to reduce the risk for HPV infections for young girls. However the
currently available vaccines provide protection against four HPV types of 14 considered to be
potentially carcinogenic. Although protection might be provided by herd immunity, this effect requires
a certain vaccination coverage and young women already infected with HPV do not necessarily benefit
from subsequent vaccination and still might develop cervical cancer twenty years later. Screening
programs based on Pap test in developed countries are well established, but getting women to attend
the cervical cancer screening in developing countries remains a major concern. In the light of all these
factors there is still a need for therapeutic intervention strategies. Different approaches are
conceivable, many of them are based on therapeutic vaccines based on RNA, DNA, peptides or fulllength proteins of diverse HPV-antigens.
In this part of the thesis, based on the insights that could be gained in the first part of this thesis, two
different intervention strategies involving immune modulation of the cancer environment will be
investigated. The first strategy aims at local application of a newly developed substance that might
enhance the local immune response by induction of inflammatory processes. In a second approach the
effect of regulatory T cell depletion on the efficiency of immune attack towards autologous tumor cells
shall be investigated.
5.1
Effects of TLR agonist treatment on immune cells
It has been shown in the past that TLR-agonists act as immune modifiers that, locally applied, can
positively influence the immune response and potentially reverse immune suppression. The substance
imiquimod is a well characterized immune stimulatory agent that is approved for the treatment of
condylomata accuminata, actinic keratosis and basal cell carcinoma, but is also tested in patients with
vulvar intraepithelial neoplasia. Within the scope of this thesis the potency of a new, secondgeneration immune modifier was evaluated. The substance called TMX-202 was obtained from
Telormedix SA, Bioggio, Switzerland and is a modified purine base derivative that is supposed to be
even more potent than actually available immuno-stimulatory agents such as imiquimod.
5. Results II
95
TMX also is a TLR7 agonist and was tested in vitro by measuring the effects on PBMCs of healthy
donors. It has been demonstrated in the past that TLR-9 agonist treatment increased the expression of
the corresponding TLR-9 on B cells (BOURKE et al., 2003). It is conceivable that the new TLR
agonist TMX also positively correlates with TLR expression on peripheral immune cells and thus
further enhances the innate and adaptive immune response by a positive feedback loop between
stimulation and activation of TLRs and their expression. To gain a better understanding of its mode of
action and its potency to induce immune responses the effects of TMX-202 on TLR7 mRNA and
protein levels were investigated (chapters 5.1.1 and 5.1.2).
The down-stream effect of TLR stimulation is the induction of inflammation that provokes the
attraction of further immune cells to the treated site and thus stimulates both the innate and the
adaptive immune response. To gain insight in the potency of the new TLR-agonist to induce
inflammation the cytokine release was measured (chapter 5.1.3).
As a long term goal, the TLR-agonist should be included in a combinatory drug composed of TMX202 and other immune modifiers that could be locally applied and thus is suitable for non-invasive
anogenital lesions.
5.1.1
The effect of TLR7 agonist treatment on the TLR7 mRNA expression
levels in PBMCs
To characterize the effects of the second-generation TLR7 agonist TMX on PBMCs a total number of
4 healthy donors were tested. Peripheral blood mononuclear cells were isolated from freshly drawn
blood and cultured for 72 hours in the presence of imiquimod, TMX or the vector control DMSO (as
described in 3.x). The expression of TLR7 was first measured on the transcript level by quantitative
real-time PCR. Possible effects of the substances on mRNA levels were compared between the
compounds. An additional negative control is represented by untreated cells that were frozen at day 0
and not subjected to in vitro culture. Furthermore, cells that were not treated with any substance but
cultured under the same conditions as those that received the treatment were included in the analysis.
For normalization purposes controls were included that were treated with the same amounts of DMSO
that were added with substance (dissolved in DMSO) to TMX-treated cells. Each treatment
experiment was normalized with the corresponding DMSO concentration in order to take into account
the effect of DMSO.
The mRNA levels in PBMCs that were frozen on day 0 before treatment was started were similar to
those of cultured, but untreated cells (data not shown). Therefore the values obtained for DMSOtreated cells were normalized against these untreated cells cultured under the same conditions. The
DMSO controls were then used to normalize the corresponding values obtained for PBMCs treated
with the immuno-modulatory agents by matching the DMSO concentrations used during stimulation.
In the first approach involving the first two donors, the effects of the new TLR7 agonist TMX-202 at a
concentration of 10 µM was compared with imiquimod at a concentration of 30 µM. This
concentration was reported previously in the context of immune cell in vitro vaccination approach
(FAHEY et al., 2009) while the TMX-concentration was based on preliminary in vitro data
communicated by Telormedix.
96
5. Results II
The results of the TLR7 quantitative real-time PCR for donors 1 and 2 (Figure 5.1) demonstrated that
imiquimod in both donors induced higher mRNA levels compared to the DMSO control. Donor 1
displayed high fold changes of TLR7 mRNA after treatment with both of the substances, imiquimod
and TMX-202, but mRNA expression was more up-regulated after TMX treatment. Donor 2 also
showed increased TLR7 mRNA after imiquimod treatment expression, while TMX treatment did not
show an effect on TLR7 mRNA levels. Here again, following imiquimod treatment higher foldchanges could be measured.
FIGURE 5.1
TLR7 mRNA EXPRESSION IN PBMCS TREATED WITH TMX AND IMIQUIMOD. Changes of
mRNA levels in comparison to the DMSO control are displayed on the y-axis (fold-change). The
experimental groups are displayed on the x-axis. The bars represent the results for the tested groups.
Two further donors were tested to compare the effects of TMX-treatment administered in different
concentrations. The 10µM dosage from the first experiment was compared with a reduced TMX-202
concentration (1µM). Furthermore, another aspect was investigated in this second experiment, as not
only the PBL fraction but also the adherent cell fraction representing mainly monocytes was analyzed
separately. Thus, changes in TLR7 mRNA levels were measured under two different TMX-202
concentrations separately for monocytes and lymphocytes (PBLs) (Figure 5.2). Donor 3 displayed
down-regulation of TLR7 mRNA expression in all cases except for the 1 µM concentration in the
monocyte fraction. Donor 4 showed a general TMX-induced up-regulation of TLR7 mRNA
expression levels compared with the corresponding DMSO controls. The 1µM dosage had a higher
effect on mRNA levels than the 10µM in both of the cell types, monocytes and PBLs.
FIGURE 5.2
TLR7 mRNA EXPRESSION IN MONOCYTES AND LYMPHOCYTES (PBLs) TREATED WITH
DIFFERENT TMX CONCENTRATIONS. Changes of mRNA levels in comparison to the DMSO
control are displayed on the y-axis (fold-change). The experimental groups are displayed on the x-axis.
The bars represent the results for the tested groups.
5. Results II
5.1.2
97
The effect of TLR7 agonist treatment on the TLR7 protein expression
in PBMCs
A second fraction of the same PBMCs that were tested for TLR7 mRNA expression levels was
subjected to TLR7 Western Blot in order to investigate the effect of TLR7 agonist treatment on the
protein level. As a positive control for TLR7 expression a B cell lymphoma cell line (Raji) was
included. Whole cell lysates of the same samples were tested, including the DMSO controls, the d0
uncultured PBMCs and PBMCs under treatment. As a loading control actin expression was
investigated.
The results for donors 1 and 2 treated with imiquimod and TMX-202 are shown in Figure 5.3.
The baseline TLR7 expression in uncultured and immediately stored PBMCs (d0) was difficult to
evaluate for both donors. No effects of any of the treatments (neither controls nor substances) could be
observed in donor 1. Donor 2 showed comparable TLR7 protein levels for the DMSO controls and
TMX, however, also less expression in imiquimod-treated cells.
FIGURE 5.3
TLR7 PROTEIN EXPRESSION IN PBLs TREATED WITH TMX AND IMIQUIMOD. Shown are the
results of the anti-TLR7 western blots of donor 1 (left) and donor 2 (right). TLR7 expression of treated
PBMCs is compared with uncultured control PBMCs (PBMCs d0), DMSO controls and the TLR7
positive control (Raji cells).
The results for donors 3 and 4 treated with two different TMX concentrations (1µM and 10µM) are
shown in Figure 5.4. Donor 3 showed slight baseline TLR7 expression and similar intensities of the
protein bands for the DMSO control and 1µM TMX. A strong signal for 10µM TMX treated PBLs
could be observed which might not be related to the treatment as a stronger signal can also be
observed for actin. Although lacking baseline TLR7 expression in donor 4 could be explained by very
protein concentration in the sample due to the lacking actin signal, the comparison between the highest
DMSO control (4 µl) and the TMX-treated samples revealed an induction of TLR7 protein expression
following treatment.
98
FIGURE 5.4
5. Results II
TLR7 PROTEIN EXPRESSION IN PBLs TREATED WITH TMX AND IMIQUIMOD. Shown are the
results of the anti-TLR7 western blots of donor 3 (left) and donor 4 (right). TLR7 expression of treated
PBMCs is compared with uncultured control PBMCs (PBMCs d0), DMSO controls and the TLR7
positive control (Raji cells).
In summary, the effects of different treatment approaches on the TLR7 expression on the protein level
that has been investigated in the PBMCs of 4 healthy individuals remained inconclusive. In most cases
no changes in protein expression could be observed – or could not definitively be related to the
treatment – and the observed protein expression was not concordant with changes in TLR7 mRNA
levels during treatment. The only exception is donor 4 who displayed higher protein levels for both
TMX concentrations compared to the DMSO controls. This is in concordance with the increase in
mRNA levels measured following treatment with TMX-202.
5.1.3
Release of the pro-inflammatory cytokine IL-6 of PBMCs upon
treatment with TLR7 agonists
Following the investigation of mRNA and protein levels induced by TLR agonist treatment, another,
more functional readout to investigate the effects of TMX-202 treatment was chosen based on the
quantification of interleukin (IL)-6 released by immune cells. IL-6 is a potent inducer of inflammation
and therefore indicative for the initiation of innate and adaptive immune responses. The supernatants
from PBMCs cultures that were treated with imiquimod, TMX-202 and the controls were tested in IL6 ELISA.
Although the effects of TMX treatment on mRNA and protein expression in the four tested donors
remained inconclusive, it could be shown by ELISA that the IL-6 release was consistently induced by
TMX treatment (Figures 5.5 and 5.6). The IL-6 release of stimulated PBMCs into the cell culture
medium was significantly higher than under DMSO control treatment. Massive IL-6 release was
induced with 1µM TMX compared with the DMSO control in donors 3 and 4 (p=0.0036 and
p<0.0001), but further increased in dose-dependent manner with 10µM TMX compared with the 1µM
TMX treatment (p=0.0002 and p=0.0004) (Figure 5.6). The IL-6 concentrations released under
imiquimod treatment in donors 1 and 2 did not exceed the IL-6 release measured in the DMSO
5. Results II
99
controls or untreated cells (Figure 5.5). Interestingly, in one donor (donor 4) DMSO equally induced a
slightly higher IL-6 release compared with the untreated control cells (p=0.0247).
FIGURE 5.5
IL-6 SECRETION BY PBMCS TREATED WITH IMIQUIMOD AND TMX-202. The IL-6
concentration (pg/ml) is presented on the y-axis. The experimental groups for donors 1 and 2 are
displayed on the x-axis. The colored bars represent the means for the tested groups, standard deviations
are shown as black whiskers (comparison by unpaired t-test, p-values are indicated).
FIGURE 5.6
IL-6 SECRETION BY PBMCS TREATED WITH DIFFERENT TMX-202 CONCENTRATIONS. The
IL-6 concentration (pg/ml) is presented on the y-axis. The experimental groups for donors 3 and 4 are
displayed on the x-axis. The colored bars represent the means for the tested groups, standard deviations
are shown as black whiskers (comparison by unpaired t-test, p-values are indicated).
100
5.2
5. Results II
Effects of TMX-202 treatment on the in vitro priming of
naïve T lymphocytes with HPV-associated and host cell
antigens and the generation of antigen-specific T cells
The potency of the new TLR agonist was investigated on a functional level in a large experiment
based on the in vitro priming of naïve T cells with HPV-related antigens that were loaded on dendritic
cells for antigen-presentation. This experimental setup allowed the effects of TMX-202 to be
investigated for both of the arms, the innate and the adaptive immunity. The final read-out of the
treatment, however, focused on the adaptive immune response as was evaluated by the potency of
stimulated T cells to kill tumor cells. This was measured in a heterologous system based on PBMCs of
a healthy HLA-A2 positive donor and CaSki cells. TMX-202 treatment was applied during the
complete procedure starting with the generation of dendritic cells from monocytes and continued
during the stimulation of T cells with the antigen-presenting cells until the end of the experiment. As
potentially relevant antigens in HPV-associated cancers p16INK4a, strongly overexpressed in HPVassociated tumors, and HPV16 L1, one of the most immunogenic HPV antigens, were chosen.
While for p16INK4a a peptide was available that has been demonstrated in previous experiments to bind
to HLA-A2 antigens, potential HPV16 L1 peptides had to be evaluated for their binding capacities to
HLA molecules in a T2 cell based peptide binding assay.
5.2.1
Determination of L1 peptides bound to HLA class I antigens with
high affinity for stimulation assays
In order to define out of a panel of predicted L1 peptides (for sequences see chapter 3.1.7, for
predicted peptide panel see supplementary Table S9.8) those that have the highest binding affinity to
HLA class I antigens and therefore being suitable for in vitro priming of T cells they were tested in
peptide-binding assay based on T2 cells. The mean fluorescence intensities (MFIs) for each peptide
were measured and compared with the negative and positive controls. As negative control served T2
cells incubated in absence of any peptide thus defining the baseline fluorescence intensity. To compare
the effect of beta2-microglobuline (β2m) on the MFI the negative control was performed with and
without β2m added to the culture. It could be shown that the addition of β2m to the cells, required for
stabilizing the complex built of HLA class I antigens and peptide, did not increase the MFI in absence
of any peptides (Figure 5.7). Peptides that were reported to have high binding affinities (L1_323) or
were evaluated before in the context of other experiments (p16_R1 and viral MP) were included to
obtain reference MFIs as positive controls. The values for all three positive controls (L1_323, p16_R1
and viral MP) were significantly higher than the negative control (Figure 5.7).
For the T cell in vitro priming the peptides with highest MFIs were chosen by applying the following
inclusion criteria: Only peptides that fulfilled two distinct criteria, having significantly higher MFIs
compared with the negative control and with a MFI at least as high as the positive control with the
lowest MFI. The L1-peptides L1_2, L1_12 and L1-97 had MFIs that were significantly higher than the
5. Results II
101
negative control (all p<0.0001). Furthermore, the MFIs of the L1 peptides were significantly higher
than the control peptide with the lowest MFI which was viral MP (Figure 5.7).
FIGURE 5.7
MEAN FLUORESCENCE INTENSITIES (MFIs) MEASURED FOR DIFFERENT HPV16 L1
PEPTIDES IN A T2-CELL BASED PEPTIDE BINDING ASSAY.
The peptide binding assay was repeated once and the result obtained in the first assay could be
confirmed. Again, the peptides L1_2, L2_12 and L1_97 were revealed to be the best binding ones and
therefore chosen for subsequent T cell in vitro priming (supplementary Figure S9.8).
5.2.2
The effect of TMX treatment on dendritic cell maturation
The generation of antigen-specific T lymphocytes was based on an autologous system that involved
antigen-presenting cells of the same healthy donor from whom also T cells were obtained. Being the
most potent antigen-presenting cells, dendritic cells (DCs) were generated from the adherent PBMC
fraction, the monocytes, under the influence of a basic cytokine cocktail including GM-CSF and IL-4.
To test the potency of the immune modulatory agent TMX on the innate immune system, involving
maturation of dendritic cells from monocytes, and also on the adaptive immunity in terms of
interacting with T cells and priming them towards the chosen antigens, TMX was added to the
dendritic cell culture. Following the standard protocols for dendritic cell generation from monocytes
the cells require a “maturation cocktail” consisting of different pro-inflammatory cytokines including
IL-1β, TNF-α, PGE-2 and IL-6. As TMX leads to IL-6 secretion of peripheral immune cells creating a
strongly pro-inflammatory milieu as shown in section 5.1.3 one could hypothesize that TMX treatment
also might have an influence on dendritic cell maturation and that the endogenous IL-6 production
could replace the exogenously added cytokine cocktail. The effect of TMX on monocytes and
generation of dendritic cells was evaluated by the cell counts obtained after dendritic cell culture,
morphology of the growing cells and expression of co-stimulatory molecules CD80 and CD86 on
dendritic cells which is a sign for DC maturation.
102
5. Results II
The cell counts of harvested monocytes and dendritic cells – although varying between different
cycles of DC generation - demonstrate that the numbers of harvested cells depends on the treatment.
Cell numbers were calculated as the percentage of full PBMCs that could be harvested after 6 days
culture period. Obviously the number of monocytes that became adherent and thus were separated
from the non-adherent lymphocytes varied from one generation cycle to another. However, out of the
cells that initially became adherent, more cells could be obtained after TMX stimulation compared
with DMSO controls. The difference was most pronounced after the second and third round of
dendritic cell generation with a 1.8- and 2.1-fold increase in cell numbers (Figure 5.8).
FIGURE 5.8
CELL NUMBERS OBTAINED DURING THE FOUR DENDRITIC CELL GENERATION CYCLES.
The percentage of dendritic cells that could be harvested from total PBMCs subjected to adherence for
monocyte isolation is displayed on the y-axis. The 4 cycles of DC generation are shown on the x-axis
with the bars representing the different tested groups (DMSO and TMX).
Also, the morphology of monocyte culture is indicative for the maturation of dendritic cells: while
newly adhered monocytes are regular and round, growing and maturing dendritic cells display the
typical, longish and branched, dendrite-like morphology. The cultures that obtained TMX treatment in
comparison with the DMSO controls showed faster, at an earlier time point, and to a higher extent
cells with a dendrite-like morphology. The morphologic changes became obvious 48 hours after
treatment with TMX had started and could be observed in more cells than in the culture containing
DMSO treated cells. After 96 hours under TMX treatment the monocyte culture displayed clear
morphologic signs of dendritic cells. Still, these cells were more frequent than in the DMSO-treated
culture (Figure 5.9). These effects could be observed in all 4 successively established DC cultures,
independently of the cell density and the rate of yield of monocytes from full PBMCs.
5. Results II
FIGURE 5.9
103
REPRESENTATIVE MICROSCOPIC IMAGES OF THE MORPHOLOGY OF DENDRITIC CELLS
GENERATED FROM MONOCYTES UNDER THE INFLUENCE OF CONTROL SUBSTANCE
DMSO (A) AND TMX (B). Shown are examples for 1) early dendritic cell culture (48h) at 20x
magnification and 2) a later time point of dendritic cell generation (96h) at 40x magnification.
The expression of CD80 and CD86 is indicative for activated antigen-presenting cells – B cells and
monocytes. They are co-stimulatory molecules that bind to CD28 and CTLA-4, which are the
corresponding ligands on T cells. CD80 and CD86 together play an important role in T cell activation
and priming towards distinct antigens. They are up-regulated during the activation of monocytes and
maturation of dendritic cells (CD86 is a marker for early maturation, while CD80 is a marker for
mature DC). While morphology and cell numbers were recorded for all DC cultures the expression of
co-stimulatory molecules could only be investigated in one out of 4 DC cultures because there was not
a decent amount of cells available in the other cycles. The cell numbers were limited and in most cases
all available DC had to be used for the T cell stimulation to assure the ratio of 1:10 between antigenpresenting cells and T cells. FACS analysis of the available DCs revealed that culturing monocytes in
presence of TMX in comparison with DMSO treatment leads to higher expression of CD80 (28.8% vs.
19.92%) and CD86 (41.95% vs. 28.8%). The results are shown in Figures 5.10 and 5.11.
104
5. Results II
FIGURE 5.10
RESULTS OF THE FACS ANALYSIS FOR CD80 EXPRESSED ON DENDRITIC CELLS. The results
are shown for DC generation under DMSO treatment (top) and TMX treatment (bottom). The
fluorescence intensities for CD80 are given on the x-axis. Region borders (R2) were defined based on the
isotype control with the FI for mouse IgG1 given on the x-axis. The percentage of cells that are CD80+ is
given in R2.
FIGURE 5.11
RESULTS OF THE FACS ANALYSIS FOR CD86 EXPRESSED ON DENDRITIC CELLS. The results
are shown for DC generation under DMSO treatment (top) and TMX treatment (bottom). The
fluorescence intensities for CD86 are given on the y-axis. Region borders (R2) were defined based on the
isotype control with the FI for mouse IgG2b given on the y-axis. The percentage of cells that are CD86+
is given in R2.
5. Results II
5.2.3
105
The effect of TMX treatment on stimulation of naïve T cells with
HPV-associated antigenic peptides
The priming of naïve T cells with peptide-loaded dendritic cells was carried out in 4 cycles over 24
days. Functional analyses during the stimulation period were not possible as T cell numbers were
limited and all available cells were used for the final killing assay.
However, the appearance of the T cells in culture and their morphology was recorded. Also, after each
stimulation cycle that has been completed, T cell numbers were determined upon harvesting and
reseeding cells with newly generated dendritic cells. From the photos taken of the T cell cultures
(Figure 5.12) it becomes obvious that, although the same T cell numbers were initially seeded, T cells
under TMX developed differently from those treated with DMSO only. On day 10 of the stimulation T
cells that were treated with DMSO were less dense compared with the TMX-treated T cells (Figure
5.12. 1A and 1B). Although they seemed to recover until day 21 they still appeared to be less close to
each other and more scattered over the well than the T cell culture treated with TMX (Figure 5.12 2A
and 2B).
FIGURE 5.12
APPEARANCE OF T CELLS DURING STIMULATION WITH PEPTIDE-LOADED DENDRITIC
CELLS UNDER THE INFLUENCE OF CONTROL SUBSTANCE DMSO (A) AND TMX (B). Shown
are examples for 1) an earlier time point of T cell priming (day 10) and 2) a later time point of T cell
stimulation (day 21) at 20x magnification.
The morphologic appearance of the T cell cultures was confirmed by the cell numbers recorded upon
harvesting and re-stimulation. Figure 5.13 demonstrates the development of T cell numbers over time
during the stimulation. While T cells stimulated under TMX treatment with TMX-generated DCs
continuously grew until day 17, T cells numbers under DMSO conditions decreased until day 11.
Nonetheless, they recovered until day 21 and finally both cultures were harvested with more than
3x106 cells and thus globally showed a positive growing tendency.
106
FIGURE 5.13
5.2.4
5. Results II
DEVELOPMENT OF T CELL NUMBERS DURING THE IN VITRO PRIMING. Shown are the cell
numbers for T cells stimulated in presence of TMX and in presence of the control substance DMSO.
The effect of TMX treatment on the killing potency of stimulated T
cells against CaSki cells
The final read-out of the T cell in vitro priming was the killing assay of CaSki cells in a heterologous
tumor cell – immune cell system. To minimize the reactivity of T cells against tumor cells due to HLA
mismatching, a PBMC donor expressing the HLA-A2 allele was chosen.
The reactivity of T cells stimulated with peptides against L1 and p16INK4a and cultured either under
TMX or DMSO treatment was compared. Therefore they were co-incubated with tumor cells and the
degranulation rate as measured by CD107a expression on the cell surface was evaluated.
First, the gate for T cells was defined by T cell cultured alone. Its suitability was also checked for T
cells that were co-incubated with CaSki cells (Figure 5.14). Then two gates containing CD8+CD107a+
T cells (R2) and the total fraction of CD8+ cells irrespective of CD107a expression (R4) were defined.
The percentage of cells that upon co-incubation with tumor cells expressed CD107a on their cell
surface and that simultaneously expressed CD8 (cytotoxic T lymphocytes) were higher in the T cell
culture that had undergone a treatment with TMX compared with the cells that were treated with
DMSO (p=0.1353). In a second step the degranulation rate of CD8+ T cells measured upon coincubation with CaSki cells was calculated by diving the fraction of CD8+/CD107+ T cells by total
amount of CD8+ T cells measured in the corresponding well. The percentage of CD107a-expressing
CTLs among all CD8+ T cells again tended to be higher in TMX-treated T cells (8.26%) compared
with DMSO-treated T cell culture (6.75%) (p=0.2202) (Figure 5.15).
In conclusion, a slightly higher CD107a release could be obtained by TMX-treatment compared with
untreated cells. This is true for CD8+CD107a+ T cells and the fraction of CD107a+ T cells among all
CD8+ T cells (degranulation rate of CD8+ T cells).
5. Results II
107
FIGURE 5.14
EXEMPLARY RESULTS OF THE FACS ANALYSIS FOR CD8 AND CD107A. The gating strategy in
the FSC/SSC is shown for A) T cells and B) T cells with CaSki cells. One of the duplicates is shown for
C) the co-incubation of T cells with CaSki cells under DMSO treatment and D) TMX treatment. The
fluorescence intensities for CD8 (x-axis) and CD107a (y-axis) are given. Region borders were defined
based on the isotype controls (not shown). The percentage of cells that are CD8+CD107a+ is given in R2
and the percentage of CD8+ cells in R4.
FIGURE 5.15
EVALUATION OF THE CD8+ T CELLS FOR THE DEGRANULATION MARKER CD107A AND
DEGRANULATION RATE. The percentage of positive cells is presented on the y-axis. The
experimental groups are displayed on the x-axis. The blue bars represent the results for the tested groups,
standard deviations are shown as black whiskers (comparison by Student’s t-test, p-values are indicated).
108
5.3
5. Results II
Establishment of an autologous system for the development
and evaluation of therapeutic intervention strategies in
HPV-associated diseases
The previous results (chapter 4.2) demonstrated that regulatory T cells might play a role in the
carcinogenesis of cervical cancers and that immuno-modulatory treatment might reverse the
immunosuppressive state of the host’s immune system and lead to better killing of cervical cancer
cells (CaSki) (chapter 5.2.4). Other strategies, such as cell-based approaches, might also be of
importance in the battle against HPV-cancers and will be considered in this thesis. For the
investigation of immunological questions autologous models based on tumor cells and immunocytes
deriving from the same donor are of special interest as they provide advantages in terms of avoidance
of cross-reactivity and cytotoxicity due to unmatched HLA allelic phenotypes. However, autologous
HPV-associated tumor models for the cervix as well as for other sites are lacking. One major part of
this thesis therefore was to establish a HPV-positive tumor cell line for these purposes. This was based
on tumor samples of head and neck squamous cell carcinoma (HNSCC) patients that could be
obtained from collaboration partners of the University Hospitals Giessen and Muenster. As cervical
cancer and HPV-positive HNSCC have the same underlying mechanisms of tumorigenesis, HNSCC
tumor might also function as a reliable model for HPV-related diseases.
In the course of this project, one HNSCC cell line from a HPV-positive patient could be established
and used for further immunological studies.
5.3.1
The cell line HN038M: general features and patient’s characteristics
In the course of this project tumors samples of 31 HPV-positive HNSCC patients, primary tumors
together with or without the corresponding metastatic lymph nodes, were obtained. The tissue was
prepared and cultured as described in section 3.2.4. After many attempts, one cell line out of these 31
primary cultures could successfully be established by explant culture. This cell line derives from a
lymph node metastasis of a male patient who underwent his first surgery in March 2013.
The patient’s and the tumor’s characteristics as well as the clinical course of the disease are
summarized in Table 5.1.
The tissue of the primary tumor and the metastasis was prepared as described in section 3.2.4.
Following the enzymatic digestion of the tissue two tumor explant cultures were initiated, one
containing the primary tumor cells and the other containing the metastatic cell material. Regular
microscopic evaluation showed that the culture containing the primary tumor cells in contrast to the
metastasis seven weeks after tumor preparation still did not contain adherent and growing tumor cell
cluster and therefore was discarded.
5. Results II
Parameter
Description
Sex
male
Age
58 years at diagnosis (march 2013)
exposure to noxa
heavy smoker
109
primary tumor
HPV-association
localization
Size
p16INK4a status as determined before surgery: positive
oropharyngeal cancer of the palatine/lingual tonsil
resected mucosal tissue (7,5 x 5,5 x 1,5 cm3), with a ulcerous area
of about 2,1 x 1,5 cm2 in the center of the tissue
cTNM staging
cT3, cN2b, cM0
pTNM staging
pT2, pN2b (14/18), L1, V0
10 of 11 lymph nodes on the right side affected (level IIb)
metastatic LN
4 of 7 lymph nodes on the right side affected (Level V)
grade of malignity
G2
R0
ICD-O code
8070/3
further clinical course
recurrent disease, relapse within one year: detection of multiple
metastases
beginning of January 2014
end of January 2014
TABLE 5.1





second surgery: macroscopically recurrent disease could not
be observed; removal of a lymph node conglomerate
lymph node metastasis could be identified
partially necrotic tissue, moderately differentiated (G2)
squamous cell epithelium
ICD-O-Code: 8070/6




third surgery with removal of further lymph nodes
in 1/25 “metastasis of the known primary tumor”
poorly differentiated (G3)
ICD-O-Code: 8070/6
OVERVIEW OF THE MAIN CHARACTERISITCS AND THE CLINICAL COURSE OF THE
PATIENT FROM WHOM THE CELL LINE IS DERIVED.
The explant culture of the metastasis after 3 weeks has already shown macroscopically and
microscopically detectable tumor cell clusters within the fibroblast layer (Figure 5.16 A). After the
fibroblasts had undergone apoptosis, the tumor cell nests remained stably attached to the cell culture
flask. However, they did not further expand across their initial “borders” determined by the outer cells
and, although cells proliferated, only the minority of the newly generated cells adhered to the free
space of the bottom of the flask (Figure 5.16 B). In this state the tumor cells remained stable over 11
months. The culture was subjected to repeated trypsinization in order to detach the cells from the
bottom and allow them to adhere again but in a more homogeneously distributed pattern.
110
5. Results II
Finally, cell proliferation and adherence of newly generated cells to the flask could be stimulated by
this treatment (Figure 5.16 C and D). The culture after 13 months became 90% confluent, could be
split and analyzed by FACS staining and cytometry analysis in order to determine the content of
epithelial cells (Figure 5.17).
FIGURE 5.16
MORPHOLOGIC APPEARANCE OF THE CELL LINE HN038M. Shown are A) an initial tumor cell
nest (arrow) embedded in fibroblasts (week3), B) expanded tumor cell nest (month 8) and C) tumor cells
of the established cell line at 100x magnification and B) at200x magnification.
The analysis revealed that the culture contained ~ 99% of BerEP4+ cells, a marker for epithelial cells
that have been stable for more than 13 months and still proliferate autonomously. The FACS results
could be confirmed several times and the tumor cells were subjected to further characterization which
is described in sections 5.3.2 to 5.3.4
To date the culture is stable, continuously growing and has until now undergone 43 passages.
FIGURE 5.17
EXAMPLARY RESULTS OF THE FACS ANALYSIS FOR BEREP4 OF THE NEWLY GENERATED
TUMOR CELL LINE. One of the replicates of tumor cells harvested at confluence (passage x) is shown.
The fluorescence intensity (FI) for BerEP4 is given on the x-axis. Region borders (R2) were defined
based on the isotype control with the FI for mouse IgG1 given on the x-axis. The percentage of cells that
are BerEP4+ is given in R2.
5. Results II
5.3.2
111
Determination of the HPV-status and oncogene activity
To further characterize the established cell line and to validate the clinical finding in terms of HPVassociation of the tumor, the formalin-fixed paraffin-embedded tissue of the metastasis was ordered to
compare the characteristics of the tumor cell line with the archived tumor material. Therefore, tissue
sections were stained for p16INK4a to confirm the original diagnosis of the pathologist. The original
FFPE material of the lymph node metastasis showed a strong and diffuse staining for p16INK4a (Figure
5.18).
FIGURE 5.18
p16INK4a IMMUNOHISTOCHEMISTRY OF FORMALIN-FIXED PARAFFIN-EMBEDDED
METASTATIC TUMOR TISSUE OF THE PATIENT FROM WHOM THE CELL LINE IS DERIVED.
Shown is A) on overview of the lymph node metastasis at 20x magnification and B) details at 40x
magnification (p16INK4a-positive tumor is marked by an arrow).
In order to assure that the cultured cells still have this feature equally and were not selected for
p16INK4a-negative cell clones, p16INK4a cytology staining on cultured cells was performed. Therefore,
tumor cells were harvested and spun down onto a microscopy glass slide. The p16 INK4a staining for
cytological preparations revealed that virtually all cells contained in the sample strongly stained for
p16INK4a indicating viral oncogene activity (Figure 5.19 A,B).
FIGURE 5.19
p16INK4a CYTOLOGY OF THE HNSCC CELL LINE HN038M AND HPV DNA STATUS
VISUALIZED BY GP5+/6+ LUMINEX PCR. A) and B) Staining of tumor cells of the HN038M cell line
(passage 14) shows a clear p16INK4a-staining (brown signal). C) Agarose gel following GP5+/6+ PCR
shows amplification of HPV DNA in the HN038M tumor cells and in the positive controls (HeLa and
SiHa) but not in the negative controls.
112
5. Results II
In order to proof the underlying HPV-infection in the cells and the oncogene activity the tumor cells
were subjected to HPV-genotyping and viral oncoprotein expression of HPV16 E7.
The GP5+/6+ PCR demonstrated that HPV DNA was amplified (Figure 5.19 C) and the subsequent
Luminex-based HPV genotyping revealed that the sample was positive for HPV16 DNA. The HPV
status was also compared with the original FFPE tissue samples of the primary tumor and the
metastases to validate these findings. HPV genotyping demonstrated that the archived tumor material
also harbored HPV16 DNA (supplementary Table 9.9).
In order to examine whether p16INK4a overexpression was linked to viral oncogene activity, the viral
oncogene expression was investigated by western blot analysis for HPV16 E7 expression.
For the characterization of the protein expression the viral oncoprotein E7 was investigated. Samples
of different subcultures that have undergone varying numbers of passages (7 and 16 passages) were
analyzed for viral oncoprotein E7 expression and compared with each other. They were also compared
with HPV16-positive SiHa cells which were used as positive control for HPV oncoprotein expression.
As shown in Figure 5.20 the tumor cell line HN038M displayed a strong staining for the viral
oncoprotein E7 (located at 17 kDa) at earlier passages as well as at a later time point when the tumor
cell had undergone more passages.
FIGURE 5.20
WESTERN BLOT ANALYSIS OF DIFFERENT FRACTIONS OF THE HNSCC CELL LINE HN038M
FOR HPV16E7. Tumor cells of different passages (passage 16 and passage 7) were tested and compared
with HPV16-positive cell line SiHa used as control.
The tumor cell line was further characterized for HLA class I antigen expression, which is an
important factor for immunological studies. Expression of HLA class I antigens is the prerequisite for
the recognition of cells by T cells and therefore required for by CD8+ T cells.
Tumor cells of the cell line HN038M were characterized for HLA class I expression by flow
cytometry analysis. It could be demonstrated in two independent experiments that virtually all cells
were positive for HLA class I antigens (Figure 5.21).
5. Results II
FIGURE 5.21
5.3.3
113
REPRESENTATIVE RESULTS OF THE FACS ANALYSIS FOR HLA CLASS I ANTIGENS OF THE
NEWLY GENERATED TUMOR CELL LINE. One of the replicates of tumor cells harvested at
confluence (passage x) is shown. The fluorescence intensity (FI) for HLA class I antigens is given on the
x-axis. Region borders (R2) were defined based on the isotype control with the FI for mouse IgG1 given
on the x-axis. The percentage of cells that are HLA class I positive is given in R2.
Cell line validation via short-tandem-repeat profiling
The detection of misidentification of standard cell lines and the increasing awareness of the danger for
cross-contamination, the proof of authenticity of established and newly generated cell lines that are
used in experiments has become indispensable. Short-tandem-repeat (STR) profiling is a DNA
fingerprinting method based on the characterization of hypervariable DNA sequences, so called
microsatellites, and recommended for cell line authentication. It allows the determination of a unique,
cell-line specific profile based on 8 different STR loci. The comparison with database comprising all
characterized and registered cell lines allows to authenticate the cell line and to exclude crosscontamination with other cell lines.
Cell line authentication was carried out by Multiplexion GmbH, Heidelberg. STR profiling and
comparison with database revealed that the newly generated HNSCC cell line HN038M has a unique
sequence, showing only 90% identity with already known cell lines (less than 96% identity is defined
as a cell line being not identical with the compared “best hit” cell line). The search for the best hit
among cell lines registered in the database identified the cell line UACC-257. This is a melanotic
melanoma cell line of non-epithelial origin which is not in use in our laboratory. The established
HNSCC cell line has not been present in database to date and shows a genotype code that is unique to
this cell line and does not match to any of the cell lines contained in the database.
best database hit
identity
genotype code
UACC-257
90%
AATTAAAAAATTAAAAATAAAWA
TTTTTTTAAWTWTATTTAATTATWT
(W= uncertain signal)
TABLE 5.2
CHARACTERISITICS OF THE TUMOR CELL LINE HN038M.
114
5. Results II
All these characteristics revealed by cell line characterization exclude cross-contamination of the
primary culture with additional cells from other cell line (established cell lines). In conclusion, the
identity of the cell line was confirmed with a unique sequence being revealed for the sample.
Furthermore, the characteristic genotype code, which represents a 48-letter code for 24 single
nucleotide polymorphism (SNP) locations, was identified. The main characteristics are summarized in
Table 5.2, for more detailed information provided by the company see also supplementary Figure 9.9.
As the newly generated cell line is currently not present in the Multiplex Cell Authentication (MCA)
database (CASTRO et al., 2013) and does not show identity with any other cell lines reported therein,
the novelty could be proofed and cross-contamination was excluded.
5.4
Effect of regulatory T cell depletion on the cellular immune
response against autologous tumor cells
The presence of regulatory T cells in low-grade lesions and their increasing frequencies in high-grade
lesion and invasive cervical cancer (chapter 4.2) is a hint for the role they play in all steps of cervical
carcinogenesis. Their contribution to tumor progression and metastasis and the resulting poor
prognosis for patients has been, apart from cervical cancer, also been demonstrated in other tumor
entities (reviewed in HALVORSEN et al., 2014). With the availability of the above described
autologous model system that could successfully be established the idea was prompted to test the
immunosuppressive effects of Tregs in vitro and measure the cell-mediated cytotoxicity in presence
and absence of Tregs. Therefore, peripheral blood lymphocytes could be obtained from the patient
who gave rise to the cell line that were subjected to Treg depletion and used for the killing assay.
5.4.1
T cell purity and Treg depletion
The efficiency of Treg depletion was monitored by flow cytometry analysis by comparing the total
(undepleted) T cell fraction with the T cells following Treg depletion. The results are shown in Figure
5.22 and demonstrate that depletion of CD25+ T cells by magnetic labelling decreased the amount of
CD4+CD25+ T cells from 1.93% in undepleted T cells to 0.75% in Treg depleted T cells.
5. Results II
FIGURE 5.22
115
RESULTS OF THE FACS ANALYSIS OF CD4+CD25+ T CELLS CONTAINED IN THE T CELL
FRACTIONS USED FOR CD107a DEGRANULATION ASSAY BEFORE AND AFTER MAGNETIC
DEPLETION OF TREG CELLS. The gating strategy in the FSC/SSC is shown in the upper part of the
figure and was applied for both T cell fractions. The frequencies of Tregs before (total T cells) and after
Treg depletion are shown in lower part of the figure. The fluorescence intensities for CD4 (x-axis) and
CD107a (y-axis) are displayed. The percentage of cells that are CD4+CD107a+ are given in R5.
5.4.2
Characterization of the effect of Treg depletion on the killing potency
of autologous T cells against the tumor cell line HN038M
The cytotoxic effect of T cells against tumor cells that were either depleted from regulatory T cells or
not was measured by CD107a expression on the cell surface as described in section 3.x. CD107a
degranulation in T effector cells is induced upon recognition of and activation by tumor cells. As the
Treg depletion via magnetic labelling (chapter 3.2.4) targets the CD4+ T cell population of T cells
isolated from PBMCs the killing effect also was measured by analyzing the CD4+ T cell population.
Although an additional staining for CD8+ T cells was not possible due to restricted cell numbers, the
fraction of non-CD4+ T cells can be considered to reflect effects of CD8+ T cells. The gating was
performed on T cells as shown in chapter 5.2 and the defined gate was then also checked for samples
consisting of T cells co-incubated with tumor cells. T cells were analyzed by plotting CD107a
expression against CD4 expression and defining a gate for CD107a+ cells among the CD4+ and nonCD4+ T cells respectively which represented two clearly distinguishable cell populations (Figure
5.23). The values for CD4+CD107a+ T cells were obtained by applying the same gates for all
samples.
116
FIGURE 5.23
5. Results II
RESULTS OF THE FACS ANALYSIS FOR CD4 AND CD107A. One of the duplicates is shown for the
co-incubation of autologous tumor cells with total T cells (left) and Treg depleted T cells (right). The
fluorescence intensities for CD4 (x-axis) and CD107a (y-axis) are given. Region borders were defined
based on the isotype controls (not shown). The percentage of cells that are CD4+CD107a+ are given in
R2 and the percentage of CD4-CD107a+ cells in R3.
The results obtained from the comparison between Treg depleted and total T cells are shown in Figure
5.x. Treg depleted T cells compared with total non-depleted T cell fraction showed a slightly better
killing effect as measured by the percentage of CD107+ cells among the CD4+ T cells as defined by
region R2. Interestingly, this effect can also be observed in the non-CD4+ T cell fraction (R3).
FIGURE 5.24
EVALUATION OF THE CD4+ AND NON-CD4+ T CELLS FOR THE DEGRANULATION MARKER
CD107A. The percentage of positive cells is presented on the y-axis. The experimental groups are
displayed on the x-axis. The blue bars represent the results for the tested groups, standard deviations are
shown as black whiskers (comparison by Student’s t-test, p-values are indicated).
In summary, a higher degranulation rate could be observed in the CD4+ T cell fraction and also in the
non-CD4+ T cell population after Treg depletion. The effect was even more pronounced in the nonCD4+ T cell fraction where the proportion of CD107a+ T cells following Treg depletion was 3 times
higher compared with the total T cell fraction.
5. Results II
5.4.3
117
The killing capacities of T cells co-incubated with autologous tumor
cells can also be monitored in real-time
The effect of Treg depletion on tumor cell killing was monitored by a second experimental approach.
Thereby changes in impedance caused by cytotoxic effects mediated by T cells against tumor cells
were measured as explained in section 3.2.4. These effects are displayed as cell indices, a unit that
reflects changes in size and morphology of the cells, grade of adherence of the cells to the plate as well
as cell density (PEPER et al., 2014). The results obtained from this measurement are shown in Figure
5.x. While the ascending curves represent the growing phase of tumor cells during the first 96 hours
(represented by the dotted line), the descending curves represent the co-incubation of tumor cells with
the effector cells during the following 96 hours (continuous line). T cells were added following the
adherence and growing of tumor cells, 96 hours after the experiment has been started (marked by an
arrow).
FIGURE 5.25
DYNAMIC REAL-TIME MONITORING OF T CELL-MEDIATED CYTOTOXICTY AGAINST
AUTOLOGOUS TUMOR CELLS MEDIATED BY TOTAL T CELLS AND TREG DEPLETED T
CELLS. The values recorded by the xCELLigence system are displayed as dimensionless cell indices.
Controls (grey and blue) were also measured and compared with the co-cultures of tumor cells and T cells
(green and red) (top). Slope values were defined in distinct phases of the killing marked by the brackets
A, B and C and visualized as bar graphs (bottom).
118
5. Results II
The spikes interrupting the curve during the tumor cell growing phase can be explained by the daily
removal of the plate from the analyzing unit for change of the media. A massive decrease of cell index
values directly after addition of the T cells to the culture could be observed at t=96 hours. This
coincides with the time point when T cells were added and thereby half of the tumor cell medium was
replaced by lymphocyte medium. This decrease in some samples is followed by a recovery phase
accompanied be a re-increase of the cell index (between t=96 hours and t=105 hours).
As depicted in Figure 5.25 the addition of non-adherent T cells to the wells did not have any effect on
the impedance and the resulting cell index (blue control curve). Slight differences in the growing
behavior of tumor cells during the first 96 hours is reflected by higher or lower cell index values of the
tumor cell cultures that were then subjected to different treatments. At that time point the tumor cells
subsequently co-incubated with the total T cell fraction had a higher cell index than tumor cells coincubated with Treg depleted T cells (cell indices for different time points are summarized in Table
5.3).
Time point
description
T cell total
cell index
Treg depleted
cell index
p-value
96:00
after tumor cell growing phase
4.34
5.23
0.0569
105:00
after media change and recovering
4.02
4.47
0.2544
132:00
crossing point of both curves
4.21
4.21
0.9273
192:00
end of experiment
3.38
2.38
0.0619
(hours after start
of experiment)
TABLE 5.3
CELL INDICES FOR TUMOR CELLS CO-INCUBATED WITH TOTAL T CELLS AND TREG
DEPLETED T CELLS. The values recorded by the xCELLigence system are displayed for different time
points beginning after tumor cell adherence and proliferation.
The starting point for measuring the real effect of T cells on tumor cells was set to 105 hours after start
of the experiment which represent the end of the recovery phase. Here, the indices for the Treg
depletion experiment were still higher than for the total T cell experiment. These differences in the cell
indices underscore even more the effects that the different T cell fractions had on the tumor cells
which will be explained below.
During the following 96 hours of co-incubation of total T cells with tumor cells the curve showed a
slight overall decrease its course is comparable with the grey control curve (tumor cells without T
cells). The tumor cell culture that later on was treated with the Treg depleted T cells had reached a
higher cell index after 96 hours growing. After change of the media the cells did not show a recovering
phase but from that time point on a constantly decreasing curve which, although initially higher,
crossed the curve of the tumor cells treated with total T cells at approximately 132 hours. At the end of
the measurement the cell index of this curve was far lower (2.38) than that of tumor cells treated with
total T cells (3.38) (p=0.0619).
The trends of the curves can be better characterized by determining the slope (in 1/h) over the
complete co-incubation period (starting from the recovering phase, phase A) and also in single
sections (B, C) (Figure 5.25). The analysis of the overall slope demonstrated that the values for total T
cells and Treg depleted T cells are negative, but the values for the “Treg depleted” curve show a
greater descending slope. The analysis of the slope in the first killing phase (B) demonstrated, that the
5. Results II
119
slope was positive for the total T cell curve (+0.0068) while the Treg curve was decreasing (-0.0074).
In the last section (C) from the crossing point until the end of the experiment (132 hours - 192 hours)
both curves displayed negative slopes, the slope for the Treg curve (-0.0297) however is twice the
value of the total T cell curve (-0.0136). The graphical visualization of the slope values calculated for
the different cultures and the control also demonstrated that the curve for total T cells (red) is similar
to the control curve (grey) and that the curve for Treg depletion (green) behaves completely different.
The slope values explain the differences observed for the cell indices for the both co-cultures, with the
Treg curve starting at a higher cell index and finally falling below the total T cell curve.
In summary, the real-time measurement of the T cell mediated cytotoxicity against autologous tumor
cells demonstrated that Treg depletion enhances the killing of tumor cells and thus confirms the results
obtained in the first experiment by CD107a degranulation assay.
120
6. Discussion and Conclusion
6.
6.1
DISCUSSION AND
CONCLUSION
Overview of the results obtained during the thesis
The central goals of this thesis were to generate a deeper understanding of the immune status of
patients with HPV-associated precancerous lesions and cancers and evaluate possible intervention
strategies to enhance anti-tumoral immune responses.
In the first part (chapter 4) immune markers that might contribute to tumor immune evasion were
investigated on the immune cell side and on the tumor cell side. It could be shown that immune
infiltrates in cervical lesions are denser in high-grade lesions compared to low-grade lesions. This was
observed for different immune cell markers (CD3, CD8, GranB, Foxp3 and CD3ζ) and does not point
to a clear immune activation or suppression (chapter 4.2). Invasive cervical cancer, however, was
characterized by a further significant increase in Foxp3+ regulatory T cells accompanied by
significantly decreased CD8/CD3 and CD3/CD3ζ ratios which might be a hint for the
immunosuppressive state of patients with invasive disease. Although the changes between different
infection and histomorphological stages in precancers were not significant large variances in T cells
densities in all histomorphological CIN grades could be observed, for example for Tregs and also
CD8+ T cells. This might indicate that more or less infiltration with distinct T cell subtypes - effector
T cells or immune suppressive T cells - is associated with progression or regression of the lesions. To
also gain a deeper insight in the immunological modification on the tumor side contributing to
immune evasion mechanisms the expression of HLA antigens was investigated within this thesis
(chapter 4.3). Alterations in terms of HLA class I antigen losses and down-regulation, especially of
HLA class I heavy chain A, and HLA class II de novo expression in precancerous lesions and cancers
were common. The selective down-regulation of HLA class I antigens could represent another
effective immune evasion mechanism. Interestingly, it could be demonstrated in a longitudinal setting
(chapter 4.4) that immune infiltrates in CIN can be influenced by local immune modulatory drug
treatment based on imiquimod and that a response to the immune stimulatory treatment with
imiquimod is associated with increasing immune cell densities of CD3+ and CD8+ T cells. The major
methodological approach of this first part (chapter 4.1) was the establishment of automated cell
detection and quantification platform for immune cell infiltrates in cervical precancerous lesions. This
tool allows high-throughput screening of larger cohorts on the search for immunological prognostic
markers and also for monitoring of treatment strategies.
Based on the findings obtained from the immunological characterization of CIN lesions and cervical
carcinoma samples in the first part therapeutic strategies could be deduced for the second part of this
work (chapter 5). The approach based on immuno-modulatory drug treatment was pursued in this
chapter and also depletion of regulatory T lymphocytes was evaluated in different experimental
settings. Immune modulation by TLR-agonist treatment was further investigated by comparing two
6. Discussion and Conclusion
121
different compounds: the approved substance imiquimod and a new purine base derivative called
TMX-202 were tested for efficiency regarding immune stimulation (chapter 5.1). It was demonstrated
that TMX-202 in comparison to imiquimod induces massive IL-6 secretion. In an in vitro priming
experiment of naïve T cells it was also shown that this substance can stimulate the adaptive immune
response and enhance the killing of CaSki cells. One aim of the second part was also to establish a
HPV-positive HNSCC tumor cell line as an autologous model for HPV-associated cancers (chapter
5.2). Autologous systems are of special interest for immunological studies involving tumor cell killing
as alloreactivity of immune cells against an incompletely HLA-matched tumor cell lines can be a
problem. This model was used to test another strategy aiming at the circumvention of possible immune
suppressive effects mediated by regulatory T cells (chapter 5.3). The killing effects of T cells after
Treg depletion and without Treg depletion against the autologous tumor cell line were compared and
found to be enhanced with Treg depleted T cells in two independent experimental approaches.
FIGURE 6.1
THE HPV-RELATED CANCER PROGRESSION MODEL INTERPRETED IN THE CONTEXT OF
THE RESULTS OBTAINED IN THE COURSE OF THIS THESIS.
122
6.2
6. Discussion and Conclusion
An automated cell quantification tool allows the analysis of
the immune cell contexture of cervical precancerous
lesions in high-throughput approaches
The tumor micromilieu is thought to be highly important for a better understanding of the factors that
influence tumor development and progression. Parameters of interest are the immune cell composition
in term of densities and different phenotypes of immune cells entering the tumor area but also features
inherent to these cells such as production of enzymes or cytokines that might be released in the tumor
environment. The importance of the “immune cell contexture” in primary tumors of different sites and
their metastases has frequently been reported and it has been demonstrated that the quality of this
tumor micromilieu impacts the clinical outcome of the patients (reviewed in FRIDMAN et al., 2011;
FRIDMAN et al., 2012; FRIDMAN et al., 2014).
The idea to search for biomarkers that might be relevant for prognosis and for the prediction of the
patient’s clinical outcome is widespread in the field of oncology and tumor biology and not restricted
to any tumor entity (LLOYD et al., 2010) (GALON et al., 2006). On the hunt for suitable prognostic
cancer biomarkers whole slide imaging and automated quantification tools are the approaches that
scientists currently strive for. The impact of this methodological approach is reflected by the number
of up-to-date publications related to this topic (IRSHAD et al., 2014) and the multitude of reviews that
address not only general aspects of digitalized pathology and the state-of-the-art of this relatively new
field but also the challenges for imaging informatics and the needs of pathologists (KOTHARI et al.,
2013; TAYLOR, 2014; WEBSTER and DUNSTAN, 2014).
The characterization and definition of prognostic biomarkers is also highly important in the screening
of cervical intraepithelial neoplasia (CIN). They are frequently detected especially in young women,
but they often remain without any clinical consequence due to a high regression rate. Virtually all
women diagnosed with a high-grade CIN are surgically treated resulting in over-treatment. To solve
this problem reliable biomarkers are necessary for prognosis and risk-adapted treatment strategies.
The first step in the direction of virtual microscopy of cervical precancers was done with the
establishment of a platform used for the scanning and evaluation of cervical cytology slides (GRABE
et al., 2010). At that time an algorithm was developed that allows the automated detection of p16 INK4a
stained cells and reliable discrimination from unstained cells.
In the context of immune cell characterization aiming at identification of potentially immune
suppressive or cell-mediated cytotoxic mechanisms that could represent progression and regression
markers the developed tool for automated cell detection and quantification is highly important. It is a
prerequisite for defining immunological markers indicative for the clinical course of cervical
intraepithelial neoplasia. One major aim of the presented thesis was the development of such a
platform for the quantification of immune cell infiltrates in cervical precancerous lesions.
As outlined in chapter 4.1 this aim could be achieved in a cooperation project with the TIGA Center,
Heidelberg. Continuous feedback given between computer scientists, pathologists and immunologist
allowed the method to be brought to perfection. The use of available server structures for data storage
allowed all cooperation partners involved in the project to have access to the whole slide scans. The
establishment of the method was based on CIN samples stained for CD3 and CD8 and involved
continued control of the cell detection rate and manual comparison of computed cell signals with the
6. Discussion and Conclusion
123
brown DAB staining signals of digitalized slides. Several rounds of improvement including adaption
of annotation tools and defining the threshold for signal recognition also for weaker stains were
undergone until the algorithm finally was applied to the patient samples.
The established system has several advantages in comparison with existing manual approaches in the
investigation of immunological approaches. (1) It offers high reproducibility as the procedure is based
on standardized immunohistochemical staining protocols and application of a defined cell detection
and quantification algorithm. (2) It represents an objective way for cell quantification and assures
reliable discrimination of positively stained cells from negative cells. Especially the precise
demarcation of the basal membrane allows immune cells in this region exactly to be determined: this
is very challenging during manual quantification and requires the utmost concentration of the
investigator as high immune cell densities can be found in this region. (3) Manual quantification
methods mostly are restricted to smaller areas. This novel method which is based on whole-slideimages of the affected tissue allows the comprehensive assessment of not only the tumor or lesion
itself but also of the surrounding tissue, the whole microenvironment which seems to largely impact
the course of the disease. (4) It offers time-effectiveness and allows high-throughput screening of large
cohorts which could be analyzed for a variety of different (immunological) markers.
The advantages of the automated system over manual counting of positively stained cells under the
microscope in terms of objectivity reproducibility of the results are widely recognized (FUCHS et al.,
2008). Nonetheless, the established platform shall be applied to larger sample sets in the near future to
validate the method. Thereby, the intra-observer variability by comparison of repeated manual
counting and automated quantification shall be evaluated as well as the inter-observer variability by
comparison of manually and automatically quantified immune cell counts of different investigators.
In the long run the technological basis of this study shall be further developed, the parameters to be
analyzed shall be expanded and the underlying algorithms have then to be adapted to these needs.
Many other markers could be interesting and as their diversity is high and the sample material limited
immune-fluorescence allowing double staining appears as an interesting option. Additionally, further
immune-cell related information, such as definition of T cell clusters, that frequently are observed in
the tumor microenvironment (EDWARDS et al., 1995; HALAMA et al., 2009), might be of interest.
Therefore the coordinates of each cell have to be offset against the positions of the surrounding cells to
evaluate how many cells are in direct contact with each other and how far they are located from the
epithelium.
The actually available operations shall also be extended and one major aim is rendering the annotation
of the lesions easier and faster. Currently the exact annotation of the lesion has to be performed
manually by drawing the exact borders at the basal and superficial cell layers. The improvement could
be achieved by including a previously described algorithm that allows the automated separation of
stromal and tumor tissue based on a DAPI staining of the nuclei (LAHRMANN et al., 2011). This
algorithm could be adapted to CIN lesions allowing the automated annotation of the basal membrane.
This procedure would then require only an approximate demarcation of the abnormal epithelium to
initiate the annotation, the basal membrane, however, would then be automatically detected in this
region and the ROIs would be calculated as described in chapter 4.1.2. This method would be even
more time efficient and exact.
124
6. Discussion and Conclusion
Another challenge of the project and a step towards even more automation regarding the annotating
step of the process is the usage of registration methods (MOLES LOPEZ et al., 2014). By elaborating
this method, the lesion regions would have to be defined only once on the p16 INK4a reference slide,
either manually or automatically, which would then be used as template to automatically transfer the
annotated region to all other stained and digitalized slides by pattern recognition. The p16INK4a staining
could be reliably used for the detection of transformed high-grade lesions, as it is a surrogate for
oncogene overexpression and in these cases displays as strong and diffuse staining pattern. A few
slides, however, do not express p16INK4a and still would have to be annotated manually. Nevertheless,
image registration could reduce the workload as an annotated p16 INK4a stained slide could be used as a
reference template for several other consecutive stained slides.
Also the morphological appearance could be used and integrated in the computational detection of the
lesion as CIN lesions are graded according to the degree of morphologically atypical epithelial cells.
While in CIN1 the basal third of the epithelium is affected, CIN2 is defined as showing abnormalities
until the middle third and in CIN3 atypical cells can also be found in superficial third of the
epithelium. For p16INK4a-negative low-grade lesion that have not yet entered the transforming infection
stage a combination of information gained from the morphological appearance and an epithelial
marker would be conceivable to identify the p16INK4a-negative lesions (KEENAN et al., 2000; WANG
et al., 2007). An algorithm could be developed that recognizes the lesion by computational analysis of
the cellular morphology and p16INK4a positivity in parallel and by specifically distinguishing the lesion
from the background and also the stromal tissue. This would allow an objective and standardized
definition of the intraepithelial neoplasia and thus the corresponding p16INK4a reference slide could be
used as template for the automated annotation of the following serial slides stained with different
immune markers.
In conclusion, this methodological approach is in accordance with the actual needs of the classical
pathology and the contemporary trend towards whole slide imaging that replaces visual inspection and
evaluation of glass slides under the microscope and allows for high-throughput analyses. Automated
quantification platforms allow the classical pathological discipline to be transferred in to the world of
digitalization. The establishment of such a system for CIN lesions was a necessary step to make up
leeway and close the gap to the achievements already made for other tumor entities (GALON et al.,
2006; KUNZ et al., 2014; LLOYD et al., 2010). This new approach facilitates sharing of sample
material between pathologists and scientists as it did also in the here described cooperation project and
might improve the reproducibility not only in terms of pathological diagnosis (BUENO et al., 2014)
but especially in the scientific investigation of biomarkers predicting the clinical course of CIN
lesions. The established method allows the immune cell contexture in the whole affected area to be
taken into account and immune cells to be quantified in a standardized and objective way and
therefore is highly relevant for the prediction of biomarkers and as guidance for immunotherapy.
6. Discussion and Conclusion
6.3
125
Immune cell densities and composition are different in
high-grade lesions and cancers compared with low-grade
lesions
A non-negligible proportion of morphologically defined low-grade CIN1 overexpresses p16INK4a
(TSOUMPOU et al., 2009) indicating that within these lesions there is to a certain extent already viral
oncogene overexpression constituting the initial transforming event. Only a small fraction of these
early transforming infections stages, however, progresses towards higher stages (CIN2/CIN3) (WANG
et al., 2004). The study presented herein addressed the question whether density and phenotype of
infiltrating immune cells are different in low-grade CIN that have already entered the transforming
stage (p16INK4a-positive) and those that are still in the permissive stage (p16 INK4a-negative) and thus
correlates with the early induction of transformation in low-grade lesions or whether this shift rather is
associated with established and morphologically advanced high-grade dysplasia that may have
accumulated chromosomal instability. To answer these questions, different T cell phenotypes in CIN
were quantified in a cross-sectional study cohort and analyzed in relation to the p16INK4a-expression of
the lesions. In addition to well characterized immune cell markers (CD3, CD8, Granzyme B and
Foxp3) also CD3 ζ-chain was included as here data for CIN are scarcer.
T cell infiltrates were compared between non-transforming and transforming infection stages in lowgrade CIN (chapter 4.2.2). The analysis revealed that p16ÌNK4a-positivity in a substantial proportion of
low-grade CIN (CIN1) representing the early transforming infection stages, was not associated with
significant changes in densities of different T cell phenotypes infiltrating the lesion-adjacent stroma
and the lesion. T cell densities in these p16INK4a-positive low-grade CIN were similar to those in
p16INK4a-negative low-grade CIN demonstrating that the onset of p16INK4a expression which represents
the beginning of the transforming infection state was demonstrated is not associated with changes in T
cells densities or in the composition of the infiltrate.
Changes obviously occur at a later time point after transforming processes have been initiated: in later
CIN stages (CIN2/3) compared to low-grade CIN the T cell infiltrate densities fundamentally changed
irrespective of the T cell phenotype (chapter 4.2.3). The finding of increased T cell density in highgrade lesions – observed for all investigated T cell subtypes in both compartments (except epithelial
Treg cells) - is in accordance with other studies describing also denser infiltration and altered immune
cell composition in increasing clinicopathologic CIN stages (BONTKES et al., 1997; JAAFAR et al.,
2009; MONNIER-BENOIT et al., 2006).
Additionally, the higher absolute T cell infiltration in high-grade CIN compared to low-grade lesion
was accompanied by lesion/stroma ratios tending to be decreased for all T cell subtypes except for
GranB which was slightly increased. This demonstrates that despite dense infiltration with immune
cells attracted to the adjacent stromal compartment the recruitment of T cells into the lesion, where T
cell effectors should do their job and eliminate transformed cells, is hampered in high-grade lesions.
This effect may - in combination with immunological tolerance – favor progression and the outgrowth
of the lesion although dense immune cell infiltrates are present at the lesion site.
Invasive cervical carcinomas were compared with high-grade CIN (chapter 4.2.3) and showed a
further increase in total T cell numbers. Significantly higher densities were observed for CD3+,
GranB+ and Foxp3+ T cells. This is also in agreement with several other studies reporting on higher
126
6. Discussion and Conclusion
densities of CD4+ T lymphocytes (ADURTHI et al., 2008; LODDENKEMPER et al., 2009), CD8+ T
cells (ADURTHI et al., 2008; EDWARDS et al., 1995; LODDENKEMPER et al., 2009) and also
Foxp3+ regulatory T lymphocytes (ADURTHI et al., 2008; HOU et al., 2012; WU et al., 2011).
One explanation for the denser T cell infiltration in high-grade CIN and cancer could be the increased
antigenicity due to the permanent viral oncogene expression as proposed by Loddenkemper et al.
(LODDENKEMPER et al., 2009) or the potential expression of tumor-associated cellular antigens.
It has been described in literature that genomic alterations are induced following viral oncogene
overexpression initiating transformation of the host cells and that the accumulation of distinct
secondary alterations is driving the progression of a lesion (DUENSING and MUNGER, 2004).
However, studies based on comparative genomic hybridization revealed that these genomic alterations
are rare in low-grade lesions irrespective of their p16INK4a-status (THOMAS et al., 2013). The results
obtained from the presented study show that despite p16INK4a-positivity indicating viral E6 and E7
overexpression, transforming low-grade CIN are not yet characterized by marked immune cell
infiltrate changes, which only occur – as well as the accumulation of genomic alterations
(chromosomal alterations) - in later high-grade stages of CIN.
Interestingly, the lesion/stroma ratio for GranB+ activated CTLs and also the epithelial CD8+/CD3+
ratio were significantly decreased in invasive cancer samples, a finding confirmed by other studies
speculating on the ineffectiveness of effector T lymphocyte responses despite a strong infiltration due
to immunoregulation mechanisms provoking further T cell recruitment to the lesion/tumor while
disease progression is unhampered at the same time (ADURTHI et al., 2008; EDWARDS et al., 1995;
LODDENKEMPER et al., 2009; MONNIER-BENOIT et al., 2006). Immune suppression mechanisms
seem to be more important in high-grade lesions where all sorts of immune cells are attracted to a
greater extent. In cervical cancer the highest Treg density could be found, which was constantly
increasing with disease severity. The presence of regulatory T cells, however, could also be observed
in low-grade lesions. This finding together with the fact that they show large variances in each of the
diagnostic categories point to the role they could play in disease progression and clinical outcome of
the patients. Also one might speculate that the observed increase of total CD3+ T cell infiltration
correlates with an increased proportion of other types of immune regulating, inhibiting cells. This
would have to be tested with other markers. With only one immune regulation T cell type, represented
by Treg cells, investigated in this study, the exact mode of action of immune control mechanisms
enabling HPV-transformed cells to evade the immune system and allow disease progression, remains
still to be identified in prospective studies. A comprehensive overview will be given in chapter 6.4
where different immune regulation mechanisms as possible markers for future studies will be
discussed. Also strategies adapted by the tumor cells themselves might be involved in the
circumvention of host immune responses and enable the cells to further grow out to high-grade lesions
and invasive cancers. One of these mechanisms is the alteration of the antigen-presentation capacity
(chapter 4.3) which prevents potential HPV-associated antigens to be presented to immune cells. This
could hamper the activation of cytotoxic T lymphocytes – that as demonstrated in the course of this
analysis are frequently are present in the tumor environment – and thus favor disease progression.
T cell infiltration of CIN lesions and the adjacent stromal compartment is highly heterogeneous with
regard to the T cell densities and also phenotypes and was shown to increase with histomorphological
lesion grades. The correlation of T cell infiltrates with the p16 INK4a status and thereby with biologically
defined progression steps of precancerous lesions, which was done for the first time in this study,
6. Discussion and Conclusion
127
demonstrated that there are no differences in the T cell numbers between p16 INK4a-negative and
p16INK4a-positive low-grade CIN. Only in later, morphologically more advanced high-grade CIN
(p16INK4a-positive CIN2/3) remarkable alterations of T cell densities could be found. This is in
agreement with the idea of the local selection and outgrowth of more advanced abnormal subclones
that have acquired genomic alterations and the influence that these aberrations have on the local
immune milieu during progression of established lesions.
The above described heterogeneous T cell densities within the same histomorphological category were
reported previously for example for T reg cells by Adhurti et al. who argue that this T cell phenotype
varies over time and is dependent of persisting HPV infections (ADURTHI et al., 2008). It has been
demonstrated that also a proportion of established high-grade CIN (CIN2/3) regress spontaneously
(MUNK et al., 2007) and one could speculate that the dynamics of progression and regression
correlates with the variation in T cell densities and that this could be a valuable progression marker
especially for high-grade lesions that all are p16INK4a-positive for which reason p16INK4a alone cannot
predict progression. The density and phenotype of infiltrating immune cells could be a source of
predictors for the natural course of CIN and the clinical outcome as it has been described for various
other cancer types (CUNHA et al., 2012; DAVIDSSON et al., 2013; GALON et al., 2006; KIM et al.,
2013). Single longitudinal studies reported on higher GranB-expressing cytotoxic T cells in regressing
CIN (TRIMBLE et al., 2010; WOO et al., 2008) and this might also be true for other T cell types in
both outcome groups.
With the samples deriving from the Austrian imiquimod trial described in section 4.4 and based on
automated high-throughput screening methods as described previously (chapters 4.1 and 6.2) these
analyses can be transferred to a prospective study of high clinical relevance. T cell densities and
phenotypes there can be investigated in relation to the clinical outcome and the correlation with
regression or progression of the lesions is likely to contribute to a better understanding of the here
discussed heterogeneity in T cell densities. This might allow the definition of the “immune evasion
phenotype” – an immunological phenotype associated with immune evasion. Once this combination of
immune characteristics is defined it could also be used as a clinically relevant immune cell marker
panel to estimate the progression risk of patients.
6.4
HLA class I and class II antigen expression is altered in
cervical intraepithelial neoplasia and cancers
Alterations of HLA class I antigens on tumor cells have been reported in different tumor entities and
are believed to play – in addition to the variation in T cell infiltrate densities – an important role in the
battle of the host’s immune system against cancer cells. Modulation of the antigen presentation
capacities of the tumor cells is an elaborated mechanism by which tumor cells adopt to the host’s
immune system to possibly evade an immune attack (CHANG and FERRONE, 2007). HLA class I
antigen expression is reported to be associated with the clinical outcome of the patients in different
cancers such as HNSCC (MEISSNER et al., 2005), rectal cancer (REIMERS et al., 2014) and
melanoma (HICKLIN et al., 1998). For cervical cancer patients a negative correlation between absent
HLA class I heavy chain expression and a poor clinical outcome (MEHTA et al., 2008) has been
128
6. Discussion and Conclusion
shown which is explained by presence or absence of recruitment of distinct T cell phenotypes to the
tumor (JORDANOVA et al., 2008). The vast majority of these analyses have been performed in
cancer patients while the role of HLA class I and class II antigen modulation in earlier stages of
cervical carcinogenesis is less well characterized. Also studies involving both components of HLA
class I complexes, the heavy chain and the light chain, together with HLA class II antigen expression
have been lacking. To close these gap CIN and cervical cancer samples (n=40) were analyzed for
HLA class I and HLA class II antigen expression (chapter 4.3).
With regard to HLA class I antigen expression, the analyses performed during this thesis demonstrated
that normal, non-dysplastic epithelium adjacent to the lesions showed strong, homogenous and
membranous expression for HLA class I heavy chains and β2m in all observed regions. In contrast,
CIN and invasive cancers are characterized by a high frequency of alterations in HLA class I antigen
expression.
Importantly, the observed losses of HLA class I expression in the majority of the analyzed samples do
not represent total HLA class I loss, but often affect only parts of the lesion/tumor defined as
heterogeneous expression pattern. About 40 % of lesions (45.0% of CIN and 35.3% of cancers) still
retain the expression of HLA class I heavy chain A of the cell surface. With regard to cytoplasmic
expression the percentage of lesions that express HLA class I heavy chain A is about 75% (80.0% of
CIN and 70.6% of cancers).
Alterations of HLA class I heavy chain expression is more frequently were observed of for the staining
with HCA-2 representing HLA class I heavy chain A epitopes while staining results for HC-10 (heavy
chains B and C) and β2m less frequently showed alterations.
Possible mechanisms explaining the higher frequencies of losses observed with the HCA-2 antibody
mainly recognizing HLA-A heavy chains could be discussed as following: One potential explanation
might be that in these cervical lesions a selective loss of the HLA-A locus occurs more frequently as
compared with the HLA-B and HLA-C loci which visualized by the HC-10 antibody (reviewed in
SELIGER et al., 2002). Selective loss of HLA class I allospecificities in malignant cells has also been
reported in melanoma (PASCHEN et al., 2003), renal cell carcinoma (LUBOLDT et al., 1996) and
colorectal cancer (KLOOR et al., 2005). This alteration potentially reflects immune selection caused
by the massive immune infiltrates entering the tumor microenvironment (chapter 4.2) and might be
involved in down-regulated presentation of tumor-associated antigens. Additionally, tumor antigens
with a higher antigenic potential might be bound by HLA heavy chain A compared with the other
classical heavy chains of the HLA class I complex. Selective loss or down-regulation of HLA-A could
then be considered as an adaption of the tumor cells under the immune selective pressure of the host’s
immune system (CHANG et al., 2003). It is known from other tumor types that the presence or
absence of distinct HLA haplotypes, not only HLA class I but also class II antigens, contribute to a
higher susceptibility for cancer (RAZMKHAH and GHADERI, 2013), and it is conceivable that this is
also true for the development of cervical and other HPV-associated cancers.
Concerning the observed HC-10 staining pattern (less frequent alterations), a definitive conclusion
concerning HLA-B and HLA-C heavy chains in this context cannot be drawn for different reasons.
First of all, the antibody has overlapping specificity for HLA-B and HLA-C heavy chains and also for
6. Discussion and Conclusion
129
some HLA-A epitopes. If one of the antigens is down-regulated, the presence of the other heavy chain
subtype would still result in a positive staining signal (STAM et al., 1986).
Furthermore, other underlying mechanisms such as defects in the antigen-processing in the cytosol and
endoplasmatic reticulum might also be involved resulting in disturbed antigen loading and transport to
the cell membrane. This could be caused by loss of transporter-associated with antigen processing
(TAP) (BANDOH et al., 2010) or tapasin (HAN et al., 2008) which are involved in the transport of
antigenic peptides into the endoplasmatic reticulum and loading on HLA class I molecules.
Total loss of HLA class I antigens is causally linked to complete loss of β2m expression due to
structural defects of one of the β2m locus on chromosome 15. In this case, HLA class I heavy chains
cannot any longer be trafficked by the endoplasmatic reticulum and golgi apparatus to be finally
expressed on the cell membrane (reviewed in SELIGER et al., 2002). Cytoplasmic β2m expression is
retained in 100% of CIN2 samples and heterogeneous expression of β2m expression could only be
observed in the minority of CIN3 and cancer samples. None of the samples were negative for
cytoplasmic β2m expression. This implies that loss of β2m is not the major mechanism of immune
evasion contributing to the cervical carcinogenesis. This is in contrast to other tumor types such as
melanoma or microsatellite unstable colorectal cancers where the β2m wild-type allele is lost
(PASCHEN et al., 2003; TIKIDZHIEVA et al., 2012).
The fact that expression in most of the regions is retained argues against a total functional disruption.
Selective loss or down-regulation could be mediated by the interaction of HPV with the expression of
HLA class I molecules. It has been shown that HPV16 E7 induces HLA class I down-regulation
(BOTTLEY et al., 2008) as well as HPV16 E5 (CAMPO et al., 2010). This might represent a
mechanism developed by the virus to circumvent immune attack of virally infected cells by preventing
antigen-presentation of viral antigen and thus to establish the infection and promote the completion of
the viral life cycle (ASHRAFI et al., 2005).
Once the underlying mechanisms are clear, the re-induction of full HLA class I antigen expression by
therapeutic intervention (LANZA et al., 1995) may be a goal and naturally occurring immune
responses might then be successfully eradicate the lesion. In addition, treatment strategies based on
vaccines or other immune enhancing therapeutics can probably restore or further enhance the immune
attack against tumor cells.
The method based of immunohistochemical analyses of HLA class I complexes certainly has
limitations. The formalin fixation process of the tumor samples leads to dissociation of assembled
complexes into free heavy chains and β2m. In contrast to fresh, unfixed tissue material or cells, where
functional HLA class I complexes can be detected for example with the W6/32 antibody recognizing
assembled HLA-A/B/C complexes, on paraffin-embedded tissue the heavy and light chains have to be
stained separately by distinct antibodies as described previously (KLOOR et al., 2005). This is the
reason why this approach does not allow functional conclusions to be drawn from the analysis.
Although membranous expression of the components are considered to be a surrogate for the potential
antigen-presentation capacity by HLA I antigen complexes, this method is remains of limited
accuracy. However, the observed higher frequency of HLA-A losses are not thought to be caused by
deficient antibody specificity as the antibodies used in this study are well characterized and widely
130
6. Discussion and Conclusion
accepted for use in immunohistochemical analyses of HLA class I expression patterns (SERNEE et al.,
1998 and STAM et al., 1986).
HLA class II antigen expression can be detected in different solid tumors of non-lymphoid origin
(ALTOMONTE et al., 2003) and also cervical precancerous stages and cancers were found to be
positive for HLA class II molecules (CHIL et al., 2003; GLEW et al., 1992). The biological function
in the context of antigen-presentation and activation of effector T cells still remains unclear. The
investigation of HLA class II antigen expression was therefore included in the characterization of
antigen-presentation mechanisms with the aim to unravel a possible correlation with the classical
antigen-presentation pathway mediated by HLA class I antigens.
The staining with the monoclonal antibody LG-612.14 for HLA class II chains DP, DQ and DR
demonstrated that the majority of CIN2 and CIN3 lesions are positive for HLA class II antigens.
Around 80% of them displayed membranous HLA class II antigen staining. Similarly, cervical
carcinoma samples also are positive for membranous HLA class II molecules in around 85% of all
cases.
This is a strikingly high percentage of precancerous lesions and invasive cancers compared with other
solid tumors of different origins that are reported to express HLA class in tumor cells. Among these
are melanoma, gastric, colorectal and breast cancer which to a lesser extent show HLA class II antigen
expression, 50-60% of melanoma for example (reviewed in ALTOMONTE et al., 2003).
These observations raise the question of the biological significance and the functional relevance in
terms of antigen-presentation. In consideration of the fact that around 85% of CIN2, CIN3 and
invasive cancers express HLA class II with 38.4% to 52.6% being scored “positive” and showing
membranous expression on virtually all tumor cells, this might probably not contribute to efficacious
antigen-presentation directly mediated by tumor cells and a stimulation of anti-tumoral immune
responses.
Although it has been shown in the past that tumors – under inflammatory processes - might present
peptides via the HLA class II antigen complex to CD4+ T cells and that these can mediate cytotoxicity
leading to tumor rejection (DENGJEL et al., 2006; EKKIRALA et al., 2014) the sole binding and
presentation of peptides does not necessarily lead to the induction of a cell-mediated immune
response. This also requires the presence of co-stimulatory molecules, such as CD28, and their
absence rather induces antigen-specific immune tolerance mechanisms favoring disease progression
(BAL et al., 1990; GASPARI et al., 1988; HARDING et al., 1992).
With the results seen in this light one might speculate whether HLA class II antigen-negative lowgrade lesions therefore represent those that are more likely to regress as they would not – as described
in this scenario – induce immune suppression. This question however can only be addressed in a study
providing information about the functional role of HLA class II antigen expression for example by
correlating it with different immune cell phenotypes present in these lesions and with the clinical
outcome of the patients which requires a longitudinal setting such as the Austrian Imiquimod trial.
By interpreting the alterations in HLA class I and class II alterations as adaptions of the tumor cells
under the immunoselective pressure of the host’s anti-tumoral immune responses, the roles of HLA
6. Discussion and Conclusion
131
class I and II in enabling CD8+ T lymphocytes and NK cells to recognize, bind and kill tumor cells
have also to be taken into consideration. The frequently observed HLA class I down-regulation or
complete loss in tumors was early associated with impaired CD8+ CTL-mediated anti-tumoral
responses (reviewed in GARRIDO et al., 1997). The absence of HLA class I molecules on the tumor
cell however is associated with the induction of NK-cell mediated killing (BOTTINO et al., 2004).
From the developing tumor’s point of view this would be a weak immune evasion mechanism. It was
demonstrated that HLA class II molecules expressed on tumor cells protect them from being attacked
and lysed by NK cells (JIANG et al., 1996). Expression of HLA class II antigens might therefore also
be considered as – secondary – evolutionary development allowing tumor progression. The combined
alterations, HLA class I down-regulation and HLA class II expression on tumor cells could therefore
represent mechanisms that play together to circumvent cell-mediated cytotoxicity.
HLA class II expression could be caused by HPV infections and interference of the virus the host
cell’s antigen-presentation machinery. Such a correlation, however, could not be demonstrated
(GLEW et al., 1992). One could speculate that a so far unknown event that is related to the
transformation processes in high-grade lesions be associated with HLA class II antigen expression.
The observed staining pattern in precancerous lesions could also represent the phenotypical heritage of
the initially infected keratinocytes that did not resolve the HPV infection and further grew out to
precancerous lesions. This hypothesis is supported by the observed peculiar staining pattern of
columnar epithelium in the squamocolumnar junction in combination with the absence of HLA class II
expression in normal squamous epithelium. It has recently been shown that a distinct cell population
present in the squamocolumnar junction zone is susceptible to HPV infections and furthermore is
characterized by a distinct protein expression profile (HERFS et al., 2012). It is conceivable that HLA
class II expression is another characteristic of these highly metaplastic cells. If the assumption holds
true that the vast majority of cervical lesions originate in this region and develop by clonal expansion
of distinct cells, the strong HLA class II expression could be explained by the maintenance of this
phenotype in outgrowing lesions. This was hypothesized earlier in a study also observing high
expression in the metaplastic epithelium and strong expression in cervical precancerous lesions and
cancers (CHIL et al., 2003). This phenomenon then would rather be explainable by cell-intrinsic
characteristics than an adaption caused by interaction of the viral infection with the host cell’s antigenpresentation machinery. Interestingly, only half of the CIN1 were positive for HLA class II antigen
staining and this did not correlate with p16INK4a expression representing the transforming infection
stage. Considering the high frequencies of HLA class II expression in later stages, one could speculate
that low-grade lesions positive for HLA class II antigen expression are more likely to progress which
could not addressed in this study but requires a longitudinal approach.
132
6.5
6. Discussion and Conclusion
The density and composition of immune cell infiltrates can
be influenced by immuno-modulatory drugs
Several studies have demonstrated that the composition of immune infiltrates and the behavior of
immune cells such as migration can be influenced by immune modifying agents such as imiquimod
(HACKSTEIN et al., 2012; HUANG et al., 2009b; SUZUKI et al., 2000). Imiquimod is TLR7/8
agonist and its potential to enhance the patient’s immune response prompted physicians to initiate a
multitude of trials in order to investigate is efficacy in off-label indications such HPV-associated
vulvar intraepithelial neoplasia (WESTERMANN et al., 2013) (VAN SETERS et al., 2002) (VAN
SETERS et al., 2008). Although imiquimod is known to cause local and systemic side effects, it
appears to be a promising alternative to surgical standard treatment. In particular, in women affected
by multifocal VIN imiquimod treatment can replace cold knife excision as a first intervention option
(FREGA et al., 2013). Also in CIN patients there is a non-negligible need for conservative treatment
strategies as the surgical standard treatment, LEEP conization, is supposed to affect the outcome of
subsequent pregnancies and provoke pre-term birth (ARBYN et al., 2008; SIMOENS et al., 2012).
The Austrian imiquimod trial was the first randomized, placebo-controlled trial performed to test the
efficacy of topical imiquimod treatment in patients with high-grade CIN (GRIMM et al., 2012). Three
biopsies per patient were taken over 20 weeks during the treatment and after the completion of the
treatment protocol and the clinical outcome of each patient was defined based on the last biopsy taken.
These tissue specimens allow the investigation of changes in immune cells densities under treatment
with a TLR7/8 agonist and might give insights in how the immune modifier acts and which immune
cell composition is associated with a clinical response to the treatment. Of this unique patient cohort
cervical biopsies of 10 patients could be obtained who received imiquimod therapy over 16 weeks.
Albeit numerically restricted, these samples represent very valuable patient material allowing address
questions that have never been investigated before.
The tissue specimens that could be obtained of this trial were characterized by p16INK4a staining and
then analyzed based on the method described in chapter 4.1 for total T cell infiltration represented by
CD3+ cells and cytotoxic lymphocytes represented by CD8+ T cells. The obtained data were
comparatively evaluated as immune infiltrates in non-responders and responders to the imiquimod
treatment (described in chapter 4.4). In a first approach the question was addressed whether CD3 and
CD8 T cells in initial CIN2/3 biopsies are different between lesions that subsequently regressed
(responders to imiquimod) and those that persisted or even progressed (non-responders). Interestingly,
in non-responders a higher initial infiltration with CD3+ T cells was observed compared with
responders. The fact that these patients do not respond to the imiquimod treatment might be an
evidence for the presence of cell types others than effector cells present in the lesion
microenvironment. Although only one single T cell subtype was investigated (represented by CD8+ T
cells) and a definitive conclusion cannot be drawn from these results, the preliminary results allow the
speculation that the difference between non-responders and responders lies in a higher densities of T
cells phenotypes eventually responsible for immune regulation such as Treg cells. The higher absolute
T cell densities represented by CD3+ T cells could possibly be explained by a higher fraction of these
6. Discussion and Conclusion
133
“unfavorable” cell types which are absent in responders and thus having a lower CD3+ T cell
infiltration. Responders were characterized by lower total CD3 T cell infiltration. They had, however,
higher total CD8+ T cell densities and CD8/CD3 T cell ration compared with non-responders. The
higher proportion of cytotoxic T cells in responders before treatment might constitute a better initial
situation probably leading to an enhanced response to imiquimod.
Additionally, the cohort is predestined to answer the question whether the immune cell densities and
composition are different between non-responders and responders at the end of the imiquimod
treatment. It allows also the analysis of possible changes of the local immune cell composition that
occur during the treatment and their effect on the clinical response of the patients. The total T cell
infiltration with CD3+ T cells in responders after treatment compared with the initial biopsy indicated
that imiquimod locally applied to the cervix attracted immune cells to the lesion site. Responders also
showed a further increased infiltration with CD8+ T cells that could not be observed in nonresponders. The densities after treatment exceeded the CD8+ cell counts of the initial biopsies taken
before the treatment was started. However, the CD8/CD3 ratios were not higher after treatment
compared with the initial CD8/CD3 ratio in the biopsies taken before treatment, which might indicate
that together with CD8+ cytotoxic T cells also other T cell subtypes must have been attracted to the
lesion site in a proportional way representing a non-negligible proportion of T cells. This might be an
explanation why the CD8/CD3 ratio is not influenced to the extent one would expect from the absolute
CD8+ cell counts. In non-responders the total T cell infiltration represented by CD3+ T cells was not
different at the end of the treatment compared with the initial biopsies. Interestingly, non-responders
after treatment compared with week 0 showed a further decreased CTL infiltration regarding the
absolute cell counts leading also to even lower CD8/CD3 ratios than before the treatment. The exact
composition of the initially dense CD3+ T cell infiltrate in non-responders, aside from the
characterization of CD8+ T cell densities, remains largely unclear and warrants further investigation of
immune cell phenotypes possibly responsible for the unfavorable immune cell composition that might
be associated with treatment resistance. On the other hand, it is also worthwhile to characterize in
more detail T cell subtypes others than CTLs in responders and the underlying mechanisms
contributing to a clinical response to imiquimod.
It has been demonstrated before that generally low immune cell densities in the tumor environment are
associated with a poorer prognosis in cervical cancer (NEDERGAARD et al., 2007), especially low
CD8+ T cell counts in combination with high regulatory T cells infiltration correlates with poor
prognosis (SHAH et al., 2011). A major aim of cancer immunotherapy therefore is to enhance the antitumoral immune responses and to attract immune cells to the lesion site. It has been shown in the past
that imiquimod treatment in patients with vulvar intraepithelial neoplasia contributes to the
normalization of immune cell counts, for example by maturation of immature Langerhans cells, and
thus induces histological regression of the lesions (TERLOU et al., 2010). Changes in immune cell
counts are conceivable to be also the underlying reason for regression of the proportion of CIN
patients that had responded to imiquimod therapy. It appears to be obvious that the enhanced CD8+ T
cell infiltration into the lesion might contribute to the better outcome of these patients as it has been
reported before (DE VOS VAN STEENWIJK et al., 2013; PIERSMA et al., 2007) and that imiquimod
also in cervical intraepithelial neoplasia might be able to expand pre-existing CD8+ T cell response
(TODD et al., 2004).
134
6. Discussion and Conclusion
As discussed above (chapter 1.3.3), p16INK4a is a reliable marker routinely used in clinical practice and
specifically highlighting the stage of infection as it is a surrogate that indicates the presence of HPV
oncogene activity and induction of transformation of the cell (VON KNEBEL DOEBERITZ et al.,
2012). Its overexpression, however, only proofs the presence of HPV transformed cells and does not
indicate if the lesion will progress into high-grade CIN and cancer or regress, which happens in a nonnegligible proportion of all cases (SCHIFFMAN and WENTZENSEN, 2010). Until now the
prediction of possible regression and progression therefore has remained an unsolved diagnostic
problem and consequently in the clinical practice all high-grade CINs are routinely treated by surgical
intervention irrespective of the individual risk for progression. The characterization of distinct immune
cell phenotypes and the combination of different immune markers to define a biomarker tool appears
as an interesting option for the prediction of the progression risk. Although the here presented analysis
was based on a small sample size of only 10 patients and differences were not yet significant the data
argue for a consistently differential T cell distribution in non-responders compared with responders.
The both markers analyzed so far representing the total T cell infiltration as measured by CD3+ T cells
and possible cytotoxic responses as indicated by CD8+ T cells already provided interesting insights in
their potential prognostic role (as measured by responsiveness to imiquimod treatment). However,
other immunological markers might be of interest to gain more insight in the mode of action of
imiquimod and to identify further prognostically relevant mechanisms. In the past, imiquimod has
been reported in the context of Langerhans cell migration (SUZUKI et al., 2000) and recruitment of
CD8+ T cells via the integrin CD49a (SOONG et al., 2014). It induces furthermore the expression of
cytokines and chemokines, such as CXCR3, IFN-γ and reduces IL-10 and TGF-β expression
(HUANG et al., 2009b; SOONG et al., 2014; WENZEL et al., 2005). Interestingly, imiquimod also
seems to turn myeloid and plasmacytoid dendritic cells into effector cells by inducing them to express
perforin, Granzyme B and TRAIL (STARY et al., 2007). These markers only represent a restricted
selection of possible markers that could be analyzed to determine the effects of TLR7/8 agonist
treatment and to characterize the typical immune cell phenotypes of regressing or progression lesions.
Also markers that represent immune evasion mechanisms would have to be taken into consideration to
characterize the “immune evasion phenotype”. Markers that might contribute to progression of lesions
or mediate resistance to immuno-modulatory treatments are discussed in more detail in the following
section 6.6.
The samples obtained from the Austrian imiquimod trial demonstrate how important longitudinal
information is for the understanding of immune cell densities and phenotypes influencing the course of
the disease. As the trial included also a placebo-controlled patient group the natural course of CIN
without therapeutic intervention (Figure 6.2), T cell infiltrates can be associated with progressions of
treated and untreated patients also a T cell infiltration profile important for spontaneous regression
(under placebo) might be identified. By its longitudinal setting and the placebo controlled patient
group the Austrian trial is a precious study cohort to better understand immune cell composition in
regressing and progressing CIN lesions in the placebo group or in the imiquimod group. This allows
also the determination of immune cell compositions associated with spontaneous regression or
susceptibility to imiquimod therapy. On the long run this cohort could be the clue for a better
understanding of the factors that influence progression and regression and might be decisive for the
definition of immune markers for a more risk adapted treatment of patients. Thus, not all patients with
6. Discussion and Conclusion
135
high-grade lesions might have to undergo classical surgical treatment but could alternatively be tightly
monitored and wait for spontaneous regression or non-surgically be treated with an immune
modulator.
FIGURE 6.2
OVERVIEW OF THE TREATMENT SCHEDULE OF THE AUSTRIAN IMIQUIMOD TRIAL AND
OUTCOME OF THE PATIENTS. The setup of this trial does not only allow to compare effects of
imiquimod treatment in responders and non-responders but also to define T cell phenotypes associated
spontaneous regression (placebo group) or disease progression (under treatment or placebo).
With preliminary results (chapter 4.4) obtained in the here described characterization of CD3+ and
CD8+ T cell infiltrates clinical impact of this study begins to show even though only few samples and
only two T cell markers were investigated so far. Despite the relatively small sample size differences
in immune cell densities could be observed when the patients were stratified for clinical outcome.
Furthermore, it could be demonstrated that the local therapy of cervical intraepithelial neoplasia with
the immuno-modulatory drug imiquimod can influence the T cell infiltration in terms of density and
composition. Significant differences remain to be shown in a larger setting where all patients of both
treatment arms, in total 59 patients, shall be included.
As discussed formerly (section 6.1) the densities and phenotypes of tissue infiltrating immune cells is
investigated in various cancer types to define prognostic markers. Furthermore, the characterization of
the immune cell contexture (FRIDMAN et al., 2014) is indispensable for the mechanistic
understanding of cancer immunotherapy playing a more and more important role in clinical praxis.
The whole-slide-imaging and quantification platform implemented in CIN histopathology (chapter
4.1) will be used to characterize the patient cohort of the imiquimod trial. The developed method will
allow quantifying in a standardized way the immune cell composition of the complete
microenvironment and provides a highly information-rich profile which can be used for the definition
of a prognostic biomarker tool.
136
6.6
6. Discussion and Conclusion
The search for the prognostic markers characterizing the
immune evasion phenotype has to be continued
The characterization of immune infiltrates in cervical intraepithelial neoplasia (chapter 4.2) has
demonstrated that despite a generally higher infiltration with different T cell phenotypes in high-grade
CIN, these lesions have progressed to a certain extent and may further progress into invasive cancer
(SCHIFFMAN and WENTZENSEN, 2010). In addition to the presence of Treg cells in the cancer
environment, decreasing CD3ζ-expression possibly leading to a lack in T cell activation and changed
ratios of effector cells (decreased CD8/CD3 ratio) that were observed in the CIN studies, other
markers still may be of interest and contribute to progression (Table 6.1). This chapter discusses
possible markers and reviews mechanisms described also in other than HPV-related diseases that
could be considered in the further analyses of the Austrian imiquimod cohort (chapter 6.5) in order to
define the immune profile that characterizes the immune evasion phenotype of progressing CIN.
Of course, not all immune cell phenotypes, receptors and ligands or cytokines represent markers
exclusively associated with immune suppression and evasion. The majority of the here listed
mechanism are originally associated with cytotoxic immune response (Table 6.x). However, their
altered expression, down-regulation or changes in their ratios to other markers or cell phenotypes
harbor potential immune inhibiting effects.
The following list does not claim to be complete but rather is a try to summarize the most important
players in immune evasion that exert the effect on different levels. The different aspects were
classified in different categories in dependence on whether immune cell phenotypes or rather signaling
molecules or receptor and ligands are involved.
Evasion mechanism
Effect
References
CD4+CD25+Foxp3+ Treg cells
promote progression of primary tumors, possibly
also involved in promoting metastasis
(HALVORSEN et al.,
2014)
CD4+CD69+CD25- Treg cells
express CD122 and membrane-bound TGF-β1 by
which they mediate immune escape and tumor
progression
(HAN et al., 2009)
reversal of the CD4/CD8
T cell ratio
Together with presence of Treg cells has a negative
impact on clinical outcome
(SHAH et al., 2011)
immature dendritic cells
convert anergic T cells into immune suppressive
Treg cells
(PLETINCKX et al.,
2014)
loss of co-stimulatory
molecules CD27 and CD28 on
T cells
Senescent T cell phenotype induced by tumor cells;
leads to suppression of responder T cells
proliferation and promotes tumor progression
(MONTES et al.,
2008)
γδ-T17 cells
accumulation of myeloid-derived suppressor cells
(WU et al., 2014)
myeloid derived suppressor
cells (MDSCs)
suppression of T-cell and NK cell function
reviewed in (DIAZMONTERO et al.,
2014)
A) immune cell phenotypes
6. Discussion and Conclusion
137
lack of Langerhans cells
impaired antigen-presentation in the epithelium
(FAUSCH et al.,
2002)
CCR8(+) inflammatory
myeloid cells (monocytes and
CCL1 secreted by tumors binds CCR8  tumor-
(ERUSLANOV et al.,
induced inflammation  immune evasion
2013)
granulocytes)
B) cytokine and chemokine microenvironment
TGF-β
immunosuppressive cytokine that hampers the Th1
response
(PALOMARES et al.,
2014)
IL-10
immunosuppressive cytokine that hampers the Th1
response
(SYRJANEN et al.,
2009)
IL-13
immunosuppressive cytokine that hampers the Th1
response
(DEEPAK et al.,
2010)
chemokine CXCL12 and
promote tumor growth, invasion, metastasis and
reviewed in
chemokine receptor CXCR4
therapeutic resistance
(CHATTERJEE et al.,
2014)
C) endothelial factors, T cell homing and migration
decreased mucosal addressin
cell adhesion molecule
(MAdCAM) expression
decreased CD8+ T cell access to cervical
tissue
(TRIMBLE et al.,
2010)
vascular cell adhesion
expression by tumor cells promotes T cell migration
(WU, 2007)
molecule-1 (VCAM-1)
away from the tumor
E-cadherin down-regulation
associated with decreased numbers of Langerhans
cells in the epithelial and viral immune evasion
(LEONG et al., 2010)
D) antigen processing and presentation in tumor cells
HLA class I antigen downregulation
promotes escape of tumor cells from recognition
and destruction by HLA class I-restricted, antigenspecific cytotoxic T lymphocytes
reviewed in (CHANG
et al., 2003)
increased non-classical HLA
class I antigen (HLA-G)
expression
impairs the cell-mediated anti-tumoral immune
response
(RODRIGUEZ et al.,
2012)
dysregulation of transporter
associated with antigen
processing (TAP)
Disturbed antigen loading on HLA class I heavy
chains resulting in impaired antigen presentation
(BANDOH et al.,
2010)
E) altered immune cell ligand/receptor expression
up-regulation of CD94/NKG2A
inhibitory NK receptors
(SHEU et al., 2005)
MHC class I chain-related
molecule A (MICA) down-
CTL and NK cell ligand, impaired effector cell
activation
(LU et al., 2011)
abnormal CTLA-4 expression
and dysregulation
down-regulation of T cell proliferation and effector
function
(MAO et al., 2010)
PD-L1 expression in tumors
binding to PD-1 on TILs leads to impaired T cell
functions through suppression of T cell receptor
signaling
(MAINE et al., 2014),
reviewed in
(MCDERMOTT and
ATKINS, 2013)
Fas and FasL expression on
Changes in Fas expression promotes tumor growth
(ABRAMS, 2005)
regulation
138
6. Discussion and Conclusion
tumor cells
by reduced apoptosis sensitivity, FasL expression
on tumor cells mediates killing of T cells entering
the tumor
F) other mechanisms
micro-RNAs (miRNA-155)
reduced levels of miRNA-155 results in decreased
(DUDDA et al., 2013)
numbers of CD8+ effector T cells
IDO, TDO expression in tumor
cells
catalyzes immunosuppressive kynurenine leading to
cell cycle arrest and functional anergy of effector
cells, Treg differentiation and activation
reviewed in (MUNN
and MELLOR, 2013;
PLATTEN et al.,
2012)
matrix-metalloproteinase
NK cell dysfunction; down-regulation of IL-2
(PENG et al., 2014)
(MMP-1, MMP-2 and MMP-9)
expression
receptor a (IL-2Ra) expression on activated tumorinfiltrating lymphocytes
(SHEU et al., 2001)
increased inducible nitric oxide
synthase (iNOS) expression,
high levels of nitric oxide (NO)
nitric oxide acts as signaling molecule and promotes
cancer formation, progressions and metastasis
(CHENG et al., 2014)
microparticles (subtype of
extracellular vesicles
containing nucleic acids and
proteins)
involved in immune evasion, angiogenesis, tumor
invasion and metastasis
(VOLOSHIN et al.,
2014)
expression of sialic acids on
tumor cells
promote immune evasion via interaction with the
inhibitory receptor Siglec
(BULL et al., 2014)
TLR4 expression on tumor
cells
TLR stimulation induces synthesis of IL-6, iNOS
and other factors, mediates resistance of tumor cells
to CTL attack and promotes immune evasion
(HUANG et al., 2005)
TABLE 6.1
POTENTIAL MARKERS FOR THE DEFINITION OF THE “IMMUNE EVASION PHENOTYPE”. A
combination of markers that contribute to immune evasion in HPV-related precancerous stages and
cancers could be used as a diagnostic biomarker tool.
The immunohistochemical analyses of cervical precancerous and cancerous lesions performed in the
first part of this work gave hints of immunosuppressive and immune evasion mechanisms that might
play a role in the progression of HPV-associated diseases. The presence of Tregs in these lesions,
varying from low densities to high infiltration within one diagnostic category and increased in cancers
compared to precancerous stages as well as varying densities of effector T cell phenotypes in the
lesions implies that these variations might contribute to either the progression or regression of the
lesions. Furthermore, this indicates that the quality of the immune infiltrates might correlate with the
clinical outcome and could be the basis for defining prognostic markers. To better understand which
combination of markers is the most relevant for prediction of progression or regression. Here again the
longitudinal nature of the Austrian imiquimod trial is extremely valuable as it allows to decipher
distinct immunological constitutions associated with progression of high-grade lesions – in untreated
patients and under the influence of imiquimod. In combination with the automated cell quantification
method allowing high-throughput screening of larger patient cohorts and a broad variety of different
immune markers the identification of a prognostically relevant biomarker tool usable for treatment
decision appears as a realistic goal.
6. Discussion and Conclusion
139
The positive effect of immuno-modulatory drugs on the density and composition of the T cell infiltrate
could be demonstrated for imiquimod for CD3+ and CD8+ T lymphocytes (chapter 4.4).
In the second part of this thesis different intervention strategies were investigated in more detail in
immune and tumor cell based in vitro assays. Here, immuno-modulatory agents (chapter 5.1) and also
manipulations directly on the cellular levels in terms of Treg depletion (chapter 5.3) were analyzed to
explore the potential of different immunological treatment strategies.
6.7
A new immune modulatory drug, TMX-202, shows
promising effects the priming of naïve T cells to HPVassociated antigens
Immune modulation has been shown to be one mechanism that potentially leads to tumor eradication
by enhancing the host’s immune responses against abnormal cells. Aldara, the 5% imiquimod cream
formulation, is an immuno-modulatory TLR7/8 ligand-based substance approved for the treatment of
warts, actinic keratosis and basal squamous carcinoma (chapter 1.x). Because of lacking conservative
treatment options it is also given as an off-label drug to patients with anal and vulvar intraepithelial
lesions and melanoma in situ (DAVID et al., 2011) and investigated in a multitude of trials to prove its
efficacy in these off-label indications. Its efficacy could also be demonstrated in the first randomized,
controlled trial enrolling high-grade CIN patients (GRIMM et al., 2012). Although imiquimod is
considered to be safe, it causes local and systemic adverse effects which require the treatment protocol
to be interrupted (chapter 1.5.3). Considering the potent immuno-modulatory capacity of imiquimod
by induction of a strong cytokine release and a Th1-dominant anti-tumor immune response and the
non-deniable need for such an immune stimulating treatment it is worthwhile to consider alternate
drugs for TLR activation.
In cooperation with a company specialized in immuno-modulatory drugs, Telormedix S.A., which
provided a new substance for initial tests, the potency of TMX-202 a purine-like TLR7 agonist
bioconjugated to a phospholipid (Figure 1.11) agonist could be tested in different experiments
(CRAIN et al., 2013). Within the presented thesis, its immuno-stimulatory effects were tested in vitro
in the HPV-setting as it could be an interesting substance for a combinatorial drug approach that
increases the immune response to papillomaviruses. The results obtained in these experiments
contributed to a patent application.
It has been shown in the past that the TLR expression levels in B cells can be up-regulated by both
activation of the antigen-receptor or stimulation of the TLR itself by treatment with a TLR-ligand
(BOURKE et al., 2003). This finding implies that external stimuli simulating infection could regulate
the expression levels of TLRs by a positive feedback loop. Therefore, possible regulatory mechanisms
on the expression levels of TLRs were also tested under the influence of imiquimod and the new TLRligand TMX-202 and measured on the transcript and protein level. The PBMCs of four healthy donors
were treated, two of them with imiquimod and TMX-202 to compare the effects of the approved and
the newly developed drug and two of them were treated with different TMX-202 concentrations.
140
6. Discussion and Conclusion
Donor 1 showed high changes in TLR7 mRNA expression which have to be interpreted with caution,
as RNA concentrations following isolation were low. These low mRNA levels might have resulted in
low Ct values and high-fold changes when visualized in a log2 scale. Also the effects of imiquimod on
the PBLs of the first two donors were contradictory as donor 2 displayed higher changes in TLR7
mRNA expression following imiquimod treatment while donor 1 responded to TMX-202 treatment
with mRNA up-regulation. Whether this is a specific effect of imiquimod or rather induced by
potential side effects that could be caused by the imidazoquinoline cannot be deduced from only
donors tested. Also the results obtained for donors 3 and 4 were inconsistent with donor 3 showing no
effect or even decreasing mRNA expression levels and donor 4 increasing TLRL7 mRNA expression.
On the protein level for donors 1 to 3 no effect of any of the treatments could be observed. The
PBMCs of donor 4, however, displayed increased TLR7 protein expression following treatment with
TMX. In summary, only donor 4 showed a convincing influence of TMX-202 on both mRNA and
protein levels that were consistently up-regulated after treatment and under both applied
concentrations of 1µM and 10µM. The fact that natural infections lead to up-regulation of the
corresponding TLRs in vivo accompanied by an enhanced cytokine release in a time dependent
manner (HUANG et al., 2009a; KAUR et al., 2014) might be a reason for the responsiveness of donor
4. A previous infection might have induced immune cell activation and thus enhanced their
responsiveness. Secondary stimulation with a TLR-ligand might faster and to a higher extent than in
the PBMCs of the other donors have up-regulated TLR7 expression on mRNA and protein level. The
immune cells might still have been in an activated state and thus shown a greater reaction to the
external stimulation.
However, most importantly the down-stream effect of TLR agonist treatment which is considered to
be the release of pro-inflammatory cytokines stimulating both innate and adaptive immunity
(STANLEY, 2002). The pro-inflammatory cytokine plays a pivotal role in linking both arms of the
immune system and mediates the transitions from inflammatory processes to the acquired immune
response (reviewed in JONES, 2005). The pro-inflammatory processes after imiquimod and TMX-202
treatment were therefore measured by IL-6 ELISA using the supernatants deriving from PBMC
stimulation and were compared between the different treatment groups. It could be shown that
imiquimod induced significantly higher IL-6 levels compared with the controls. TMX treatment,
however, further increased the IL-6 release (by approximately two powers of ten) which was
extremely significant compared with controls. These results confirmed data published for dosedependent IL-6 release measured in whole blood following TMX-202 treatment (CRAIN et al., 2013).
In addition, it could be shown that the substance has a strong potential for the induction of a proinflammatory cytokine milieu and this is not dependent on the TLR7 mRNA or protein expression
levels but rather dose-dependent.
With these insights gained in the mechanisms how TMX-202 could link the innate with the adaptive
immune response its potential to probably enhance T cell responses against HPV-associated antigens
were tested additionally in an in vitro approach. This experiment is based on the priming of naïve T
cells with HPV-associated antigens loaded on dendritic cells in order generate antigen-specific T cells
by bringing them repeatedly in contact with antigens presented by professional APCs. Therefore a
well-established protocol used in our department was used (KAUFMANN et al., 2001).
6. Discussion and Conclusion
141
One of the peptides used in this experimental approach were p16 INK4a, a host cell protein which by its
specific overexpression in HPV-transformed lesions and all HPV-induced cancers is a potential target
protein for secondary vaccination approaches. In contrast to the viral proteins it is not HPV-type
specific. The second antigen is the major capsid protein L1 of HPV16 which is known to be a strongly
antigenic protein on which the prophylactic vaccines are based. The antigenicity of p16INK4a as well as
of HPV16 L1 were demonstrated in our phase I/IIa p16INK4a vaccination clinical trial and by a
therapeutic vaccine based on chimeric virus-like particles consisting of a L1p16INK4a fusion protein
(FAULSTICH, 2014). Both studies demonstrated that p16INK4a-specific and L1-specific cellular
immune responses can be developed following vaccination.
For p16INK4a the peptide sequence used in the clinical trial was used. For L1 a series of 9-mer and 10mer HLA-A2 restricted peptides were predicted and chosen as described in section 3.x and were tested
in a peptide-binding assay. One L1 sequence that was reported to induce L1-specific T cells following
in vitro priming was included as positive control (KAUFMANN et al., 2001), together with an
influenza matrix protein sequence, to evaluate the binding capacities to HLA antigens in the peptide
binding assay. The p16INK4a peptide used in the clinical trial also was considered to have a high
binding affinity and therefore was considered to be a control peptide for the newly synthesized L1 9mer peptides.
Of the tested L1 peptides three (L1_2, L_12 and L1_97; sequences in chapter 3.x) were demonstrated
to meet criteria defined to identify the best binding antigens: they had a significantly higher MFI
reflecting the binding capacity as compared with the background control and furthermore had a higher
MFI compared with the lowest “positive” control.
The effects of TMX-202 stimulation during T cell priming were compared with DMSO as vector
control and investigate on different levels: on the one hand dendritic cells were characterized in more
detail and on the other hand T lymphocytes were evaluated by their potency to kill tumor cells as
measured by CD107a degranulation rates.
The experiment was based on an autologous but HLA-A2 matched cell system involving CaSki cells
as targets and T lymphocytes obtained from a HLA-A2 positive healthy donor as described formerly
(RESSING et al., 1996).
The effect of TMX-202 treatment during maturation of dendritic cell from monocytes was generally
monitored by changes in the morphology and also cell numbers. In comparison to DMSO treated
control cells monocytes under the influence of the immuno-modulatory drug earlier and to a larger
extent showed a changing morphology from regularly shaped and round adherent monocytes, the
plasmacytoid morphology, to the dendrite-like morphology with branched cell appendices
(SOUMELIS and LIU, 2006). The better effect of TMX-202 treatment in comparison to the DMSO
control could also be demonstrated by higher cell numbers obtained from the original fraction of
PBMCs used for adherence of monocytes. Under TMX-202 influence consistently higher cell numbers
could be harvested demonstrating the higher rate of surviving and maturing cells under immunomodulatory drug treatment. Furthermore, it could be demonstrated that TMX-202 treated dendritic
cells extent expressed the co-stimulatory molecules CD80 and CD86 to higher extent compared with
the DMSO treated cells. CD80 and CD86 become expressed during maturation of dendritic cells and
are functionally relevant for T cell activation (DILIOGLOU et al., 2003). These results imply that
dendritic cells generated from monocytes and treated with a basic mixture of GM-CSF and IL-4
become functionally mature under the treatment with TMX-202. This is in accordance with recently
142
6. Discussion and Conclusion
published data also demonstrating that TLR-ligands can induce dendritic cell maturation (DEIFL et
al., 2014) (MASSA and SELIGER, 2013). This further indicates that the standard protocol based on
GM-CSF and IL-4 supplemented with a cytokine cocktail consisting of IL-1β, TNF-α, PGE-2 and IL-6
for the final maturation of dendritic cells (COLIC et al., 2004) might be substituted by one single
agent, the immune modifier TMX-202, which by inducing high levels of IL-6 can also lead to
functional maturation of dendritic cells.
With regard to the TMX-202 mediated effects on T cells during the repeated stimulations with
peptide-loaded dendritic cells the only parameters that could be investigated were the T cell
morphology and viability as estimated by microscopic inspection and T cell numbers representing T
cell proliferation as counted before each re-stimulation. As described in section 5.x at day x of the in
vitro priming the T cells numbers in the DMSO experiment decreased, which was also reflected by
less dense T cell culture upon visual inspection, while the TMX-treated T lymphocytes showed a
continuously increasing growth curve. However, DMSO-treated T cells recovered over time until the
end of the experiment and reached almost the number of the T cells under TMX treatment. This could
be interpreted as a first hint for the efficacy of the immuno-modulatory drug to promote adaptive T
cell responses either indirectly by stimulation with dendritic cells matured under TMX-treatment or by
direct effects of the immuno-modulatory treatment on T cell differentiation and proliferation.
The final readout of the 22 days lasting stimulation experiment was the measurement of T cell
mediated killing of CaSki cells as indicated by CD107a released upon co-incubation of the effector
cells with the target cells. Recognition of tumor cells induces cytolytic vesicles in effector T cells to
localize to the membrane and to release lytic enzymes by fusion with the outer cell membrane and
thereby CD107a contained in the inner membrane of the vesicles is transferred to the cell surface. The
killing rates of T cells stimulated under DMSO treatment were compared with the killing potential of
the T cell culture that were treated with the new immune modulator TMX-202. It could be
demonstrated that T cells cultured in presence of TMX-202 led to better killing rates as represented by
a higher fraction of CD107a-expressing CD8+ T cells. The treatment seems to promote a better
stimulation of T lymphocytes and induction of antigen-specificity. It is conceivable that this is either
related to more potent antigen-presentation mechanisms mediated by dendritic cells matured with
TMX-202 substituting the classical pro-inflammatory cytokine mix (DEIFL et al., 2014), or caused
directly by TMX-202 affecting T cell proliferation and differentiation into effector T cells (JIN et al.,
2012).
Although the killing rates are not significantly different between the DMSO and the TMX-202 T cell
experiment and the fractions of CD107a-expressing T cells killings were relative small, one should
consider that the frequencies of naïve CD8+ T cells in the peripheral blood in general is very low.
They represent only approximately 2.5% of all leukocytes contained in the peripheral blood
(CHEVALLIER et al., 2013).
It cannot completely ruled out that alloreactivity of the T lymphocytes against the heterologous tumor
cell line also contributed to the killing effect, but the differences in the killing effects of the two T cell
cultures (DMSO and TMX-202) still remain obvious. Spontaneous, non-antigen-specific reactivity of
T cells in the presence of CaSki cells was not included in the experimental setup as all T cells were
6. Discussion and Conclusion
143
stimulated with the mixture of possible antigens and this approach was solely focused on the effects
that an immune modulator could contribute to such an in vitro “vaccination” of naïve T cells.
To minimize the risk for alloreactivity of T cells against tumor cells, albeit matched for the HLA-A
allele (HLA-A2), killing assays preferably should be carried out in an autologous cell system as it was
established in the course of this thesis (chapter 5.2). Unfortunately, at the time point when the here
discussed experiment was carried out the tumor cell line was not yet established.
6.8
The generation of a HPV-associated head and neck
squamous cell carcinoma cell line for immunological
studies based on an autologous system
Although a restricted number of HPV-positive head and neck squamous cell carcinoma cell lines are
available and sporadically new HPV16-positive cell lines deriving form head and neck cancers are
published (TANG et al., 2012) they might not be optimal for certain immunological studies. The
establishment of an autologous system that provides a HPV-positive cancer cell line and at the same
time - preferentially freshly isolated - immune cells of the same patients is invaluable. In our
department this goal could be achieved for a colorectal cancer cell line in the past but a model until
now was still lacking for HPV-associated cancers. Therefore the establishment of a HPV-positive cell
line deriving from a HNSCC patient can be considered the major methodological approach of the
second part of this thesis. Once a tumor cell culture was continuously growing and had undergone
several passages without losing its adherence and proliferation capacities, which could be observed in
1 out of 31 tumor explant cultures, it was subjected to further analyses to proof its association with
HPV infection. The metastatic tumor cell line HN038M was stable for 11 months, showed continued
proliferation and contained nearly 100% tumor cells after 13 months and 2 passages. The portion of
epithelial cells contained in the culture was determined using BerEP4 antibody directed against the
epithelial cell adhesion molecule (EpCAM) by which cells of epithelial origin can be stained
specifically (BREZICKA, 2005) as the antibody does not bind to fibroblasts which are of mesodermal
origin. This demonstrates that the underlying tumor preparation protocol established during this work
(section 3.4.2) successfully eradicates contaminating fibroblast over time by sequential trypsinization
and the maintenance of tumor cells keeping their proliferating potential.
Importantly, only a fraction of about 20% of all HNSCC tumors is contributable to HPV (GILLISON
et al., 2000). The primary tumor from which the metastasis derived that could be established as cell
line was an oropharyngeal cancer located in the area of the palatine/lingual tonsil where most of the
typical HPV-associated oropharyngeal cancers occur. Nonetheless, the cell line had to be characterized
for clear signs of HPV presence and contribution of the virus to the tumorigenesis. These results were
also compared with the analysis performed in the cooperating clinic. Such investigations allowed the
non-HPV induced cancers clearly to be distinguished from those who are caused by underlying HPV
infection and transformation caused by viral oncoproteins E6 and E7 that interfere with the host cell
pathways.
The staining of cultured tumor cells for p16INK4a revealed a strongly positive staining pattern for the
cells harvested from the culture of HN038M. As p16INK4a is a surrogate for viral oncogene
144
6. Discussion and Conclusion
overexpression in transforming infection (chapter 1.3.3), this result indicates the underlying HPVinfection and transforming processes induced by the virus – more specifically by the activity of the
viral oncogene E7 – in the cell line HN038M (VON KNEBEL DOEBERITZ et al., 2012). However,
as p16INK4a in the head and neck occasionally is expressed without any relation to HPV (PRIGGE et
al., 2014) the sample was further subjected to HPV genotyping and viral oncogene expression was
analyzed by western blot analysis.
The GP5+/6+ primer-based PCR for amplification of HPV DNA clearly revealed amplified DNA
located between the 100bp and 200bp marker bands for the tumor cell samples and the positive
controls. It could be shown by Luminex-based HPV genotyping (SCHMITT et al., 2006) that the
tumor cells of the HN038M cell line harbor HPV16 DNA (SCHMITT et al., 2006). These results were
compared with the characteristics of the tissue material that was directly formalin-fixed and paraffinembedded following surgery. The paraffin-embedded tissue was stained for p16INK4a by
immunohistochemistry and it could be demonstrated that the conserved material of the metastasis
equally shows a strong and diffuse staining pattern of p16INK4a-positive cells.
Furthermore, HPV genotyping was also performed with DNA obtained from the original formalinfixed tissue samples it was demonstrated that both primary tumor and metastasis of this patient also
are positive for HPV16.
To rule out the possibility of an underlying permissive HPV infection that would not contribute to the
transformation of the tumor cells but rather represent a secondary effect, the cell lines was tested for
HPV16 E7 oncogene expression. Western blot analysis of samples collected at different time points
representing an earlier and a later passage, revealed that the cell line strongly expresses the E7
oncoprotein; with the same total amount of protein loaded on the gel, the cell line expresses even
higher E7 levels as the SiHa control. It can therefore be concluded that in this cell line HPV16
infection and oncoprotein activity was the driving mechanisms for carcinogenesis (MCLAUGHLINDRUBIN and MUNGER, 2009).
In the context of the planned immunological experiments involving killing of tumor cell by autologous
immune cells, HLA class I antigen expression was an important characteristic of this cells line to be
determined. HLA class I antigen expression and other antigen-processing components are frequently
reported to be altered in HNSCC and might represent a major mechanism that contributes to immune
evasion and thus tumor progression and metastasis (BANDOH et al., 2010; TANG et al., 2009)
(MANDIC et al., 2004; NÄSMAN et al., 2013; PRIME et al., 1987). Flow cytometry analysis of the
tumor cell line HN038M for HLA class I expression was performed with a monoclonal antibody
(clone W6/32) detects functional HLA class I antigens expressed on the cell surface by recognizing
heavy chains A, B and C. The analysis demonstrated that virtually all tumor cells expressed HLA class
I molecules on their cell surface (97.02 %). In conclusion, there were no concerns to use this cell in
subsequent immunological studies investigating the potential effect of regulatory T cells on the killing
efficiency of effector T cells. The high HLA class I expression was considered to be the prerequisite
for tumor cells to be theoretically recognized, bound and killed by cytotoxic T lymphocytes.
Finally, the cell line was characterized by short tandem repeat (STR) profiling to exclude crosscontamination by established and frequently used cell lines. The awareness of the rising frequency of
falsely identified cell lines and cross-contamination of cultures by standard cell lines, led the American
Type Culture Collection (ATCC) Standards Development organization workgroup to initiate a
6. Discussion and Conclusion
145
consensus standard on the authentication of cell lines based on STR profiling which should be applied
to standardize the procedure of cell line characterization and to assure the reliability of published
results (BARALLON et al., 2010; CONNEXIN et al., 2010). The STR profiling allows cell lines to be
identified on the individual level, to compare them with and distinguish them from cell lines contained
in the database (NIMS et al., 2010). The STR profiling, carried out by Multiplexion GmBH,
Heidelberg, showed that the new cell line HN038M is not identic with any of the cell lines contained
in the database which is defined as less than 96% identity with the best fitting comparison sequence.
Therefore, cross-contamination with other cell lines frequently used in the same laboratory room, such
as the HPV-associated cell lines CaSki, HeLa, SiHa and the colorectal cancer cell line HCT116, can be
excluded and the originality of the new HNSCC is demonstrated.
With the tumor preparation and treatment protocol adapted to head and neck squamous cell carcinoma,
the sampling of HPV+ tumors preparation and establishment of cultures will be continued in order to
establish further autologous HPV-associated cell lines in the future. Enlarging the numbers of HPVassociated tumor cell lines of patients that are alive is a valuable enrichment for the scientific
community and would allow performing - as long as patient does well - further immunological studies
based on autologous tumor and immune cells.
With the cell line in hands a cell-based immuno-modulatory intervention strategy was tested by
applying depletion of regulatory T lymphocytes from the T cell fraction and evaluating their killing
potential against autologous tumor cells in comparison to undepleted T cells.
6.9
Regulatory T cells seem to have an inhibitory effect on
anti-tumoral immune responses against autologous tumor
cells of a HPV-positive HNSCC patient
The contribution of regulatory T lymphocytes to cancer progression is one non-negligible mechanism
frequently discussed and considered as a major concern. The presence of dense Treg infiltrates are
reported in different tumor entities and their frequent occurrence in cancers is causally linked to tumor
development at different sites of the body (KIM et al., 2013; MICHEL et al., 2008; SHAH et al., 2011;
WOLF et al., 2003). Regulatory T cells are thought to hamper different kinds of therapeutic
vaccination approaches or other strategies elaborated to induce T cell mediated anti-tumoral responses
– not only in the HPV-setting. In the context of cancer immunotherapy Treg depletion therefore plays
a crucial role (reviewed in CURIEL, 2007 and NISHIKAWA and SAKAGUCHI, 2014). Results from
the here described study (chapter 4.2) and also published data demonstrate that regulatory T
lymphocytes play a non-negligible role in HPV-associated cervical cancers and the precursor lesions
(LODDENKEMPER et al., 2009; MOLLING et al., 2007; VISSER et al., 2007; WU et al., 2011).
The regulatory T cell phenotype characterized as CD4+CD25+Foxp3+ T cells has been shown to
contribute to suppression of cytotoxic responses and in vitro Treg depletion is reported to T cell
mediated immune responses (CHEN et al., 2012). The effect of Treg depletion, however, is rarely
investigated in the setting of HPV-related diseases (CHUANG et al., 2009; TUVE et al., 2007) and
mainly demonstrated in mouse models. Only one study could be found investigating Treg depletion in
146
6. Discussion and Conclusion
the context of nasopharyngeal carcinoma – without, however, considering possible underlying HPV
infections (FOGG et al., 2013). Data concerning the role of Tregs in HPV-associated OSCC and the
impact of Treg depletion is scarce.
The established HPV16-positive tumor cell line deriving from an OSCC patient was used for an initial
experiment which aimed at the investigation of the potential immunosuppressive effect mediated by
regulatory T lymphocytes. The killing potential of PBMCs isolated from the blood of the OSCC
patients against the autologous tumor cell line HN038M was measured in two different experimental
approaches and was based on the comparison between Treg depleted T cells and the total (undepleted)
T cell fraction. This last chapter thus completes the circle with regard to the immunohistochemical
analyses performed in the first part of this work. Although it could not be demonstrated that regulatory
T cells contribute to progression of precancerous lesions - because of the cross-sectional nature of the
study -, the enormous variances in different diagnostic CIN grades might imply a functional role in
tumor development.
In a broad general approach - without considering possible underlying mechanisms - the effect of Treg
depletion was measured by the impedance-based Roche Xcelligence System. The assay principle is
explained in detail in section 3.2.4 and is based on impedance measurement reflecting changes in cell
density, adherence and morphology. These changes for example can be caused by manipulations such
as drug treatment inducing apoptosis. The convincing argument in favor of this system is the
possibility to monitor the effects of treatments on tumor cells in real-time and in a label-free manner. It
has been demonstrated to be applicable for monitoring vaccine-based cytotoxicity on tumor cells
(PHAM et al., 2014), T cell mediated killing (PEPER et al., 2014) and also was compared with 51Crrelease assay (measuring the release of 51Cr from labelled target cells upon cytolysis) to demonstrate
that the impedance-based assay can detect changes in the levels of antigen-specific cytotoxic T cells
with increased sensitivity compared with the standard chromium release assay (ERSKINE et al.,
2012). It therefore represents an attractive alternative assay to established experiments as exposure to
gamma radiation or other labelling reagents can be avoided, high reproducibility can be obtained and
fewer cells are required for the experimental setup.
For the above describe experiment the cell index values were measured during the growing phase and
the killing phase of tumor cells and for Treg depleted T cells and the total T cell fraction. The cell
index calculated for the T cell control wells demonstrated that the addition of effector cells to the
adherent tumor cells did not affect the impedance. Therefore the impedance curves for the coincubated T cells and tumor cells can be considered to represent the true killing effect on tumor cells
without requiring further normalization for T cell impedance. It could be shown that the T cells after
Treg depletion induces a stronger decrease in cell index values and also in the slopes calculated for the
cell index curves compared with the total T cell fraction (Figure 5.25). Decreasing cell indices can be
interpreted as being caused by tumor cell lysis and T cell mediated cytotoxicity. The differences
between Treg depleted T cells and the total T cell fraction are significant over the total killing period
and also in the two defined sub-phases representing the first and second killing phase.
To take into account the effects mediated by T cells with or without previous Treg depletion more
specifically, CD107a degranulation assay was performed to gain information about the cytotoxic
potential of the effector cells. This assay was first described in 2003 as a “novel technique to
enumerate antigen-specific CD8+ T cells using a marker expressed on the cell surface following
6. Discussion and Conclusion
147
activation induced degranulation, a necessary precursor of cytolysis” (BETTS et al., 2003). Although
the here described experiment was performed in an antigen-independent manner it has been
demonstrated that CD8+ T cells expressing CD107a are involved in antigen-specific cytotoxicity
(BETTS et al., 2003) and that such assays allows the identification and analysis of tumor-reactive T
cells in vitro (RUBIO et al., 2003).
The performed CD107a degranulation assay showed that among T cells stained for CD4 and CD107a
those subjected to Treg depletion showed a higher fraction expression CD107a compared with T cells
that were not depleted from regulatory T cells. It has been reported that CD4+ T cells also can
cytotoxic potential and contribute to elimination of tumor cells (reviewed in MARSHALL and
SWAIN, 2011 and APPAY, 2004). This same trend could also been seen in the fraction of non-CD4+
T lymphocytes. This opens the question of the nature of the phenotype of the T cells contained in this
population. Based on the assay principle for T cell isolation (chapter 3.2.4) which selects for CD3+ T
cells the presence of natural killer cells (CD3-negative) in the isolated T cell fraction can be ruled out.
Furthermore, the presence of HLA class I antigens on the tumor cells has been demonstrated and it is
unlikely that NK cells, if present, would have any killing effect against HLA class I antigen expressing
tumor cells (MORETTA et al., 1996). It can therefore be assumed that the non-CD4+T cell population
is composed of a fraction of CD8+ CTLs and also natural killer T cells (NKT cells) which in contrast
to NK cells express CD3.
The percentage of CD107-expressing cells in both fractions, non-CD4+ and CD4+ T cells, are
relatively low. It has been demonstrated in different settings that the frequencies of antigen-specific T
cells are relatively low, and one can speculate that antigen-specific T cells circulating in the blood are
even less frequent than tumor-infiltrating lymphocytes at the tumor site (reported frequencies range
from 0.01% to 0.4% for CD8+ T cells) (HE et al., 1999; POLLACK et al., 2014). For HPV-associated
antigens, due to effective immune evasion mechanisms the frequencies for antigens specific T cells
might even be lower as reported for low levels of E7-specific precursor T cells (1 of 3947 T cells) in
the blood (HOFFMANN et al., 2006). However, although changes remain low with regard to the
absolute CD107+ T cell frequencies, the killing rate is 3 times higher after Treg depletion in the nonCD4+ T cell fraction and such changes might have tremendous effects in vivo in respect to the low
frequencies of potential antigen-specific T cells.
In conclusion, both assays by addressing different parameters, the changes in impedance caused lysed
tumor cells and the CD107a expression on the cell surface of T cells, demonstrated that Treg depletion
enhances the killing efficiency of the remaining T cell fraction and that Treg mediated suppression
might play a role in the investigated OSCC tumor probably having participated in disease progression.
This finding might also be an explanation for disease recurrence after surgical treatment in this patient
although HPV-associated HNSCC in general have a better prognosis and clinical outcome. Depletion
of regulatory T cells might therefore be an important treatment option to be considered for HPVassociated diseases in general and in OSCC in particular where data so far have been lacking and
allows the circumvention of immunosuppressive effects.
Intervention strategies in this context for example might be based on drugs that specifically target
Tregs (FOGG et al., 2013). Recently also therapeutic approaches involving chemotherapeutic agents
for control and reversal of the immunosuppressive effects mediated by regulatory T lymphocytes
148
6. Discussion and Conclusion
might be applicable in anti-cancer therapy (reviewed in ALIZADEH and LARMONIER, 2014;
D'ARENA et al., 2011; OHKURA et al., 2011).
A combined therapy is conceivable involving immune stimulating agents such as TLR ligands and
drugs combatting immune suppression, in the same way as today classical chemotherapeutic agents
are combined such as cytostatic and cytotoxic drugs or combinations of antibody-based anti-cancer
treatments. Probably here again, the combination of different strategies might be more effective than
one single therapy alone by addressing the variability of mechanisms developed by HPV-associated
diseases to circumvent the host’s immune attack
Importantly, the single treatment strategies described in this thesis show a tendency to contribute to a
reduced tumor growth. The combination of these strategies, however, is conceivable to improve and
potentiate the effects obtained with each of these strategies alone.
6.10
Future prospects
The initial immunohistochemical analysis of immune cell infiltrated in cervical intraepithelial
neoplasia and cancers demonstrated that changes of immune cell infiltrate are not associated with the
onset of transforming infections in histomorphological low-grade lesions. The observed T cell
densities are not yet different compared with non-transforming low-grade lesions. However, as lesions
of the same histomorphological grade with different biological and clinical behavior are pooled within
one diagnostic group - an approach which could not be avoided due to nature of most of the available
patient cohorts - the T cells infiltrate data cannot be related to the clinical outcome of the patients.
Interestingly, broad ranges of immune cell densities could be observed for distinct T cell subtypes, e.g.
regulatory T cells and CD8+ cytotoxic T lymphocytes indicating that samples are characterized by a
large heterogeneity. This might reflect samples of patients with either progressing or regressing
disease. Only patients samples stratified for the clinical behavior of the lesions could unravel the
impact of distinct immune cell phenotypes on disease outcome. The changes in immune cell densities
and the phenotypic composition have to be in a prospective setting in order to gain a better
understanding of how these changes are related with the clinical outcome of the patients.
These aspects could perfectly be addressed in the patient cohort of the Austrian imiquimod trial. In the
course of the 20 weeks treatment and observation protocol three biopsies per patient were sampled.
The effect of a topical immuno-modulatory drug, imiquimod, was tested in a randomized, placebocontrolled setting and the efficiency of the treatment was determined by comparing the imiquimodtreated arm with the placebo-group. The small number of patient samples that could be obtained for
the first analyses of T cell infiltrates in imiquimod treated patients represents extremely precious
material and served as a basis for the first analyses of immune infiltrates in imiquimod-treated lesions.
The first step was taken towards a deeper understanding of the immunophenotypic reversal mediated
by immune modifiers such as TLR-ligands. However, if access will be gained to the placebo-treated
patient samples, the natural course of untreated high-grade CIN lesions over time can be monitored
and immune cell infiltrate data correlated with the course of the disease, e.g. progression or regression.
The analysis of these samples is therefore considered to help answering the questions that could not be
addressed in the cross-sectional study. T cell phenotypes that contribute to spontaneous regression
6. Discussion and Conclusion
149
without previous therapeutic intervention can be investigated, as well as distinct immune phenotypes
that rather are associated with disease persistence or progression. Also the analyses that have been
initiated with the patients of the imiquimod treatment arm will be expanded to enlarge the sample size
and validate the results obtained so far.
In this context further immune markers might be relevant and should be included to enlarge the
immune phenotypic characterization. Based on the list shown in section 6.6 the best marker
combinations can be defined after having been evaluated in preliminary immunohistochemical
analyses as it was done for some of the most important T cell markers in the first cross-sectional
approach. It has to be demonstrated that the chosen markers are reliable predictors of the biological
behavior of the disease and the clinical outcome of the patients. The final biomarker set might also be
a combination of immune cell and tumor cell markers as long as they alone and even more in
combination are predictive for the clinical course of the disease.
On the long run this precious cohort will allow to define an “immune score”, a biomarker-based tool
that could be applicable in the clinical routine to predict the risk for progression of CIN lesions and
also the chance to respond to non-surgical interventions such as topical treatment with TLR agonists.
This tool might help to make individualized and risk-adapted treatment decisions, minimize overtreatment of a clinically heterogeneous disease and permit at least a distinct proportion of young
women to obtain conservative treatment.
HLA class I and class II antigens is a potential component of this novel “immune score”. However,
with the results obtained so far, their biological relevance and their contribution to immune evasion or
effective anti-tumoral immune responses is still not clear. Their impact on the quality of the immune
response that might be initiated has to be elucidated by additional analyses. In a first approach T cell
densities and also different T cell phenotypes infiltrating the tumor microenvironment should be
correlated with the expression pattern of antigen-presenting molecules in the lesions. The best,
clinically most relevant approach again would be a longitudinal one allowing the correlation with the
patients’ outcome to reveal the role of these alterations in the context of immune evasion or either
immune attack of the host. First analyses demonstrated that HLA class II expression indeed seems to
impact immune densities in terms of CD3+ and CD8+ T cells. Thereby, a higher proportion of cells
expressing HLA class II antigens as well as a higher fraction of cells showing membranous expression
were associated with a trend towards denser T cell infiltrates (data not shown).
All these investigations that finally should lead to the definition of an “immune score” for use in the
diagnosis and prognosis will be based on the newly developed automated quantification method
described in this thesis. It represents a highly standardized and objective method to quantify immune
cells as the results of cell counting are not biased by subjective criteria defined by the investigator.
Especially for the development of a clinically relevant immune cell based biomarker tool, the
reliability of the results has to be demonstrated and they need to be validated in a larger sample cohort.
Here, the established quantification platform is the method of choice as it allows high-throughput
screening of large cohorts.
.
150
FIGURE 6.3
6. Discussion and Conclusion
GRAPHICAL OVERVIEW OF THE POSSIBILITIES TO COMBINE THE RESULTS AND
ESTABLISHED METHODS IN THE FUTURE IN ORDER TO DEVELOP NEW DIAGNOSTIC
TOOLS AND TREATMENT STRATEGIES. #
The Austrian imiquimod trial together with the multitude of imiquimod trials performed in vulvar
intraepithelial neoplasia (VIN) patients demonstrated the clinical efficacy of TLR-agonist based
treatment that aims at immune modulation of the lesion microenvironment. The promising results
obtained in this study make TLR-agonist treatment a strategy to be pursued. However, in consideration
of the known side effects of imiquimod further immuno-modulatory treatment strategies were
evaluated in the second part of the thesis. A second generation TLR-agonist, TMX-202, was tested in
the in vitro priming of naïve T cells to p16INK4a and HPV16 L1 peptides.
# REFERENCES FOR IMAGES USED IN FIGURE 6.x
http://emedicine.medscape.com/article/2086864-overview (30.11.2014)
http://de.wikipedia.org/wiki/CD3-Rezeptor (30.11.2014)
http://www.mskcc.org/blog/cancer-immunotherapy-named-science-magazine-breakthrough-year (30.11.2014)
http://www.zejournal.info/infos-insolites/1-articles-infos-insolites/1944-il-braque-une-pharmacie-avec-un-tube-de-creme-de-beaute (30.11.2014)
http://www.duden.de/rechtschreibung/Tablette (30.11.2014)
http://www.mikroelektronik.fraunhofer.de/de/presse-und-medien/vue-nachrichten/article/zeiss-setzt-bei-lichtblattmikroskop-auf-technologie-des-fraunhoferipms.html (30.11.2014)
http://www.rcsb.org/pdb/101/motm.do?momID=143 (30.11.2014)
http://www.biooncology.com/therapeutic-targets/cd79b (30.11.2014)
http://www.soc.ucsb.edu/sexinfo/article/cervical-cancer (30.11.2014)
http://www.genengnews.com/insight-and-intelligenceand153/cancer-cell-line-identity-crisis/73502226/ (30.11.2014)
6. Discussion and Conclusion
151
It could be shown that the substance has effects on dendritic cells maturation and T cell proliferation
and also contributes to a slightly increased killing potential of T cells in the context of HPV-associated
diseases. These first results warrant further studies to better characterize the potential of the new TLRagonist. The planning of a vaccination experiment based on p16 INK4aL1 chimeric virus-like particles in
a mouse model is under way. Here, TMX can be evaluated as a potential adjuvant. It is well
conceivable that the cellular immune responses against L1 and p16 INK4a in mice vaccinated with the
chimeric VLPs are potentiated by using TMX as an adjuvant, as effects were observed on both levels,
the innate immunity and also adaptive immune responses
With the oral squamous cell carcinoma cell line generated in the course of this thesis a valuable
autologous model was established which can serve as basis for further immunological studies. For
example the effect of TMX-202 treatment could now be validated in an autologous system to allay
concerns regarding the alloreactivity between immune cells and an allogeneic tumor cell line. It
represents a perfect model for the generation of antigen-specific T cells by in vitro priming with the
auto-antigen p16INK4a and viral antigens such as L1 and to test the killing potential of T cells against
autologous tumor cells. The patient recruitment will be continued to obtain further tumor tissue
samples for the generation of more cell lines. These are necessary to validate the results obtained from
one single patient and to evaluate whether or not the findings are representative for HPV-associated
diseases and the conclusions that were drawn can be generalized
Although the results look promising, the effect of regulatory T cell depletion demonstrated in the
autologous cell line HN038M will have to be tested in further cell lines to validate the results obtained
in one patient. A combination strategy consisting of Treg depletion along with immuno-modulatory
drug treatment would be highly interesting as the better killing effect observed after Treg depletion
could further be enhanced if anti-tumoral response of the remaining cell fraction would be “enhanced”
by TLR-ligand treatment. The experiment demonstrated how important strategies aiming at Treg
depletion might be for the improvement of cancer immunotherapy approaches.
This treatment strategy in general could be further refined by using drugs specifically targeting Tregs
in order to deplete them from the total T cell fraction. A multitude of new and already established
drugs are actually discussed to selectively eliminate the immunosuppressive effected mediated by
regulatory T lymphocytes. Among these cyclophosphamide (Cytoxan) (CAMISASCHI et al., 2013),
denileukin diftitox (TELANG et al., 2011) and ipilimumab (HODI et al., 2010) represent interesting
therapeutic
drugs.
152
7.
7. References
REFERENCES
ABRAMS, S.I. (2005). Positive and negative consequences of Fas/Fas ligand interactions in the
antitumor response. Front Biosci 10, 809-821.
ADURTHI, S., KRISHNA, S., MUKHERJEE, G., BAFNA, U.D., DEVI, U., and JAYSHREE, R.S. (2008).
Regulatory T cells in a spectrum of HPV-induced cervical lesions: cervicitis, cervical intraepithelial
neoplasia and squamous cell carcinoma. Am J Reprod Immunol 60, 55-65.
ADURTHI, S., MUKHERJEE, G., KRISHNAMURTHY, H., SUDHIR, K., BAFNA, U.D., UMADEVI, K., et al.
(2012). Functional tumor infiltrating TH1 and TH2 effectors in large early-stage cervical cancer are
suppressed by regulatory T cells. Int J Gynecol Cancer 22, 1130-1137.
AKIRA, S., and TAKEDA, K. (2004). Toll-like receptor signalling. Nat Rev Immunol 4, 499-511.
ALBERS, A.E., and KAUFMANN, A.M. (2009). Therapeutic human papillomavirus vaccination. Public
health genomics 12, 331-342.
ALIZADEH, D., and LARMONIER, N. (2014). Chemotherapeutic targeting of cancer-induced
immunosuppressive cells. Cancer Res 74, 2663-2668.
ALTOMONTE, M., FONSATTI, E., VISINTIN, A., and MAIO, M. (2003). Targeted therapy of solid
malignancies via HLA class II antigens: a new biotherapeutic approach? Oncogene 22, 6564-6569.
AMADOR-MOLINA, A., HERNANDEZ-VALENCIA, J.F., LAMOYI, E., CONTRERAS-PAREDES, A., and
LIZANO, M. (2013). Role of innate immunity against human papillomavirus (HPV) infections and
effect of adjuvants in promoting specific immune response. Viruses 5, 2624-2642.
ANTINORE, M.J., BIRRER, M.J., PATEL, D., NADER, L., and MCCANCE, D.J. (1996). The human
papillomavirus type 16 E7 gene product interacts with and trans-activates the AP1 family of
transcription factors. EMBO J 15, 1950-1960.
APPAY, V. (2004). The physiological role of cytotoxic CD4(+) T-cells: the holy grail? Clin Exp Immunol
138, 10-13.
ARBYN, M., KYRGIOU, M., SIMOENS, C., RAIFU, A.O., KOLIOPOULOS, G., MARTIN-HIRSCH, P., et al.
(2008). Perinatal mortality and other severe adverse pregnancy outcomes associated with treatment
of cervical intraepithelial neoplasia: meta-analysis. BMJ 337, a1284.
ASHRAFI, G.H., HAGHSHENAS, M.R., MARCHETTI, B., O'BRIEN, P.M., and CAMPO, M.S. (2005). E5
protein of human papillomavirus type 16 selectively downregulates surface HLA class I. Int J Cancer
113, 276-283.
BAIS, A.G., BECKMANN, I., LINDEMANS, J., EWING, P.C., MEIJER, C.J., SNIJDERS, P.J., et al. (2005). A
shift to a peripheral Th2-type cytokine pattern during the carcinogenesis of cervical cancer becomes
manifest in CIN III lesions. J Clin Pathol 58, 1096-1100.
7. References
153
BAL, V., MCINDOE, A., DENTON, G., HUDSON, D., LOMBARDI, G., LAMB, J., et al. (1990). Antigen
presentation by keratinocytes induces tolerance in human T cells. Eur J Immunol 20, 1893-1897.
BANDOH, N., OGINO, T., KATAYAMA, A., TAKAHARA, M., KATADA, A., HAYASHI, T., et al. (2010). HLA
class I antigen and transporter associated with antigen processing downregulation in metastatic
lesions of head and neck squamous cell carcinoma as a marker of poor prognosis. Oncol Rep 23, 933939.
BARALLON, R., BAUER, S.R., BUTLER, J., CAPES-DAVIS, A., DIRKS, W.G., ELMORE, E., et al. (2010).
Recommendation of short tandem repeat profiling for authenticating human cell lines, stem cells,
and tissues. In Vitro Cell Dev Biol Anim 46, 727-732.
BASLER, C.F., and GARCIA-SASTRE, A. (2002). Viruses and the type I interferon antiviral system:
induction and evasion. Int Rev Immunol 21, 305-337.
BECKMAN, R.A., and LOEB, L.A. (2005). Negative clonal selection in tumor evolution. Genetics 171,
2123-2131.
BERGERON, C., RONCO, G., REUSCHENBACH, M., WENTZENSEN, N., ARBYN, M., STOLER, M., et al.
(2014). The clinical impact of using p16 immunochemistry in cervical histopathology and cytology: An
update of recent developments. Int J Cancer doi: 10.1002/ijc.28900. [Epub ahead of print].
BERNARD, H.U. (2005). The clinical importance of the nomenclature, evolution and taxonomy of
human papillomaviruses. J Clin Virol 32 Suppl 1, S1-6.
BETTS, M.R., BRENCHLEY, J.M., PRICE, D.A., DE ROSA, S.C., DOUEK, D.C., ROEDERER, M., et al. (2003).
Sensitive and viable identification of antigen-specific CD8+ T cells by a flow cytometric assay for
degranulation. J Immunol Methods 281, 65-78.
BEUCHER, S. (1992). The watershed transformation applied to image segmentation. Scanning Microsc
Int 6, 299–314.
BIGNOLD, L.P. (2002). The mutator phenotype theory can explain the complex morphology and
behaviour of cancers. Cell Mol Life Sci 59, 950-958.
BIGNOLD, L.P. (2003). The mutator phenotype theory of carcinogenesis and the complex
histopathology of tumours: support for the theory from the independent occurrence of nuclear
abnormality, loss of specialisation and invasiveness among occasional neoplastic lesions. Cell Mol Life
Sci 60, 883-891.
BONTKES, H.J., DE GRUIJL, T.D., WALBOOMERS, J.M., VAN DEN MUYSENBERG, A.J., GUNTHER, A.W.,
SCHEPER, R.J., et al. (1997). Assessment of cytotoxic T-lymphocyte phenotype using the specific
markers granzyme B and TIA-1 in cervical neoplastic lesions. Br J Cancer 76, 1353-1360.
BOSCOLO-RIZZO, P., DEL MISTRO, A., BUSSU, F., LUPATO, V., BABOCI, L., ALMADORI, G., et al. (2013).
New insights into human papillomavirus-associated head and neck squamous cell carcinoma. Acta
Otorhinolaryngol Ital 33, 77-87.
BOTTINO, C., MORETTA, L., PENDE, D., VITALE, M., and MORETTA, A. (2004). Learning how to
discriminate between friends and enemies, a lesson from Natural Killer cells. Mol Immunol 41, 569575.
154
7. References
BOTTLEY, G., WATHERSTON, O.G., HIEW, Y.L., NORRILD, B., COOK, G.P., and BLAIR, G.E. (2008). Highrisk human papillomavirus E7 expression reduces cell-surface MHC class I molecules and increases
susceptibility to natural killer cells. Oncogene 27, 1794-1799.
BOURKE, E., BOSISIO, D., GOLAY, J., POLENTARUTTI, N., and MANTOVANI, A. (2003). The toll-like
receptor repertoire of human B lymphocytes: inducible and selective expression of TLR9 and TLR10 in
normal and transformed cells. Blood 102, 956-963.
BREZICKA, T. (2005). Expression of epithelial-cell adhesion molecule (Ep-CAM) in small cell lung
cancer as defined by monoclonal antibodies 17-1A and BerEP4. Acta Oncol 44, 723-727.
BRINKMAN, J.A., HUGHES, S.H., STONE, P., CAFFREY, A.S., MUDERSPACH, L.I., ROMAN, L.D., et al.
(2007). Therapeutic vaccination for HPV induced cervical cancers. Dis Markers 23, 337-352.
BUCK, C.B., DAY, P.M., and TRUS, B.L. (2013). The papillomavirus major capsid protein L1. Virology
445, 169-174.
BUENO, G., DENIZ, O., FERNANDEZ-CARROBLES MDEL, M., VALLEZ, N., and SALIDO, J. (2014). An
automated system for whole microscopic image acquisition and analysis. Microsc Res Tech 77, 697713.
BULL, C., DEN BROK, M.H., and ADEMA, G.J. (2014). Sweet escape: Sialic acids in tumor immune
evasion. Biochim Biophys Acta.
BURD, E.M. (2003). Human papillomavirus and cervical cancer. Clin Microbiol Rev 16, 1-17.
CAMISASCHI, C., FILIPAZZI, P., TAZZARI, M., CASATI, C., BERETTA, V., PILLA, L., et al. (2013). Effects of
cyclophosphamide and IL-2 on regulatory CD4+ T cell frequency and function in melanoma patients
vaccinated with HLA-class I peptides: impact on the antigen-specific T cell response. Cancer Immunol
Immunother 62, 897-908.
CAMPO, M.S., GRAHAM, S.V., CORTESE, M.S., ASHRAFI, G.H., ARAIBI, E.H., DORNAN, E.S., et al.
(2010). HPV-16 E5 down-regulates expression of surface HLA class I and reduces recognition by CD8 T
cells. Virology 407, 137-142.
CASTRO, F., DIRKS, W.G., FAHNRICH, S., HOTZ-WAGENBLATT, A., PAWLITA, M., and SCHMITT, M.
(2013). High-throughput SNP-based authentication of human cell lines. Int J Cancer 132, 308-314.
CHAIWONGKOT, A., VINOKUROVA, S., PIENTONG, C., EKALAKSANANAN, T., KONGYINGYOES, B.,
KLEEBKAOW, P., et al. (2013). Differential methylation of E2 binding sites in episomal and integrated
HPV 16 genomes in preinvasive and invasive cervical lesions. International journal of cancer Journal
international du cancer 132, 2087-2094.
CHAN, M., HAYASHI, T., KUY, C.S., GRAY, C.S., WU, C.C., CORR, M., et al. (2009). Synthesis and
immunological characterization of toll-like receptor 7 agonistic conjugates. Bioconjug Chem 20, 11941200.
CHAN, P.K., LIU, S.J., CHEUNG, J.L., CHEUNG, T.H., YEO, W., CHONG, P., et al. (2011). T-cell response
to human papillomavirus type 52 L1, E6, and E7 peptides in women with transient infection, cervical
intraepithelial neoplasia, and invasive cancer. J Med Virol 83, 1023-1030.
CHANG, C.-C., CAMPOLI, M., and FERRONE, S. (2003). HLA class I defects in malignant lesions: What
have we learned? The Keio Journal of Medicine 52, 220-229.
7. References
155
CHANG, C.-C., and FERRONE, S. (2007). Immune selective pressure and HLA class I antigen defects in
malignant lesions. Cancer immunology, immunotherapy : CII 56, 227-236.
CHATTERJEE, A., PULIDO, H.A., KOUL, S., BELENO, N., PERILLA, A., POSSO, H., et al. (2001). Mapping
the sites of putative tumor suppressor genes at 6p25 and 6p21.3 in cervical carcinoma: occurrence of
allelic deletions in precancerous lesions. Cancer Res 61, 2119-2123.
CHATTERJEE, S., BEHNAM AZAD, B., and NIMMAGADDA, S. (2014). The intricate role of CXCR4 in
cancer. Adv Cancer Res 124, 31-82.
CHATURVEDI, A.K., ANDERSON, W.F., LORTET-TIEULENT, J., CURADO, M.P., FERLAY, J., FRANCESCHI,
S., et al. (2013). Worldwide trends in incidence rates for oral cavity and oropharyngeal cancers. J Clin
Oncol 31, 4550-4559.
CHATURVEDI, A.K., ENGELS, E.A., PFEIFFER, R.M., HERNANDEZ, B.Y., XIAO, W., KIM, E., et al. (2011).
Human papillomavirus and rising oropharyngeal cancer incidence in the United States. J Clin Oncol
29, 4294-4301.
CHEN, Y.L., CHANG, M.C., CHEN, C.A., LIN, H.W., CHENG, W.F., and CHIEN, C.L. (2012). Depletion of
regulatory T lymphocytes reverses the imbalance between pro- and anti-tumor immunities via
enhancing antigen-specific T cell immune responses. PLoS ONE 7, e47190.
CHENG, H., WANG, L., MOLLICA, M., RE, A.T., WU, S., and ZUO, L. (2014). Nitric oxide in cancer
metastasis. Cancer Lett 353, 1-7.
CHEVALLIER, P., ROBILLARD, N., ILLIAQUER, M., ESBELIN, J., MOHTY, M., BODIN-BRESSOLLETTE, C., et
al. (2013). Characterization of various blood and graft sources: a prospective series. Transfusion
(Paris) 53, 2020-2026.
CHIL, A., SIKORSKI, M., BOBEK, M., JAKIEL, G., and MARCINKIEWICZ, J. (2003). Alterations in the
expression of selected MHC antigens in premalignant lesions and squamous carcinomas of the
uterine cervix. Acta Obstet Gynecol Scand 82, 1146-1152.
CHOW, L.T., BROKER, T.R., and STEINBERG, B.M. (2010). The natural history of human papillomavirus
infections of the mucosal epithelia. APMIS 118, 422-449.
CHUANG, C.M., HOORY, T., MONIE, A., WU, A., WANG, M.C., and HUNG, C.F. (2009). Enhancing
therapeutic HPV DNA vaccine potency through depletion of CD4+CD25+ T regulatory cells. Vaccine
27, 684-689.
COLIC, M., MOJSILOVIC, S., PAVLOVIC, B., VUCICEVIC, D., MAJSTOROVIC, I., BUFAN, B., et al. (2004).
Comparison of two different protocols for the induction of maturation of human dendritic cells in
vitro. Vojnosanit Pregl 61, 471-478.
CONGER, K.L., LIU, J.S., KUO, S.R., CHOW, L.T., and WANG, T.S. (1999). Human papillomavirus DNA
replication. Interactions between the viral E1 protein and two subunits of human dna polymerase
alpha/primase. J Biol Chem 274, 2696-2705.
CONNEXIN, J.S., CONNEXIN, P.D., and CONNEXIN, A.L. (2010). Cell line misidentification: the
beginning of the end. Nat Rev Cancer 10, 441-448.
156
7. References
CRAIN, B., YAO, S., KEOPHILAONE, V., PROMESSI, V., KANG, M., BARBERIS, A., et al. (2013). Inhibition
of keratinocyte proliferation by phospholipid-conjugates of a TLR7 ligand in a Myc-induced
hyperplastic actinic keratosis model in the absence of systemic side effects. Eur J Dermatol 23, 618628.
CUNHA, L.L., MORARI, E.C., GUIHEN, A.C., RAZOLLI, D., GERHARD, R., NONOGAKI, S., et al. (2012).
Infiltration of a mixture of immune cells may be related to good prognosis in patients with
differentiated thyroid carcinoma. Clin Endocrinol (Oxf) 77, 918-925.
CURIEL, T.J. (2007). Tregs and rethinking cancer immunotherapy. J Clin Investig 117, 1167-1174.
D'ARENA, G., DEAGLIO, S., LAURENTI, L., DE MARTINO, L., DE FEO, V., FUSCO, B.M., et al. (2011).
Targeting regulatory T cells for anticancer therapy. Mini Rev Med Chem 11, 480-485.
DARRAGH, T.M., COLGAN, T.J., THOMAS COX, J., HELLER, D.S., HENRY, M.R., LUFF, R.D., et al. (2013).
The Lower Anogenital Squamous Terminology Standardization project for HPV-associated lesions:
background and consensus recommendations from the College of American Pathologists and the
American Society for Colposcopy and Cervical Pathology. Int J Gynecol Pathol 32, 76-115.
DAVID, C.V., NGUYEN, H., and GOLDENBERG, G. (2011). Imiquimod: a review of off-label clinical
applications. J Drugs Dermatol 10, 1300-1306.
DAVIDSSON, S., OHLSON, A.L., ANDERSSON, S.O., FALL, K., MEISNER, A., FIORENTINO, M., et al.
(2013). CD4 helper T cells, CD8 cytotoxic T cells, and FOXP3(+) regulatory T cells with respect to lethal
prostate cancer. Mod Pathol 26, 448-455.
DE GIORGI, V., SALVINI, C., CHIARUGI, A., PAGLIERANI, M., MAIO, V., NICOLETTI, P., et al. (2009). In
vivo characterization of the inflammatory infiltrate and apoptotic status in imiquimod-treated basal
cell carcinoma. Int J Dermatol 48, 312-321.
DE MARTEL, C., FERLAY, J., FRANCESCHI, S., VIGNAT, J., BRAY, F., FORMAN, D., et al. (2012). Global
burden of cancers attributable to infections in 2008: a review and synthetic analysis. Lancet Oncol 13,
607-615.
DE VILLIERS, E.M., FAUQUET, C., BROKER, T.R., BERNARD, H.U., and ZUR HAUSEN, H. (2004).
Classification of papillomaviruses. Virology 324, 17-27.
DE VOS VAN STEENWIJK, P.J., HEUSINKVELD, M., RAMWADHDOEBE, T.H., LOWIK, M.J., VAN DER
HULST, J.M., GOEDEMANS, R., et al. (2010). An unexpectedly large polyclonal repertoire of HPVspecific T cells is poised for action in patients with cervical cancer. Cancer Res 70, 2707-2717.
DE VOS VAN STEENWIJK, P.J., PIERSMA, S.J., WELTERS, M.J., VAN DER HULST, J.M., FLEUREN, G.,
HELLEBREKERS, B.W., et al. (2008). Surgery followed by persistence of high-grade squamous
intraepithelial lesions is associated with the induction of a dysfunctional HPV16-specific T-cell
response. Clin Cancer Res 14, 7188-7195.
DE VOS VAN STEENWIJK, P.J., RAMWADHDOEBE, T.H., GOEDEMANS, R., DOORDUIJN, E.M., VAN
HAM, J.J., GORTER, A., et al. (2013). Tumor-infiltrating CD14-positive myeloid cells and CD8-positive
T-cells prolong survival in patients with cervical carcinoma. Int J Cancer 133, 2884-2894.
DEEPAK, P., KUMAR, S., JR., KISHORE, D., and ACHARYA, A. (2010). IL-13 from Th2-type cells
suppresses induction of antigen-specific Th1 immunity in a T-cell lymphoma. Int Immunol 22, 53-63.
7. References
157
DEIFL, S., KITZMULLER, C., STEINBERGER, P., HIMLY, M., JAHN-SCHMID, B., FISCHER, G.F., et al.
(2014). Differential activation of dendritic cells by toll-like receptors causes diverse differentiation of
naive CD4(+) T cells from allergic patients. Allergy 69, 1602-1609.
DENGJEL, J., NASTKE, M.D., GOUTTEFANGEAS, C., GITSIOUDIS, G., SCHOOR, O., ALTENBEREND, F., et
al. (2006). Unexpected abundance of HLA class II presented peptides in primary renal cell
carcinomas. Clin Cancer Res 12, 4163-4170.
DENNY, L.A., FRANCESCHI, S., DE SANJOSE, S., HEARD, I., MOSCICKI, A.B., and PALEFSKY, J. (2012).
Human papillomavirus, human immunodeficiency virus and immunosuppression. Vaccine 30 Suppl 5,
F168-174.
DHARMAPURI, S., AURISICCHIO, L., NEUNER, P., VERDIRAME, M., CILIBERTO, G., and LA MONICA, N.
(2009). An oral TLR7 agonist is a potent adjuvant of DNA vaccination in transgenic mouse tumor
models. Cancer Gene Ther 16, 462-472.
DIAZ-MONTERO, C.M., FINKE, J., and MONTERO, A.J. (2014). Myeloid-derived suppressor cells in
cancer: therapeutic, predictive, and prognostic implications. Semin Oncol 41, 174-184.
DIEBOLD, S.S., KAISHO, T., HEMMI, H., AKIRA, S., and REIS E SOUSA, C. (2004). Innate antiviral
responses by means of TLR7-mediated recognition of single-stranded RNA. Science 303, 1529-1531.
DILIOGLOU, S., CRUSE, J.M., and LEWIS, R.E. (2003). Function of CD80 and CD86 on monocyte- and
stem cell-derived dendritic cells. Exp Mol Pathol 75, 217-227.
DOEBERITZ, M., and VINOKUROVA, S. (2009a). Host factors in HPV-related carcinogenesis: cellular
mechanisms controlling HPV infections. Arch Med Res 40, 435-442.
DOEBERITZ, M.K., and VINOKUROVA, S. (2009b). Host factors in HPV-related carcinogenesis: cellular
mechanisms controlling HPV infections. Arch Med Res 40, 435-442.
DOORBAR, J. (2005). The papillomavirus life cycle. J Clin Virol 32 Suppl 1, S7-15.
DOORBAR, J. (2006). Molecular biology of human papillomavirus infection and cervical cancer. Clin
Sci (Lond) 110, 525-541.
DOORBAR, J. (2013). The E4 protein; structure, function and patterns of expression. Virology 445, 8098.
DUDDA, J.C., SALAUN, B., JI, Y., PALMER, D.C., MONNOT, G.C., MERCK, E., et al. (2013). MicroRNA155 is required for effector CD8+ T cell responses to virus infection and cancer. Immunity 38, 742753.
DUENSING, S., and MUNGER, K. (2002). The human papillomavirus type 16 E6 and E7 oncoproteins
independently induce numerical and structural chromosome instability. Cancer Res 62, 7075-7082.
DUENSING, S., and MUNGER, K. (2004). Mechanisms of genomic instability in human cancer: insights
from studies with human papillomavirus oncoproteins. Int J Cancer 109, 157-162.
DUESBERG, P., MANDRIOLI, D., MCCORMACK, A., and NICHOLSON, J.M. (2011). Is carcinogenesis a
form of speciation? Cell Cycle 10, 2100-2114.
DUNNE, E.F., UNGER, E.R., STERNBERG, M., MCQUILLAN, G., SWAN, D.C., PATEL, S.S., et al. (2007).
Prevalence of HPV infection among females in the United States. JAMA 297, 813-819.
158
7. References
DWORACKI, G., MEIDENBAUER, N., KUSS, I., HOFFMANN, T.K., GOODING, W., LOTZE, M., et al.
(2001). Decreased zeta chain expression and apoptosis in CD3+ peripheral blood T lymphocytes of
patients with melanoma. Clin Cancer Res 7, 947s-957s.
EDWARDS, R.P., KUYKENDALL, K., CROWLEY-NOWICK, P., PARTRIDGE, E.E., SHINGLETON, H.M., and
MESTECKY, J. (1995). T lymphocytes infiltrating advanced grades of cervical neoplasia. CD8-positive
cells are recruited to invasion. Cancer 76, 1411-1415.
EKKIRALA, C.R., CAPPELLO, P., ACCOLLA, R.S., GIOVARELLI, M., ROMERO, I., GARRIDO, C., et al.
(2014). Class II Transactivator-Induced MHC Class II Expression in Pancreatic Cancer Cells Leads to
Tumor Rejection and a Specific Antitumor Memory Response. Pancreas 43, 1066-1072.
ERSKINE, C.L., HENLE, A.M., and KNUTSON, K.L. (2012). Determining optimal cytotoxic activity of
human Her2neu specific CD8 T cells by comparing the Cr51 release assay to the xCELLigence system. J
Vis Exp, e3683.
ERUSLANOV, E., STOFFS, T., KIM, W.J., DAURKIN, I., GILBERT, S.M., SU, L.M., et al. (2013). Expansion
of CCR8(+) inflammatory myeloid cells in cancer patients with urothelial and renal carcinomas. Clin
Cancer Res 19, 1670-1680.
FAHEY, L.M., RAFF, A.B., DA SILVA, D.M., and KAST, W.M. (2009). Reversal of human papillomavirusspecific T cell immune suppression through TLR agonist treatment of Langerhans cells exposed to
human papillomavirus type 16. Journal of immunology (Baltimore, Md : 1950) 182, 2919-2928.
FAULSTICH, F. (2014). Generation and evaluation of chimeric particles consisting of HPV16 L1 and
p16INK4a for second generation HPV vaccines. Dissertation.
FAUSCH, S.C., DA SILVA, D.M., RUDOLF, M.P., and KAST, W.M. (2002). Human Papillomavirus VirusLike Particles Do Not Activate Langerhans Cells: A Possible Immune Escape Mechanism Used by
Human Papillomaviruses. The Journal of Immunology 169, 3242-3249.
FAUSCH, S.C., FAHEY, L.M., DA SILVA, D.M., and KAST, W.M. (2005). Human Papillomavirus Can
Escape Immune Recognition through Langerhans Cell Phosphoinositide 3-Kinase Activation. The
Journal of Immunology 174, 7172-7178.
FELLER, L., WOOD, N.H., KHAMMISSA, R.A., CHIKTE, U.M., MEYEROV, R., and LEMMER, J. (2010). HPV
modulation of host immune responses. SADJ 65, 266-268.
FERLAY, J., SHIN, H.R., BRAY, F., FORMAN, D., MATHERS, C., and PARKIN, D.M. (2010). Estimates of
worldwide burden of cancer in 2008: GLOBOCAN 2008. Int J Cancer 127, 2893-2917.
FOGG, M., MURPHY, J.R., LORCH, J., POSNER, M., and WANG, F. (2013). Therapeutic targeting of
regulatory T cells enhances tumor-specific CD8+ T cell responses in Epstein-Barr virus associated
nasopharyngeal carcinoma. Virology 441, 107-113.
FREGA, A., SESTI, F., SOPRACORDEVOLE, F., BIAMONTI, A., SCIRPA, P., MILAZZO, G.N., et al. (2013).
Imiquimod 5% cream versus cold knife excision for treatment of VIN 2/3: a five-year follow-up. Eur
Rev Med Pharmacol Sci 17, 936-940.
FRIDMAN, W.H., GALON, J., PAGÈS, F., TARTOUR, E., SAUTÈS-FRIDMAN, C., and KROEMER, G. (2011).
Prognostic and predictive impact of intra- and peritumoral immune infiltrates. Cancer Res 71, 56015605.
7. References
159
FRIDMAN, W.H., PAGES, F., SAUTES-FRIDMAN, C., and GALON, J. (2012). The immune contexture in
human tumours: impact on clinical outcome. Nat Rev Cancer 12, 298-306.
FRIDMAN, W.H., REMARK, R., GOC, J., GIRALDO, N.A., BECHT, E., HAMMOND, S.A., et al. (2014). The
immune microenvironment: a major player in human cancers. Int Arch Allergy Immunol 164, 13-26.
FUCHS, T.J., WILD, P.J., MOCH, H., and BUHMANN, J.M. (2008). Computational pathology analysis of
tissue microarrays predicts survival of renal clear cell carcinoma patients. Medical image computing
and computer-assisted intervention : MICCAI International Conference on Medical Image Computing
and Computer-Assisted Intervention 11, 1-8.
FUENTES-GONZALEZ, A.M., CONTRERAS-PAREDES, A., MANZO-MERINO, J., and LIZANO, M. (2013).
The modulation of apoptosis by oncogenic viruses. Virol J 10, 182.
GALON, J., COSTES, A., SANCHEZ-CABO, F., KIRILOVSKY, A., MLECNIK, B., LAGORCE-PAGES, C., et al.
(2006). Type, density, and location of immune cells within human colorectal tumors predict clinical
outcome. Science 313, 1960-1964.
GANGULY, N., and PARIHAR, S.P. (2009). Human papillomavirus E6 and E7 oncoproteins as risk
factors for tumorigenesis. J Biosci 34, 113-123.
GARCIA-CHACON, R., VELASCO-RAMIREZ, S.F., FLORES-ROMO, L., and DANERI-NAVARRO, A. (2009).
Immunobiology of HPV Infection. Arch Med Res 40, 443-448.
GARCIA-LORA, A., ALGARRA, I., GAFORIO, J.J., RUIZ-CABELLO, F., and GARRIDO, F. (2001).
Immunoselection by T lymphocytes generates repeated MHC class I-deficient metastatic tumor
variants. Int J Cancer 91, 109-119.
GARRIDO, F., RUIZ-CABELLO, F., CABRERA, T., PÉREZ-VILLAR, J.J., LÓPEZ-BOTET, M., DUGGAN-KEEN,
M., et al. (1997). Implications for immunosurveillance of altered HLA class I phenotypes in human
tumours. Immunol Today 18, 89-95.
GASPARI, A., TYRING, S.K., and ROSEN, T. (2009). Beyond a decade of 5% imiquimod topical therapy. J
Drugs Dermatol 8, 467-474.
GASPARI, A.A., JENKINS, M.K., and KATZ, S.I. (1988). Class II MHC-bearing keratinocytes induce
antigen-specific unresponsiveness in hapten-specific Th1 clones. J Immunol 141, 2216-2220.
GASPARINI, R., and PANATTO, D. (2009). Cervical cancer: from Hippocrates through Rigoni-Stern to
zur Hausen. Vaccine 27 Suppl 1, A4-5.
GEORGOPOULOS, N.T., PROFFITT, J.L., and BLAIR, G.E. (2000). Transcriptional regulation of the major
histocompatibility complex (MHC) class I heavy chain, TAP1 and LMP2 genes by the human
papillomavirus (HPV) type 6b, 16 and 18 E7 oncoproteins. Oncogene 19, 4930-4935.
GIBB, R.K., and MARTENS, M.G. (2011). The impact of liquid-based cytology in decreasing the
incidence of cervical cancer. Rev Obstet Gynecol 4, S2-S11.
GILLISON, M.L., CASTELLSAGUE, X., CHATURVEDI, A., GOODMAN, M.T., SNIJDERS, P., TOMMASINO,
M., et al. (2014). Eurogin Roadmap: comparative epidemiology of HPV infection and associated
cancers of the head and neck and cervix. Int J Cancer 134, 497-507.
160
7. References
GILLISON, M.L., KOCH, W.M., CAPONE, R.B., SPAFFORD, M., WESTRA, W.H., WU, L., et al. (2000).
Evidence for a causal association between human papillomavirus and a subset of head and neck
cancers. J Natl Cancer Inst 92, 709-720.
GLEW, S.S., DUGGAN-KEEN, M., CABRERA, T., and STERN, P.L. (1992). HLA class II antigen expression
in human papillomavirus-associated cervical cancer. Cancer Res 52, 4009-4016.
GONZALES, R.W., RE; EDDINS, SL (2009). Digital Image Processing using Matlab. (Tennessee:
Gatesmark Publishing).
GRABE, N., LAHRMANN, B., POMMERENCKE, T., VON KNEBEL DOEBERITZ, M., REUSCHENBACH, M.,
and WENTZENSEN, N. (2010). A virtual microscopy system to scan, evaluate and archive biomarker
enhanced cervical cytology slides. Cell Oncol 32, 109-119.
GRIMM, C., POLTERAUER, S., NATTER, C., RAHHAL, J., HEFLER, L., TEMPFER, C.B., et al. (2012).
Treatment of cervical intraepithelial neoplasia with topical imiquimod: a randomized controlled trial.
Obstet Gynecol 120, 152-159.
GUL, N., GANESAN, R., and LUESLEY, D.M. (2004). Characterizing T-cell response in low-grade and
high-grade vulval intraepithelial neoplasia, study of CD3, CD4 and CD8 expressions. Gynecol Oncol
94, 48-53.
HACKSTEIN, H., HAGEL, N., KNOCHE, A., KRANZ, S., LOHMEYER, J., VON WULFFEN, W., et al. (2012).
Skin TLR7 triggering promotes accumulation of respiratory dendritic cells and natural killer cells. PLoS
ONE 7, e43320.
HALAMA, N., ZOERNIG, I., SPILLE, A., WESTPHAL, K., SCHIRMACHER, P., JAEGER, D., et al. (2009).
Estimation of immune cell densities in immune cell conglomerates: an approach for high-throughput
quantification. PLoS ONE 4, e7847.
HALVORSEN, E.C., MAHMOUD, S.M., and BENNEWITH, K.L. (2014). Emerging roles of regulatory T
cells in tumour progression and metastasis. Cancer Metastasis Rev.
HAMID, N.A., BROWN, C., and GASTON, K. (2009). The regulation of cell proliferation by the
papillomavirus early proteins. Cell Mol Life Sci 66, 1700-1717.
HAN, L.Y., FLETCHER, M.S., URBAUER, D.L., MUELLER, P., LANDEN, C.N., KAMAT, A.A., et al. (2008).
HLA class I antigen processing machinery component expression and intratumoral T-Cell infiltrate as
independent prognostic markers in ovarian carcinoma. Clinical cancer research : an official journal of
the American Association for Cancer Research 14, 3372-3379.
HAN, S., ZHANG, C., LI, Q., DONG, J., LIU, Y., HUANG, Y., et al. (2014). Tumour-infiltrating CD4(+) and
CD8(+) lymphocytes as predictors of clinical outcome in glioma. Br J Cancer 110, 2560-2568.
HAN, Y., GUO, Q., ZHANG, M., CHEN, Z., and CAO, X. (2009). CD69+ CD4+ CD25- T cells, a new subset
of regulatory T cells, suppress T cell proliferation through membrane-bound TGF-beta 1. J Immunol
182, 111-120.
HARDING, F.A., MCARTHUR, J.G., GROSS, J.A., RAULET, D.H., and ALLISON, J.P. (1992). CD28mediated signalling co-stimulates murine T cells and prevents induction of anergy in T-cell clones.
Nature 356, 607-609.
7. References
161
HARWOOD, C.A., SURENTHERAN, T., SASIENI, P., PROBY, C.M., BORDEA, C., LEIGH, I.M., et al. (2004).
Increased risk of skin cancer associated with the presence of epidermodysplasia verruciformis human
papillomavirus types in normal skin. Br J Dermatol 150, 949-957.
HE, X.S., REHERMANN, B., LOPEZ-LABRADOR, F.X., BOISVERT, J., CHEUNG, R., MUMM, J., et al. (1999).
Quantitative analysis of hepatitis C virus-specific CD8(+) T cells in peripheral blood and liver using
peptide-MHC tetramers. Proc Natl Acad Sci U S A 96, 5692-5697.
HEINE, H., and LIEN, E. (2003). Toll-like receptors and their function in innate and adaptive immunity.
Int Arch Allergy Immunol 130, 180-192.
HEMMI, H., KAISHO, T., TAKEUCHI, O., SATO, S., SANJO, H., HOSHINO, K., et al. (2002). Small anti-viral
compounds activate immune cells via the TLR7 MyD88-dependent signaling pathway. Nat Immunol 3,
196-200.
HERFS, M., YAMAMOTO, Y., LAURY, A., WANG, X., NUCCI, M.R., MCLAUGHLIN-DRUBIN, M.E., et al.
(2012). A discrete population of squamocolumnar junction cells implicated in the pathogenesis of
cervical cancer. Proceedings of the National Academy of Sciences 109, 10516-10521.
HICKLIN, D.J., WANG, Z., ARIENTI, F., RIVOLTINI, L., PARMIANI, G., and FERRONE, S. (1998). beta2Microglobulin mutations, HLA class I antigen loss, and tumor progression in melanoma. J Clin Investig
101, 2720-2729.
HODI, F.S., O'DAY, S.J., MCDERMOTT, D.F., WEBER, R.W., SOSMAN, J.A., HAANEN, J.B., et al. (2010).
Improved survival with ipilimumab in patients with metastatic melanoma. N Engl J Med 363, 711723.
HOFFMANN, T.K., ARSOV, C., SCHIRLAU, K., BAS, M., FRIEBE-HOFFMANN, U., KLUSSMANN, J.P., et al.
(2006). T cells specific for HPV16 E7 epitopes in patients with squamous cell carcinoma of the
oropharynx. Int J Cancer 118, 1984-1991.
HOLLDACK, J. (2014). Toll-like receptors as therapeutic targets for cancer. Drug Discov Today 19, 379382.
HOPMAN, A.H., THEELEN, W., HOMMELBERG, P.P., KAMPS, M.A., HERRINGTON, C.S., MORRISON,
L.E., et al. (2006). Genomic integration of oncogenic HPV and gain of the human telomerase gene
TERC at 3q26 are strongly associated events in the progression of uterine cervical dysplasia to
invasive cancer. J Pathol 210, 412-419.
HORN, J., DAMM, O., KRETZSCHMAR, M.E., DELERE, Y., WICHMANN, O., KAUFMANN, A.M., et al.
(2013). Estimating the long-term effects of HPV vaccination in Germany. Vaccine 31, 2372-2380.
HOU, F., LI, Z., MA, D., ZHANG, W., ZHANG, Y., ZHANG, T., et al. (2012). Distribution of Th17 cells and
Foxp3-expressing T cells in tumor-infiltrating lymphocytes in patients with uterine cervical cancer.
Clin Chim Acta 413, 1848-1854.
HUANG, B., ZHAO, J., LI, H., HE, K.-L., CHEN, Y., MAYER, L., et al. (2005). Toll-Like Receptors on Tumor
Cells Facilitate Evasion of Immune Surveillance. Cancer Res 65, 5009-5014.
HUANG, S., WEI, W., and YUN, Y. (2009a). Upregulation of TLR7 and TLR3 gene expression in the lung
of respiratory syncytial virus infected mice. Wei Sheng Wu Xue Bao 49, 239-245.
162
7. References
HUANG, S.J., HIJNEN, D., MURPHY, G.F., KUPPER, T.S., CALARESE, A.W., MOLLET, I.G., et al. (2009b).
Imiquimod enhances IFN-gamma production and effector function of T cells infiltrating human
squamous cell carcinomas of the skin. J Investig Dermatol 129, 2676-2685.
HUH, K., ZHOU, X., HAYAKAWA, H., CHO, J.Y., LIBERMANN, T.A., JIN, J., et al. (2007). Human
papillomavirus type 16 E7 oncoprotein associates with the cullin 2 ubiquitin ligase complex, which
contributes to degradation of the retinoblastoma tumor suppressor. J Virol 81, 9737-9747.
IRSHAD, H., VEILLARD, A., ROUX, L., and RACOCEANU, D. (2014). Methods for nuclei detection,
segmentation, and classification in digital histopathology: a review-current status and future
potential. IEEE reviews in biomedical engineering 7, 97-114.
JAAFAR, F., RIGHI, E., LINDSTROM, V., LINTON, C., NOHADANI, M., VAN NOORDEN, S., et al. (2009).
Correlation of CXCL12 expression and FoxP3+ cell infiltration with human papillomavirus infection
and clinicopathological progression of cervical cancer. Am J Pathol 175, 1525-1535.
JANEWAY, C.A., JR., and MEDZHITOV, R. (2002). Innate immune recognition. Annu Rev Immunol 20,
197-216.
JIANG, Y.Z., COURIEL, D., MAVROUDIS, D.A., LEWALLE, P., MALKOVSKA, V., HENSEL, N.F., et al. (1996).
Interaction of natural killer cells with MHC class II: reversal of HLA-DR1-mediated protection of K562
transfectant from natural killer cell-mediated cytolysis by brefeldin-A. Immunology 87, 481-486.
JIN, B., SUN, T., YU, X.H., YANG, Y.X., and YEO, A.E. (2012). The effects of TLR activation on T-cell
development and differentiation. Clin Dev Immunol 2012, 836485.
JONES, E.E., and WELLS, S.I. (2006). Cervical cancer and human papillomaviruses: inactivation of
retinoblastoma and other tumor suppressor pathways. Curr Mol Med 6, 795-808.
JONES, S.A. (2005). Directing transition from innate to acquired immunity: defining a role for IL-6. J
Immunol 175, 3463-3468.
JORDANOVA, E.S., GORTER, A., AYACHI, O., PRINS, F., DURRANT, L.G., KENTER, G.G., et al. (2008).
Human leukocyte antigen class I, MHC class I chain-related molecule A, and CD8+/regulatory T-cell
ratio: which variable determines survival of cervical cancer patients? Clinical cancer research : an
official journal of the American Association for Cancer Research 14, 2028-2035.
JUNG, C., and KIM, C. (2010). Segmenting clustered nuclei using H-minima transform-based marker
extraction and contour parameterization. IEEE Trans Biomed Eng 57, 2600-2604.
KAJITANI, N., SATSUKA, A., KAWATE, A., and SAKAI, H. (2012). Productive Lifecycle of Human
Papillomaviruses that Depends Upon Squamous Epithelial Differentiation. Frontiers in microbiology
3, 152.
KANODIA, S., DA SILVA, D.M., and KAST, W.M. (2008). Recent advances in strategies for
immunotherapy of human papillomavirus-induced lesions. International journal of cancer Journal
international du cancer 122, 247-259.
KANODIA, S., FAHEY, L.M., and KAST, W.M. (2007). Mechanisms used by human papillomaviruses to
escape the host immune response. Curr Cancer Drug Targets 7, 79-89.
KAUFMANN, M., NIELAND, J., SCHINZ, M., NONN, M., GABELSBERGER, J., MEISSNER, H., et al. (2001).
HPV16 L1E7 chimeric virus-like particles induce specific HLA-restricted T cells in humans after in vitro
vaccination. International journal of cancer Journal international du cancer 92, 285-293.
7. References
163
KAUR, R., CASEY, J., and PICHICHERO, M. (2014). Cytokine, chemokine, and toll-like receptor
expression in middle ear fluids of children with acute otitis media. Laryngoscope.
KAWANA, K., ADACHI, K., KOJIMA, S., KOZUMA, S., and FUJII, T. (2012). Therapeutic Human
Papillomavirus (HPV) Vaccines: A Novel Approach. The open virology journal 6, 264-269.
KAWANA, K., YASUGI, T., and TAKETANI, Y. (2009). Human papillomavirus vaccines: current issues &
future. Indian J Med Res 130, 341-347.
KEENAN, S.J., DIAMOND, J., MCCLUGGAGE, W.G., BHARUCHA, H., THOMPSON, D., BARTELS, P.H., et
al. (2000). An automated machine vision system for the histological grading of cervical intraepithelial
neoplasia (CIN). J Pathol 192, 351-362.
KERSEMAEKERS, A.M., VAN DE VIJVER, M.J., KENTER, G.G., and FLEUREN, G.J. (1999). Genetic
alterations during the progression of squamous cell carcinomas of the uterine cervix. Genes
Chromosomes Cancer 26, 346-354.
KIM, S.T., JEONG, H., WOO, O.H., SEO, J.H., KIM, A., LEE, E.S., et al. (2013). Tumor-infiltrating
lymphocytes, tumor characteristics, and recurrence in patients with early breast cancer. Am J Clin
Oncol 36, 224-231.
KLAES, R., FRIEDRICH, T., SPITKOVSKY, D., RIDDER, R., RUDY, W., PETRY, U., et al. (2001).
Overexpression of p16(INK4A) as a specific marker for dysplastic and neoplastic epithelial cells of the
cervix uteri. Int J Cancer 92, 276-284.
KLOOR, M., BECKER, C., BENNER, A., WOERNER, S.M., GEBERT, J., FERRONE, S., et al. (2005).
Immunoselective pressure and human leukocyte antigen class I antigen machinery defects in
microsatellite unstable colorectal cancers. Cancer Res 65, 6418-6424.
KORZENIEWSKI, N., SPARDY, N., DUENSING, A., and DUENSING, S. (2011). Genomic instability and
cancer: lessons learned from human papillomaviruses. Cancer Lett 305, 113-122.
KOTHARI, S., PHAN, J.H., STOKES, T.H., and WANG, M.D. (2013). Pathology imaging informatics for
quantitative analysis of whole-slide images. J Am Med Inform Assoc 20, 1099-1108.
KUNZ, P., FELLENBERG, J., MOSKOVSZKY, L., SAPI, Z., KRENACS, T., POESCHL, J., et al. (2014).
Osteosarcoma microenvironment: whole-slide imaging and optimized antigen detection overcome
major limitations in immunohistochemical quantification. PLoS ONE 9, e90727.
KURIMOTO, A., OGINO, T., ICHII, S., ISOBE, Y., TOBE, M., OGITA, H., et al. (2004). Synthesis and
evaluation of 2-substituted 8-hydroxyadenines as potent interferon inducers with improved oral
bioavailabilities. Bioorg Med Chem 12, 1091-1099.
KUSS, I., SAITO, T., JOHNSON, J.T., and WHITESIDE, T.L. (1999). Clinical significance of decreased zeta
chain expression in peripheral blood lymphocytes of patients with head and neck cancer. Clin Cancer
Res 5, 329-334.
LAHRMANN, B., HALAMA, N., SINN, H.P., SCHIRMACHER, P., JAEGER, D., and GRABE, N. (2011).
Automatic tumor-stroma separation in fluorescence TMAs enables the quantitative high-throughput
analysis of multiple cancer biomarkers. PLoS ONE 6, e28048.
LANZA, L., PEIRANO, L., BOSCO, O., CONTINI, P., FILACI, G., SETTI, M., et al. (1995). Interferons upregulate with different potency HLA class I antigen expression in M14 human melanoma cell line.
Possible interaction with glucocorticoid hormones. Cancer Immunol Immunother 41, 23-28.
164
7. References
LAURSON, J., KHAN, S., CHUNG, R., CROSS, K., and RAJ, K. (2010). Epigenetic repression of E-cadherin
by human papillomavirus 16 E7 protein. Carcinogenesis 31, 918-926.
LEONG, C.M., DOORBAR, J., NINDL, I., YOON, H.S., and HIBMA, M.H. (2010). Deregulation of Ecadherin by human papillomavirus is not confined to high-risk, cancer-causing types. Br J Dermatol
163, 1253-1263.
LIGGETT, W.H., JR., and SIDRANSKY, D. (1998). Role of the p16 tumor suppressor gene in cancer. J
Clin Oncol 16, 1197-1206.
LLOYD, M.C., ALLAM-NANDYALA, P., PUROHIT, C.N., BURKE, N., COPPOLA, D., and BUI, M.M. (2010).
Using image analysis as a tool for assessment of prognostic and predictive biomarkers for breast
cancer: How reliable is it? Journal of pathology informatics 1, 29.
LODDENKEMPER, C., HOFFMANN, C., STANKE, J., NAGORSEN, D., BARON, U., OLEK, S., et al. (2009).
Regulatory (FOXP3+) T cells as target for immune therapy of cervical intraepithelial neoplasia and
cervical cancer. Cancer Sci 100, 1112-1117.
LONGWORTH, M.S., WILSON, R., and LAIMINS, L.A. (2005). HPV31 E7 facilitates replication by
activating E2F2 transcription through its interaction with HDACs. EMBO J 24, 1821-1830.
LU, J., AGGARWAL, R., KANJI, S., DAS, M., JOSEPH, M., POMPILI, V., et al. (2011). Human ovarian
tumor cells escape gammadelta T cell recognition partly by down regulating surface expression of
MICA and limiting cell cycle related molecules. PLoS ONE 6, e23348.
LUBOLDT, H.J., KUBENS, B.S., RUBBEN, H., and GROSSE-WILDE, H. (1996). Selective loss of human
leukocyte antigen class I allele expression in advanced renal cell carcinoma. Cancer Res 56, 826-830.
MAINE, C.J., AZIZ, N.H., CHATTERJEE, J., HAYFORD, C., BREWIG, N., WHILDING, L., et al. (2014).
Programmed death ligand-1 over-expression correlates with malignancy and contributes to immune
regulation in ovarian cancer. Cancer Immunol Immunother 63, 215-224.
MANDIC, R., LIEDER, A., SADOWSKI, M., PELDSZUS, R., and WERNER, J.A. (2004). Comparison of
surface HLA class I levels in squamous cell carcinoma cell lines of the head and neck. Anticancer Res
24, 973-979.
MAO, H., ZHANG, L., YANG, Y., ZUO, W., BI, Y., GAO, W., et al. (2010). New insights of CTLA-4 into its
biological function in breast cancer. Curr Cancer Drug Targets 10, 728-736.
MARSHALL, N.B., and SWAIN, S.L. (2011). Cytotoxic CD4 T cells in antiviral immunity. J Biomed
Biotechnol 2011, 954602.
MARTIN, C.M., and O'LEARY, J.J. (2011). Histology of cervical intraepithelial neoplasia and the role of
biomarkers. Best Pract Res Clin Obstet Gynaecol 25, 605-615.
MASSA, C., and SELIGER, B. (2013). Fast dendritic cells stimulated with alternative maturation
mixtures induce polyfunctional and long-lasting activation of innate and adaptive effector cells with
tumor-killing capabilities. J Immunol 190, 3328-3337.
MATKOWSKI, R., GISTEREK, I., HALON, A., LACKO, A., SZEWCZYK, K., STASZEK, U., et al. (2009). The
prognostic role of tumor-infiltrating CD4 and CD8 T lymphocytes in breast cancer. Anticancer Res 29,
2445-2451.
7. References
165
MATSUI, S., AHLERS, J.D., VORTMEYER, A.O., TERABE, M., TSUKUI, T., CARBONE, D.P., et al. (1999). A
model for CD8+ CTL tumor immunosurveillance and regulation of tumor escape by CD4 T cells
through an effect on quality of CTL. J Immunol 163, 184-193.
MCBRIDE, A.A. (2013). The papillomavirus E2 proteins. Virology 445, 57-79.
MCCREDIE, M.R., SHARPLES, K.J., PAUL, C., BARANYAI, J., MEDLEY, G., JONES, R.W., et al. (2008).
Natural history of cervical neoplasia and risk of invasive cancer in women with cervical intraepithelial
neoplasia 3: a retrospective cohort study. Lancet Oncol 9, 425-434.
MCDERMOTT, D.F., and ATKINS, M.B. (2013). PD-1 as a potential target in cancer therapy. Cancer
medicine 2, 662-673.
MCLAUGHLIN-DRUBIN, M.E., CRUM, C.P., and MÜNGER, K. (2011). Human papillomavirus E7
oncoprotein induces KDM6A and KDM6B histone demethylase expression and causes epigenetic
reprogramming. Proc Natl Acad Sci U S A 108, 2130-2135.
MCLAUGHLIN-DRUBIN, M.E., and MUNGER, K. (2009). The human papillomavirus E7 oncoprotein.
Virology 384, 335-344.
MCLAUGHLIN-DRUBIN, M.E., and MUNGER, K. (2013). Biochemical and functional interactions of
human papillomavirus proteins with polycomb group proteins. Viruses 5, 1231-1249.
MCLAUGHLIN-DRUBIN, M.E., PARK, D., and MUNGER, K. (2013). Tumor suppressor p16 INK4A is
necessary for survival of cervical carcinoma cell lines. Proc Natl Acad Sci U S A 110, 16175-16180.
MEDZHITOV, R. (2007). Recognition of microorganisms and activation of the immune response.
Nature 449, 819-826.
MEHTA, A.M., JORDANOVA, E.S., KENTER, G.G., FERRONE, S., and FLEUREN, G.J. (2008). Association
of antigen processing machinery and HLA class I defects with clinicopathological outcome in cervical
carcinoma. Cancer Immunol Immunother 57, 197-206.
MEISSNER, M., REICHERT, T.E., KUNKEL, M., FERRONE, S., and SELIGER, B. (2005). Defects in the
Human Leukocyte Antigen Class I Antigen Processing Machinery in Head and Neck Squamous Cell
Carcinoma : Association with Clinical Outcome Association with Clinical Outcome. 2552-2560.
MELSHEIMER, P., VINOKUROVA, S., WENTZENSEN, N., BASTERT, G., and VON KNEBEL DOEBERITZ, M.
(2004). DNA aneuploidy and integration of human papillomavirus type 16 e6/e7 oncogenes in
intraepithelial neoplasia and invasive squamous cell carcinoma of the cervix uteri. Clin Cancer Res 10,
3059-3063.
MICHEL, S., BENNER, A., TARIVERDIAN, M., WENTZENSEN, N., HOEFLER, P., POMMERENCKE, T., et al.
(2008). High density of FOXP3-positive T cells infiltrating colorectal cancers with microsatellite
instability. Br J Cancer 99, 1867-1873.
MIGHTY, K.K., and LAIMINS, L.A. (2014). The role of human papillomaviruses in oncogenesis. Recent
Results Cancer Res 193, 135-148.
MOGENSEN, T.H. (2009). Pathogen recognition and inflammatory signaling in innate immune
defenses. Clin Microbiol Rev 22, 240-273, Table of Contents.
166
7. References
MOLES LOPEZ, X., BARBOT, P., VAN EYCKE, Y.R., VERSET, L., TREPANT, A.L., LARBANOIX, L., et al.
(2014). Registration of whole immunohistochemical slide images: an efficient way to characterize
biomarker colocalization. J Am Med Inform Assoc.
MOLHO-PESSACH, V., and LOTEM, M. (2007). Viral carcinogenesis in skin cancer. Curr Probl Dermatol
35, 39-51.
MOLLING, J.W., DE GRUIJL, T.D., GLIM, J., MORENO, M., ROZENDAAL, L., MEIJER, C.J.L.M., et al.
(2007). CD4(+)CD25hi regulatory T-cell frequency correlates with persistence of human
papillomavirus type 16 and T helper cell responses in patients with cervical intraepithelial neoplasia.
International journal of cancer Journal international du cancer 121, 1749-1755.
MONNIER-BENOIT, S., MAUNY, F., RIETHMULLER, D., GUERRINI, J.S., CAPILNA, M., FELIX, S., et al.
(2006). Immunohistochemical analysis of CD4+ and CD8+ T-cell subsets in high risk human
papillomavirus-associated pre-malignant and malignant lesions of the uterine cervix. Gynecol Oncol
102, 22-31.
MONTES, C.L., CHAPOVAL, A.I., NELSON, J., ORHUE, V., ZHANG, X., SCHULZE, D.H., et al. (2008).
Tumor-induced senescent T cells with suppressor function: a potential form of tumor immune
evasion. Cancer Res 68, 870-879.
MORETTA, A., BOTTINO, C., VITALE, M., PENDE, D., BIASSONI, R., MINGARI, M.C., et al. (1996).
Receptors for HLA class-I molecules in human natural killer cells. Annu Rev Immunol 14, 619-648.
MUNK, A.C., KRUSE, A.J., VAN DIERMEN, B., JANSSEN, E.A., SKALAND, I., GUDLAUGSSON, E., et al.
(2007). Cervical intraepithelial neoplasia grade 3 lesions can regress. APMIS 115, 1409-1414.
MUNN, D.H., and MELLOR, A.L. (2013). Indoleamine 2,3 dioxygenase and metabolic control of
immune responses. Trends Immunol 34, 137-143.
NAKAGOMI, H., PETERSSON, M., MAGNUSSON, I., JUHLIN, C., MATSUDA, M., MELLSTEDT, H., et al.
(1993). Decreased expression of the signal-transducing zeta chains in tumor-infiltrating T-cells and NK
cells of patients with colorectal carcinoma. Cancer Res 53, 5610-5612.
NAKAMURA, T., SHIMA, T., SAEKI, A., HIDAKA, T., NAKASHIMA, A., TAKIKAWA, O., et al. (2007).
Expression of indoleamine 2, 3-dioxygenase and the recruitment of Foxp3-expressing regulatory T
cells in the development and progression of uterine cervical cancer. Cancer Sci 98, 874-881.
NÄSMAN, A., ANDERSSON, E., NORDFORS, C., GRÜN, N., JOHANSSON, H., MUNCK-WIKLAND, E., et al.
(2013). MHC class I expression in HPV positive and negative tonsillar squamous cell carcinoma in
correlation to clinical outcome. International journal of cancer Journal international du cancer 132,
72-81.
NATALE, C., GIANNINI, T., LUCCHESE, A., and KANDUC, D. (2000). Computer-assisted analysis of
molecular mimicry between human papillomavirus 16 E7 oncoprotein and human protein sequences.
Immunol Cell Biol 78, 580-585.
NEDERGAARD, B.S., LADEKARL, M., THOMSEN, H.F., NYENGAARD, J.R., and NIELSEN, K. (2007). Low
density of CD3+, CD4+ and CD8+ cells is associated with increased risk of relapse in squamous cell
cervical cancer. Br J Cancer 97, 1135-1138.
NGUYEN, H.P., RAMIREZ-FORT, M.K., and RADY, P.L. (2014). The biology of human papillomaviruses.
Curr Probl Dermatol 45, 19-32.
7. References
167
NIMS, R.W., SYKES, G., COTTRILL, K., IKONOMI, P., and ELMORE, E. (2010). Short tandem repeat
profiling: part of an overall strategy for reducing the frequency of cell misidentification. In Vitro Cell
Dev Biol Anim 46, 811-819.
NISHIKAWA, H., and SAKAGUCHI, S. (2014). Regulatory T cells in cancer immunotherapy. Curr Opin
Immunol 27, 1-7.
NOYA, F., CHIEN, W.M., BROKER, T.R., and CHOW, L.T. (2001). p21cip1 Degradation in differentiated
keratinocytes is abrogated by costabilization with cyclin E induced by human papillomavirus E7. J
Virol 75, 6121-6134.
OH, J.M., KIM, S.H., CHO, E.A., SONG, Y.S., KIM, W.H., and JUHNN, Y.S. (2010). Human papillomavirus
type 16 E5 protein inhibits hydrogen-peroxide-induced apoptosis by stimulating ubiquitinproteasome-mediated degradation of Bax in human cervical cancer cells. Carcinogenesis 31, 402-410.
OHKURA, N., HAMAGUCHI, M., and SAKAGUCHI, S. (2011). FOXP3+ regulatory T cells: control of
FOXP3 expression by pharmacological agents. Trends Pharmacol Sci 32, 158-166.
OLDSTONE, M.B. (1998). Molecular mimicry and immune-mediated diseases. FASEB J 12, 1255-1265.
OSTOR, A.G. (1993). Natural history of cervical intraepithelial neoplasia: a critical review. Int J
Gynecol Pathol 12, 186-192.
OTSU, N. (1979). A threshold selection method from grey level histograms. IEEE Trans Syst Man
Cybern 9, 62–66.
PALEFSKY, J. (2009). Human papillomavirus-related disease in people with HIV. Curr Opin HIV AIDS 4,
52-56.
PALOMARES, O., MARTIN-FONTECHA, M., LAUENER, R., TRAIDL-HOFFMANN, C., CAVKAYTAR, O.,
AKDIS, M., et al. (2014). Regulatory T cells and immune regulation of allergic diseases: roles of IL-10
and TGF-beta. Genes Immun.
PARKER, K.C., BEDNAREK, M.A., and COLIGAN, J.E. (1994). Scheme for ranking potential HLA-A2
binding peptides based on independent binding of individual peptide side-chains. J Immunol 152,
163-175.
PARKIN, D.M., and BRAY, F. (2006). Chapter 2: The burden of HPV-related cancers. Vaccine 24 Suppl
3, S3/11-25.
PASCHEN, A., MENDEZ, R.M., JIMENEZ, P., SUCKER, A., RUIZ-CABELLO, F., SONG, M., et al. (2003).
Complete loss of HLA class I antigen expression on melanoma cells: a result of successive mutational
events. Int J Cancer 103, 759-767.
PASSMORE, J.A., BURCH, V.C., SHEPHARD, E.G., MARAIS, D.J., ALLAN, B., KAY, P., et al. (2002). Singlecell cytokine analysis allows detection of cervical T-cell responses against human papillomavirus type
16 L1 in women infected with genital HPV. J Med Virol 67, 234-240.
PATEL, S., and CHIPLUNKAR, S. (2009). Host immune responses to cervical cancer. Curr Opin Obstet
Gynecol 21, 54-59.
PENG, Y.P., ZHANG, J.J., LIANG, W.B., TU, M., LU, Z.P., WEI, J.S., et al. (2014). Elevation of MMP-9 and
IDO induced by pancreatic cancer cells mediates natural killer cell dysfunction. BMC Cancer 14, 738.
168
7. References
PEPER, J.K., SCHUSTER, H., LOFFLER, M.W., SCHMID-HORCH, B., RAMMENSEE, H.G., and
STEVANOVIC, S. (2014). An impedance-based cytotoxicity assay for real-time and label-free
assessment of T-cell-mediated killing of adherent cells. J Immunol Methods 405, 192-198.
PETER, M., STRANSKY, N., COUTURIER, J., HUPE, P., BARILLOT, E., DE CREMOUX, P., et al. (2010).
Frequent genomic structural alterations at HPV insertion sites in cervical carcinoma. J Pathol 221,
320-330.
PETT, M.R., ALAZAWI, W.O., ROBERTS, I., DOWEN, S., SMITH, D.I., STANLEY, M.A., et al. (2004).
Acquisition of high-level chromosomal instability is associated with integration of human
papillomavirus type 16 in cervical keratinocytes. Cancer Res 64, 1359-1368.
PHAM, P.V., NGUYEN, N.T., NGUYEN, H.M., KHUAT, L.T., LE, P.M., PHAM, V.Q., et al. (2014). A simple
in vitro method for evaluating dendritic cell-based vaccinations. OncoTargets and therapy 7, 14551464.
PIERSMA, S.J., JORDANOVA, E.S., VAN POELGEEST, M.I., KWAPPENBERG, K.M., VAN DER HULST, J.M.,
DRIJFHOUT, J.W., et al. (2007). High number of intraepithelial CD8+ tumor-infiltrating lymphocytes is
associated with the absence of lymph node metastases in patients with large early-stage cervical
cancer. Cancer Res 67, 354-361.
PLATTEN, M., WICK, W., and VAN DEN EYNDE, B.J. (2012). Tryptophan Catabolism in Cancer: Beyond
IDO and Tryptophan Depletion. Cancer Res 72, 5435-5440.
PLETINCKX, K., VAETH, M., SCHNEIDER, T., BEYERSDORF, N., HUNIG, T., BERBERICH-SIEBELT, F., et al.
(2014). Immature dendritic cells convert anergic non-regulatory T cells into Foxp3 IL-10 regulatory T
cells by engaging CD28 and CTLA-4. Eur J Immunol.
POLLACK, S.M., JONES, R.L., FARRAR, E.A., LAI, I.P., LEE, S.M., CAO, J., et al. (2014). Tetramer guided,
cell sorter assisted production of clinical grade autologous NY-ESO-1 specific CD8(+) T cells. Journal
for immunotherapy of cancer 2, 36.
PRIGGE, E.S., TOTH, C., DYCKHOFF, G., WAGNER, S., MULLER, F., WITTEKINDT, C., et al. (2014). p16
/Ki-67 co-expression specifically identifies transformed cells in the head and neck region. Int J Cancer.
PRIME, S.S., PITIGALA-ARACHCHI, A., CRANE, I.J., ROSSER, T.J., and SCULLY, C. (1987). The expression
of cell surface MHC class I heavy and light chain molecules in pre-malignant and malignant lesions of
the oral mucosa. Histopathology 11, 81-91.
RAMMENSEE, H., BACHMANN, J., EMMERICH, N.P., BACHOR, O.A., and STEVANOVIC, S. (1999).
SYFPEITHI: database for MHC ligands and peptide motifs. Immunogenetics 50, 213-219.
RAZMKHAH, M., and GHADERI, A. (2013). HLA Class I Allele Frequencies in Southern Iranian Women
with Breast Cancer. Iranian journal of basic medical sciences 16, 140-143.
REIMERS, M.S., ENGELS, C.C., PUTTER, H., MORREAU, H., LIEFERS, G.J., VAN DE VELDE, C.J., et al.
(2014). Prognostic value of HLA class I, HLA-E, HLA-G and Tregs in rectal cancer: a retrospective
cohort study. BMC Cancer 14, 486.
RESSING, M.E., VAN DRIEL, W.J., CELIS, E., SETTE, A., BRANDT, M.P., HARTMAN, M., et al. (1996).
Occasional memory cytotoxic T-cell responses of patients with human papillomavirus type 16positive cervical lesions against a human leukocyte antigen-A *0201-restricted E7-encoded epitope.
Cancer Res 56, 582-588.
7. References
169
REUSCHENBACH, M., TRAN, T., FAULSTICH, F., HARTSCHUH, W., VINOKUROVA, S., KLOOR, M., et al.
(2011). High-risk human papillomavirus in non-melanoma skin lesions from renal allograft recipients
and immunocompetent patients. Br J Cancer 104, 1334-1341.
RICHART, R.M. (1973). Cervical intraepithelial neoplasia. Pathol Annu 8, 301-328.
ROCCO, J.W., and SIDRANSKY, D. (2001). p16(MTS-1/CDKN2/INK4a) in cancer progression. Exp Cell Res
264, 42-55.
RODRIGUEZ, J.A., GALEANO, L., PALACIOS, D.M., GOMEZ, C., SERRANO, M.L., BRAVO, M.M., et al.
(2012). Altered HLA class I and HLA-G expression is associated with IL-10 expression in patients with
cervical cancer. Pathobiology 79, 72-83.
ROJO, M.G., GARCIA, G.B., MATEOS, C.P., GARCIA, J.G., and VICENTE, M.C. (2006). Critical comparison
of 31 commercially available digital slide systems in pathology. International journal of surgical
pathology 14, 285-305.
RUBIO, V., STUGE, T.B., SINGH, N., BETTS, M.R., WEBER, J.S., ROEDERER, M., et al. (2003). Ex vivo
identification, isolation and analysis of tumor-cytolytic T cells. Nat Med 9, 1377-1382.
RUIFROK, A.C., and JOHNSTON, D.A. (2001). Quantification of histochemical staining by color
deconvolution. Anal Quant Cytol Histol 23, 291-299.
SCHAPIRO, F., SPARKOWSKI, J., ADDUCI, A., SUPRYNOWICZ, F., SCHLEGEL, R., and GRINSTEIN, S.
(2000). Golgi alkalinization by the papillomavirus E5 oncoprotein. J Cell Biol 148, 305-315.
SCHIFFMAN, M., CASTLE, P.E., JERONIMO, J., RODRIGUEZ, A.C., and WACHOLDER, S. (2007). Human
papillomavirus and cervical cancer. Lancet 370, 890-907.
SCHIFFMAN, M., and WENTZENSEN, N. (2010). From human papillomavirus to cervical cancer. Obstet
Gynecol 116, 177-185.
SCHMITT, M., BRAVO, I.G., SNIJDERS, P.J., GISSMANN, L., PAWLITA, M., and WATERBOER, T. (2006).
Bead-based multiplex genotyping of human papillomaviruses. J Clin Microbiol 44, 504-512.
SELIGER, B., CABRERA, T., GARRIDO, F., and FERRONE, S. (2002). HLA class I antigen abnormalities and
immune escape by malignant cells. Semin Cancer Biol 12, 3-13.
SERNEE, M.F., PLOEGH, H.L., and SCHUST, D.J. (1998). Why certain antibodies cross-react with HLA-A
and HLA-G: epitope mapping of two common MHC class I reagents. Mol Immunol 35, 177-188.
SHAH, W., YAN, X., JING, L., ZHOU, Y., CHEN, H., and WANG, Y. (2011). A reversed CD4/CD8 ratio of
tumor-infiltrating lymphocytes and a high percentage of CD4(+)FOXP3(+) regulatory T cells are
significantly associated with clinical outcome in squamous cell carcinoma of the cervix. Cell Mol
Immunol 8, 59-66.
SHEPHERD, P.S., ROWE, A.J., CRIDLAND, J.C., COLETART, T., WILSON, P., and LUXTON, J.C. (1996).
Proliferative T cell responses to human papillomavirus type 16 L1 peptides in patients with cervical
dysplasia. The Journal of general virology 77 ( Pt 4), 593-602.
SHEU, B.-C., CHIOU, S.-H., LIN, H.-H., CHOW, S.-N., HUANG, S.-C., HO, H.-N., et al. (2005). Upregulation of Inhibitory Natural Killer Receptors CD94/NKG2A with Suppressed Intracellular Perforin
Expression of Tumor-Infiltrating CD8+ T Lymphocytes in Human Cervical Carcinoma. Cancer Res 65,
2921-2929.
170
7. References
SHEU, B.-C., HSU, S.-M., HO, H.-N., LIEN, H.-C., HUANG, S.-C., and LIN, R.-H. (2001). A Novel Role of
Metalloproteinase in Cancer-mediated Immunosuppression. Cancer Res 61, 237-242.
SILVA, C.S., MICHELIN, M.A., ETCHEBEHERE, R.M., ADAD, S.J., and MURTA, E.F. (2010). Local
lymphocytes and nitric oxide synthase in the uterine cervical stroma of patients with grade III cervical
intraepithelial neoplasia. Clinics (Sao Paulo) 65, 575-581.
SIMOENS, C., GOFFIN, F., SIMON, P., BARLOW, P., ANTOINE, J., FOIDART, J.M., et al. (2012). Adverse
obstetrical outcomes after treatment of precancerous cervical lesions: a Belgian multicentre study.
BJOG 119, 1247-1255.
SMOLA, S. (2014). Human papillomaviruses and skin cancer. Adv Exp Med Biol 810, 192-207.
SNIJDERS, P.J., STEENBERGEN, R.D., HEIDEMAN, D.A., and MEIJER, C.J. (2006). HPV-mediated cervical
carcinogenesis: concepts and clinical implications. J Pathol 208, 152-164.
SOONG, R.S., SONG, L., TRIEU, J., KNOFF, J., HE, L., TSAI, Y.C., et al. (2014). Toll-like Receptor Agonist
Imiquimod Facilitates Antigen-Specific CD8+ T-cell Accumulation in the Genital Tract Leading to
Tumor Control through IFNgamma. Clin Cancer Res 20, 5456-5467.
SOUMELIS, V., and LIU, Y.J. (2006). From plasmacytoid to dendritic cell: morphological and functional
switches during plasmacytoid pre-dendritic cell differentiation. Eur J Immunol 36, 2286-2292.
STAM, N.J., SPITS, H., and PLOEGH, H.L. (1986). Monoclonal antibodies raised against denatured HLAB locus heavy chains permit biochemical characterization of certain HLA-C locus products. J Immunol
137, 2299-2306.
STANLEY, M. (2006). Immune responses to human papillomavirus. Vaccine 24 Suppl 1, S16-22.
STANLEY, M. (2008). Immunobiology of HPV and HPV vaccines. Gynecol Oncol 109, S15-21.
STANLEY, M. (2012a). Epithelial Cell Responses to Infection with Human Papillomavirus. Clin
Microbiol Rev 25, 215-222.
STANLEY, M. (2012b). Perspective: Vaccinate boys too. Nature 488, S10.
STANLEY, M.A. (2002). Imiquimod and the imidazoquinolones: mechanism of action and therapeutic
potential. Clin Exp Dermatol 27, 571-577.
STARY, G., BANGERT, C., TAUBER, M., STROHAL, R., KOPP, T., and STINGL, G. (2007). Tumoricidal
activity of TLR7/8-activated inflammatory dendritic cells. J Exp Med 204, 1441-1451.
STEBEN, M., and GARLAND, S.M. (2014). Genital warts. Best Pract Res Clin Obstet Gynaecol.
STEVENSON, M.M., and RILEY, E.M. (2004). Innate immunity to malaria. Nat Rev Immunol 4, 169-180.
SUZUKI, H., WANG, B., SHIVJI, G.M., TOTO, P., AMERIO, P., TOMAI, M.A., et al. (2000). Imiquimod, a
topical immune response modifier, induces migration of Langerhans cells. J Investig Dermatol 114,
135-141.
SYRJANEN, S., NAUD, P., SARIAN, L., DERCHAIN, S., ROTELI-MARTINS, C., LONGATTO-FILHO, A., et al.
(2009). Immunosuppressive cytokine Interleukin-10 (IL-10) is up-regulated in high-grade CIN but not
associated with high-risk human papillomavirus (HPV) at baseline, outcomes of HR-HPV infections or
incident CIN in the LAMS cohort. Virchows Arch 455, 505-515.
7. References
171
TANG, A.L., HAUFF, S.J., OWEN, J.H., GRAHAM, M.P., CZERWINSKI, M.J., PARK, J.J., et al. (2012). UMSCC-104: a new human papillomavirus-16-positive cancer stem cell-containing head and neck
squamous cell carcinoma cell line. Head Neck 34, 1480-1491.
TANG, Q., ZHANG, J., QI, B., SHEN, C., and XIE, W. (2009). Downregulation of HLA class I molecules in
primary oral squamous cell carcinomas and cell lines. Arch Med Res 40, 256-263.
TAYLOR, C.R. (2014). Issues in using whole slide imaging for diagnostic pathology: "routine" stains,
immunohistochemistry and predictive markers. Biotech Histochem 89, 419-423.
TELANG, S., RASKU, M.A., CLEM, A.L., CARTER, K., KLARER, A.C., BADGER, W.R., et al. (2011). Phase II
trial of the regulatory T cell-depleting agent, denileukin diftitox, in patients with unresectable stage
IV melanoma. BMC Cancer 11, 515.
TERLOU, A., VAN SETERS, M., KLEINJAN, A., HEIJMANS-ANTONISSEN, C., SANTEGOETS, L.A.M.,
BECKMANN, I., et al. (2010). Imiquimod-induced clearance of HPV is associated with normalization of
immune cell counts in usual type vulvar intraepithelial neoplasia. International journal of cancer
Journal international du cancer 127, 2831-2840.
THOMAS, L.K., BERMEJO, J.L., VINOKUROVA, S., JENSEN, K., BIERKENS, M., STEENBERGEN, R., et al.
(2013). Chromosomal gains and losses in human papillomavirus-associated neoplasia of the lower
genital tract - A systematic review and meta-analysis. Eur J Cancer.
TIKIDZHIEVA, A., BENNER, A., MICHEL, S., FORMENTINI, A., LINK, K.H., DIPPOLD, W., et al. (2012).
Microsatellite instability and Beta2-Microglobulin mutations as prognostic markers in colon cancer:
results of the FOGT-4 trial. Br J Cancer 106, 1239-1245.
TODD, R.W., STEELE, J.C., ETHERINGTON, I., and LUESLEY, D.M. (2004). Detection of CD8+ T cell
responses to human papillomavirus type 16 antigens in women using imiquimod as a treatment for
high-grade vulval intraepithelial neoplasia. Gynecol Oncol 92, 167-174.
TRIMBLE, C.L., CLARK, R.A., THOBURN, C., HANSON, N.C., TASSELLO, J., FROSINA, D., et al. (2010).
Human Papillomavirus 16-Associated Cervical Intraepithelial Neoplasia in Humans Excludes CD8 T
Cells from Dysplastic Epithelium. The Journal of Immunology 185, 7107-7114.
TSOUMPOU, I., ARBYN, M., KYRGIOU, M., WENTZENSEN, N., KOLIOPOULOS, G., MARTIN-HIRSCH, P.,
et al. (2009). p16(INK4a) immunostaining in cytological and histological specimens from the uterine
cervix: a systematic review and meta-analysis. Cancer Treat Rev 35, 210-220.
TUVE, S., CHEN, B.M., LIU, Y., CHENG, T.L., TOURE, P., SOW, P.S., et al. (2007). Combination of tumor
site-located CTL-associated antigen-4 blockade and systemic regulatory T-cell depletion induces
tumor-destructive immune responses. Cancer Res 67, 5929-5939.
VAN SETERS, M., FONS, G., and VAN BEURDEN, M. (2002). Imiquimod in the treatment of multifocal
vulvar intraepithelial neoplasia 2/3. Results of a pilot study. J Reprod Med 47, 701-705.
VAN SETERS, M., VAN BEURDEN, M., TEN KATE, F.J., BECKMANN, I., EWING, P.C., EIJKEMANS, M.J., et
al. (2008). Treatment of vulvar intraepithelial neoplasia with topical imiquimod. N Engl J Med 358,
1465-1473.
VERNON, S.D., UNGER, E.R., MILLER, D.L., LEE, D.R., and REEVES, W.C. (1997). Association of human
papillomavirus type 16 integration in the E2 gene with poor disease-free survival from cervical
cancer. Int J Cancer 74, 50-56.
172
7. References
VISSER, J., NIJMAN, H.W., HOOGENBOOM, B.N., JAGER, P., VAN BAARLE, D., SCHUURING, E., et al.
(2007). Frequencies and role of regulatory T cells in patients with (pre)malignant cervical neoplasia.
Clin Exp Immunol 150, 199-209.
VOLOSHIN, T., FREMDER, E., and SHAKED, Y. (2014). Small But Mighty: Microparticles as Mediators of
Tumor Progression. Cancer Microenviron.
VON KNEBEL DOEBERITZ, M. (2002). New markers for cervical dysplasia to visualise the genomic
chaos created by aberrant oncogenic papillomavirus infections. Eur J Cancer 38, 2229-2242.
VON KNEBEL DOEBERITZ, M., REUSCHENBACH, M., SCHMIDT, D., and BERGERON, C. (2012).
Biomarkers for cervical cancer screening: the role of p16( INK4a) to highlight transforming HPV
infections. Expert Rev Proteomics 9, 149-163.
WANG, S.S., TRUNK, M., SCHIFFMAN, M., HERRERO, R., SHERMAN, M.E., BURK, R.D., et al. (2004).
Validation of p16INK4a as a marker of oncogenic human papillomavirus infection in cervical biopsies
from a population-based cohort in Costa Rica. Cancer Epidemiol Biomarkers Prev 13, 1355-1360.
WANG, Y., CROOKES, D., DIAMOND, J., HAMILTON, P., and TURNER, R. (2007). Segmentation of
squamous epithelium from ultra-large cervical histological virtual slides. Conf Proc IEEE Eng Med Biol
Soc 2007, 775-778.
WEBSTER, J.D., and DUNSTAN, R.W. (2014). Whole-slide imaging and automated image analysis:
considerations and opportunities in the practice of pathology. Vet Pathol 51, 211-223.
WENTZENSEN, N., VINOKUROVA, S., and VON KNEBEL DOEBERITZ, M. (2004). Systematic review of
genomic integration sites of human papillomavirus genomes in epithelial dysplasia and invasive
cancer of the female lower genital tract. Cancer Res 64, 3878-3884.
WENZEL, J., UERLICH, M., HALLER, O., BIEBER, T., and TUETING, T. (2005). Enhanced type I interferon
signaling and recruitment of chemokine receptor CXCR3-expressing lymphocytes into the skin
following treatment with the TLR7-agonist imiquimod. J Cutan Pathol 32, 257-262.
WESTERMANN, C., FISCHER, A., and CLAD, A. (2013). Treatment of vulvar intraepithelial neoplasia
with topical 5% imiquimod cream. Int J Gynaecol Obstet 120, 266-270.
WESTRA, W.H. (2012). The morphologic profile of HPV-related head and neck squamous carcinoma:
implications for diagnosis, prognosis, and clinical management. Head and neck pathology 6 Suppl 1,
S48-54.
WHITESIDE, T.L. (2004). Down-regulation of zeta-chain expression in T cells: a biomarker of prognosis
in cancer? Cancer Immunol Immunother 53, 865-878.
WISE-DRAPER, T.M., and WELLS, S.I. (2008). Papillomavirus E6 and E7 proteins and their cellular
targets. Front Biosci 13, 1003-1017.
WOLF, A.M., WOLF, D., STEURER, M., GASTL, G., GUNSILIUS, E., and GRUBECK-LOEBENSTEIN, B.
(2003). Increase of regulatory T cells in the peripheral blood of cancer patients. Clin Cancer Res 9,
606-612.
WOO, Y.L., STERLING, J., DAMAY, I., COLEMAN, N., CRAWFORD, R., VAN DER BURG, S.H., et al. (2008).
Characterising the local immune responses in cervical intraepithelial neoplasia: a cross-sectional and
longitudinal analysis. BJOG 115, 1616-1621; discussion 1621-1612.
7. References
173
WU, M.Y., KUO, T.Y., and HO, H.N. (2011). Tumor-infiltrating lymphocytes contain a higher
proportion of FOXP3(+) T lymphocytes in cervical cancer. J Formos Med Assoc 110, 580-586.
WU, P., WU, D., NI, C., YE, J., CHEN, W., HU, G., et al. (2014). γδT17 Cells Promote the Accumulation
and Expansion of Myeloid-Derived Suppressor Cells in Human Colorectal Cancer. Immunity 40, 785800.
WU, T.C. (2007). The role of vascular cell adhesion molecule-1 in tumor immune evasion. Cancer Res
67, 6003-6006.
YIM, E.K., and PARK, J.S. (2005). The role of HPV E6 and E7 oncoproteins in HPV-associated cervical
carcinogenesis. Cancer research and treatment : official journal of Korean Cancer Association 37,
319-324.
ZEHBE, I., PILCH, H., TOMMASINO, M., and MÄURER, M. (2002). Reduced T-cell receptor ( TCR ) zeta
chain expression in cervical cancer. 8, 2-11.
ZHAO, K.N., and CHEN, J. (2011). Codon usage roles in human papillomavirus. Rev Med Virol 21, 397411.
ZHENG, Z.M., and BAKER, C.C. (2006). Papillomavirus genome structure, expression, and posttranscriptional regulation. Front Biosci 11, 2286-2302.
ZHOU, Q., ZHU, K., and CHENG, H. (2013). Toll-like receptors in human papillomavirus infection. Arch
Immunol Ther Exp 61, 203-215.
ZUR HAUSEN, H. (2011). Infections causing cancer (Weinheim: WILEYl-VCH).
ZUR HAUSEN, H., GISSMANN, L., STEINER, W., DIPPOLD, W., and DREGER, I. (1975). Human papilloma
viruses and cancer. Bibl Haematol, 569-571.
ZUR HAUSEN, J., SCHULTE-HOLTHAUSEN, H., WOLF, H., DORRIES, K., and EGGER, H. (1974). Attempts
to detect virus-specific DNA in human tumors. II. Nucleic acid hybridizations with complementary
RNA of human herpes group viruses. Int J Cancer 13, 657-664.
174
8.
8. Publications, Presentations, Poster
PUBLICATIONS,
PRESENTATIONS AND
POSTERS
Publications
VINOKUROVA S, VON KNEBEL DOEBERITZ M, SAUER M, REUSCHENBACH M.
Compounds and Methods for increasing the Immune Response to Papillomavirus. Patent UH12178EP
AD/NH (submitted 12.2.2014).
SAUER, M., SCHÄFER, K., SINN, P., SCHMIDT, D., KLOOR, M., NELIUS, N., SOHN, C.,
EICHBAUM, M., VON KNEBEL DOEBERITZ, M., REUSCHENBACH, M. Immune cell
infiltration in relation to p16INK4a expression in cervical intraepithelial neoplasia. Submitted.
SAUER, M., REUSCHENBACH, M., WENTZENSEN, N., FERRONE, S., LAHRMANN, B.,
GRABE, N., SCHMIDT, D., VON KNEBEL DOEBERITZ, M., KLOOR, M. HLA class II antigen
expression in cervical intraepithelial neoplasia and invasive cancer. Manuscript in preparation.
Presentations and Posters
Sauer M, Hampl M, Nehls N, Schlotfeldt I, Wentzensen N, Sinn P, von Knebel Doeberitz M,
Reuschenbach M. Local and systemic immune parameters in CIN and VIN patients
27th International Papillomavirus Conference, Berlin, Germany, 2011 (Poster).
Sauer M, Hampl M, Schaefer K, Schlotfeldt I, Wentzensen N, Sinn P, von Knebel Doeberitz M,
Reuschenbach M. Characterization of the local immune response in cervical and vulvar
intraepithelial lesions. EUROGIN, Prague, Czech Republic, 2012 (Oral presentation).
Sauer M, Schaefer K, Schlotfeldt I, Wentzensen N, Sinn P, Schmidt D, von Knebel Doeberitz M,
Reuschenbach M. Immune cell infiltration in HPV-induced carcinogenesis. CIMT, Mayence,
Germany, 2013 (Poster).
Sauer M, Schaefer K, Schlotfeldt I, Wentzensen N, Sinn P, Schmidt D, von Knebel Doeberitz M,
Reuschenbach M. Immune cell infiltration in HPV-induced carcinogenesis. Tumorimmunology
meets Oncology (TIMO IX), Halle (Saale), Germany (Oral presentation).
Sauer M, Schaefer K, Schlotfeldt I, Wentzensen N, Sinn P, Schmidt D, Nelius N, von Knebel
Doeberitz M, Reuschenbach M. Immune cell infiltration in relation to p16 INK4a expression in
cervical intraepithelial neoplasia and cancer. EUROGIN, Florence, Italy. 2013 (Oral presentation).
Sauer M, Reuschenbach M, Wentzensen N, Ferrone S, Lahrmann B, Grabe N, Schmidt D, von Knebel
Doeberitz M, Kloor M. HLA class II antigen expression in cervical intraepithelial neoplasia and
invasive cancer. 29th International Papillomavirus Conference, Seattle, USA, 2014 (Poster)
9. Supplementary Material
9.
175
SUPPLEMENTARY MATERIAL
176
9. Supplementary Material
9.1
Supplementary Figures
FIGURE S9.1
Distribution of CD3+ and CD8+ T cell counts/mm² and the Ratio CD8/CD3 in the epithelium and stromal compartments in non-responders compared with responders in week 0
(before treatment). The line in the center of each box represents the median value of the distribution; the borders of the box represent the upper and lower quartiles (25%-75%).
9. Supplementary Material
FIGURE S9.2
177
Distribution of CD3+ and CD8+ T cell counts/mm² and the Ratio CD8/CD3 in the stromal compartments in non-responders compared with responders in week 0 (before
treatment). The line in the center of each box represents the median value of the distribution; the borders of the box represent the upper and lower quartiles (25%-75%).
178
FIGURE S9.x
9. Supplementary Material
Distribution of CD3+ and CD8+ T cell counts/mm² and the Ratio CD8/CD3 in the stromal compartments in non-responders compared with responders in week 20 (after
treatment). The line in the center of each box represents the median value of the distribution; the borders of the box represent the upper and lower quartiles (25%-75%).
9. Supplementary Material
FIGURE S9.x
179
Distribution of CD3+ and CD8+ T cell counts/mm² and the Ratio CD8/CD3 in the stromal compartments in non-responders compared with responders in week 20 (after
treatment). The line in the center of each box represents the median value of the distribution; the borders of the box represent the upper and lower quartiles (25%-75%).
180
FIGURE S9.5
13. Supplementary Material
Distribution of ratios for epithelial to stromal cell counts non-responders compared with responders in
week 20 (after treatment). The line in the center of each box represents the median value of the
distribution; the borders of the box represent the upper and lower quartiles (25%-75%).
13. Supplementary Material
FIGURE S9.6
181
Distribution of CD3+ and CD8+ T cell counts/mm² and the Ratio CD8/CD3 in the epithelial and stromal
compartments in non-responders compared with responders. Data for week 0 (before treatment) and week
20 (after treatment) are shown next to each other. The line in the center of each box represents the median
value of the distribution; the borders of the box represent the upper and lower quartiles (25%-75%).
182
FIGURE S9.7
9. Supplementary Material
Distribution of CD3+ and CD8+ T cell counts/mm² and the Ratio CD8/CD3 in the stromal compartments
in non-responders compared with responders. Data for week 0 (before treatment) and week 20 (after
treatment) are shown next to each other. The line in the center of each box represents the median value of
the distribution; the borders of the box represent the upper and lower quartiles (25%-75%).
9. Supplementary Material
FIGURE S9.8
Peptide binding assay – repetition and validation of the results obtained in the first assay.
183
184
FIGURE S9.9
9. Supplementary Material
Human Cell Line Authentication Report for the HNSCC cell line HN038.
9. Supplementary Material
9.1
Supplementary Tables
mean
intraepithelial
range
SD
CD3
Low-grade CIN
p16- CIN1
p16+ CIN1
17.6
13.3
9.92
12.14
8.3-43.0
4.3-99.3
9.98
25.52
2.0-161.1
42.72
0.0-23.0
0.3-20.0
6.72
5.22
3.7-32.3
3.3-39.7
9.68
11.10
0.7-37.9
12.55
0.0-1.7
0.0-4.7
0.67
1.29
0.0-5.3
0.0-8.3
1.66
2.29
0.0-5.3
1.63
0.3-9.7
0.0-9.0
3.72
2.49
2.0-17.0
0.0-5.7
4.05
1.82
0.0-36.0
8.80
3.7-107.0
3.7-31.0
32.30
7.42
8.3-22.7
5.3-38.9
5.22
9.00
2.0-45.3
13.09
High-grade CIN
CIN2
CIN3
22.3
25.2
CxCa
43.8
CD8
8.5
4.8
14.8
14.1
15.7
GranB
p16- CIN1
p16+ CIN1
0.5
0.5
1.1
2.2
1.5
Foxp3
3.4
2.2
5.8
2.1
4.9
p16- CIN1
p16+ CIN1
25.2
10.4
15.6
16.3
TABLE S9.1
17.0
77.58
intraepithelial/stromal
range
SD
p-value
17.6
14.4
1.8-39.5
3.0-27.3
13.89
8.57
37.0
38.0
7.8-56.8
11.5-73.5
16.90
19.34
58.3
9.8-149.3
42.72
0.9
0.5
0.0-3.5
0.0-2.3
1.21
0.70
0.8
2.7
0.0-2.5
0.8-12.3
1.01
2.75
5.7
0.0-20.0
5.98
17.4
7.6
0.0-46.0
1.5-16.8
17.13
5.58
21.7
19.3
6.0-37.8
1.5-45.3
12.29
13.11
42.1
3.3-97.8
31.20
38.6
26.5
5.0-83.5
8.5-49.0
23.96
13.97
56.5
58.6
16.0-81.8
13.5-141.5
21.94
37.57
60.8
14.0-115.0
34.58
* p=0.6498
0.46
0.41
0.07-1.04
0.07-0.89
0.330
0.249
0.42
0.26
0.13-1.33
0.06-0.64
0.371
0.212
0.41
0.02-1.52
0.403
0.72
0.42
0.00-1.90
0.01-1.14
0.655
0.351
0.59
0.39
0.14-1.85
0.11-0.67
0.642
0.204
0.36
0.02-0.92
0.294
0.68
0.46
0.00-1.67
0.00-1.33
0.689
0.582
0.88
0.73
0.00-1.70
0.00-1.93
0.715
0.542
0.40
0.00-1.67
0.427
0.25
0.32
0.04-0.67
0.00-1.33
0.200
0.355
0.39
0.16
0.07-1.06
0.00-0.71
0.343
0.198
0.17
0.00-1.58
0.384
0.64
0.42
0.14-1.98
0.14-0.73
0.542
0.203
0.37
0.38
0.14-1.38
0.12-1.21
0.383
0.281
0.39
0.04-1.31
0.363
** p=0.0799
*** p=0.0414
*** p=0.9244
* p=0.7856
* p=0.2012
** p<0.0001
** p=0.4774
*** p=0.2045
*** p=0.3929
* p=0.5262
* p=0.9523
** p=0.0014
** p=0.1838
*** p=0.0095
*** p=0.0467
* p=0.6444
* p=0.7237
** p=0.0076
** p=0.2136
*** p=0.0243
* p=0.1910
*** p=0.0464
* p=0.2282
** p=0.0286
Invasive disease
CxCa
34.3-336.0
*** p=0.5933
High-grade CIN
CIN2
CIN3
140.5
** p=0.2558
Low-grade CIN
mean
** p<0.0001
* p=0.5375
Invasive disease
CxCa
33.28
63.10
*** p=0.9193
High-grade CIN
CIN2
CIN3
18.0-125.0
33.5-260.5
** p=0.0028
Low-grade CIN
p16- CIN1
p16+ CIN1
73.0
110.1
* p=0.9039
Invasive disease
CxCa
43.30
19.39
*** p=0.8802
High-grade CIN
CIN2
CIN3
15.8-136.8
11.5-84.8
** p=0.0012
Low-grade CIN
p-value
* p=0.2035
56.3
32.6
* p=0.0864
Invasive disease
CxCa
stromal
range
SD
*** p=0.2968
High-grade CIN
CIN2
CIN3
mean
** p=0.0273
Low-grade CIN
p16- CIN1
p16+ CIN1
p-value
* p=0.1438
7.3-37.3
1.3-45.3
Invasive disease
CD3ζ
185
* p=0.3897
** p=0.0022
*** p=0.7623
** p=0.0672
*** p=0.6801
*** p=0.6801
Mean cell numbers, ranges and standard deviations (SD) for all T cell phenotypes in correlation with the
lesion grades and p16INK4a expression status: intraepithelial and stromal cell numbers (per 0.0625mm²)
and ratio lesion/lesion-adjacent stroma.
(1) p-values (comparison p16INK4a-negative and p16INK4a-positive low-grade CIN)
(2) p-values (comparison low-grade CIN vs. high-grade CIN)
(3) p-values (comparison high-grade CIN vs. invasive disease)
186
9. Supplementary Material
mean
intraepithelial
range
SD
CD8/CD3
Low-grade CIN
p16- CIN1
p16+ CIN1
CIN2
CIN3
0.70
0.79
GranB/CD3
0.265
0.215
0.216-1.906
0.11-1.94
0.464
0.532
p16- CIN1
p16+ CIN1
0.04
0.03
0.04-1.69
0.479
0.09
0.19
0.00-0.11
0.00-0.20
0.049
0.066
0.00-0.64
0.00-0.77
0.197
0.230
0.00-0.53
0.153
p16- CIN1
p16+ CIN1
0.20
0.20
0.03-0.68
0.00-0.67
0.237
0.197
0.26
0.18
0.06-0.40
0.00-1.31
0.100
0.301
0.18
0.00-0.87
0.266
1.10
0.99
0.35-1.97
0.32-1.75
0.658
0.459
0.80
0.96
0.53-1.00
0.14-1.77
0.174
0.458
TABLE S9.2
0.67
0.23-1.88
0.525
0.226
0.183
0.40
0.18-0.82
0.186
0.02
0.01
0.00-0.05
0.00-0.06
0.020
0.021
0.01
0.03
0.00-0.04
0.00-0.14
0.017
0.038
0.04
0.00-0.18
0.044
0.26
0.23
0.00-0.54
0.07-0.46
0.164
0.117
0.32
0.22
0.20-0.55
0.02-0.54
0.102
0.138
0.28
0.28
0.08-0.60
0.149
0.79
0.88
0.26-1.49
0.49-1.60
0.434
0.340
0.82
0.59
0.60-1.27
0.10-1.23
0.206
0.260
0.47
0.16-0.80
0.176
** p=0.1314
*** p=0.2000
* p=0.8950
** p=0.0508
*** p=0.0418
* p=0.6498
** p=0.8833
*** p=0.5747
* p=0.8576
* p=0.8980
** p=0.5522
Invasive disease
CxCa
0.275-1.018
0.13-0.77
*** p=0.2318
High-grade CIN
CIN2
CIN3
0.53
0.41
** p=0.7690
Low-grade CIN
p16- CIN1
p16+ CIN1
0.139
0.159
* p=0.8403
Invasive disease
CxCa
0.09-0.52
0.26-0.70
*** p=0.3762
High-grade CIN
CIN2
CIN3
0.31
0.43
** p=0.0041
Low-grade CIN
p-value
* p=0.1861
* p=0.9039
Invasive disease
0.09
SD
*** p=0.0090
High-grade CIN
CIN2
CIN3
stromal
range
** p=0.0258
Low-grade CIN
CxCa
Foxp3/CD3
0.00-1.00
0.01-0.87
Invasive disease
0.52
mean
* p=0.1347
High-grade CIN
CxCa
CD3ζ/CD3
0.48
0.39
p-value
** p=0.0700
*** p=0.1171
*** p=0.0090
Means, ranges and standard deviations (SD) for the ratios calculated between different T cell phenotypes
and CD3+ T cells in correlation with the lesion grades and p16 INK4a expression status (means per
0.0625mm²).
(1) p-values (comparison p16INK4a-negative and p16INK4a-positive low-grade CIN)
(2) p-values (comparison low-grade CIN vs. high-grade CIN)
(3) p-values (comparison high-grade CIN vs. invasive disease)
9. Supplementary Material
progressing/persistent CIN2/3
CD3
Epithel
M100
M500
M1000
M0-500
regressing CIN2/3
mean
min
max
STD
mean
min
max
STD
p=
V1
V4
V7
537,0
36,2
1194,3
570,45
160,8
20,2
610,1
252,08
0.190
751,3
114,5
1388,1
900,61
427,6
56,4
1846,7
698,63
1.000
287,8
135,8
439,8
215,02
371,1
140,4
645,6
231,96
0.429
V1
V4
V7
1606,1
669,7
3460,0
1270,71
1386,8
722,5
1890,1
541,14
0.905
4593,7
1960,1
6790,1
2444,51
1978,2
483,2
6193,9
2176,92
0.167
1843,5
1675,7
2011,2
237,26
2166,2
1172,4
4934,5
1375,77
1.000
V1
V4
V7
1883,9
946,7
3699,8
1236,12
945,9
420,9
1936,5
663,05
0.190
4565,1
2230,5
5750,2
2021,84
3079,7
237,7
7985,2
2800,21
0.381
1532,7
1320,3
1745,2
300,47
2401,4
991,4
5600,2
1689,64
0.643
V1
V4
V7
V1
V4
V7
M0-1000 V1
V4
V7
TABLE S9.3
187
883,0
396,2
1838,1
675,40
597,9
53,0
1831,8
719,44
0.286
3308,3
1907,0
5408,2
1852,21
2159,1
173,9
6842,0
2478,56
0.381
1056,9
639,0
1474,7
590,91
2008,0
432,1
5818,5
1948,59
0.643
3490,1
1860,3
7159,8
2491,96
2332,6
1155,8
3703,6
1092,75
0.730
9158,8
4190,7
12540,3
4395,14
5057,9
734,3
14179,1
4926,73
0.381
3376,2
3331,5
3420,9
63,21
4567,6
2598,7
10534,6
3021,80
1.000
4373,0
2275,1
8042,6
2532,27
2930,6
1208,8
4346,4
1413,45
0.730
12467,1
6097,6
16153,7
5538,95
7217,0
908,2
21021,0
7240,28
0.381
4433,0
4059,9
4806,2
527,70
6575,6
3066,1
16353,1
4913,47
0.643
Mean cell numbers, minima, maxima (per mm²) and standard deviations for CD3+ cell counts in
progressing/persistent and regressing CIN2/3. The results are shown separately for the epithelium and all
stromal compartments and all time points from week 0 over week 8 until week 20.
188
9. Supplementary Material
progressing/persistent CIN2/3
mean
min
max
STD
mean
min
max
STD
p=
82,1
10,3
173,1
82,76
113,8
26,3
318,7
120,26
0,730
w8
169,0
44,6
293,3
175,84
128,2
23,5
397,0
142,67
0,643
w 20
58,2
58,1
58,2
0,08
174,1
40,6
407,3
164,20
0,643
CD8 Epithelium w 0
CD8 M100
CD8 M500
CD8 M1000
CD8 M0-500
CD8 M0-1000
TABLE S9.4
regressing CIN2/3
w0
394,2
99,2
784,0
319,65
973,3
236,1
1883,2
749,23
0,286
w8
1082,1
446,1
1603,0
586,98
640,6
160,7
1421,2
572,09
0,262
w 20
246,9
127,4
366,4
169,03
852,2
191,5
1222,2
487,61
0,286
w0
598,8
308,8
1234,5
432,27
616,5
75,0
1882,3
723,28
0,730
w8
947,4
377,7
1276,5
495,33
964,1
144,4
1815,8
635,94
1,000
w 20
201,3
80,8
321,8
170,38
985,8
176,9
2802,9
960,89
0,286
w0
346,2
136,9
661,3
247,53
471,6
21,5
1136,3
501,54
1,000
w8
694,9
422,1
861,4
238,14
691,8
173,2
1712,1
630,56
0,714
w 20
201,9
192,5
211,2
13,17
1220,1
171,4
4159,6
1508,64
0,429
w0
993,0
445,8
2018,5
734,29
1589,8
311,1
3765,5
1350,08
0,730
w8
2029,4
823,8
2790,9
1056,11
1604,7
316,8
3125,7
1173,83
1,000
w 20
448,2
208,2
688,2
339,42
1838,0
368,4
4021,3
1330,51
0,286
w0
1339,2
614,9
2444,4
779,36
2061,4
332,5
4901,8
1831,64
0,730
w8
2724,4
1625,1
3335,0
953,97
2296,4
490,0
4837,8
1791,11
0,548
w 20
650,0
419,3
880,7
326,24
3058,1
563,5
8181,0
2781,83
0,286
Mean cell numbers, minima, maxima (per mm²) and standard deviations for CD3+ cell counts in
progressing/persistent and regressing CIN2/3. The results are shown separately for the epithelium and all
stromal compartments and all time points from week 0 over week 8 until week 20.
9. Supplementary Material
189
progressing/persistent CIN2/3
CD3
CD3
CD3
CD3
CD3
Epithelium/M100
Epithelium/M500
Epithelium/M1000
Epithelium/M0-500
Epithelium/M0-1000
TABLE S9.5
regressing CIN2/3
mean
min
max
STD
mean
min
max
STD
p=
w0
0,34
0,05
0,91
0,403
0,09
0,03
0,32
0,128
0,063
w8
0,13
0,06
0,20
0,103
0,20
0,06
0,44
0,148
0,857
w 20
0,15
0,08
0,22
0,098
0,19
0,09
0,40
0,140
0,643
w0
0,32
0,03
0,88
0,396
0,24
0,04
0,98
0,411
1,000
w8
0,15
0,05
0,24
0,134
0,23
0,03
0,92
0,348
1,000
w 20
0,21
0,08
0,33
0,180
0,23
0,07
0,65
0,243
1,000
w0
0,90
0,02
2,01
0,946
0,27
0,11
0,39
0,143
0,730
w8
0,30
0,06
0,53
0,334
0,31
0,07
1,26
0,469
0,857
w 20
0,26
0,21
0,30
0,061
0,28
0,08
0,67
0,221
1,000
w0
0,17
0,02
0,45
0,200
0,07
0,02
0,24
0,099
0,413
w8
0,07
0,03
0,11
0,059
0,09
0,02
0,30
0,107
1,000
w 20
0,09
0,04
0,13
0,065
0,10
0,04
0,25
0,091
0,643
w0
0,14
0,01
0,37
0,165
0,04
0,02
0,14
0,054
0,413
w8
0,06
0,02
0,09
0,052
0,07
0,02
0,24
0,088
1,000
w 20
0,06
0,03
0,09
0,042
0,07
0,03
0,18
0,063
1,000
Ratios for epithelial to stromal cells counts (means, minima, maxima and standard deviations) for CD3+
cell in progressing/persistent and regressing CIN2/3. The results for each stromal compartment are given
and are shown separately for all time points from week 0 over week 8 until week 20.
progressing/persistent CIN2/3
CD8
CD8
CD8
CD8
CD8
Epithelium/M100
Epithelium/M500
Epithelium/M1000
Epithelium/M0-500
Epithelium/M0-1000
TABLE S9.6
regressing CIN2/3
mean
min
max
STD
mean
min
max
STD
p=
w0
0,16
0,08
0,25
0,084
0,11
0,07
0,17
0,040
0,413
w8
0,11
0,04
0,18
0,103
0,24
0,05
0,62
0,207
0,429
w 20
0,31
0,16
0,46
0,211
0,22
0,05
0,37
0,145
0,429
w0
0,12
0,03
0,26
0,106
0,24
0,09
0,45
0,154
0,286
w8
0,14
0,04
0,25
0,150
0,20
0,03
0,74
0,274
1,000
w 20
0,45
0,18
0,72
0,382
0,29
0,02
0,65
0,241
0,429
w0
0,33
0,02
0,82
0,363
0,47
0,14
1,22
0,438
0,556
w8
0,37
0,05
0,70
0,455
0,21
0,09
0,62
0,207
1,000
w 20
0,29
0,28
0,30
0,018
0,32
0,01
0,89
0,309
1,000
w0
0,07
0,02
0,13
0,050
0,07
0,04
0,09
0,019
1,000
w8
0,06
0,02
0,11
0,062
0,12
0,03
0,34
0,116
0,429
w 20
0,18
0,08
0,28
0,139
0,12
0,01
0,22
0,089
0,429
w0
0,05
0,01
0,11
0,046
0,05
0,04
0,08
0,017
0,905
w8
0,05
0,01
0,09
0,055
0,09
0,02
0,22
0,081
0,643
w 20
0,10
0,07
0,14
0,052
0,09
0,01
0,18
0,065
0,643
Ratios for epithelial to stromal cells counts (means, minima, maxima and standard deviations) for CD8+
cell in progressing/persistent and regressing CIN2/3. The results are shown separately for the epithelium
and all stromal compartments and all time points from week 0 over week 8 until week 20.
190
9. Supplementary Material
progressing/persistent CIN2/3
CD8/CD3
CD8/CD3
CD8/CD3
CD8/CD3
CD8/CD3
CD8/CD3
TABLE S9.7
regressing CIN2/3
mean
min
max
STD
mean
min
max
STD
p=
Epithelium w 0
0,19
0,11
0,29
0,074
2,50
0,20
5,84
2,576
0,063
w8
0,30
0,21
0,39
0,127
0,46
0,20
0,79
0,256
0,643
w 20
0,28
0,13
0,43
0,209
0,47
0,12
0,78
0,260
0,429
w0
0,32
0,12
0,86
0,359
0,72
0,28
1,13
0,391
0,063
w8
0,31
0,09
0,61
0,269
0,39
0,21
0,77
0,210
0,714
w 20
0,14
0,06
0,22
0,110
0,43
0,15
0,75
0,278
0,286
w0
0,47
0,14
1,30
0,555
0,68
0,17
1,43
0,586
0,730
w8
0,28
0,07
0,57
0,261
0,44
0,12
0,62
0,219
0,262
w 20
0,12
0,06
0,18
0,087
0,42
0,10
0,63
0,241
0,286
w0
0,48
0,18
1,03
0,374
0,82
0,41
1,89
0,632
0,190
w8
0,25
0,15
0,45
0,172
0,46
0,25
1,00
0,283
0,262
w 20
0,22
0,14
0,30
0,112
0,51
0,16
0,72
0,198
0,143
w0
0,40
0,14
1,09
0,404
0,71
0,23
1,26
0,456
0,190
w8
0,30
0,08
0,59
0,297
0,42
0,15
0,68
0,201
0,714
w 20
0,13
0,06
0,20
0,132
0,43
0,12
0,70
0,254
0,286
w0
0,42
0,15
1,07
0,438
0,71
0,24
1,37
0,481
0,413
w8
0,29
0,10
0,55
0,232
0,40
0,16
0,58
0,170
0,548
w 20
0,15
0,09
0,22
0,092
0,45
0,13
0,64
0,232
0,286
M100
M500
M1000
M0-500
M0-1000
Ratios for CD8 to CD3 cell counts (means, minima, maxima and standard deviations) in
progressing/persistent and regressing CIN2/3. The results for the epithelium and all stromal
compartments and all time points from week 0 over week 8 until week 20 are shown.
9. Supplementary Material
191
SYFPEITHI
BIMAS
x-mer
Rank
Start position
aa-sequence
score
rank
Start position
aa-sequence
score
9-mer
1
60
ILVPKVSGL
30
1
67
GLQYRVFRI
139.17
2
97
RLVWACVGV
23
2
249
YLRREQMFV
133.74
1
12
YLPPVPVSKV
30
2
12
YLPPVPVSKV
735.86
2
2
SLWLPSEATV
27
1
2
SLWLPSEATV
577.28
10-mer
TABLE S9.8
Results of HLA-A2 epitope prediction for HPV16 L1. Two different databases, SYFPEITHI and BIMAS, were used for HLA epitope prediction in order to obtain L1 peptide
sequences as potential antigens to be used for in vitro priming of T cells. The two peptides with the highest score of each database and for 9-mer and 10-mer peptides respectively
were chosen for peptide synthesis. Because of total concordance between the two databased with regard to the results obtained for 10-mer peptides, in total 6 potentially antigenic
L1 peptides were synthesized. In addition, the L1 peptide with the starting position 323 which is known from literature was synthesized as positive control.
192
13. Supplementary Material
Program
Build
Date
SN
TemplateName
TemplateDescription
Luminex 100 IS
2.3
5/13/201411:52:44 AM
LX10000266007
Optiplex HPV Genotyping Kit
Beadmix-Zusammensetzung entspr der Progen-Beschichtung
SampleVolume
DDGate
SampleTimeout
50 uL
7000 to 20000
70 sec
DataType:
Mean Fluorescence Intensity
HPV
HPV06- HPV11- HPV16- HPV18- HPV26- HPV31- HPV33- HPV35- HPV39- HPV42- HPV43- HPV44- HPV45- HPV51- HPV52- HPV53- HPV56- HPV58- HPV59- HPV66- HPV68- HPV70- HPV73- HPV82- β-globinSample
R02
R03
R10
R11
R12
R13
R14
R15
R23
R24
R25
R27
R28
R29
R40
R41
R42
R43
R45
R46
R48
R49
R66
R67
R64
HN038 M tumor cells
14.5
1
2627.5 0
1
0
3
2
3
1
2
1
1
1
2
1
2
1
1
2
2
1
4
3
17
FFPE metastasis
8
0
2590
1
1
1
1
1
5
1
1
0
1
1
1.5
3
1
1
2
1
1
2
4
2
8.5
FFPE primary tumor
10
2
1299
3
3
0
1
2
6
1.5
1
2
0
2
1
2.5
1
4
3.5
3
1
4
4
4
31
HeLa positive control
16
0
87
415
0
1
1
1
3
1
1
1
4
1
1
2
1
1
2
2
1
2
4
4
62
Caski positive control
17
0
3007
0
1
0
3
1
4
1
1
1
1
1
1
2
1.5
1
1
2
1
2
4
3
10
H2O control
17
0
1
1
1
1
1
1
5
1
1
0
1
1
2
2
1
2
2
1
2
1
3
3.5
8
no tissue/cells (empty tube) 19
0
0
0
1
0
0
1.5
4
1
1
1
1
1
2
2
2
2
1
2
2
2
4
2
9
hybridization control
17
1
1
0.5
0
0
0
2
4
2
1
1
1
1
3
2
1
1
2
2
2
1.5
3
3.5
8
TABLE S9.9
HPVHYB1R70
4
4
4
4
3
3.5
3
775.5
Total
Events
2272
1095
1570
2246
2621
2442
2483
2449
HPV Genotyping results for the tumor cell line HN038M and the corresponding archived (FFPE) tissue of the corresponding lymph node metastasis and primary tumor. Results
are displayed as mean intensity values (MFIs) for the samples and controls and for all tested HPV-types. Samples or controls positive for the corresponding HPV types are
marked in blue, also the hybridization control is highlighted in blue.
Download PDF