niemietzthesisfinal

niemietzthesisfinal
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
Diplom-Biochemiker Thomas Niemietz
born in: Weimar
Oral-examination: ................................................
1
Referees:
First:
Second:
Prof. Dr. Philipp Beckhove
Prof. Dr. Jürgen Weitz
2
3
Phenotypic and molecular characterization of
colorectal cancer-derived circulating and disseminated
tumor cells
4
Table of contents
Table of contents
Abstract ............................................................................................................................ 6
Zusammenfassung............................................................................................................ 7
Aim .................................................................................................................................. 8
Introduction ...................................................................................................................... 9
Colorectal cancer ......................................................................................................... 9
CRC development ........................................................................................................ 9
Metastasis and the epithelial mesenchymal transition (EMT) ................................... 12
Dissemination and colonization ................................................................................. 14
Tumor stem cells ........................................................................................................ 16
Cancer progression models ........................................................................................ 17
Disseminated tumor cells ........................................................................................... 19
Clinical significance of CTCs and DTCs in CRC ..................................................... 21
Molecular characterization of DTCs and CTCs......................................................... 22
Results ............................................................................................................................ 26
Compartmental differences of CTCs in CRC ............................................................ 26
Establishment of a reliable CTC enrichment and detection method.......................... 32
An mRNA expression study on CRC - associated CTCs and DTCs ......................... 37
Genomic characterization of CTCs ............................................................................ 46
An orthotopic mouse model of cancer cell dissemination ......................................... 57
Discussion ...................................................................................................................... 69
Compartmental differences of CTCs in CRC ............................................................ 69
Establishment of a reliable CTC enrichment and detection method.......................... 71
A mRNA expression study on CRC - associated CTCs and DTCs ........................... 72
Genomic characterization of CTCs ............................................................................ 75
An orthotopic mouse model of cancer cell dissemination ......................................... 79
Conclusion and outlook ............................................................................................. 81
Material and methods ..................................................................................................... 82
Methods...................................................................................................................... 82
Material ...................................................................................................................... 88
Appendix ........................................................................................................................ 94
Literature ...................................................................................................................... 108
Acknowledgment ......................................................................................................... 118
5
Abstract
Abstract
Tumors shed large numbers of cells into the vasculature. It is supposed that these cells
give rise to metastases that are the major cause of death from cancer. However, the
isolation and characterization of the so called circulating tumor cells (CTCs) is still
challenging since cells are rarely to find among millions of normal leukocytes. The
objective of this thesis was to characterize colorectal cancer (CRC) - associated CTCs
and disseminated tumor cells (DTCs) from bone marrow samples. We studied CTC
incidence in blood samples of CRC patients with the help of the FDA-cleared
CellSearchTM system. We found that the number of patients with CTCs and the amount
of CTCs was significantly correlated with the stage of disease. In addition, a significant
higher rate of patients with CTCs in tumor-draining venous blood compared to the
central venous blood was found. Furthermore, genomic analyses of single CTCs have
been performed. This comprised the evaluation of typical CRC-associated mutations
such as point mutations in TP53, BRAF, KRAS as well as the detection of microsatellite
instability (MSI). Additionally, some CTCs were used to study global chromosomal
aberrations by comparative genomic hybridization technique. Results revealed a
remarkable genomic heterogeneity among the CTCs of single patients. Moreover, we
detected several cases of genomic disparity among CTCs and the prevailing clone of
matched cancer tissue. To evaluate differential gene expression that might enable tumor
cell dissemination, we studied CTCs that were obtained from patients’ blood samples
and from an othotopic mouse model of metastatic CRC. Gene expression profiles of
CTCs and cell samples from cancer tissue were compared. The down-regulation of cell
adhesion molecules such as E-cadherin might be involved in tumor cell dissemination.
However, a significant epithelial-mesenchymal transition (EMT) in CTCs could not be
confirmed. DTCs isolated from human bone marrow samples seem to adopt a dedifferentiated phenotype and lack expression of CK20 and CK19. Absence of EGFR
and the proliferation marker Ki67 confirmed previous reports of dormancy in DTCs.
With the present work we were able to demonstrate that the molecular characterization
of single CTCs is feasible. However, further studies are required to increase the
knowledge about the molecular traits of CTCs which might help to improve therapy
and prognosis in CRC.
6
Abstract
Zusammenfassung
Krebszellen werden von Tumoren zahlreich ins Blut abgegeben. Man vermutet, dass
diese Zellen Metastasen, die Hauptursache krebsbedingter Todesfälle, initiieren
können. Jedoch ist die Isolation und Charakterisierung dieser so genannten
zirkulierenden Tumorzellen (ZTZs) immer noch eine Herausforderung, da sie unter
Millionen von Leukozyten nur schwer zu detektieren sind. Diese Arbeit wurde mit dem
Ziel erstellt ZTZs sowie ins Knochenmark disseminierte Tumorzellen (DTZ) im
kolorektalen Karzinom (KRK) zu charakterisieren. Das Vorkommen von ZTZs in
Blutproben von KRK Patienten wurde mit dem CellSearchTM System untersucht.
Sowohl Anzahl der Patienten mit ZTZs als auch Menge von ZTZs korrelierten
signifikant mit dem Krankheitsstadium. Es fanden sich signifikant mehr Patienten mit
ZTZs in tumordrainierenden venösem Blut im Vergleich zu zentralvenösem Blut.
Zusätzlich wurden genomische Einzelzellanalysen an ZTZs durchgeführt. Dies
umfasste die Analyse von typischen KRK-assoziierten Punktmutationen in TP53,
KRAS und BRAF sowie das Detektieren von Mikrosatelliteninstabilität. Einige ZTZs
wurden für die Analyse von chromosomalen Aberrationen mittels CGH untersucht.
Eine bemerkenswerte genomische Heterogenität innerhalb der ZTZ Populationen
einzelner Patienten wurde gefunden. Außerdem traten häufig genomische Unterschiede
zwischen ZTZs und dem dazugehörigen Krebsgewebe auf. Da differenzielle
Genexpression vermutlich die Dissemination ermöglicht, wurden ZTZs, die aus
Patientenblut und aus einem orthotopen Metastasierungsmodel in der Maus isoliert
wurden, analysiert. Die Genexpressionsprofile der ZTZs und Zellproben des Krebsgewebes wurden verglichen. Die Runterregulation von Zelladhäsionsmolekülen wie Ecadherin könnte an der Tumorzelldissemination beteiligt sein. Allerdings konnte eine
signifikante Epitheliale-Mesenchymale-Transition der ZTZs nicht bestätigt werden.
DTZs, die aus menschlichen Knochenmarkproben isoliert wurden, scheinen einen
dedifferenzierten Phänotyp anzunehmen. Die Differenzierungsmarker CK20 und CK19
wurden nicht detektiert. Berichte über eine niedrige Expression von Ki67 und EGFR,
die auf eine proliferative Quieszenz in DTZs hindeuten, ließen sich hingegen
bestätigen. Die vorliegende Arbeit zeigt, dass eine umfassende molekulare
Charakterisierung von ZTZs und DTZs möglich ist. Um mehr über die molekularen
Eigenschaften dieser Zellen zu erfahren sind jedoch weitere Studien erforderlich. Diese
könnten die Therapie und Prognose des KRK verbessern helfen.
7
Aim
Aim
The objectives of this thesis were the detection, isolation and characterization of
colorectal cancer (CRC) - derived circulating and disseminated tumor cells (CTCs
/DTCs)). To increase our knowledge about the early steps of metastasis we were
focused on the genomic and transcriptional profiles of CTCs and DTCs. Due to the lack
of reliable and standardized techniques to isolate and analyze CTCs obtained from
human blood samples, an efficient CTC isolation protocol had to be designed.
Transcriptional profiles of solid tumor samples and CTCs had to be ascertained to
analyze disparities of the mRNA expression profile during and after the dissemination
process. Additionally, an orthotopic mouse model was established to mimic cancer cell
dissemination. CTCs that were obtained from the mouse model ought to provide further
information about the molecular mechanisms involved in tumor cell dissemination.
8
Introduction
Introduction
Colorectal cancer
CRC is ranked as the third most common cancer worldwide 1. In Europe about 250.000
new cases of colon cancer are diagnosed each year 2. However, the majority of the
patients does not succumb to the primary tumor but dies due to metastases with a fiveyear survival rate of only 30-40% 3. Surgery is the primary treatment for CRC;
nevertheless, CRC seems to be curable only when diagnosed and treated at an early
stage of disease 4, 5.
CRC progression is mostly staged by the tumor-node-metastasis (TNM) classification.
This system is based on the depth of invasion of the bowel wall, the extent of lymph
node involvement, and the presence of distant metastasis. Beyond that, an additional
system for overall staging was suggested by the Union International contre le Cancer
(UICC). It is referred as Roman Numeral staging that classifies increased tumor
progression with roman numerals from 0 – IV (Figure 1) 6.
Figure 1: UICC staging of tumor progression.
CRC development
CRC arises as a result of the progressive accumulation of genetic and epigenetic
changes within cells from the epithelium lining the colon and rectum 7. Usually, CRC
occurs late in lifetime. 92% of patients are diagnosed after the age of 50 3. More than
80% of CRC cases are regarded as sporadic whereas familial syndromes account for
about 8-15% of cases 7. There are two major forms of hereditary CRC, the familial
adenomatous polyposis syndrome (FAP) and hereditary nonpolyposis colorectal cancer
(HNPCC or Lynch Syndrome) 8. FAP, a disease in which thousands of polyps develop
along the colon and harbor the potential to be the founder of a tumor, was traced to a
9
Introduction
deletion or inactivation of the adenomatous polyposis coli (APC) gene. Its protein
(APC) plays a major role in the Wnt / β-catenin pathway (Figure 2) 9. If APC function
is lost, degradation of β-catenin is abolished. B-catenin enriches in the cytosol and
translocates into the nucleus where it forms a complex with the transcription factor
TCF/LEF leading to the expression of proto-oncogenes like c-myc or cyclin D1.
Interestingly, 70 - 80% of sporadic CRC carry inactivating mutations of the APC gene,
too. CRC with intact APC were often found to harbor activating mutations of β-catenin
instead 9. Thus, the majority of CRC cases are elicited by the aberrant activation of the
Wnt / β-catenin pathways 10.
Figure 2: The Wnt signaling pathway. In the absence of a Wnt signal, the destruction complex
containing APC and other proteins targets the degradation of cytoplasmic β-catenin in a proteasomedependent manner. In the nucleus, Wnt target genes are kept silent by the repressor Groucho interacting
with DNA-bound T cell factor (TCF). In the presence of a Wnt ligand, the destruction complex is
inactivated through its phosphorylation dependent recruitment to the Wnt receptor Frizzled (Frz) and its
co-receptor LRP. Degradation of β-catenin no longer occurs. Cytoplasmic β-catenin translocates to the
nucleus, where the transcription of multiple genes is initiated through displacement of Groucho and the
interaction of β-catenin with TCF/ LEF family of transcription factors. Figure taken from Pino and
Chung 11.
Inactivating mutations in various mismatch-repair (MMR) genes lead to the
aforementioned HNPCC disease or Lynch syndrome respectively. Silenced or
inactivated MMR genes are found in about 12% of sporadic CRC as well
12
. They
10
Introduction
maintain genomic integrity and mediate DNA damage-induced cell death. Loss of
MMR function can be detected by microsatellite instability (MSI). Microsatellites (MS)
are short repetitive sequences in DNA that are prone to a higher rate of mutation during
DNA replication compared to other DNA regions. If MS repeat errors remain unfixed
due to defects in MMR function, the length of MS can change which might lead to
frameshift mutations that inactivate or alter tumor suppressor genes such as APC, TP53
and SMAD4. MS sequences of the TGF-β receptor II (TGFBR2) are particularly
affected by MSI. Frameshift mutations are found in the TGFBR2 gene in 80% of CRC
with MSI, leading to alterations in the TGF-β/SMAD signaling
10
. TGF-β signaling
leads to the formation of a hetero-oligomeric complex of SMAD proteins that
translocates to the nucleus and interacts with key transcription factors (c-jun,
p300/CBP, c-myc). Moreover, several cell cycle checkpoint genes (p21, p27, p15) were
found to be TGF-β signaling targets as well. Consequently, defects in TGF-β will
contribute to loss of cell cycle control 13, 14.
The process of colorectal tumorigenesis is complex and involves sequential mutations
of several key signaling pathways (Figure 3)
15
. The inactivation of the APC gene is
believed to be an early event in CRC. Further mutations in oncogenes such as KRAS
and in tumor suppressor genes like TP53 are required for a complete malignant
progression. KRAS, a member of the small G-protein family, mediates cellular signal
transduction and was found to be constitutively active in up to 30-40% of CRC
patients. 95% of KRAS mutations are found in codons 12, 13 and 61. If KRAS is
constitutively active, the cellular proliferation is uncoupled from extracellular signals
such as epidermal growth factors (EGFs) 16. TP53 encodes the transcription factor p53
that is involved in various cellular pathways including cell cycle regulation and
apoptosis. It is believed that loss of p53 function might allow evasion from cell cycle
arrest and apoptosis. Inactivating mutations of TP53 often mediate the transition of
adenomas to invasive cancers
17
. These mutations are located mainly in the DNA
binding domain of the p53 protein and are common in about 50% of CRC cases 18.
Furthermore, the PI3K signaling cascade is frequently found to be constitutively active
in CRC
19
. This is often caused by mutations in PI3KCA, the gene encoding the
catalytic subunit of PI3K. One of the most prominent downstream targets of the PI3K
pathway is Akt. Akt is critically involved in the regulation of apoptosis, gene
transcription by NF-kB pathway and cell cycle progression
20
. Interestingly,
immunohistochemical analyses showed that some tumors expressing high levels of
11
Introduction
activated Akt are significantly associated with poor prognosis
21-23
. However, for CRC
the strong expression of activated Akt was found to be associated with low cancer stage
and favorable outcome 24.
Beyond genomic mutations increasing data imply that miRNAs are involved in the
deregulation of CRC associated genes, too
25
. For instance, a decreased expression of
miR-34a was found in human colon cancer tissues compared to normal tissues and
might be linked to a reduced expression of TP53.
Figure 3: Multistep genetic model of CRC development. The progression of CRC is accompanied by
sequential mutational events in key signaling pathways. The activation of the Wnt signaling pathway
results from mutations in the APC gene and is an early event in tumor initiation. Further cancer
progression to late adenomas and early carcinomas requires mutations in KRAS and TP53 and loss of
heterozygosity at chromosome 18q. Aberrant TGF-β signaling and activation of the PI3K/Akt signaling
pathway through activating mutations in the PIK3CA gene lead to full malignancy. Figure taken from
Saif and Chu. 7.
Metastasis and the epithelial mesenchymal transition (EMT)
Most common tumors, like CRC, are of epithelial origin 2. Such carcinomas derive
from epithelial cells, which are characterized by tight cell–cell interactions, the basalapical polarization and the inability to migrate. Genetic damages lead to increased cell
proliferation and to the formation of an initially benign tumor. Within a multistep
process neoplastic cells accumulate further genetic and phenotypic changes and switch
into malignancy. They become invasive, break through the basal lamina, enter the
lymphatic vessels or penetrate the blood vasculature and reach the circulation. The
mechanism of how tumor cells can disseminate and spread through the body is
currently not fully understood. A widely accepted but still controversial discussed
hypothesis is that tumor cells undergo an epithelial-mesenchymal transition (EMT)
27, 28
26,
. EMT is a crucial and highly conserved process in embryo development. It is
12
Introduction
characterized by the down-regulation of cell adhesion molecules, the loss of a rigid
cytoskeleton and an increased cell motility which is essential during gastrulation
movements and neural crest formation. Growth factors including TGF-β, EGF,
HGF/SF, bFGF and PDGF were found to be initiators of EMT. Diverse signal
transduction mechanisms mediate EMT, e. g. receptor tyrosine kinases / Ras, Wnt,
Notch, Hedghog and NFkB pathways
29
. It has been proposed that invasion and
metastasis of carcinoma cells is linked to EMT 30. This is supported by the finding that
in most epithelial cancers E-cadherin, a cell-cell adhesion molecule, is inactivated or its
expression is down-regulated. Various studies have shown that this seems to correlate
with cancer grade and patient survival
31-34
different levels of gene expression
35
. E-cadherin expression can be regulated at
. A major mechanism, however, is the
transcriptional repression through the binding of transcriptional repressors to E-boxes
located at the proximal promoter site. Most prominent transcription factors that downregulate E-cadherin expression include Snail, Slug, SIP-1, ZEB1 and TWIST1
Various studies link these transcription factors generally to EMT
37-40
36
.
. For instance,
Snail was found to down-regulate several epithelial markers such as claudins,
occludins, desmoplactin and cytokeratins and up-regulate the expression of the
mesenchymal markers fibronectin and vitronectin. TWIST1 has been associated with
increased N-cadherin expression which was found to support cell motility.
In summary, EMT leads to large changes in the gene expression profile and might be
responsible for loosing cellular assembly. Concomitantly, increased cell motility might
enable tumor cells undergoing EMT to enter the vasculature and to spread over the
body (Figure 4).
13
Introduction
Blood vessel
Figure 4: EMT and tumor cell dissemination. Epithelial mesenchymal transition (EMT) seems to be
critically involved in tumor cell dissemination. Extracellular signaling mediates the up-regulation of
EMT-inducing transcription factors in the primary tumor. The down-regulation of epithelial markers and
the up-regulation of mesenchymal proteins facilitate tumor cell migration and invasion into the blood
vessels.
Dissemination and colonization
Tumor cells that have detached from the primary tumor and circulate in the blood are
termed CTCs. It is estimated that large numbers of CTCs enter the circulation but only
a small portion is able to survive and to establish metastasis at secondary organs
The main portion of CTCs can not sustain the shear stress in the blood stream
41, 42
43
.
and
undergoes apoptosis. Moreover, for survival CTCs have to develop mechanisms to be
protected from anoikis, which is a kind of programmed cell death that is induced when
anchorage-dependent cells detach from the extracellular matrix
observed CTC clusters or multicellular tumor cell aggregates
45
44
. Occasionally
might be a way to
protect CTCs from mechanical burdens and anoikis. Furthermore, CTCs need to be
protected from immunosurveillance. Besides a decreased expression of MHC I
proteins 46 it has been reported that CTCs might avoid NK cell-mediated lysis through
the aggregation with platelets
47, 48
. In addition, Pawelek et al. have suggested that
CTCs might fuse with bone marrow-derived cells (BMDC) in the blood 49. The BMDC
– CTC fusion might increase CTC survival and could provide an explanation for
acquired mesenchymal traits in CTCs.
14
Introduction
Liver metastases are pre-dominantly observed in CRC 50-52. CTCs, which are shed from
the primary tumor, reach the liver through the mesenteric and portal venous system.
The majority of the CTCs is filtered by the liver before they are able to enter the
systemic circulation. According to a cascade theory
50
of cancer progression that
suggests the step-wise progression of malignancy, the CTCs that are trapped in the liver
grow out to metastases and increase the amount of CTCs in the blood. This might
explain why CTCs in peripheral blood are primarily found in patients with overt
metastases 53. In contrast to that hypothesis it is also conceivable that numerous CTCs
pass the liver and distribute throughout the body. In this alternative scenario the usually
observed homing of CTCs to liver or lung would be attributed to adhesion molecules
and certain cytokines. Already in 1889 Stephen Paget hypothesized that metastasis is
not a random process and presumed that there must be a strong impact of the
interaction between tumor cells and the microenvironment
54
. Paget developed the
“seed and soil” hypothesis claiming that certain tumor cells (the seed) metastasize to
certain organs (the soil). It is widely accepted that tumors and CTCs are a very
heterogeneous cell population. The question which properties of seed and soil
efficiently support metastasis is still largely unknown. However, the local expression of
chemoattractants and cell adhesion molecules might explain the organ tropism of
disseminated cancer cells. It was found that the expression of the chemokine receptor
CXCR4 on breast cancer cells guides them to tissues with high expression levels of the
CXCR4 ligand SDF-1. These tissues include bone marrow, liver and lung
55
. In
addition, a receptive microenvironment seems to be a prerequisite to successfully
establish a metastatic lesion at a distant site. It was found that VEGFR-1+ bone marrowderived hematopoietic progenitor cells can be mobilized to the invasive front of the
primary tumor as well as to the designated target organs of metastasis where they
support angiogenesis and tumor progression
56
. Already before the arrival of tumor
cells VEGFR-1+ bone marrow-derived hematopoietic progenitor cells together with
endothelial and stromal cells prepare the pre-metastatic niche as the future site of
metastasis through modification of the extracellular matrix and complex chemokine
signaling. Once, CTCs/DTCs reach the pre-metastatic site, they extravasate from the
blood vessels and become sessile again. It is supposed that the process of EMT is
reverted and the epithelial traits need to be re-gained to from metastases
57
. Probably,
the DTCs stay initially in a dormant state. Metastatic relapse in patients with CRC can
occur years after diagnosis and resection of the primary tumor 58. Therefore, it is
15
Introduction
assumed that many cancer patients harbor dormant tumor cells that even resist
chemotherapy due to their low proliferation rate, appropriate detoxifying enzymes and
presumable stem cell properties. Currently unknown factors can trigger tumor cell
proliferation again. In this case the metastatic niche undergoes an angiogenic switch by
recruiting endothelial progenitor cells. Finally, increasing vascularization enables the
progressive growth to marcometastases.
Tumor stem cells
Although a tumor develops from a single, mutated cell, tumor cells within a tumor are
not identical 59. One explanation could be the genomic instability of tumor cells 60 and a
constant selective pressure for cells to adapt to the tumor microenvironment.
Alternatively, it was proposed that a hierarchy exists in tumors with only a minority of
cells that are capable of regenerating the tumor. These cells are termed cancer stem
cells
61
. The bulk of the tumor, however, is formed by cells with a limited capacity to
divide. Currently it is not clear if tumor stem cells originate from normal stem cells or
from progenitor cells with a blocked differentiation program or even from
differentiated cells that somehow acquired self renewing capacity. The hypothesis that
cancers like all other tissues originate from stem cells is not new and is supported by
the finding that only a small subpopulation of cancer cells is able to induce tumor
growth in mouse models. Biomarkers to identify and define cancer stem cells, among
them CD44, CD133 and CD26, are still not standardized and under intensive
investigation 62, 63. Nevertheless, the cancer stem cell hypothesis is still controversial 6466
. Critics state that it might be an in vitro assay-associated artifact. Supporting
evidences rely mainly on xenotransplantion studies in immunosuppressed mice.
Furthermore, they allude that the low numbers of human tumor cells producing tumors
in mice arise from difficulties of the tumor cells to adapt to the novel and foreign
microenvironment. Hence, common assays rather select for cells that are tumorigenic in
mice but neglect that tumor formation depends on complex interactions with
extracellular matrix components and multiple non-tumor cells that might not be present
in the xenotropic mouse model.
16
Introduction
Is metastasis initiated by tumor stem cells?
If tumor stem cells exist, then they might be already present in benign lesions.
Consequently, additional features have to be acquired to develop malignant growth. An
intriguing concept of malignant tumor progression was presented by Brabletz and
colleagues67. They propose the existence of a mobile cancer stem cell fraction in
tumors and metastases they term migrating cancer stem cells (MCS). According to their
hypothesis, these cells might originate from stationary cancer stem cells that
additionally gained cellular mobility through transient acquisition of EMT. The concept
comprises that a migrating front of MCS extends the tumor mass at the tumor-host
interface and the formation of metastases is caused by long distance migration of MCS.
Cancer progression models
Early or late dissemination of tumor cells
The resection of the primary tumor is often not curative in CRC since tumor cells might
have already disseminated and are spread throughout the body
68
. Systemic treatments
to prevent the formation of distant metastases were developed to target DTCs
69
.
However, DTCs are poorly characterized and we know hardly anything about the
metastatic founder cells so far. From this it follows that predictions about the properties
of DTCs require at least models of cancer progression. In particular it is important to
elucidate if primary tumors can be used to predict the success of systemic treatments to
defeat DTCs (Figure 5). Two basic models of metastasis are discussed
70
. Basically,
they are focused on the question if tumor cell dissemination is an early or late event in
tumor progression. The linear progression model states that cancer cells disseminate
not until the tumor reaches full malignancy. Consequently, DTCs resemble the primary
lesion. The second model suggests a parallel progression of DTCs and the primary
tumor. This means that DTCs leave the primary lesion already at an early point of time.
In this scenario DTCs develop independently and divergent from the primary tumor.
There are arguments for both models. The linear progression model is supported by the
finding that the appearance of metastases correlates with tumor size and that an early
resection of the primary lesion prevents dissemination of cancer cells and increases the
patients’ survival. However, the model can not explain the occurrence of metastases in
early stage cancer or of metastases of unknown primary site. Moreover, based on the
assumption that primary tumors and metastases have a comparable growth rate, one can
17
Introduction
estimate the point of time when a metastasis started to grow. Data from registries, that
report the time from resection to the appearance of distant metastasis, contradict the
late dissemination of tumor cells because metastases would be expected to emerge
much later in time. In addition, genetic evidence favors the parallel progression model.
Techniques for whole genome analysis of single DTCs revealed that significant fewer
genetic aberrations can be found in DTCs compared to primary tumor cells indicating
an early dissemination of tumor cells. Furthermore, genomic disparities in matched
primary tumors and manifested metastases rather support the parallel progression
model.
A
B
Figure 5: Early or late dissemination of cancer cells. A The late dissemination and metastatic cascade
model predicts that disseminating tumor cells (DTCs) resemble the most aggressive clone and therefore
the tumor at time of diagnosis. Therapies can be selected on the basis of analysis of the primary tumor. B
Parallel progression complicates therapy selection. Genomes may vary between different sites and
undergo independent progression. Specific genetic changes might be selected at different sites and
different targeting strategies have to be applied for DTC, small clusters and small colonies. Figure and
description was taken from Klein 70.
18
Introduction
Disseminated tumor cells
Definition
The term minimal residual disease (MRD) comprises single tumor cells and
micrometastases throughout the body. According to the location in the body further
distinctions are required:
-
Disseminated tumor cells (DTCs) comprise tumor cells that are found in lymph
nodes or in the bone marrow.
-
Circulating tumor cells (CTCs) are located in the blood.
Methods of detection
Metastatic spread due to tumor cell dissemination is the most threatening aspect of
cancer. To increase our knowledge about the metastatic process several techniques
have been developed to capture CTCs in blood and DTCs in bone marrow and lymph
nodes for enumeration and characterization. The presence of lymph node - associated
DTCs is usually confirmed immunohistologically by epithelial specific staining, wheras
diverse techniques for the detection of CTCs and bone marrow-associated DTCs exist.
Techniques for CTC and DTC detection are highly analog; therefore, this chapter is just
focused on CTC identification methods.
Since CTCs are very rare and hard to find in a strong background of leukocytes (often
less than one in 106 peripheral blood mononucleated cells (PBMC)), sophisticated
approaches are needed for CTC detection 71, 72. Most protocols apply a density gradient
centrifugation in combination with an immunomagnetic isolation step to enrich CTCs
from a blood sample. However, CTCs seem to be very heterogeneous in their
morphology and in the expression of cell-surface antigens which probably leads to a
bias in all used enrichment methods. Most data were obtained from the CellSearchTM
(Veridex) system
73
which represents the only FDA- approved device for CTC
detection in colorectal, prostate and breast cancers that allows the enumeration of CTCs
in whole blood specimens of cancer patients. CTCs can be distinguished from
leukocytes due to the expression of epithelial protein markers which are normally
absent on blood cells. The purification / detection process starts with the
immunomagnetical enrichment of EpCAM-expressing CTCs with the aid of an
EpCAM-conjugated ferrofluid. Afterwards the specimens are stained with a nuclear
stain to prove cellular integrity and fluorescent antibody conjugates directed against
CD45, cytokeratin 8, 18 and 19. A trained operator can distinguish cytokeratin-positive
19
Introduction
CTCs from residual CD45- positive leukocytes to determine the CTC amount in a
7.5 ml blood sample (Figure 6).
Figure 6: The CellSearch. The CellSearchTM system is used for the detection of CTCs. The figure
summarizes the processing of a blood sample as explained in the text.
Several other techniques for CTC detection have been developed recently. Most of
them rely on EpCAM and cytokeratin expression of the target cells. For instance, the
CTC chip is a microfluidic device that captures EpCAM-expressing CTCs within an
array of antibody coated microposts
74
. Subsequent cytokeratin staining enables the
identification of CTCs. EpCAM expression of target cells is also required for the
automated cell separator MagSweeperTM
75
that is designed to perform efficient
immunomagnetic CTC enrichment. Other surface-marker-based CTC detection devices
include standard flow cytometry and laser scanning cytometric based systems, such as
MaintracTM, IkonoskopeTM and AriolTM.
A different approach to isolate CTCs is the ISET system
76
. Since CTCs usually are
larger than peripheral blood leukocytes, it is possible to separate them with the ISET
system by size filtration. However, the method is controversial because the CTC size is
highly variable and sometimes comparable to leukocytes 77.
Another technique, the EPISPOT (Epithelial immunoSPOT), was designed to detect
viable CTCs
78
. Leukocyte-depleted blood or bone marrow specimens, that potentially
contain CTCs, are cultured to accumulate a sufficient amount of marker proteins. By
adapting the enzyme linked immunospot technology CTC detection is getting feasible.
Other techniques use polymerase chain reaction (PCR) based systems. The AdnatestTM
method relies on the immunomagnetic CTC isolation and the subsequent detection of
20
Introduction
tumor marker expression in a multiplex PCR 79. A big drawback of that method is that
CTCs can not be quantified. The lack of quantification is caused by the impossibility to
determine from how many cells the detected transcripts are derived.
In summary, despite several CTC detection methods are available a standardized CTC
detection system without bias is still lacking. CTCs conserve characteristics of the
epithelial cancer tissue from which they originate. This is the basis for the common
detection methods; however, it is also the main point of critics. In the course of EMT
and the dissemination process CTCs might down-regulate the expression of epithelial
markers, such as EpCAM, a protein that is also involved in cell adhesion. Standardized
and more unbiased methods for CTC detection would be favorable to get a
comprehensive insight in tumor cells shedding.
Clinical significance of CTCs and DTCs in CRC
Routine histopathological methods have confirmed that lymph node involvement is one
of the most important prognostic factors for colon cancer patients
80
. However, these
methods frequently miss smaller lesions and micrometastases. Increasing sensitivity of
detection due to molecular detection methods revealed that lymphatic spread of tumor
cells seems to be common already in early stage CRC. Since lymph node - associated
CTCs are found in N0-staged lymph nodes (analyzed as negative in conventional
pathology) in up to 86% of cases, it is not clear to what extent micrometastases and
single lymph node - associated CTCs contribute to a worse prognosis in CRC 81 82 83.
Numerous studies addressed the question if the detection of CTCs and DTCs are a
prognostic factor for progression free survival and overall survival in CRC 84. Due to an
increasing amount of data obtained from the CellSearchTM system the case for
metastatic CRC (mCRC) seems to be quite clear
85 86, 87
. Cohen showed that CTC
counts above the baseline of 2 CTCs per 7.5 ml of peripheral blood indicate an about
50% reduced progression free survival and overall survival in mCRC patients
(Figure 7).
21
Introduction
Figure 7: CTC count and prognosis in mCRC. Progression-free survival (panel A) and overall
survival (panel B) for metastatic CRC patients by favorable (<3) versus unfavorable (≥3) baseline CTC
count. Figure was taken from Cohen et al. 86
However, the impact of CTCs and DTCs in non-metastasized CRC patients is still
controversial. The main problem is that the available studies vary dramatically in study
design and in the applied methods. Standardized guidelines are required to obtain
comparable study results.
Only a few studies address the prognostic significance of DTCs in the bone marrow of
CRC patients
84
. It is supposed that the bone marrow might serve as a reservoir of
DTCs and would be a good indicator for MRD. Lifespan of CTCs in the blood is
short 88 and the fact that CTCs can still be found in the blood months after removal of
the primary tumor, indicates that CTCs might circulate among different metastatic
sites 89. The bone marrow might be a source of CTCs even years after curative
resection. DTCs usually stay there in dormancy until they are released again due to
unknown stimuli. Since the available studies are very heterogeneous in sample size,
staining methods and follow-up-times, further studies are required to finally assess the
role of DTCs in CRC 90.
Molecular characterization of DTCs and CTCs
Understanding the process of metastasis requires the molecular analysis of CTCs and
DTCs. However, their rareness is the major difficulty for a detailed characterization.
Fortunately, along with improvements in single cell techniques the characterization of
the metastases precursor cells is becoming more and more feasible.
22
Introduction
Immunocytochemistry
The phenotypic heterogeneity of DTCs isolated from bone marrow samples was shown
by immunocytochemical double or triple stainings
91
(Figure 8). DTCs detected in the
bone marrow are often dormant which was indicated by the low expression of
proliferation markers such as Ki67, p120 and PCNA. This might explain the
appearance of clinical metastases late after the curative resection of the primary
tumor 58. With regard to immune escape mechanisms the down-regulation of major
histocompatibility complexe I (MHC I) was found on DTCs of breast cancer patients 92.
Recent studies also report that the over-expression of Her2, a receptor for EGFR, in
breast cancer-associated DTCs and CTCs is linked to a poor clinical outcome
93
. This
finding might support the use of trastuzumabTM, a humanized anti-Her2 monoclonal
antibody, for systemic treatment in Her2 over-expressing breast cancer patients.
Furthermore, urokinase-type plasminogen activator receptor (uPar) expression on
DTCs correlates with metastatic relapse in gastric cancer
94
. It was speculated that
signaling mediated by Her2 and uPar might be involved in growth induction in dormant
DTCs. Since heterogeneity of CTCs and DTCs becomes more and more evident 45, the
search for the metastasis founder cell fraction is proceeding. Based on CD44, CD24
and ALDH1 the existence of a potential cancer stem cell subpopulation was reported
for breast cancer-associated CTCs
95
. However, so far nothing is known about the
prognostic significance of this finding.
Figure 8: Heterogeneity of DTCs. Phenotypic profile of DTCs in bone marrow determined by
immuno-cytochemical double staining using anti-cytokeratin antibodies for tumor cell detection. Figure
was taken from Pantel et al. 91
23
Introduction
Genomic characterization
The genomic analyses of DTCs and CTCs were initially necessary to confirm the
malignant origin of cytokeratin-positive stained cells in the bone marrow. By means of
fluorescence-in-situ-hybridization
96
it was possible to show chromosomal aberrations
in CTCs / DTCs. When later whole genome amplification techniques became available,
single cell comparative genomic hybridization (CGH) was applied to screen for
chromosomal gains and losses. It was reported that DTCs generally display fewer
genetic aberrations than their matched primary tumors in breast, prostate and
esophageal cancer supporting the theory of an early dissemination of tumor cells
97
.
Growing evidence indicates that DTCs might develop independently from the primary
tumor. This may lead to primary tumors and metastases with different mutations in
oncogenes or tumor suppressor genes. For instance, disparate KRAS mutations of
tumors and matched metastases were observed in up to 60 % of CRC
98
. Since
constitutively activating KRAS mutations in CRC have been associated with failure of
anti-EGFR therapy, the molecular profile of CTCs/ DTCs is becoming increasingly
important to defeat systemic cancer and to predict therapy response 99.
Transcriptome analysis
Array-based gene expression analysis could provide crucial information about the
dissemination and survival mechanisms of CTCs and DTCs and might also permit the
identification of novel therapeutic targets fighting systemic cancer. However, these
approaches are mainly limited by the low number of CTCs / DTCs and impurity of the
applied isolation strategies. Most of the available studies were solely able to perform
RNA expression analyses of CTC or DTC - enriched cell fractions in a background of
several thousands of leukocytes 100-102. Despite these limitations Watson et al. provided
supporting evidence for EMT in DTCs by means of global gene expression
profiling 102. They detected the expression of TWIST1, a putative inducer of EMT, in
DTCs in early stage breast cancer patients and found TWIST1 expression correlated
with early disease relapse.
Klein et al.
103
reported the transcriptome analysis of single micrometastic cells from
bone marrow samples. With the help of a sophisticated RNA extraction and subsequent
cDNA amplification strategy it was possible to compare the gene expression profiles of
DTCs and to conclude that most of DTCs adapt a dormant cell cycle status.
Additionally, they detected the expression of EMMPRIN (extracellular matrix
24
Introduction
metalloproteinase inducer) in the analyzed micrometastatic cells. EMMPRIN is thought
to
facilitate
tumor
invasion
by
stimulation
of
matrix
metalloproteinases.
Immunocytostaining of cytokeratin positive DTCs of diverse cancers revealed
EMMPRIN expression in 82 % of DTCs indicating their invasive properties.
25
Results
Results
The objective of this thesis was to characterize CRC-associated CTCs on the
transcriptional and genomic level. Following the trail of hematogenous tumor cell
dissemination we recorded the CTC amount and distribution in patients’ blood obtained
from local tumor draining veins and central veins. CTCs that have been detected were
isolated for their genomic characterization with regard to point mutations, MSI and
chromosomal instability. In addition, intact and freshly processed CTC samples were
required to enable mRNA expression profiling. Therefore, we had to establish a reliable
CTC enrichment and isolation protocol. Finally, a mouse model of metastatic CRC was
developed. The CTCs that were obtained from the mouse model ought to provide
further insights into the dynamic regulation of mRNA expression involved in tumor
cell dissemination.
Compartmental differences of CTCs in CRC
Due to the clinical significance and prognostic value of CTCs in CRC, we were deeply
interested if the presence and the detection rate of CTCs differs among different blood
compartments. For that purpose we used the US Food and Drug Agency (FDA)
approved CellSearchTM system to quantify the amount of CTCs in our patient cohort
and to investigate the distribution of CTCs in the mesenteric venous blood
compartment (MVBC) compared to the central venous blood compartment (CVBC).
CRC - associated CTCs that are shed from the primary tumor reach the liver via the
mesenteric and portal venous system before they enter the systemic circulation
50
. Our
study ought to provide data about primary hematogenous CTC dissemination and
systemic hematogenous spread with regard to an assumed CTC filtering function of the
liver. Furthermore, the presence of CTCs was correlated to various clinicopathological
parameters. The study was conducted and supervised by our group member Dr. Rahbari
and has been published 104.
Blood samples of 200 CRC patients who underwent surgical resection for CRC in the
department of surgery of the University Clinic Heidelberg between May 2009 and
April 2011 were enrolled prospectively. Table 1 gives an overview about the study
design and clinicopathologic characteristics of the study population.
26
Results
Central venous blood of 200 patients was analyzed for the presence of CTCs whereas
83 % of the analyzed blood samples showed no detectable CTCs. Univariate analysis
(Table 2) confirmed that the number of patients with CTCs as well as the CTC amount
correlated with the stage of disease. CTCs were detected in 10.2 % (19 of 142) of
patients with UICC stage I-III and in 35.8 % (19 of 53) of UICC stage IV patients
(p<0.0001). Stage I-III patients showed a mean amount of 0.2 (range: 0-5) CTCs per
7.5 ml blood sample and 2.4 (range: 0-83) CTCs were detected in stage IV specimens
(p<0.0001). Serum levels of CEA and CA 19-9 were associated significantly with
detection of CTCs in central blood (Table 2).
Regarding the detection of CTCs in MVBC data of 80 patients (40%) were available. A
significant higher rate of patients with CTCs in the MVBC (35%, n=28) compared to
the CVBC (17.5%, n=14) (p = 0.01) was found. Furthermore, the average number of
CTCs was higher in the MVBC (mean 1.5, median 0, range: 0-32) than in the CVBC
(mean 0.3, median 0, range 0-5) (p = 0.006) (Figure 9). CTCs in the CVBC were more
frequently found (p = 0.01) and with higher number of tumor cells (p = 0.01) in patients
with detectable CTCs in the MVBC. However, the presence of CTCs in the MVBC did
not correlate with the staging of the disease (Table 3). Univariate association of
clinicopathological variables with CTC presence and amount in MVBC showed that
CTCs in the MVBC were detected at a higher rate (p = 0.02) and with higher quantity
(p = 0.01) in patients with colon compared to patients with rectal cancer (Table 2).
Correlations of tumor cell detection in MVBC to other assessed parameters were not
found.
In summary, we found that the number of patients with CTCs and the amount of CTCs
was significantly correlated with the stage of disease. In addition, a significant higher
rate of patients with CTCs in tumor-draining venous blood compared to the central
venous blood was detected.
27
Results
104
Tabel 1: Clinicopathological characteristics of the study population.
Clinicopathological parameter
N (%) or median (range)
Age
65 (27-86)
Sex
Male
129 (64.5)
Female
71 (35.5)
Site of disease
Colon
96 (48)
Rectum
104 (52)
T stage
T0
5 (2.5)
T1
13 (6.5)
T2
50 (25.0)
T3
107 (53.5)
T4
25 (12.5)
N stage
N0
114 (57.0)
N1
52 (26.0)
N2
34 (17.0)
UICC stage
0
5 (2.5)
I
45 (22.5)
II
47 (23.5)
III
50 (25)
IV
53 (26.5)
Tumor differentiation
Moderate (G2)
112 (77.2)
Poor (G3)
33 (22.8)
CEA
Elevated(> 2.5 U/l)
65 (34.4)
CA-19-9
Elevated (> 37 U/l)
29 (15.3)
% of patients
40
MVBC
CVBC
30
20
10
50

10
5


4

3
2


1
0
CTC number in 7.5 ml blood
Figure 9: CTC amount in central venous blood compartment (CVBC) and mesenteric venous
blood compartment (MVBC) of CRC patients. 83% of patients were free of detectable CTCs in CVBC
and 65% of patients were without CTCs in MVBC.
28
Table 2: Clinicopathologic correlations. Associations of clinicopathologic variables with tumor cell detection in the central and mesenteric venous blood compartment of
patients with CRC (stage I – IV) 104
Central venous blood
Characteristics
Pt. with CTC
No. of CTC
0.13
Age
Mesenteric venous blood
Pt. with CTC
0.15
No. of CTC
0.54
0.57
≤ 65
20 (21.3)
1.2 (0; 0 – 83)
16 (38.1)
1.4 (0; 0 – 14)
> 65
14 (13.3)
0.3 (0; 0 – 5)
12 (31.6)
1.5 (0; 0 – 32)
0.25
Gender
0.23
0.09
0.04
Male
19 (14.7)
0.3 (0; 0 – 5)
14 (28.0)
1.0 (0; 0 – 14)
Female
15 (21.1)
1.6 (0; 0 – 83)
14 (46.7)
2.1 (0; 0 – 32)
0.31
Site of disease
0.23
0.02
0.01
Colon
19 (19.8)
1.4 (0; 0 – 83)
17 (48.6)
2.4 (0.5; 0 – 32)
Rectum
15 (14.4)
0.2 (0; 0 – 4)
11 (24.4)
0.7 (0; 0 – 14)
<0.0001
Stage of disease
<0.0001
0.04
0.09
I – III
15 (10.2)
0.2 (0; 0 – 5)
16 (28.1)
1.4 (0; 0 – 32)
IV
19 (35.8)
2.4 (0; 0 – 83)
12 (52.2)
1.6 (1; 0 – 14)
0.07
T stage
0.04
0.65
0.48
T0-2
7 (10.3)
0.1 (0; 0 – 1)
18 (33.3)
1.0 (0; 0 – 14)
T3-4
27 (20.4)
1.1 (0; 0 – 83)
10 (38.4)
2.4 (0; 0 – 32)
0.12
N stage
0.07
0.52
0.45
N0
14 (13.2)
0.2 (0; 0 – 2)
13 (31.7)
1.3 (0; 0 – 32)
N1/2
20 (21.3)
1.4 (0; 0 – 83)
15 (38.5)
1.6 (0; 0 – 14)
29
0.33
Tumor differentiation
0.26
0.13
0.06
Moderate (G2)
16 (14.3)
0.3 (0; 0 – 5)
14 (30.4)
0.8 (0; 0 – 14)
Poor (G3)
7 (21.2)
0.7 (0; 0 – 5)
6 (54.5)
5.0 (1; 0 – 32)
0.004
Resection margin
0.003
0.09
0.19
R0
28 (14.9)
0.7 (0; 0 – 83)
24 (32.4)
1.4 (0; 0 – 32)
R1/2
6 (46.1)
1.0 (1; 0 – 5)
4 (66.7)
1.8 (1; 0 – 8)
0.81
Neoadjuvant therapy
0.89
0.62
0.45
Yes
10 (16.7)
0.2 (0; 0 – 4)
7 (25.9)
0.5 (0; 0 – 5)
No
18 (15.2)
0.34 (0; 0 – 5)
12 (31.6)
1.8 (0; 0 – 32)
0.0001
CEA
<0.0001
0.16
0.23
Normal
11 (8.9)
0.1 (0; 0 – 4)
16 (31.4)
1.6 (0; 0 – 32)
Elevated
20 (30.7)
2.0 (0; 0 – 83)
12 (48.0)
1.3 (0.5; 0 – 14)
<0.0001
CA 19-9
<0.0001
0.10
0.15
Normal
18 (11.3)
0.2 (0; 0 – 5)
22 (33.3)
1.5 (0; 0 – 32)
Elevated
13 (44.8)
3.9 (0; 0 – 83)
6 (60.0)
1.2 (1; 0 – 4)
30
Table 3: Detection of CTCs in mesenteric and central venous blood compartments and correlation to clinical staging.
Pt. with CTC
No. of CTC
0.01
Central venous blood vs. mesenteric venous blood (n = 80)
Central venous blood
Mesenteric venous blood
Detection of CTC in mesenteric venous blood (n = 80)
14 (17.5)
28 (35.0)
Patients with CTC in central venous blood
Patients without CTC in central venous blood
Detection of CTC in central venous blood (n = 80)
9 (64.3)
19 (29.2)
Patients with CTC in mesenteric venous blood
Patients without CTC in mesenteric venous blood
Stage of disease (UICC)
9 (32.1)
5 (9.8)
Central venous blood (n = 200)
0
I
II
III
IV
P
0.3 (0; 0 – 5)
1.5 (0; 0 – 32)
0.01
0.02
2.0 (1; 0 – 14)
1.3 (0; 0 – 32)
0.01
0.01
0.5 (0; 0 – 5)
0.1 (0; 0 – 2)
0.0007
0.23
0.2 (0; 0 – 1)
0.1 (0; 0 – 1)
0.2 (0; 0 – 2)
0.3 (0; 0 – 5)
2.4 (0; 0 – 83)
1 (20.0)
3 (6.8)
6 (12.7)
5 (10.0)
19 (35.8)
1 (33.3)
6 (37.5)
5 (23.8)
4 (23.5)
12 (52.1)
P
0.006
0.27
Mesenteric venous blood (n = 80)
0
I
II
III
IV
104
0.68
0.3 (0; 0 – 1)
2.7 (0; - 32)
0.7 (0; 0 – 4)
1.3 (0; 0 – 14)
1.6 (1; 0 – 14)
31
Results
Establishment of a reliable CTC enrichment and detection method
We aimed to establish a protocol for an efficient enrichment and detection of CTCs that
should provide the opportunity to analyze gene expression of the captured CTCs. The
previously described study on CTC enumeration with the CellSearch TM system showed
that CTCs are hardly to find in blood samples of CRC patients. Therefore, high
efficiency and sensitivity of the enrichment method is inevitably necessary.
Magnetic beads fail to directly isolate EpCAM - expressing tumor cells
A first attempt was designed to directly target EpCAM-expressing CTCs in the blood
with the help of magnetic beads. Hence, DynaBeads coated with a set of different
antibodies directed against EpCAM (Moc31, VU1D9, HEA125, BerEP4, KS1/4) and
CEA (B1.1/CD66) were used. All antibodies were titrated to obtain the optimal
concentration and ratio. Through the use of different antibody clones that were
expected to target different epitops we supposed to increase the efficiency of the CTC
enrichment. Spiking experiments with HT29 CRC cell line showed that direct
enrichment with anti-EpCAM-coated magnetic beads is more efficient after density
gradient centrifugation (Ficoll) compared to a direct use of magnetic beads in whole
blood specimen. Efficiency was determined by RT-PCR for EpCAM and CK20
expression (data not shown). However, the direct enrichment of spiked HT29 from the
peripheral blood mononuclear cells (PBMC) was not superior to direct RNA extraction
of the spiked PBMC pellet. Moreover, the high detection limit of 1000 spiked cancer
cells and the strong carry-over of leukocytes made the method too inefficient for further
considerations (data not shown).
Leukocyte depletion with magnetic beads results in 10fold enrichment
CD45 is a pan-leukocyte marker, which is normally absent on epithelial cells. It was
decided to use CD45 to deplete leukocytes in blood samples and to enrich the residual
CTC fraction. Figure 10 shows that the differential CD45 and EpCAM expression of
leukocytes and tumor cells enables efficient leukocyte depletion and concomitant tumor
cell enrichment. Even when low numbers of tumor cells were spiked an enrichment of
more than 10 fold was possible (Figure 10). However, leukocyte carryover is still
32
Results
strong; therefore, further techniques are required to obtain pure CTC samples for
mRNA expression analysis.
A
B
C
D
Figure 10: Surface antigen expression allows identification of CTCs A: PBMC were spiked with
SW480 colorectal tumor cells and stained for EpCAM and CD45. Tumor cells and leukocytes are
detectable in clearly separated populations. B: Leukocytes that highly express CD45 are preferentially
depleted by anti-CD45 coated magnetic beads leading to tumor cell enrichment. C: Different CRC cell
lines (HT29, SW480, HD482, and HD835) were spiked in PBMC and subsequently enriched to
demonstrate the efficiency of the described method. D: Low amounts of tumor cells were spiked in
PBMC. The tumor cell enrichment by CD45 based leukocyte depletion was tested. DA: undepleted
control, DB: unspiked, depleted control DC: 0.5 % spiked, depleted, DD: 0.05% spiked, depleted, DE:
0.01% spiked, DF: 0.005% spiked, DG: 0,001% spiked
33
Results
Fluorescence-activated cell sorting (FACS) fails to isolate CTCs
FACS sorting was considered to isolate CTCs out of leukocyte - depleted and EpCAM
/ CD45 stained blood samples. Therefore, leukocyte specimens were spiked with 0.5%
tumor cells. After leukocyte depletion with the help of magnetic beads and fluorescence
staining a defined number of EpCAM+ /CD45- tumor cells was sorted and mRNAexpression analyzed. Despite a significant carryover of leukocytes, indicated by the
detected CD45 mRNA in the samples (data not shown), it was decided to apply this
method for patient material. Subsequently 17 patients’ blood samples, including ten
patients that had overt liver metastasis and three sample of mesenterial blood were
prepared and sorted with FACS Aria IITM. In each case between 20 and 100 cells were
sorted as EpCAM+ /CD45- by flow cytometry. Subsequent analysis by quantitative
polymerase chain reaction (qPCR), however, revealed that these objects were probably
unspecific fluorescent signals or no intact cells. Β-actin expression was detected in
twelve samples, however, late in qPCR (Ct > 20) and it correlated in most cases to
CD45 expression indicating a carryover of leukocytes. Only one sample showed a
putative EpCAM expression close to the detection limit of the qPCR (Ct > 37).
The micromanipulator enables the isolation of single CTCs
Since cell sorting by flow cytometry was not successful and bears the risk of losing the
rare CTCs, it was decided to apply a micromanipulator for CTC isolation. After density
centrifugation and the aforementioned anti-CD45 beads based pre-enrichment strategy
CTCs can be identified in stained blood samples with the help of fluorescence
microscopy by using EpCAM-directed antibodies. Subsequently, the micromanipulator
that is mounted on the fluorescence microscope allows the direct pipetting of single
cells with the help of a small glass capillary (Figure 11).
34
D
Results
Figure 11: The micromanipulator. Cells in suspension are picked with a micromanipulator (A)
mounted on a microscope. Fluorescence microscopy allows the identification of stained cells. (B=Bright
field, C=Nuclear stain /DAPI, D=Cytokeratin-PE staining) Subsequently, cells can be aspirated with the
help of an exchangeable glass capillary (E +F). The picking of two aggregated CTCs is shown.
A similar approach to isolate DTCs from bone marrow samples
To assess if the developed CTC isolation protocol can be applied for the isolation of
DTCs from bone marrow specimens as well we analyzed the expression of EpCAM
and CD45 in a bone marrow sample of a healthy donator (Figure 12 A). The majority
of bone marrow cells was found to express the leukocyte marker CD45. Therefore,
depletion of bone marrow cells with the help of anti-CD45-beads is suitable for DTC
enrichment. Despite a small portion of the bone marrow cells seem to express EpCAM
to a low extent (EpCAMlow) the identification of cancer cells spiked in the bone
marrow samples is still possible because EpCAM expression in tumor cells
(EpCAMhigh) was much stronger than in the BMDCs. Flow cytometry analysis
displayed the cancer cell population clearly apart from the bone marrow cells indicating
that the strong EpCAM expression of cancer cells allows the visual identification of
DTCs by fluorescence microscopy (Figure 12 B).
35
Results
A
B
Spiked HT29
tumor cells
Figure 12: Surface antigen expression enables identification of DTCs: A: Bone marrow samples
were analyzed for the expression of CD45 and EpCAM. B: A bone marrow sample was spiked with
HT29 cancer cells. Tumor cells can be distinguished from bone marrow cells due to the high expression
of EpCAM.
36
Results
An mRNA expression study on CRC - associated CTCs and DTCs
We assume that tumor cell dissemination is associated with distinct changes in the
mRNA expression profile. The present study ought to show if CTCs and DTCs display
a specific transcriptional signature that differs from the corresponding cancer tissue.
Validation of the cDNA amplification strategy for the analysis of single cells
Despite remarkable progress during the last years, mRNA expression analysis of low
cell numbers is still challenging and only possible if sophisticated amplification
strategies are applied. To make sure that the intended amplification kit is suitable for
the analysis of the low cell numbers ten and lesser cells were isolated for mRNA
extraction. After cDNA amplification mRNA expression was assessed by quantitative
RT-PCR. Figure 13 A shows that target gene mRNA expression of even single cells
was detectable. For validation of the mRNA expression analysis ten tumor cells were
isolated repeatedly in six independent experiments. The obtained mRNA was extracted
and amplified. Subsequently, the relative expression of EpCAM, CK19 and CK18
normalized to β-actin was examined (Figure 13). Since cells were cultured under equal
conditions, we expected low inter-experimental variations of the analyzed genes.
Despite single outliers the results showed comparable expression levels indicating the
reliability of the used amplification technique.
A
B
relative expression
2.5
2.0
1.5
1.0
0.5
K
18
C
K
19
C
Ep
C
A
M
0.0
Figure 13: Validation of the cDNA amplification strategy. A RT-PCR analyses of low cell numbers.
EpCAM expression of even a single HT 29 tumor cell was detectable. B Six independent target gene
expression analyses of ten pooled tumor cells by RT-PCR normalized to β-actin. The tumor cells that
were cultured under equal conditions show a comparable gene expression resulting in a low interexperimental standard deviation.
37
Results
Unambiguous mRNA expression profiles of CTCs and DTCs reveal biological and
therapeutic implications
Between June 2010 and October 2011 72 blood samples and 9 bone marrow samples of
CRC patients who underwent surgical resection for CRC in the department of surgery
of the University Clinic Heidelberg were analyzed. Preferentially patients with overt
liver metastases were selected to increase the chances to obtain blood samples
containing CTCs. According to the results obtained from the CellSearchTM system, that
indicate a higher number of CTCs in the tumor draining veins, the blood was taken
from a mesenteric vein or in case of liver metastases from the liver vein. CTCs were
enriched and isolated as previously described by a manual enrichment protocol to
enable the isolation of living cells. Figure 14 shows two images of CTCs as observed
after EpCAM staining in a processed patient’s blood sample. Additional blood samples
were analyzed by the CellSearchTM system as an independent control of the used
enrichment and isolation method.
Anti- EpCAM
Alexa488
Bright field
Figure 14: Human CTCs. CTC specimens from patient HD2095 stained with anti-EpCAM-Alexa488
At least one CTC was detected in 56 % (40) of the samples by the CellSearchTM system
in blood samples obtained either from the mesenteric or the liver vein. However, in
only 14 % (10) of the cases it was possible to successfully enrich CTCs from the blood
and in 22% (2) of cases to isolate DTCs out of bone marrow samples. Mostly, CTCs
were found with the manual enrichment approach only when high numbers of CTCs
were detected by the CellSearchTM system in parallel. CTCs/DTCs appeared usually as
single cells. Surprisingly, cell clusters (tumor microemboli) were only rarely observed
probably due to the applied enrichment procedure. Subsequently to the detection and
38
Results
isolation CTCs/DTCs of single patients were pooled for RNA extraction and
amplification (Table 4).
Table 4: Patients characteristics. Overview about patients from whom CTCs or DTCs were obtained
for mRNA expression analysis. No data about DTC incidence in bone marrow samples can be obtained
from the CellSearchTM system (indicated by X) since the device does not allow the processing of bone
marrow samples.
patient
sex
age
UICC stage
Blood
sample
(LV= liver
vein, MV=
mesenteric
vein)
CTC
HD 2091
HD 2095
HD 2130
HD 2201
HD 2231
HD 2257
HD 2288
HD 2334
HD 2340
male
female
female
male
male
female
female
male
male
57
69
57
64
62
55
59
73
62
IV
IV
IV
IV
IV
IV
IV
III
III
LV
LV
LV
LV
LV
LV
LV
MV
MV
DTC
HD 2401
male
39
II
HD 2417
male
70
IV
bone
marrow
bone
marrow
No. of CTCs
detected by
CellSearch
(7.5 ml
blood)
Manually
isolated
CTCs for
mRNA
expression
analysis
5
469
0
2
8
5
82
208
3
4
2 x 15
3
4
4
4
7
3x8
2
X
16
X
5
Although mRNA expression analysis of single cells would be feasible, we decided to
pool CTC populations to obtain more valid data and to increase the chances to detect
even lowly expressed genes. Robust signals in qPCR were obtained when at least four
cells were pooled for analysis. In addition, a similar number of cells from matched
tissue of either primary tumor or liver metastases was taken to compare mRNA
expression. Results are depicted in heat maps (Figures 15 – 18). A large-scale
expression screen comprising 47 genes related to CTC/DTC identification, common
CRC biomarkers, EMT, stemness and invasiveness were selected to describe
CTC/DTC characteristics. In addition, we included CD45, a pan leukocyte marker that
is not expressed on epithelial cells, in the study. Although it was a major aim to analyze
CTC/DTC specimen without any disturbing background of leukocyte contaminations, it
seemed to be impossible to obtain samples that were free of any CD45 signal in the
qPCR. The ubiquitary presence of free leukocyte-derived CD45 mRNA in the patients’
samples probably leads to an unavoidable detection of CD45 mRNA in qPCR. Since
we detected CD45 always in a comparable range among CTCs/DTCs and cancer tissue
39
Results
samples, the similar CD45 signal in qPCR should be rather considered as techniquederived noise. However, leukocyte carryover can not be excluded in general. Therefore,
the supposed impurity of the specimens requires a cautious evaluation of the obtained
expression profiles. In particular the expression of genes that are significantly
expressed in the control leukocyte and bone marrow samples including CD44,
Vimentin, CXCR4 and CD47 are difficult to assess. The robust signal of those genes in
qPCR analysis indicates that they are expressed on tumor cells, however, it is difficult
to state about their up or down-regulation during the dissemination process.
CTC expression profiles reflect a fairly heterogeneous picture
To compare general mRNA expression levels of CTCs and matched cancer tissue
samples the Ct values obtained from qPCR data are depicted as heat map in figure 15.
Since carryover of leukocytes might have introduced bias in the results; the mRNA
expression profiles of three control leukocyte samples are shown in addition. The
mRNA expression profile of CTCs is unambiguously distinct form leukocytes. The
expression of EpCAM, CK18, CK19 and CEA in the CTCs and the absence of these
markers in control leukocyte samples clearly confirmed the epithelial origin of the CTC
specimens. EpCAM, CK18 and CK19 appeared to be well suited for CTC
identification. CK20, however, seemed rather unstably expressed in CTCs. In addition,
Ki67 was found to be absent or only lowly expressed in CTCs, which might indicate a
proliferative arrest in CTCs. In general, gene expression of CTCs was found to be very
heterogeneous. Unsupervised hierarchical clustering analyses using the Spearman rank
correlation did not show a correlation among CTCs or among tumor and metastases and
no matched correlation between CTCs and the corresponding cancer tissues (data not
shown).
With the present approach we were only able to evaluate the expression of a small
panel of genes. Within this panel common traits of CTC samples seem to be limited to
a few epithelial markers. Other genes seem to be expressed only occasionally in
individual CTC specimens. For instance, the expression of stem cell markers like Bmi1 and Ascl2 was limited to CTCs of HD2201 or HD2095 respectively. The same
applies for the pro-apoptotic protein Bax that was detected in CTCs of patients HD2095
and HD2334 but not in other CTC specimens. Therefore, apoptosis seems to be
attributed only to some CTCs, however, not a general feature.
40
Results
Interestingly, it was not possible to confirm an enhanced expression of EMT-associated
genes in CTCs. Although Vimentin seems to be expressed in CTCs, EMT-related
transcription factors Snail and Twist as well as hFN, N-cadherin were mostly not
detectable.
Figure 15: mRNA expression of CTCs. mRNA expression analysis of isolated CTC samples and
matched samples of tumor (Tu.) or metastasis (Met.). Additionally, three leukocyte samples were
analyzed to evaluate the influence of possible background carryover. Color scale represents ΔCt values
of target and house keeping genes (β-actin). Low ΔCt values (yellow and green) indicate highly
expressed genes. High ΔCt values (blue and purple) indicate lowly expressed genes. A white area
indicates cases without available data.
41
Results
CTCs rather resemble metastases than primary tumors in gene expression
To compare gene expression between CTCs and the matched tissue samples two further
heat maps are shown in figure 16. The representation of ΔΔCt values as color code
facilitates the identification of differentially regulated genes among CTCs and matched
cancer tissue samples. Interestingly, evaluated genes seem to be rather up-regulated in
CTCs when compared to matched tumor samples, however, rather down-regulated or
equally expressed when compared to liver metastases. For instance CD47, which was
reported to be involved in immune escape
105
, was found up-regulated in CTCs only
when compared to tumor samples, however, equally expressed referring to metastases.
Or the expression of the hyaluronic acid receptor and stem cell marker CD44s which
seemed to be up-regulated in CTCs compared to primary tumor samples, but rather
similar expressed referring to matched metastases. Other evaluated stem cell markers
(CD166, Bmi-1, Lgr5, ZEB-1, and DLG7) were only rarely detected in CTCs and in
the corresponding samples of tumor or metastasis. For CD133 no significant
differential regulation was observed.
With regard to EMT CTCs retained Vimentin expression after dissemination. If
compared to tumor samples it seems to be even up-regulated. However, as previously
mentioned Vimentin is expressed in leukocytes as well and since leukocyte carryover
can not be excluded an up-regulation of Vimentin in CTCs can only be assumed. With
further regard to EMT we were not able to show an up-regulation of mesenchymal
genes like N-cadherin, Snail or Twist in the CTCs. Fibronectin expression seemed even
to be lower expressed in CTCs than in matched metastasis samples. E-cadherin was
found to be slightly down-regulated in CTC samples compared to the primary tumors;
however, in comparison to metastases its expression was rather similar. The
proliferation marker Ki67 seemed to be rather equally expressed among CTCs and
cancer tissues. Genes attributed to apoptosis and survival, including p21, p53, Survivin
and BAX, were not found to be differentially regulated. A significant up-regulation of
metastasis-associated genes like CD44v6, CD26 or CO029 in CTCs was not observed.
42
Results
A
B
Figure 16: Compared mRNA expression of CTCs and matched tumor (A) or liver metastasis (B).
ΔΔCt values of ΔCt tumor/metastasis – ΔCt CTCs are depicted as heat map. Color scale indicates
highly expressed genes of cancer tissue in yellow and green. Similar expression in both compartments is
shown in faint green. Genes that are highly expressed in CTCs are depicted in blue and purple. A white
area indicates cases without available data.
43
Results
DTCs are probably dormant and lack EGFR expression
Unfortunately, only mRNA expression data of two samples of DTCs is available
(Figures 17 and 18). Despite this small number of specimens, a first trend seems to be
evident. Unlike the CTCs the DTCs have lost CK19 expression in the bone marrow.
Only CK18 and EpCAM expression was reliably detected in DTCs and confirmed the
epithelial origin. E-cadherin appeared to be down-regulated. However, an up-regulation
of EMT-associated genes was not detected. Analyses indicated that DTCs were
probably non-proliferative and dormant as gene expression seemed to be in general
down-regulated irrespectively if compared to tumor of metastasis (Figure 18).
Numerous genes detected in tumor or metastases samples were not expressed in DTCs
including the proliferation marker Ki67 as well as CD151, CEA, CD26, CO029 and
EGFR. A significant expression of several CRC-associated stem cell markers, such as
Lgr5, Ascl2 or ZEB1 was also not detectable. Olfm4 and DLG7 seem to be expressed
in DTCs. However, these genes are expressed in control bone marrow samples as well.
Since undesired carryover can not be excluded, expression of Olfm4 and DLG7
requires independent confirmation.
44
Results
Figure 17: mRNA expression of DTCs. mRNA expression analysis of isolated DTC and matched
samples of tumor (Tu.) or metastasis (Met.). Additionally, three bone marrow samples were analyzed to
evaluate the influence of possible background carryover. Color scale represents ΔCt values of target and
house keeping genes (β-actin). Low ΔCt values (green) indicate highly expressed genes. High ΔCt values
(blue and puple) indicate lowly expressed genes.
Figure 18: Compared mRNA expression of DTCs and matched cancer tissue. ΔΔCt values of ΔCt
tumor/metastasis – ΔCt DTCs are depicted as heat map. Color scale indicates highly expressed genes of
cancer tissue in yellow and green. Similar expression is shown in faint green. Genes that are highly
expressed in DTCs are depicted in blue and purple.
45
Results
Genomic characterization of CTCs
Today’s therapeutic approaches against cancer come away from just targeting rapid
growing cells and try to interfere with specific pathways that are genetically altered in
tumor cells. However, genetic and epigenetic heterogeneity is common in human
cancers
106, 107
. Thus, the need to determine the molecular profile of cancer cells to
predict therapy response is becoming increasingly important. With this study we want
to assess the possibilities to characterize single CRC-associated CTCs on the genomic
level in the strict sense of their mutational profile.
CTCs carry typical point mutations in CRC-associated genes
CRC-associated CTCs that were enriched and detected by the CellSearchTM system
were isolated as single cells with a micromanipulator to analyze their mutational
profile. According to the CellSearchTM guidelines CTCs were defined as EpCAM+ /
CK+ (CK8, CK18, CK19 or in combination)/ DAPI+ / CD45-. A specific method to
amplify gDNA of single cells published by Klein and colleagues
108
was applied to
yield sufficient analyzable material. We were interested in common CRC-associated
mutations including BRAF V600E, mutations in codons 12 and 13 in the KRAS gene,
and aberrations within the DNA binding domain of TP53 (Figure 19).
Figure 19: Mutation analysis. Single CTCs were analyzed for mutations in BRAF (V600E), Kras
(codon 12 and 13), and in the DNA binding domain of TP53 (exons 4-8) by direct sequencing. The
figure depicts the included genomic regions.
46
Results
It was possible to isolate and analyze between 1 and 14 CTCs from 31 individual
patients (in total 126 CTCs) (supplementary figure 1). 68.8% of included patients were
diagnosed as UICC stage IV, 21.9% UICC III, 3.1% UICC II, 6.2% UICC I
(supplementary figure 2). In CTCs of 16 patients no mutations were detected within the
evaluated genomic regions of interest. A mutation of codon 12 or 13 of the Kras gene
was detected in CTCs of six patients (19.4%). The V600E mutation of BRAF was
found in a CTC of one patient (3.2%). CTCs of eleven patients were found to carry a
mutation in the DNA binding domain of TP53 (35.5%) (Table 5).
Table 5: Overview about detected point mutations in CTCs.
No. of patients with
indicated mutations
(31 patients, 126
CTCs, 1-14 CTCs per
patient)
Patients without a
detected mutation
Kras C12/C13
Braf V600E
DNA-BD TP53
16 (51.6%)
6 (19.4%)
1 (3.2%)
11 (35.5%)
Identification of CTCs with MSI within coding and non-coding genomic regions
MSI is a common phenomenon in CRC
12
. Currently, conflicting data exist about the
relevance of MSI for prognosis and therapy response in CRC
109, 110
. We wanted to
assess if the detection of MSI in amplified gDNA specimens of CTCs is possible. Since
our single cell gDNA amplification approach depends on the sequence specific
nuclease digestion by MSE I, it was not possible to include the same panel of MS
markers that is recommended by the national cancer institute (NCI) (Bethesda
guidelines)
111
. Therefore, we typed CTCs at the three mononucleotide MS markers
NR21, NR24 and BAT 25 that are widely accepted to indicate overall MSI in a cell 112.
At least to our knowledge studies on MSI in amplified genomic specimens of single
CRC-associated CTCs have not been reported until today. To exclude false positive
typing, what might be caused by artifacts occurring during gDNA amplification, we
included 35 normal leukocytes processed similarly to the CTC samples in the study.
None of the analyzed leukocytes showed any signs of MSI within the evaluated MS.
CTCs showing MSI were found in the blood samples of two patients (HD2215 and
HD2341; 6.5%) (Figure 20). Those samples were subsequently analyzed for MSI in
MS within coding genomic regions, so called coding microsatellites (cMS). The panel
of cMS included Bax, DD5, FLT3LG and ELAVL3, because the mutation frequency of
these MS-containing genes is well-known
113
and the applied gDNA amplification
method did not interfere with their evaluation. The three analyzed CTCs of HD2341
47
Results
showed MSI in at least two MS including NR21 and the cMS within ELAVL3 (Table 6).
ELAVL3 codes for a RNA binding protein of the HuR family and is involved in the
posttranscriptional regulation of mRNA expression. According to the recommendations
of the NCI tumors with mutations in two of a panel of five MS are considered to have
high MSI (MSI-H). If only one of the five MS is altered, the tumor is considered to
have low MSI (MSI-L). Although it was not possible to evaluate the same set of MSI
markers as suggested by the NCI, we typed the tumor cells of patient HD2341 as
MSI-H because of two altered MS. In case of patient HD2215 two of nine typed CTCs
showed MSI in NR21; however, no further MSI in any of the other analyzed markers
was detected. Therefore, the aberration in only one MS marker refers to a MSI-L
tumor. MSI-L was reported for various tumors including colorectal, breast, endometrial
and ovarian cancers
114, 115
and might not be attributed to defects in cellular DNA
mismatch repair systems. Instead, some DNA replication errors that occur
stochastically and to a normal rate might have remained unrepaired due to an overload
of the mismatch repair system as caused by high rate cell divisions in cancer.
HD 2215
HD 2341
Figure 20: MSI in CTCs. Defects in DNA repair mechanisms can lead to lengthening or shortening of
microsatellites (MS). Amplified length polymorphism (AFLP) analyses of CTCs and normal control
leukocytes of patients HD2215 and HD2341 for NR21 (blue) and NR24 (green) are shown. Stutter bands
arise since polymerase chain reaction amplification creates DNA fragments that are one or several
repeats longer or shorter than the actual allele. Therefore, the highest detected peak was determined as
the reference peak. Red peaks represent internal length control markers. Asterisks indicate the unchanged
MS-length as detected in control leukocytes. Arrows indicate an altered MS length in single CTCs. For
HD2215 CTC3 and HD2341 CTC1 no signal for NR24 was detected. This might be caused by losses of
gDNA fragments during the single cell amplification procedure.
48
Results
Table 6: cMS analysis of MSI+-CTCs. Three CTCs and one leukocyte (PBMC) of patient HD 2341
were analyzed for microsatellite instability (MSI) within coding DNA regions. No data was available for
microsatellite sequence within BAX of CTC 2341 mes1 and 2341 PBMC1. MSI = microsatellite
instability, MSS = microsatellite stable
CTC
NR21
NR24
BAX
DD5
FLT3LG
ELAVL
2341mes1
MSI
-
-
MSS
MSS
MSI
2341mes2
MSI
MSI
MSS
MSS
MSS
MSI
2341mes3
MSI
MSS
MSS
MSS
MSS
MSI
2341
PBMC1
MSS
MSS
-
MSS
MSS
MSS
Inter - CTC genomic heterogeneity confirms clonal diversity
In case of 16 patients more than one CTC could be analyzed. Six of those patients had a
population of CTCs that were genomic heterogeneous (37.5%) with regard to mutations
in Kras, BRAF, TP53 or MSI (supplementary figure 2). For instance, we have analyzed
nine CTCs from the patient HD2215 for MSI (Table 7). Two CTCs were obtained from
a blood sample of the liver vein (LV-CTC) and seven were isolated from mesenteric
venous blood compartment (MV-CTC). Seven CTCs did not show any signs of MSI in
the evaluated markers but two MV-CTCs showed MSI indicated by a changed allele
length of NR21. Additionally, a Kras mutation was detected in four MS stable (MSS)
CTCs of that patient whereas the mutation was not found in the MSI-CTCs. Another
case of inter–CTC cytogenomic heterogeneity was found for patient HD2328 (Table 8).
CTCs were isolated from blood samples obtained from the mesenteric vein and the
liver vein. Astonishingly, one half of the analyzed CTCs (3/6) isolated from blood of
the liver vein carried a mutation in exon 5 of the TP53 gene whereas the residual half
did not. Intriguingly, we could not detect this mutation in CTCs obtained from the
mesenteric blood compartment.
In summary, it was possible to detect MSI and CRC-associated genomic mutations in
single CTCs. In addition, we found cases of inter-CTC genomic heterogeneity in CTC
fractions of single patients. The presence of various tumor subclones within a patient is
reflected by different CTC populations that may have derived from different metastatic
sites.
49
Results
Table 7: Genomic inter-CTC heterogeneity of CTCs from patient HD 2215.
CTCs and cancer tissue samples of patient HD 2215 were analyzed for microsatellite instability (MSI)
and mutations within the indicated genomic regions. MSS = microsatellite stable, wt = wild type allele
detected, n.d. = no data available.
Patient
sample
/CTC
2215 Tumor
2215 liver
met.
CTC LV1
CTC LV2
CTC mes 1
CTC mes 2
CTC mes 3
CTC mes 4
CTC mes 5
CTC mes 7
CTC mes 8
Microsatellite
stability
Kras C12/13
Braf V600E
TP53 E5/6
TP53 E7
TP53 E8
MSS
G12S
wt
wt
wt
wt
MSS
G12S
wt
wt
wt
wt
MSS
MSS
MSS
MSS
MSI
MSS
MSS
MSI
MSS
G12S
n.d.
n.d.
G12S
wt
G12S
wt
n.d.
n.d.
wt
n.d.
n.d.
wt
n.d.
wt
wt
n.d.
wt
n.d.
n.d.
n.d.
n.d.
n.d.
n.d.
wt
n.d.
n.d.
n.d.
n.d.
wt
wt
n.d.
n.d.
wt
n.d.
n.d.
wt
wt
n.d.
wt
n.d.
n.d.
wt
n.d.
n.d.
Table 8: Genomic inter-CTC heterogeneity of CTCs from patient HD 2328.
CTCs and a liver metastasis sample of patient HD 2328 were analyzed for microsatellite instability
(MSI) and mutations within the indicated genomic regions. MSS = microsatellite stable, wt = wild type
allele detected, n.d. = no data available.
Patient
sample
/CTC
2328 liver
met.
CTC LV1
CTC LV2
CTC LV3
CTC LV4
CTC LV5
CTC LV6
CTC LV7
CTC LV8
CTC mes 1
CTC mes 2
CTC mes 3
CTC mes 4
CTC mes 6
CTC mes 7
Microsatellite
stability
Kras C12/13
Braf V600E
TP53 E5/6
TP53 E7
TP53 E8
MSS
wt
wt
F134C
wt
wt
MSS
MSS
MSS
MSS
MSS
MSS
MSS
MSS
MSS
MSS
MSS
MSS
MSS
MSS
wt
wt
wt
wt
wt
wt
wt
wt
wt
wt
wt
n.d.
wt
wt
wt
wt
wt
wt
n.d.
wt
n.d.
wt
wt
wt
wt
n.d.
wt
wt
F134C
F134C
wt
wt
n.d.
wt
F134C
wt
n.d.
wt
wt
n.d.
wt
wt
wt
wt
wt
wt
wt
wt
n.d.
wt
n.d.
wt
wt
n.d.
wt
wt
wt
n.d.
wt
wt
wt
wt
wt
wt
wt
wt
wt
n.d.
wt
wt
50
Results
Malignant identity and chromosomal profile of single CRC-derived CTCs can be
confirmed by matrix-CGH
Chromosomal instability is a hallmark of CRC
found to be of prognostic relevance
116, 117
11
and chromosomal aberrations were
. However, if specific chromosomal
alterations exist that are significant for the progression from invasive carcinoma to
metastatic disease is still a matter of debate. We aimed to evaluate in a pilot experiment
if the amplified gDNA of single CTCs is suited for a comparative genomic
hybridization (CGH) analysis with regard to compare the genomic profiles among
CTCs and solid cancer tissue. Therefore, array-CGH analyses of provided gDNA
samples were performed in the lab of Prof. Dr. N. Stöcklein at the University Clinic
Düsseldorf. To proof if pre-amplified gDNA from single cells allows the detection of
genomic aberrations, amplified gDNA from two leukocytes were hybridized against
each other on a 60k array (Figure 21). As expected, the ratio profiles are balanced for
all autosomes. The genomic imbalance found for the sex chromosomes correctly
reflects the different sexes of the two hybridized samples.
Figure 21: Validation experiment. Amplified gDNA of single male and female leukocytes were
hybridized against each other. The balanced ratio profile found for the autosomes confirms unbiased
gDNA amplification. The genomic imbalance detected for the sex chromosomes (X, Y) correctly reflects
the different sexes (male and female) of the samples.
We analyzed in this pilot study the genomic profile of ten CTC samples with the help
of a 180 k matrix-CGH array (supplementary figure 3). Three CTC samples were
detected with a balanced CGH profile. Seven CTCs showed chromosomal aberrations
that are typically found in CRC (Figure 24) (compare data base of University of Zurich:
www.progenetix.org). These genomic aberrations comprised chromosomal gains in
chromosomes 7, 8q, 13, 20 and losses in chromosomes 4, 8p, 18. Figure 24 depicts the
penetrance plot of pooled data from all analyzed CTC specimens. The pilot study
confirms that the evaluation of chromosomal instability in single CTCs is possible.
51
Results
Comparing genomic profiles of CTCs to matched cancer tissue
Today, individual therapy decisions rely on cancer tissue biopsies
118
. To assess if the
genomic profile of CTCs should be taken into consideration as well, we aimed to
evaluate if the genomic profile of CTCs might differ from matched cancer tissue
specimens.
Patterns of mutations often differ among CTCs and matched cancer tissue
The patterns of mutation of CTCs were compared with that of macro-dissections of the
respective primary tumor or liver metastasis (Supplementary figure 1). For 26 of 31
analyzed patients at least one of the included mutations was detected either in CTCs or
in the corresponding malignant tissue. In 14 of 26 cases the detected mutations matched
between the tissue and at least one analyzed CTC. A single time MSI analysis did not
match between CTCs and malignant tissue. Sequencing results were not corresponding
to each other eight times for Kras, three times for BRAF and three times for the
sequenced region in TP53.
CGH profiles of CTCs reflect individual clones of matched cancer tissue
Additionally, array-CGH profiles of matched macro-dissected tumor or metastasis
tissues could be generated for eight of the ten analyzed CTC samples (Supplementary
figure 3). Seven tissue samples displayed typical CRC-associated genomic aberrations,
one sample showed a balanced CGH profile without significant chromosomal changes.
Twice the CGH profiles matched largely among the analyzed CTC and the matched
cancer tissue. In six cases the CGH profiles appeared to be rather disparate between
CTC and the matched cancer tissue sample.
For instance, the CGH profiles a single liver vein - derived CTC and macro-dissected
tissue from a liver metastasis of patient HD2095 showed a broadly similar
chromosomal profile (Figure 22). Despite some disparate aberrations (chromosome 1q,
2p and 17q), lots of identical chromosomal gains and losses indicate a common clonal
origin of the analyzed specimens. Disparity might be attributed to the use of macrodissected tissue from the liver metastasis that represents the average of chromosomal
aberrations found in the metastatic lesion.
For liver a vein –derived CTC and liver metastasis of patient HD2288 we found largely
disparate array-CGH profiles. We assume that the analyzed CTC does not reflect the
prevailing clone of the metastasis (Figure 23). Figure 24 compares the accumulative
52
Results
CGH penetrance plots of pooled CTC samples with pooled data obtained from arrayCGH analysis of tumor or metastasis tissue. Typical genomic aberrations associated
with CRC were found. This includes chromosomal gains at chromosomes 7, 13, 20q
and losses at chromosomes 4, 8p, 17p and 18q.
53
Results
HD 2095 CTC
HD 2095 liver metastasis tissue
Figure 22: Array-CGH profiles of a liver vein - derived CTC and liver metastasis tissue of patient
HD 2095. The human set of chromosomes is shown (1-22, X, Y). Beside each chromosome the
detected signal intensity of the array-CGH analysis is depicted. Amplitudes to the left reflect
chromosomal losses. Amplitudes to the right reflect chromosomal gains. The detected chromosomal
aberrations are broadly similar among CTC and metastasis tissue and might indicate a common clonal
origin of the analyzed specimens.
54
Results
HD 2288 CTC
HD 2288 liver metastasis tissue
Figure 23: Array-CGH profiles of a CTC and liver metastasis of patient HD 2288. The human set
of chromosomes is shown (1-22, X, Y). Beside each chromosome the detected signal intensity of the
array-CGH analysis is depicted. Amplitudes to the left reflect chromosomal losses. Amplitudes to the
right reflect chromosomal gains. The detected chromosomal aberrations are largely disparate among
CTC and metastasis tissue and might indicate genetic clonal diversity of the analyzed cancer.
55
Results
A
CTCs
Cancer tissue
B
Figure 24: Accumulative penetrance plots of array-CGH data. A The penetrance Plots of pooled
array-CGH results of CTC and cancer tissue samples are shown. Therefore, the human set of
chromosomes (1-22, X, Y) was aligned and the frequency of detected aberrations was depicted. Arrows
indicate typical CRC-associated chromosomal gains (red) and losses (green) found in both specimens. B
For comparison of CGH profiles to database entries a penetrance plot provided by the university of
Zurich (www.progenetix.org) is shown. The depicted plot includes CGH data of 1054 cases of CRC.
Significant genomic aberrations shown in database were similarly detected in own analyses.
56
Results
An orthotopic mouse model of cancer cell dissemination
As confirmed by own results, CTCs are rarely to find in blood samples of CRC
patients. Therefore, we aimed to establish a model to mimic cancer cell dissemination.
Animal experiments on cancer metastasis which inject tumor cells into the tail vein
119
or the portal vein120 are little physiologic and omit initial steps of primary tumor growth
and invasion. Ectopic animal models like subcutaneous tumor models
generate; however, most of them do not metastasize
122, 123
121
are easy to
. Thus, we decided to
establish an orthotopic mouse model of metastatic CRC because the microenvironment
in the host organ is known to have an enabling effect on metastasis 124. With the murine
model it was intended to reproducibly obtain CTCs to a higher number than from
human samples, because it is possible to take the total blood volume of a mouse
through terminal heart puncture. Moreover, the disseminated cells would be of human
origin; accordingly, possible contaminations with murine leukocytes would not
influence the results of the target gene specific mRNA expression analyses.
With the help of a microinjection system human CRC cells were injected into the cecal
wall of the animals. First trials were started with Balb/c nu/nu mice. The nude mice
lack a thymus which makes them incapable to generate mature T-cells and makes them
suited for xenotransplantation 125. We decided to test different CRC cell lines including
HCT116, LS174T, SW620, DLD1 and Colo205 for their potential of hematogenous
dissemination in the nude mice. These cell lines were chosen because their use in
orthotopic tumor models has already been described
126, 127
. More than 80 % of the
animals developed orthotopic tumors. Additionally, lymph node metastases were
frequently observed (Figure 25). No DTCs were found in the bone marrow. Solely
implantation of HCT116 lead to macroscopic liver (26%; n = 19) or lung (5% n = 19)
metastases (Figure 26).
57
Results
Figure 25: The orthotopic mouse model of CRC metastasis. The primary tumor (left), a lymph node
metastasis (middle) and macroscopically visible liver metastases (right) of a xenotransplanted animal are
displayed.
A
B
C
D
Figure 26: Histological staining. A: H/E staining of tumor tissue. B-D: EpCAM staining of tumor (B),
liver metastases (C) and micro-metastases in the lungs (D) of a HCT116 xenotransplanted animal. The
red color indicates tumor cell - associated expression of human EpCAM. Nuclei appear blue due to
counterstaining with hematoxylin.
Of the tested cell lines, HCT116 seemed to be the only one with a tendency for
hematogenous tumor cell dissemination. Therefore, further experiments were carried
out only with HCT116. Immunohistochemistry and flow cytometry analysis of stained
tissue samples from tumor, liver and lung metastases showed that the transplanted
58
Results
HCT116 cell line retains EpCAM expression after xenotransplantation which makes
EpCAM a suitable marker for the detection of CTCs and DTCs in blood and bone
marrow (Figures 26 and 27).
Figure 27: EpCAM expression of HCT116 after xenotransplantation. Single cell suspension of
primary tumor, liver and lung of a xenotransplanted mouse and a mock transplanted control animal were
stained with anti-EpCAM-Alexa-488 antibody and analyzed by flow cytometry. EpCAM-expressing
tumor cells (as indicated) were found in all analyzed tissues of the xenotransplanted mouse.
29 animals that were xenotransplated with 2 x 106 HCT116 cells developed a primary
tumor but CTCs were only observed in five mice. As we were surprised of such a low
frequency, we supposed that the level of immunodeficiency in Balb/c nu/nu mice
sufficed to clear CTCs out of the blood. Consequently, it was decided to change the
mouse strain to NOD/SCID/gamma mice. Currently, these mice are among to the most
immunodeficient strains available
128
. They lack mature T-cells, B-cells and natural
killer cells and have defects in various cytokine signaling pathways. In addition, we
decided to increase the potential for hematogenous dissemination by in vivo passaging
of the cell line since this has been shown to augment the metastatic capability 129. Thus,
instead of original HCT116 we used a sub-cell line that was obtained from an HCT116derived liver metastasis. This cell line is termed HCT116-P in the following.
To assess tumor development and the point in time of tumor cell dissemination
HCT116-P tumor cells were orthotopically transplanted into NOD/SCID/gamma mice.
Tumor growth and appearance of CTCs were monitored over time (Figure 28). All
59
Results
animals developed primary tumors, liver metastases and lung metastases. Liver and
lung metastases were assessed by histological staining (H/E staining). Liver metastases
appeared between day 14 and 21, lung metastases developed between day 21 and day
28. Beyond that, the CTC amount in cardiac blood was determined as well. No CTCs
were found before day 25 post tumor cell injection (ptci).
Tumor volume
1 x 105 HCT116 cells in NOD scid gamma mice
Tumor weight
1 x 105 HCT116 cells in NOD scid gamma mice
400
tumor weight [µg]
tumor volume [mm3]
600
400
200
300
200
100
0
0
0
10
20
30
40
0
day after tumor implantation
10
20
30
40
day after tumor implantation
Figure 28: Tumor development. Graphs show the time-dependent increase of tumor size and weight.
Within 30-35 days all animals developed tumors of a mean weight of 0.3 g (Figure 28).
At this stage lymph node metastases and macroscopic visible liver metastases were
seen in all mice. A correlation of tumor weight to CTC amount was not found. No
DTCs were found in the bone marrow.
It was possible to isolate CTCs from eight mice. The amount of CTCs varied between
the animals from zero to several hundreds of CTCs. Isolated CTCs were viable and
formed colonies when cultivated in vitro (Figure 29). Morphological characteristics and
the expression of EpCAM confirmed tumor cell identity and human origin of the
adhesive cells. Moreover, the CTC-derived cell line induced tumor growth in
NOD/SCID/ gamma mice when injected subcutaneously.
Cancer progression can lead to the accumulation of fluid in the peritoneal cavity which
is called acites. Xenotransplanted mice suffering from advanced cancer developed
ascites containing thousands of tumor cells. Tumor cells that have disseminated into the
ascites are termed AS-CTCs in the following.
60
Results
Figure 29: Colony formation of CTCs isolated from
murine blood. CTCs isolated from mouse blood form
colonies in vitro.
CTCs and AS-CTCs were stained6 with an EpCAM antibody for identification and
isolation. Figure 30 shows that the staining intensity due to different EpCAM
expression varies strongly among AS-CTCs. CTCs found in blood and ascites appeared
often as cell clusters (Figures 30 and 31).
A
B
Figure 30: Fluorescence staining of AS-CTCs. AS-CTCs of a xenotransplanted animal were stained
for membrane-bound EpCAM. A AS-CTCs and residual leukocytes in bright field. B Green fluorescence
of anti-EpCAM-AlexaFluor 488 stained AS-CTCs. Different fluorescence intensity indicates variations
in the protein expression levels among the AS-CTC.
61
Results
Figure 31: Fluorescence staining of CTCs. CTCs that were detected in murine blood are depicted in
bright field and under green fluorescence after EpCAM specific fluorescence staining. CTCs detected in
blood have a different appearance. Often different patterns of cell clusters were observed.
62
Results
mRNA expression profiling reveals different gene expression patterns
In each case about 15 cells of primary tumor tissue, of CTCs and of AS-CTCs were
isolated and processed for mRNA expression analyses. After mRNA isolation and
cDNA amplification the gene expression was evaluated with regard to cancerassociated biological traits like proliferation, cell cycle, stemness, metastasis and EMT.
Reliable results of the conducted expression analyses were available for 17 tumors, 9
CTC and 18 AS-CTC samples obtained from eight mice. As expected, gene expression
analyses of low cell numbers varied strongly. This might be attributed mainly to
fluctuations in mRNA expression within individual cells. Therefore, common changes
in gene expression require statistical evaluation and can only be detected if numerous
samples are analyzed. Figure 32 gives an overview of the relative expression values.
With regard to target gene expression CTCs resemble AS-CTCs. Comparing in vitro to
in vivo culture of HCT116-P cells revealed most changes in the transcriptional profile.
Genes that might be differentially regulated during tumor cell dissemination include Ecadherin, Snail, CD44s and EGFR as shown by comparing CTC with tumor mRNA
expression.
63
Results
40
A
EGR1
30
p53
E-cad
AS-CTC (Ct norm)
snail
CCND1
p21
CD133
Vim
20
CLDN7
Bmi1
Surv
CD166
CD151DLG7
ki67
c-myc
EpCAM
CD26 Calre
CK19
CD44s
CD44v6
10
EGFR
B-cat
CK18
R²=0.97
0
0
10
20
30
40
CTC (Ct norm)
40
B
30
cell line (Ct norm)
snail
Vim
E-cad
p53
CD133 p21
CCND1
20
CD26
EGR1
CLDN7
EGFR
CD151
Surv
CD44s
c-myc
DLG7
ki67
CD166
Calre
EpCAM
CD44v6 Bmi1
CK19 B-cat
10
CK18
R²=0.81
0
0
10
20
30
40
Tumor (Ct norm)
64
Results
40
C
E-cad
30
Maspin
p53
CTC (Ct norm)
p21
snail
CD133
CCND1
CLDN7
20
EGFR
Vim
CD26 Bmi1
Calre
Surv
c-myc
CD166
ki67
DLG7 CD44s
EpCAM CD151
CD44v6
CK19
B-cat
CK18
10
R²=0.81
0
0
10
20
30
40
Tumor (Ct norm)
Figure 32: Comparison of mRNA expression. Transcript intensities are correlated between (A) CTCs
and AS-CTCs, (B) unprocessed HCT116-P cells and tumor, and (C) CTCs and tumor. For reasons of
clarity and comprehsibility standard deviation is only shown for selected genes. CTCs and AS-CTCs are
transcriptionally similar. Unprocessed HCT116-P compared to HCT116-P-derived tumors display most
differences in transcript intensity. High Ct values indicate a low gene expression.
CTCs maintain epithelial traits
To confirm the epithelial and human identity of the isolated CTCs expression of
EpCAM, CK19 and CK18 was tested. Despite remarkable variations of transcript
intensities the expression of the three assessed epithelial markers was constantly
detectable. Figure 33 indicates that the expression of the epithelial markers is
maintained during dissemination and no significant expression changes took place.
This confirms that EpCAM, CK19 and CK18 are suitable and reliable marker genes for
the detection of disseminated tumor cells.
65
Results
CK19
4
2
lin
e
ce
ll
SC
TC
A
C
TC
0
or
relative CK18 expression
6
Tu
m
lin
e
ce
ll
SC
TC
0
A
lin
e
ce
ll
SC
TC
A
C
Tu
m
TC
0
2
TC
1
4
C
2
6
or
relative CK19 expression
3
CK18
8
Tu
m
4
or
relative EpCAM expression
EpCAM
5
Figure 33: Expression of epithelial markers in CTCs. EpCAM, CK19 and CK18 expression of tumor
tissue, CTCs and AS-CTCs normalized to β-actin are shown. No significant changes in mRNA
expression among those genes were detected.
CTCs lack signs of EMT
To support the hypothesis that disseminating tumor cells undergo a change to a rather
mesenchymal phenotype, specimens were tested for the expression of EMT-inducing
transcription factors including Twist, Snail and ZEB1. Expression of Twist and ZEB1
was not detectable in any of the samples. Snail expression was found to be slightly upregulated in CTCs (2 fold) and AS-CTCs (3 fold) samples; however, statistical
evaluation did not confirm significance. Since EMT can be induced through numerous
pathways, the expression of EMT target genes was analyzed, too. Although an average
down-regulation of E-cadherin in CTCs (106 fold) and AS-CTCs (32 fold) compared to
tumor samples was observed, data failed to reach statistical significance. N-cadherin
and hFN expression were not detectable in the samples. Vimentin expression was
detected in the specimens but expression was not significantly different among tumor
and disseminated cells.
CTCs stop proliferation and up-regulate anti-apoptotic Survivin
To assess if disseminated cells persist in a dormant, non-proliferative state and if they
are prone to undergo apoptosis the expression of Ki67, Cyclin D1, c-myc, β-Catenin,
p53, p21 and Survivin was examined. Despite strong variations of mRNA expression
levels, an up-regulation of Survivin was found in CTCs (4.6 fold, p=0.004) which was
statistical significant (Student’s t-test, p < 0.05) (Figure 34). The proliferation marker
Ki67 was significantly down-regulated (10 fold, p = 0.04) in AS-CTCs, indicating a
presumable proliferation arrest (Figure 34). No signs for differential gene expression
were found for Cyclin D1, c-myc, β-Catenin, p53 or p21.
66
Results
Ki 67
relative Ki67 expression
8
6
4
2
p=0.04
8
6
4
2
lin
e
ce
ll
A
SC
TC
C
TC
m
Tu
lin
e
ce
ll
SC
TC
A
TC
C
Tu
m
or
0
0
or
relative Survivin expression
Survivin
p=0.004
10
Figure 34: mRNA expression of Survivin and Ki67. Survivin and Ki67 expression was assessed in
tumor, CTCs, AS-CTCs and in unprocessed HCT116-P cells. Survivin was found to be significant upregulated in CTC samples potentially protecting CTCs from apoptosis. The down-regulation of the
proliferation marker Ki67 in AS-CTCs seems to indicate a proliferation arrest in AS-CTCs as a
consequence of long-term dissociation from the primary tumor site. Statistical significance was
determined by Student’s t-test. P - values below 0.05 were considered as statistical significant.
No significant changes in the expression of stem cell markers in CTCs
Samples have been analyzed for the expression of stem cell-associated markers.
Expression of Olfm4, Ascl2 and Lgr5 has not been detected in the specimens. In case
of CD133, CD166, Bmi1, EGR1 and DLG7 no significant differences in gene
expression were found.
Most metastasis-associated genes are equally expressed in CTCs and tumor samples
The expression of EpCAM, CD44v6, CO029, CD151 and CLDN7 facilitates metastasis
formation in CRC 130. Therefore, we aimed to evaluate the expression of these genes in
our mouse model. However, no significant differences in the expression of EpCAM,
CD44v6, CD151 and CLDN7 were found. The expression of the tetraspanin CO029
was not detected. Interestingly, CD44s was found to be significantly up-regulated in
CTCs (p= 0.017; 9.4 fold) and in AS-CTCs (p=0.03; 8.4 fold) as compared to tumor
samples (Figure 35). Additional metastasis-associated genes including MMP7, CXCR4
and CD26 seemed not to be differentially regulated. EGFR was found to be downregulated (5 fold) in CTCs compared to tumor samples but failed to reach statistical
significance (Figure 35).
67
EGFR
2
ce
ll
lin
e
0
SC
TC
lin
e
ce
ll
SC
TC
A
C
Tu
m
TC
0
4
A
20
6
TC
40
8
C
60
p=0.11
10
or
relative EGFR expression
CD44s
Tu
m
p=0.017
80
or
relative CD44s expression
Results
Figure 35: mRNA expression of CD44s and EGFR Relative expression of CD44s and EGFR was
measured by RT-PCR. Statistical significance was determined by Student’s t-test. P - values below 0.05
were considered to be statistical significant. Significant differences of CD44s mRNA expression were
found between tumor and CTCs specimens. EGFR expression failed to reach statistical significance
among the different subgroups.
In summary, our mouse model of metastatic CRC provides the opportunity to study
differential gene expression during the process of tumor cell dissemination. We found
that the expression of epithelial markers including CK18, CK19 and EpCAM is not
significantly altered after the dissemination. Although down-regulation of E-cadherin
seems to support metastasis, convincing signs of EMT in the disseminated cells were
not found. However, the detection of differentially expressed genes like Survivin and
CD44s, that were found to be up-regulated in CTCs, might allow first conclusions
about the mechanism of tumor cell dissemination.
68
Discussion
Discussion
This thesis provides comprehensive data about the detection, isolation and
characterization of CRC - derived CTCs and DTCs. Beside an own study on CTC
incidence in blood samples of CRC patients, we provide intriguing data about genomic
and transcriptional profiles of CTCs and DTCs. In addition, we established an
orthotopic mouse model to mimic cancer cell dissemination. To my knowledge a
comparable wide-ranging approach to elucidate the molecular mechanisms of tumor
cell dissemination has never been reported before.
Compartmental differences of CTCs in CRC
To study the CTC amount in our patient cohort we used the FDA-cleared CellSearchTM
system. We analyzed blood samples from the CVBC, which were either obtained from
the internal jugular or subclavian vein, and blood samples from the MVBC, which
were collected from a tumor-draining mesenterial vein. Blood was taken shortly before
surgical tumor resection and allowed the evaluation of compartmental differences of
CTC distribution in both venous compartments. We could confirm that the detection of
CTCs in central venous blood correlates to the stage of disease
53
. CTCs were found
rarely and in low numbers in blood samples of patients with early stages of CRC.
Therefore, the use of CTC detection systems for prognostic and therapeutic purposes
might be largely limited to advanced stages of CRC. The low number of CTCs that are
detected in early cancer stages is usually below the dynamic range required for
measuring treatment response.
Following the pathways of tumor cell dissemination in CRC we have compared the
distribution of CTCs in central and mesenteric venous blood. CTCs were found to be
more frequently present and with higher numbers in MVBC than in CVBC. Previous
studies
131-134
that have analyzed the presence of CTCs in the mesenteric venous blood
of CRC patients are in line with our findings. However, those studies relied mainly on
RT-PCR to detect the expression of CEA or cytokeratins and did not allow the exact
enumeration of CTCs. Our data supports the cascade theory in CRC assuming a
stepwise tumor progression with the liver acting as a filter for CTCs
50
. The detection
of CTCs in central venous blood indicates that the filtering function of the liver seems
69
Discussion
to be incomplete. We found an increased quantity of CTCs in CVBC in patients with
CTCs in the MVBC suggesting an amount-dependent filtering of CTCs in the liver.
The development of overt metastases causes further dissemination of CTCs into the
circulation which is supported by the findings that the CTC amount in CVBC increases
considerably in patients with stage IV disease and that the presence of liver metastases
was found to be a strong and independent factor associated with CTC detection in the
CVBC but not in the MVBC. Although various studies
53, 132
regarding the association
of clinicopathologic variables with the detection and count of CTCs in mesenteric and
central venous blood already exist, there has been no concomitant evaluation of CTC
incidence in CVBC and MVBC using the standardized CellSearchTM system. An
elevated preoperative CA 19-9 level was found to be independently correlated with the
presence of CTCs in the CVBC. CA 19-9, which has been widely used as a tumor
marker, is a sialylated Lewis blood group antigen and a ligand for the endothelial cell
adhesion molecule E-selectin
135
. Several studies have shown that the preoperative
serum CA 19-9 levels are associated with a poor prognosis for patients with CRC
137
136,
. Beside the function of CA 19-9 in adhesion of CTCs to the endothelium, our data
might suggest that CA 19-9 could possibly be involved in the detachment of tumor
cells from the primary tumor and in CTC persistence in the systemic circulation.
The prognostic relevance of the detection of CTCs in the mesenteric compartment is
still controversial. A meta-analysis
84
of the few available studies was not able to
confirm a significant association of mesenteric CTCs with disease recurrence and
survival whereas CTC detection in the peripheral/central blood was a strong predictor
of poor outcome. Our data revealed that CTC detection in the MVBC is only weakly
correlated to the assessed clinicopathological variables and is in line with the
inconsistent finding of previous studies. Therefore, we suppose that most of the CTCs
that are found in the MVBC are not of clinical relevance. Probably, most of CTCs that
are shed into the circulation can not adapt to the foreign microenvironment and die
quickly
41-43
. However, CTCs that are found in the CVBC have proved their ability to
survive in the systemic circulation and have accomplished a further decisive step on the
way to malignant progression.
70
Discussion
Establishment of a reliable CTC enrichment and detection method
mRNA expression studies on CTCs and DTCs are essential for understanding the
biology of the initial steps of the metastatic cascade. Since standardized methods for
CTC/DTC isolation are lacking, own approaches for single cell analyses had to be
developed and evaluated. The right choice of suitable CTC markers appeared to be a
major problem. Since intracellular proteins such as cytokeratins require the
permeabilization of cells for staining, they were not eligible to obtain intact cells
required for mRNA expression studies. We decided to focus on EpCAM as epithelial
surface marker. EpCAM is widely used for the detection of CTCs in blood samples of
cancer patients since it is absent on leukocytes. Also the CellSearchTM, the only FDAapproved, standardized system for CTC enumeration in various cancers makes use of
EpCAM as initial CTC marker 85. Nevertheless, EpCAM remains controversial because
it is described as a cell surface protein with functions in cell-cell adhesion that might
counteract tumor cell shedding. In 2009 Sieuwerts et al.
138
reported that some breast
cancer cell lines even lack EpCAM expression. In addition, some authors
40, 139
claim
that epithelial genes such as cytokeratins and EpCAM might be down-regulated in the
course of EMT which is assumed to be critically linked to tumor cell dissemination.
Although a general presence of EpCAM on CTCs might be uncertain, a total loss of
EpCAM on CRC - associated CTCs has not been shown so far. Despite concerns
EpCAM is presently at least to our opinion the best suited marker to distinguish
epithelial cells from leukocytes in human blood samples. Therefore, we have decided to
use EpCAM as an eligible marker for CTC detection.
Comparing the efficiency of our manual CTC enrichment approach to the standardized
CellSearchTM system revealed that CTCs were detected to a lower number and less
frequently applying the manual protocol. Whereas the CellSearchTM applies ferrofluidic
nanoparticles and strong electromagnets to enrich CTCs, our method depends on
leukocyte depletion by relative large magnetic beads that are removed with simple
ferromagnets. In addition, our approach requires searching for CTCs under a
fluorescence microscope by the unaided eye. This might not be as efficient as it can be
achieved with the sensitive camera system that is used in the CellSearch TM.
Nevertheless, our manual CTC enrichment method enabled the detection of intact
CTCs and with the help of the micromanipulator it has been possible to directly isolate
single CTCs without a strong leukocyte carryover. Establishing a functional CTC
71
Discussion
detection and isolation protocol has been decisive for the following study on mRNA
expression in CTC samples.
A mRNA expression study on CRC - associated CTCs and DTCs
At least to our knowledge, here, we describe for the first time a study on the mRNA
expression of almost 50 genes in nearly pure fractions of CRC-associated CTCs and
DTCs. CTCs/DTCs are rare and only to find in a strong background of leukocytes or
bone marrow cells, which is a major obstacle for their molecular analysis. The impurity
of the CTC/DTC fractions used in previous reports on their transcriptional signatures
has been a substantial drawback of these studies
100-102
. Analyzing rare CTCs in a
strong background of several thousands of leukocytes is mainly limited to the detection
of genes that are exclusively expressed by CTCs and can therefore only be used to
define novel CTC markers. However, an unbiased evaluation of mRNA expression in
CTCs/DTCs requires pure cell samples. With our micromanipulator-based method that
allows the pipetting of single cells, we aimed to obtain CTC/DTC samples free of any
leukocyte carryover. Despite promising results with blood samples spiked with cancer
cells, it was, however, not possible to isolate CTCs/DTCs from patients’ blood samples
lacking any CD45 mRNA signal. CD45 was actually intended to measure leukocyte
carryover; however, some researchers recently reported co-expression of CD45 with
epithelial markers in CTCs of some patients with different kinds of cancer
140
. At the
moment we can only speculate about a putative contribution of CTCs to the detected
CD45 signal. Further efforts are needed to clarify if CD45 is attributed exclusively to
leukocytes or if it is at least occasionally expressed on CTCs as well. In addition, it was
recently reported that CTCs might avoid NK cell-mediated lysis through the
aggregation with platelets
47, 48
. Moreover, Pawelek et al. have suggested that CTCs
might fuse with BMDC in the blood
49
. Both scenarios could lead to the CD45
detection in the specimens. Probably, our CTC/DTC samples are not absolutely free of
any leukocyte carryover; nevertheless, the leukocyte background has been considerably
reduced compared to other studies
100-102
. The detected CTC/DTC expression profiles
are distinct from profiles of leukocytes or cells of the bone marrow and show a unique
expression signature. Epithelial markers such as EpCAM, CEA, Calreticulin, CK18 and
CK19 indicate a substantial amount of cancer cells in the evaluated samples. Generally,
72
Discussion
a strong heterogeneity among the CTC/DTC specimens was observed which might
indicate the existence of various CTC/DTC subpopulations.
Although mRNA expression analysis of single cells would be feasible, we decided to
pool CTC populations to obtain more valid data and to increase the chances to detect
even lowly expressed genes. Studies on gene expression in single cells or small cell
pools are challenging until today. A large scattering of detected mRNA expression
levels has to be expected since transcription is rather a stochastic than a continuous
process
141
. The inherent randomness in transcription impedes the precise phenotypic
determination of individual cells. Particularly, the noisy transcription of lowly
expressed genes requires large numbers of samples and statistical evaluation to state
significant trends.
CTC samples
In this pilot study CTC samples from only ten patients were evaluated, common
findings are therefore limited but some remarkable aspects are worth discussing.
Interestingly, some genes, including CD47, CD44s and Vimentin seem to be upregulated in CTCs when compared to tumor tissue, however rather similar expressed
when referred to metastasis samples. It might be conceivable that the expression of
some metastasis-associated genes in CTCs resembles the metastases rather than the
primary tumors. For instance high expression of CD47, which might be involved in
tumor immune escape
105
, might be necessary for CTC survival in the circulation but
could be favorable for metastatic growth as well. Probably, metastases maintain at least
to some extend the expression profiles they already gained during the early steps of
tumor cell dissemination. E-cadherin is another example supporting this hypothesis. Ecadherin was found to be down-regulated in the CTC samples when compared to
primary tumors, however similarly low expressed as detected in metastases samples.
Lower expression of E-cadherin in metastases compared to matched primary tumors
was already reported for CRC
142
; hence, the reduced expression of cell-cell adhesion
molecules seems to be crucial for tumor cell detachment. However, our results do not
confirm a significant contribution of EMT to tumor cell dissemination. It is widely
reported that EMT might be a prerequisite for metastasis
39
but our CTC samples seem
to lack expression of various putative inducers of a mesenchymal transition including
Twist and Snail. Furthermore, an up-regulation of mesenchymal genes was also hardly
detectable. One might argue that our CTC identification method relies exclusively on
73
Discussion
the epithelial marker EpCAM and therefore could miss CTCs that have undergone
EMT and lack expression of epithelial proteins. Recent studies
139
with the ISET
system, a CTC detection method depending on size-dependent filtration, have shown
the presence of non-small-cell lung cancer-associated CTCs that did not express
epithelial markers. However, at least to my knowledge neither the existence of CRCassociated CTCs lacking the expression of epithelial markers nor the contribution of
such cells to cancer staging or prognosis has been reported so far. In addition, if
mesenchymal-transformed tumor cells are able to initiate metastasis is highly
controversial at all
57 143
. On the other hand the impact of EpCAM+ - CTCs on cancer
prognosis was confirmed several times 86 87.
It was reported, that CTCs are frequently apoptotic
144
. We did not detect the
differential regulation of apoptosis related genes such as Bax, Survivin, p53 and p21 in
our samples. However, our CTC samples were obtained from tumor or metastasis
draining vessels and not from peripheral blood samples; therefore most of the analyzed
CTCs have probably just detached from the solid cancer tissue and represent only a
snapshot of an assumed heterogenic CTC population. Apoptosis might occur delayed
and after a longer exposure to shear stress and the loss of cell-cell contacts.
Interestingly, we did not find an up-regulation of metastasis-associated genes, including
CD44v6, CO029 and CLDN7 in the CTC samples. Despite convincing reports
130, 145
that these genes indicate aggressive tumor growth and metastatic spread, their
expression seemed not to be attributed to the analyzed CTC samples.
We found the proliferation marker Ki-67 absent or down-regulated on CTC samples
which is in accordance to a study about breast cancer-associated CTCs
146
. Although
we can not exclude that proliferating CTCs exist, a proliferative arrest might be useful
for CTCs to adapt to the foreign microenvironment in the blood. Interestingly, we did
not detect the down-regulation of β-catenin which would additionally indicate a low
proliferation of CTCs since it is a main part of the Wnt signaling pathway 147. However,
regulation of the Wnt pathway might be achieved post-transcriptionally as well 148.
DTC samples
It was only possible to include two samples of DTCs isolated from the bone marrow of
CRC patients in this pilot study. We were able to identify DTCs due to EpCAM
expression amongst a strong excess of bone marrow cells. Detection of CK18
confirmed the epithelial origin of the DTCs. Despite CK20 is the most commonly used
74
Discussion
marker for DTC detection in RT-PCR analyses
90
, no CK20 mRNA expression was
detected in our DTC specimens. Together with the lost CK19 expression this might
indicate signs of an occurred de-differentiation in the cells and questions the use of
CK19 and CK20 as DTC markers. The gene expression profile of DTCs seems to be
distinct from their corresponding primary tumors or metastasis. In particular with
regard to systemic therapeutic approaches the molecular characterization of DTCs
might have enormous therapeutic impact. For instance, DTCs, that we have analyzed,
seem to lack EGFR expression which might mediate chemotherapy resistance and
explain limitations of anti-EGFR based therapies in systemic treatment.
The proliferation marker Ki67 was also not detected in DTCs. This is in line with other
reports that describe DTCs mostly as non-proliferative and dormant cells
146, 149
.
However, DTCs might be able to escape proliferative dormancy when supplied with
appropriate growth factors as shown by cell culture experiments
150
. DTCs that are
hidden throughout the whole body are highly suspicious to cause cancer relapse even
years after primary tumor resection 58. Thus, to reduce cancer reoccurrence therapeutic
regimens are required that effectively target DTCs. Since the molecular traits of the
primary tumor are probably distinct from DTCs, the development of successful
therapies calls for further ambitious studies on the characterization of DTCs.
Genomic characterization of CTCs
The present study combined the standardized CTC-enumeration of the CellSearchTM
system with the concomitant genotyping of clearly defined and detected CTCs. Single
CK-positive stained cells were isolated with a micromanipulator mounted on a
fluorescence microscope for the subsequent global amplification of the gDNA. To
evaluate the possibilities to characterize CTCs on the genomic level we searched for
point mutations in Kras, BRAF and TP53 as well as for signs of MSI. In addition, some
cells were used for global array-CGH analysis. Point mutations, MSI and chromosomal
instability have been associated with survival, prognosis or therapy response in CRC.
For instance, mutations of Kras and Braf indicate a lack of therapy response to antiEGFR monoclonal antibodies like cetuximab and panitumumab
by the constitutive activation of the Ras-MAPK pathway
154
151-153
which is caused
. Mutated TP53 has been
associated with survival 155, 156 and therapy response 157. MSI and specific chromosomal
75
Discussion
losses or gains were found to be of prognostic relevance
158-162
as well. Therefore, the
genomic characterization of CTCs might provide an opportunity to improve prediction
of prognosis and to adapt therapy regimes to the genetic aberrations of individual
tumors. With the present study we were able to show that epithelial cells detected with
the CellSearchTM carry CRC-associated mutations and we provide evidence that these
cells are of malignant origin. Despite a limited number of samples we detected
remarkable inter-CTC heterogeneity in CTC fractions obtained from single patients. Of
the 17 patients from whom we were able to analyze more than one CTC a genetic
disparity among CTCs of a single patient was found in six cases (37.5%). Genomic
heterogeneity of CTCs seems to be a common phenomenon and an evidence for
genomic instability and clonal diversity in CRC. A study from Fehm and colleagues 163,
that mainly included cases of breast cancer, reported heterogeneous CTC pools in 13 of
20 cases (65%). However, they used in-situ hybridization for the detection of
aneusomie. Since their approach covers more potentially mutational sites, the higher
detection rate might thus be explained. The presence of genomic heterogenic CTC
populations further complicates the selection of appropriate therapy regimens and
might require the analysis of a number of representative CTCs when individual tailored
therapy is intended.
In addition, we aimed to evaluate to what extend the detected genomic aberrations
correlate among CTCs and matched cancer tissue. In twelve of 26 cases we found
disparate genotypes for CTCs compared to cancer tissue. Intriguingly, a high portion of
disparate genotypes (CTCs vs. cancer tissue) was found for Kras and BRAF whereas
disparity was less frequently observed for TP53. Typically, inactivating mutations of
tumor suppressor genes like TP53 are found homozygous
164
. Oncogenes like Kras or
BRAF, however, often carry activating mutations in only a single allele and are
therefore heterozygous
165
. One might argue that the applied single cell gDNA
amplification technique can not reliable detect heterozygous mutations. Due to the low
amount of gDNA of a single cell some DNA fragments could fail to be amplified and
thus cannot be detected in sequence specific PCR. Therefore, sequencing results might
be misleading in heterozygous situations if only one of two different alleles can be
analyzed. Generally, it can not be excluded that allelic losses occur during the gDNA
amplification. Therefore, it is inevitable necessary to define the rate of amplificationassociated allelic losses. Since no own experiments have been performed on this, we
refer to the work of Schardt, et al.
166
who reported a rate of allelic losses due to
76
Discussion
technical issues of 9.8 % in their control. Although the processing of blood samples
with the CellSearchTM system differs from the methods that Schardt and colleagues
have used, we expect a comparable rate of allelic losses because the same amplification
protocol has been applied. The global amplification of gDNA may have caused allelic
losses that left heterozygous Kras or BRAF mutations in CTCs unrecognized. However,
the disparity between the detected mutations in CTCs and corresponding tissue cannot
be explained exclusively by technical reasons. We believe that the disparity probably
results from the use of marco-dissections to analyze the malignant tissue. Despite a
presumptive monoclonal origin tumors are heterogeneous and consist of various
subpopulations 106. Minor subclones that harbor only the wild-type alleles might remain
unrecognized when thousands of cells were pooled for analysis. In addition, it is
conceivable that specific genetic changes might have supported tumor cell
dissemination. For instance, the loss of genes involved in adherence or the
amplification of genes related to invasiveness would probably favor the dissemination
process. To our knowledge there are no comparable studies about genomic
heterogeneity in CRC associated-CTCs. Most studies analyze genetic disparity between
primary tumors and various sites of metastases. Findings indicate, despite considerable
inter-study variations, disparate point mutations for Kras in 0-60% 98, 167-169 170 and for
TP53 in 20-60% 171, 172 of cases. There are a few studies describing genetics of DTCs in
bone marrow
97, 173-175
. The frequently observed genomic disparity in DTCs in
particular in early stages of cancer, they observed, is often explained with postulated
early tumor cell dissemination 99. Tumor cell diversity is still large in early stages since
genomic instability will lead to the development of various subclones. Once the tumor
reaches full malignancy the most aggressive clone will prevail. An early dissemination
and the subsequent parallel progression of DTCs, which further contributes to the
divergent development, might explain the detected genomic disparity. However, it is
believed that CTCs can not survive for long time in the circulation. Shear stress
anoikis
44
and the easy access of the immune system
176
43
,
are thought to be the main
reasons for the short persistence of CTCs in the blood. The high number of genetic
disparity between CTCs and tumor tissue, that we have detected, is fairly surprising and
in contrast to a study from Khan et al.
177
. They studied TP53 mutations in primary
tumors, liver metastases and CTCs enriched by magnetic beads out of blood samples
from CRC patients. They reported that if a TP53 mutation was detected it was found
invariably in all tumor and liver metastasis samples from a single patient and in eight of
77
Discussion
19 cases it was found additionally in the CTC sample. However, their method could not
determine if TP53 wild-type tumor cells were also present in the enriched CTC fraction
or in TP53 mutated carcinomas since they approach did not allow single cell analysis.
Our CTC isolation protocol provides the opportunity to analyze single CTCs without
any leukocyte carryover and is, thus, well-suited for CTC- genotyping. CTCs can be
obtained repeatedly and relatively noninvasive from blood samples, therefore,
information about their genomic profile might contribute to treatment decision and the
estimation of prognosis. For instance, information about the mutational status of Kras
might be crucial for individual therapeutic intervention. The failure of anti-EGFR based
therapies was reported for colorectal tumors harboring mutant Kras
152
or Braf
153
. We
were able to show that the mutational status of Kras and BRAF can differ among CTCs
and matched cancer tissue. Cancer cells might have disseminated at an early point in
time and have developed in parallel to the tumors they have derived from. Therefore,
we conclude that genomic analysis of single cancer tissue specimens will not be
sufficient to state about the genomic profile of systemic cancer. Moreover, it is even
conceivable that genotypic changes of CTCs might occur during cancer progression.
This was shown by Meng and colleagues 178. They observed in nine of 24 breast cancer
patients an acquired Her-2 gene amplification in CTCs although the primary tumor was
initially Her-2 negative. Conversely, the emerge of Her-2 negative CTCs in Her-2
positive breast cancers subjected to anti-Her-2 therapy was reported by Hayes and
collegues 179.
In conclusion, here, we describe that the comprehensive genomic characterization of
single CRC-associated CTCs is feasible. Cancer tissue biopsies consist of an average of
malignant and non-malignant cells of a tumor and usually neglect individual cancer
subclones. However, since tumors might consist of heterogeneous cancer cells, the
single cell analysis could be favorable. We showed that point mutations, MSI and
chromosomal aberrations that might be important for prediction of survival and therapy
response can be evaluated in single cells. In addition, the identification of genotypes
that could be associated with tumor cell dissemination and ectopic survival might open
new opportunities to develop effective treatments.
78
Discussion
An orthotopic mouse model of cancer cell dissemination
CTCs are extremely rare and difficult to detect in blood samples of CRC patients. To
mimic metastatic spread, we have established an orthotopic mouse model of cancer cell
dissemination in NOD/SCID/gamma mice. Our CRC model comprises all relevant
steps of metastasis including lymphogenic and hematogenic tumor cell dissemination
and the successful formation of metastases in liver and lungs. It was possible to isolate
CTCs from cardiac blood and tumor cells from the ascites of some animals. CTCs were
viable and could be cultured and expanded in vitro. Moreover, subcutaneous injection
of CTC-derived cells confirmed their tumorigenicity.
Clearly lesser CTCs were detected in the cardiac blood in our CRC model than reported
for othotopic xenograft models of breast 180and prostate 181 cancer. Most murine models
determine CTC numbers in blood obtained from heart puncture. In case of an
orthotopic colon cancer model the venous blood that leaves the tumor needs to pass the
liver before it reaches the heart
50
. However, most of the draining blood of prostate or
mamma carcinomas is drained via the vena cava without having to pass another organ
which may act as a filter for CTCs. The detection of lower CTC numbers in our CRC
model confirms the strong filter effect of the liver that we have already postulated in
the human setting. Moreover, metastases seem to significantly contribute to overall
cancer cell dissemination since no CTCs were detected before liver metastases have
formed. We did not find any correlation of the tumor size to the CTC number in the
xenograft model. Although this is in line with previous findings
180, 181
, the presented
cancer model does not allow for the accurate determination of tumor burden throughout
the animal. Micrometastases might have established at unknown sites distributed all
over the body. Applying constantly marked tumor cell lines would be desirable for
further studies. Cancer cell lines with intrinsic GFP or Luciferase expression might
facilitate an exact tumor and metastasis quantification and might enable an unbiased
CTC count that does not depend on the expression of epithelial genes.
Most of the studied genes were found to be similarly expressed in CTCs and solid
tumor samples. This is in accordance with a previous report about CTCs in an othotopic
prostate cancer model
181
. We share the opinion of Helzer and Barnes that the
biological features necessary for metastatic spread might be already inherent in the
aggressive cell lines used in both models. This might explain the limited number of
differentially regulated genes that were detected. Interestingly, Howard and
79
Discussion
colleagues 182 established a CTC cell line from the same orthotopic PC-3 prostate
cancer model. They report a down-regulation of cell adhesion molecules including Ecadherin in the CTC-derived cell line. Although, not statistically significant, a clear
trend for the down-regulation of E-cadherin in CTCs was also observed in our study.
Together with similar findings in human CTC samples E-cadherin down-regulation
might be crucial for tumor cell detachment and migration.
Addressing the question whether EMT is involved in tumor cell dissemination we
studied the expression of EMT-related transcription factors. Although a tendency for a
slight up-regulation of Snail was observed in CTC samples, we did not find convincing
evidence that EMT is critically related to the process of tumor cell dissemination in the
present mouse model. E-cadherin down-regulation in CTCs might be achieved through
other mechanisms than EMT and since no up-regulation of mesenchymal genes was
detected, the contribution of mesenchymal transition to tumor cell dissemination in this
model remains uncertain. Interestingly, a tendency for the down-regulation of EGFR
was found in CTCs of the mouse model. Abolished expression of EGFR was also seen
in the DTCs from human patients. This is in contrast to a study by Wang et al. 183. They
report the mutual regulation of E-cadherin and EGFR. Loss of E-cadherin was found to
up-regulate EGFR expression and enhanced the proliferation of squamous cell
carcinoma of the head and neck in cell culture experiments. However, the influence of
the microenvironment and the loss of close cell-cell contacts after tumor cell
dissemination might additionally affect EGFR expression. Further studies are required
to elucidate transcriptional regulation of EGFR during the dissemination process.
CTC survival and persistence in the circulation may require cellular strategies to
develop resistance to apoptosis and anoikis. The up-regulation of the anti-apoptotic
Bcl-2 protein in CTCs has been already reported in a previous study 181. We were able
to show a significant increase in the expression of Survivin, an inhibitor of apoptosis, in
CTC specimens. Although p53 and β-catenin are linked to Survivin regulation, the
expression of these genes does not seem to be differentially regulated in CTCs
compared to solid tumor tissue. The involved regulatory pathways remain therefore
elusive. Of the metastasis-associated genes included in the present study, we found
CD44, a hyaluronic acid receptor, up-regulated in CTCs. Hyaluronic acid is the major
component of the extracellular matrix and over-expression of CD44 in CTCs might be
involved in homing of CTCs to distant organs
184
. However, increased expression was
only seen for the standard isoform CD44s. The transcription level of the putatively
80
Discussion
metastasis-associated extended isoform CD44v6 seemed not to be altered. Kuhn et
al. 130 postulated that a complex of CD44v6, CO-029, CLDN7 and EpCAM contributes
to metastasis formation. According to our results which do not indicate the differential
expression of these genes in our experiments, a correlation could not be confirmed.
This, however, might be attributed to the used HCT116-P cells, that generally do not
seem to express the tetraspanin CO-029 which might be important for complex
formation. Further studies with other cell lines might be useful to assess the role of the
CD44v6/CO-029/CLDN7/EpCAM complex in tumor cell dissemination.
Conclusion and outlook
In conclusion, our approaches provide the opportunity to characterize CTCs and DTCs
on the molecular level. We showed that the molecular profiles of CTCs and DTCs are
not necessarily the same than in the primary tumor or the site of metastasis. Therefore,
genomic analyses of cancer tissue biopsies for treatment decision might not be
sufficient for systemic cancer therapy. Moreover, analyzed CTCs appeared to be very
heterogeneous in general. Therefore, we believe that various subpopulations of CTCs
exist and only a minority is able to induce metastatic growth. However, until today the
metastasis-initiating capacity of human CTCs has never been proofed under
experimental conditions. Neither the molecular traits nor the appropriate culturing
conditions that could define the true metastatic precursors have been revealed so far. To
obtain better insights into the biology of metastasis and the resistance to established
therapies extended studies on CTCs/DTCs are required. Single cell research provides
the means to gain comprehensive knowledge about individual cancer cell
heterogeneity. Moreover, the direct characterization of single CTCs/DTCs may allow
the identification of novel therapeutic targets to defeat systemic cancer disease and may
elucidate the role of cancer stem cells in metastasis formation.
81
Material and methods
Material and methods
Methods
Blood samples
Peripheral blood samples were drawn from healthy donors from the antecubital vein
and collected in EDTA tubes. To avoid epithelial cell contamination from skin
puncture, the first 3 ml of blood were discarded. Patients’ blood samples, from patients
undergoing surgical therapy at the Department of surgery of the University Hospital
Heidelberg, were obtained after induction of general anesthesia through a central
venous catheter or from portal or liver vein. After collection, blood samples were
immediately processed. The study was approved by the ethics committee of the
University Heidelberg.
Bone marrow samples
Bone marrow samples were obtained after induction of general anesthesia by iliac crest
biopsy.
Mononuclear cell collection
Blood samples were layered over 15 ml of LSM 1077 lymphocyte gradient. The
samples were centrifuged at 4° C for 30 min at 300g without brake. The interphase
containing the PBMC fraction was collected and spun down for 5 min at 300 g and
washed once with PBS.
CTC/DTC enrichment and staining
PBMC were collected from about 27 ml blood and resuspended in 1.5 ml of beads
buffer (PBS, 2 mM EDTA, 0.1% AB serum). After washing in beads buffer 450 µl
magnetic beads were added to PBMCs and incubated in an overhead shaker for 20 min
at 4° C. According to manufacturers instruction leukocyte depletion ought to be
performed in a larger volume of buffer. Therefore, cell suspension was diluted with
additional 8 ml of beads buffer before magnetic separation. Finally, the remaining
supernatant was centrifuged 5 min 300g to obtain the CTC-enriched cell fraction. Pellet
was resuspended in 150 µl of PBS buffer containing 10% FCS and 50 U/ml penicillin/
streptomycin. Per sample 5µl of Alexa-Fluor488-anti-EpCAM antibody was added.
82
Material and methods
After 20 min of incubation on ice the stained CTCs were identified under a
fluorescence microscope. A similar protocol was applied for the bone marrow samples.
In brief, bone marrow aspirates were collected according the mononuclear cell
collection protocol. The further enrichment and staining of cell samples including
potential DTCs was performed in the same way as described for the PBMC specimens.
In case of CTC/DTC detection the identified cells were directly picked with the
micromanipulator into RNA lysis buffer.
Cancer tissue specimens
Cancer tissue specimens were minced through a 45 µm cell strainer and collected in
PBS. After centrifugation for 5 min and 300g the cell pellet was resuspended in 150 µl
of PBS buffer containing 10% FCS and 50 U/ml penicillin/ streptomycin and stained
for cancer cell identification with 5 µl anti-EpCAM Alexa488 antibody. Single cancer
cells or small cell clusters were picked with the micromanipulator directly into RNA
lysis buffer.
CellSearchTM (Veridex)
7.5 ml of patients’ blood samples, collected in CellSaveTM tubes, were used for CTC
analysis according to the manufactures instructions.
Flow cytometry
Cell sorting was performed on FACS Aria II Flow Cytometer (Beckton Dickinson).
Protein expression analysis by flow cytometry was carried out on FACS Calibur
(Beckton Dickinson). Data was analyzed by FlowJo or FACS express software.
Immunohistochemistry
1-2 µm sections of formalin-fixed, paraffin-embedded tissue samples were mounted on
object slides. After de-waxing in xylol and ethanol, antigen retrieval was achieved by
incubating in target retrieval solution (Dako # S1699) for 20 min at 95° C. After
washing in PBS, slides were incubated in methanol containing 0.3% H2O2 followed by
washing in PBS. Non-specific binging sites were blocked with 20% normal goat serum
for 1 hour. Slides were incubated with pre-labeled (Solulink All in one HRP
conjugation kit #A-9002-001) Moc31-HRP antibody (IQ Products) over night. After
83
Material and methods
washing in PBS the immunoreaction was visualized by using ABC Pro Vecta Stain kit
(Dako) according to the manual.
Cell culture
Cells were cultured at 37° C, 5% CO2 and 95% humidity in RPMI supplemented with
10% FCS, 50 U/ml penicillin and 50 U/ml streptomycin in culture flasks. Cells were
collected with trypsin EDTA and centrifugation for 5 min at 300 g.
RNA extraction
Total RNA was extracted by PicoPure RNA isolation kit (Arcturus) according to the
manual.
RNA amplification
Amplification of total RNA was achieved by the use of WT Ovation RNA
amplification kit (NuGen).
RT-PCR
Real time PCR using SYBR green was performed with LightCyclerTM (Roche)
according to the manual. An annealing temperature of 60° C and 55 cycles of
amplification were used. Amplification plots were analyzed using the second derivative
maximum method. All Primers were designed with Primer3Plus software and span
large intronic regions to exclude amplification of genomic DNA.
gDNA amplification of single cells
Single cells were transferred into PCR tubes containing 9µl of PBS. 1µl of
Proteinase K mix (1x OnePhorAll Buffer (OPA), 0.65% Tween, 0.65% Igepal, 1.3
mg/ml Proteinase K) was added. Cells were digested for 10 h at 42° C in a
thermocycler followed by an inactivation step of 10 min, 80° C. For MseI digestion
0.4µl of MseI enzyme (high concentration 10.000 U/ml) was added. After an
incubation time of 3 h at 37° C the enzyme was inactivated by heating the reaction mix
to 65° C for 5 min. LIB1 (LIB1: AGTGGGATTCCTGCTGTCAGT) and ddMse11
(ddMSE11: TAACTGACAGCdd) primers were annealed in a reaction mix of 5µl 10x
OPA , 5µl LIB1 100µM, 5µl ddMse11 100µM, 15µl H2O in a water bath. For ligation
6µl of pre-annealed adapters combined with 2µl of 10mM ATP together with 2µl T4
84
Material and methods
DNA ligase were added to the digested cell. Ligation occurred over night at 15° C.
Genomic DNA was amplified with the help of the Expand long template PCR system
(Roche) in buffer 1. Therefore, 3µl Buffer 1, 2µl dNTPs 10µM, 25µl H2O and 1µl
DNA-polymerase mix were combined and added to the reaction mix. Amplification
was performed in a thermocycler under the following conditions.
cycles
program
1x
68°C
14x
3min
8x
22x
1x
94° C
40 s
64° C
40 s
94° C
40 s
68° C
220 s
57° C
30 s
57° C
30 s +
65° C
30 s
4° C
∞
68° C
113 s +
1°
C/cycle
68° C
90s +
68° C
1 s/Cycle
150 s +
1
1
s/Cycle
s/Cycle
For quality control of gDNA amplification a multiplex PCR was used. Therefore 0.3µl
template was combined with 5µl Dream Taq Green PCR Mastermix (Fermentas), 1µl
of 10 x primer mix (see table below) and 3.7 µl H2O. PCR was run in thermocycler
(pre-incubation 2 min 95° C; 33 x amplification 30 s 95° C / 40 s 60° C / 60 s 72° C;
final elongation 5 min 72° C). Successful amplification was assessed on an agarose gel.
10x Primer mix:
Primer
LAMC1 for
LAMC1 rev
GRIK5 for
GRIK5 rev
NRK9 for
NEK9 rev
CAPS for
CAPS rev
PICK1 for
PICK1 rev
DNAH9 for
DNAH9 rev
final conc.
0.2 µM
0.2 µM
0.1 µM
0.1 µM
0.1 µM
0.1 µM
0.08 µM
0.08 µM
0.08 µM
0.08 µM
0.08 µM
0.08 µM
PCR Product
111 bp
232 bp
288 bp
175 bp
358 bp
401 bp
Gel electrophoresis of DNA using agarose gels
PCR products were verified for correct size by electrophoresis at RT using 1.5 % (w/v)
agarose gels suspended in 1x TAE buffer. GelRed was used in the agarose gel to
visualize the DNA under UV light.
85
Material and methods
Purification of PCR products
PCR clean up kit (Sigma) was used for purification of PCR products that were used for
down stream applications such as sequencing, MSI analysis or CGH.
Sequencing
PCR products that were used for sequence analysis were generated by AmpliTaq Gold
DNA Polymerase (Applied Biosystems). DNA was sequenced through sequencing
laboratory (GATC / Konstanz).
Macrodissection and DNA extraction
For the macro-dissection procedure fresh frozen cancer tissue samples were used.
Sections of 10µm were cut in a cryostat at -20° C and adhered to object slides. One
section of each macrodissected sample was hematoxylin / eosin stained and examined
by light microscopy to discriminate cancer tissue from stroma and healthy tissue.
Between ten and 20 sections, depending on the amount of cancer tissue, were collected
for gDNA extraction with the help of Qiagen DNeasy tissue kit according to
manufacturers instructions. The estimated amount of cancer tissue within the dissected
area was in every case at least 70 %.
Comparative genomic hybridization (CGH)
Array-CGH was performed externally by the group of Prof. Dr. N. Stöcklein at the
University clinic Düsseldorf. For single cell analysis Agilent SurePrint G3 Human
CGH 4x180K Oligo Microarray Kit (Cat. G4449A) was used. Amplified gDNA of
single CTCs and matched leukocytes was hybridized against each other. Array-CGH
analysis of macro-dissected snap frozen cancer tissue probed against matched healthy
tissue was carried out with Agilent SurePrint G3 Human CGH 8x60K, Oligo
Microarray Kit (Cat.: G4450A). Data processing was performed with the following
software provided by Agilent Technologies: Scan Control V 7.0.3; Feature Extraction
V 10.1.1.1 and Genomic Workbench V 5.0.14. The following parameters were used for
data analysis:
Genome build: hg18;
Evaluation algorithm ADM-2, threshold 6;
86
Material and methods
Aberration filter: minimal number of probes in area 3, minimal mean log 2 ratio for
area 0.25
MSI
Analyses of microsatellite instability were performed externally by the group of Dr. M.
Kloor in the department of pathology at the University Hospital Heidelberg. Amplified
gDNA of CTC samples as well as tumor samples were tested for MSI in NR21, NR24
and BAT 25 from the standard NCI/ICG-HNPCC marker panel
185
and analyzed as
described previously 186.Genome amplification of CTC-derived gDNA led to disruption
of the DNA while the afore mentioned markers remained unaffected and were selected
for MSI analysis.
Mice
For experiments BalbC nu/nu and NOD/SCID gamma mice were used. Authorization
number 35-9185.81/G-7/10 was obtained from the national authorities for research
experiments on animals. Mice were kept under specific pathogen free conditions in IBF
animal facility Heidelberg.
Orthotopic injection of tumor cells
Mice were anesthetized by 3-5% of isofluran. A small abdominal incision was made
and the cecum was exteriorized. 10 5 tumor cells in a total volume of 20µl of Matigel
(10mg/ml) were injected orthotopically into the coecal wall by microinjection.
Incisions were closed with 6.0 resorbable sutures. For postoperative pain relief 4mg/kg
Rimadyl was given subcutaneously.
Statistical evaluation
To identify independent factors associated with detection of CTCs in the CVBC and
MVBC the following statistical tests were applied. Categorical data were presented as
absolute and relative frequencies and compared using the χ2-test. Continuous data were
presented as median and range and compared using the Wilcoxon test. Furthermore, the
mean number of CTCs was reported. Analysis of covariance (ANOVA) was applied to
compare continuous data of more than two groups.
The significance of mRNA expression data differences was evaluated with Student’s Ttest. P – values below 0.05 were considered as statistical significant.
87
Material and methods
Material
Equipment
CellSearch (Veridex)
Ortho Clinical Diagnostics
Centrifuges
Heraeus Multifuge 3; Eppendorf Centrifuge 5810R
Electrophoresis power supply
Consort E835
Electrophoresis units
Neolab
FACS Aria II
BD
FACS Calibur
BD
Hood
Heraeus / Herasafe
Incubator
Binder
Light cycler 480
Roche
Microinjection system
World Precision Instruments
Microscopes
Leica
Nanodrop
Peqlab
PCR Thermocycler System
Eppendorf
pipettes
Gilson
Shakers
Assitent RM5
Thermomixer
Eppendorf
Transferman Micromanipulator
Eppendorf
UV- transilluminator
BioDoc-It System
Vortex
Scientific industries Vortex Genie 2
Chemicals and reagents
Agarose
Invitrogen
Antigen retrieval solution
Dako
ATP
NEB
Deoxyribonucleotide triphosphates (dNTPs)
NEB
Dimethylsulfoxide (DMSO)
Sigma
DMEM
PAA
DNA ladder
Fermentas
DNA loading dye
Fermentas
DynaBeads
Dynal / Invitrogen
EDTA
Roth
Ethanol
Roth
Fetal Calf Serum (FCS)
PAA
Gel red DNA stain
Biotium
H2O2 (30%)
Roth
Igepal-CA630
Sigma
Isofluran
CP Pharma
LSM 1077 lymphocyte
PAA
Matrigel
BD
Normal goat serum
Vector
Paraformaldehyde
Roth
PBS
PAA
88
Material and methods
Penicillin/ Streptomycin
PAA
RPMI
PAA
Trypsin EDTA
PAA
Tween
Sigma
Xylol
Roth
Kits
ABC Vecta Stain kit
Dako
Expand long template PCR system
Roche
LC 480 RT-PCR kit
Roche
PCR clean-up kit
Sigma
Pico Pure RNA extraction
Applied bioscience
Solu Link All in one HRP conjugation kit
Solu link
WT-Ovation RNA amplification kit
Nugen
Enzymes
AmpliTaq Gold
Applied biosystems
DreamTag Green
Fermentas
MseI
Roche
Proteinase K
Roche
T4 DNA ligase
Roche
Antibodies
Name of antibody
Target
company
EpCAM-AlexaFluor 488
EpCAM
Biolegend
CD45-PE
CD45
BD
EpCAM-K/S1-4
EpCAM
Santa Cruz
CD66
CEA
BD
Moc-31
EpCAM
IQ-Products
Ber-Ep4 EpCAM-FITC
EpCAM
Dako
Mouse IgG1k Con-FITC
Isotype control
BD
EpCAM-VU-1D9
mouse IgG1-PE isotype control
EpCAM
ABR
Isotype control
BD
HEA125
EpCAM
Gift of Mollberg
Oligonucleotides
All used oligonucleotides were obtained from Invitrogen.
Target
Name
sequence
Purpose
Ascl2
ASCL2_1220F
GAGGGGAGAGGATTTTCTAAGG
RT-PCR
Ascl2
ASCL2_1220R
TTATTACGCCCCAGGTCAAG
RT-PCR
Bax
Bax
Bax-Fwd
Bax-Rev
TCTGACGGCAACTTCAACTG
GGAGGAAGTCCAATGTCCAG
RT-PCR
RT-PCR
B-Catenin
Bcat_2344F
TTCCGAATGTCTGAGGACAAG
RT-PCR
B-Catenin
Bcat_2344R
TGGGCACCAATATCAAGTCC
RT-PCR
89
Material and methods
Bmi1
Bmi1_1209Fwd
GATACTTACGATGCCCAGCAG
RT-PCR
Bmi1
Bmi-1_1209Rev
GAAGTGGACCATTCCTTCTCC
RT-PCR
Calreticulin
Calr_1040F
CCACCCAGAAATTGACAACC
RT-PCR
Calreticulin
Calr_1040R
TGTCAAAGATGGTGCCAGAC
RT-PCR
CD133
CD133
CD133_2410Fa
CD133_2410Ra
AAAGTGGCATCGTGCAAAC
CCGAATCCATTCGACGATAG
RT-PCR
RT-PCR
CD151
CD151_759Fwd
AGCAACAACTCACAGGACTGG
RT-PCR
CD151
CD151_759Rev
TGCTCCTGGATGAAGGTCTC
RT-PCR
CD166
ALCAM3650Fwd
GAACACTGCACAGCGATTTC
RT-PCR
CD166
ALCAM3650Rev
CAAACACCAGTTTTCCTTTTCC
RT-PCR
CD26
CD26Fwd_2670
AGTCAGCTCAGATCTCCAAAGC
RT-PCR
CD26
CD26Rev_2670
TGTGCTGTGCTGCTAGCTATTC
RT-PCR
CD44s
CD44s
CD44v6
CD44v6
CD45
CD45
CD44s_Fa
CD44s Ra
CD44v6_Fb2
CD44v6_Rb2
CD45_3819Fb
CD45_3819Rb
AAAGGAGCAGCACTTCAGGA
TGTGTCTTGGTCTCTGGTAGC
GTACAACGGAAGAAACAGCTACC
TGTTGTCGAATGGGAGTCTTC
TCTCTTAGAAAGTGCGGAAACAG
TCCATTCTGAGCAGGGTAGG
RT-PCR
RT-PCR
RT-PCR
RT-PCR
RT-PCR
RT-PCR
CD47
CD47_1057F
ATAGCCTATATCCTCGCTGTGG
RT-PCR
CD47
CD47-1057R
CGGAGTCCATCACTTCACTTC
RT-PCR
CEA
CEA5 Fwd_2267
CATGATTGGAGTGCTGGTTG
RT-PCR
CEA
CEA5 Rev_2267
TAGGATGGTCTCGATCTCTGG
RT-PCR
CK18
CK18
CK19
CK19
CK20
CK20
CLDN7
CLDN7
CK18_1288Fwd
CK18_1288Rev
CK19TN_Fwd563
CK19TN_Rev563
CK20 1219 Fa
CK20_1219 Ra
CLDN7_656Fa
CLDN7_656 Ra
CCCTGCTGAACATCAAGGTC
TCAGACACCACTTTGCCATC
GCGAGCTAGAGGTGAAGATCC
TGTCGATCTGCAGGACAATC
ACGCCAGAACAACGAATACC
GCCATCCACTACTTCTTGCAC
TGAGCTGCAAAATGTACGACTC
CACAAACATGGCCAGGAAG
RT-PCR
RT-PCR
RT-PCR
RT-PCR
RT-PCR
RT-PCR
RT-PCR
RT-PCR
c-Myc
C-myc_1327F
TGCTCCATGAGGAGACACC
RT-PCR
c-Myc
C-myc_1327R
GATCCAGACTCTGACCTTTTGC
RT-PCR
CO029
CO-029
CO-029_444Ra
CO-029_444Fa
ACAGCTCCTAGGATACCTGTCG
TGATTGCTGTAGGTGCCATC
RT-PCR
RT-PCR
CXCR4
CXCR4_1000F
CATCATGGTTGGCCTTATCC
RT-PCR
CXCR4
CXCR4_1000R
CGATGCTGATCCCAATGTAG
RT-PCR
Cyclin D1
CCND1Fwd_933
TCCTCTCCAGAGTGATCAAGTG
RT-PCR
Cyclin D1
CCND1Rev_933
TTGGGGTCCATGTTCTGC
RT-PCR
DLG7
DLGPA5fwd_2338
TGAAAGCAGGAGCAGCATAG
RT-PCR
DLG7
DLGPA5rev_2338
ATCTGCTACTCCACCAGCAAG
RT-PCR
E-cadherin
E-cadherin
EGFR
EGFR
E-Cad2563F
E-Cad2563R
EGFR_3517Fwd
EGFR_3517Rev
AGAGGACCAGGACTTTGACTTG
TCAGTATCAGCCGCTTTCAG
TTCTTGCAGCGATACAGCTC
TGGGAACGGACTGGTTTATG
RT-PCR
RT-PCR
RT-PCR
RT-PCR
EGR1
EGR1_577F
CAGCACCTTCAACCCTCAG
RT-PCR
EGR1
EGR1_577R
AGCGGCCAGTATAGGTGATG
RT-PCR
EpCAM
EpCAM
EpCAM_733Fa
EpCAM_733Ra
CTGGATCCAAAATTTATCACGAG
GTTCCCCATTTACTGTCAGGTC
RT-PCR
RT-PCR
HDAC1
HDAC1_1042Fwd
TTAACCTGCCTATGCTGATGC
RT-PCR
90
Material and methods
HDAC1
HDAC1-1042Rev
GAAGGACTGATGTGGAGCTTG
RT-PCR
HDAC3
HDAC3_1283Fwd
ACAGGACTGATGAGGCTGATG
RT-PCR
HDAC3
HDAC3_1283Rev
CCCAACCAAGAGGTGAAAAG
RT-PCR
hFN
hFN
Ki67
Ki67
Lgr5
Lgr5
hFN1
hFN1
Ki67_8625Fwd
Ki67_8625Rev
LGR5_1685F
LGR5_1685R
CAGTGGGAGACCTCGAGAAG
TCCCTCGGAACATCAGAAAC
ACAAAAGGTGCTTGAGGTCTG
CCTTTCCCTTTCTGATTCTGC
TCATTCAGTGCAGTGTTCACC
CTGCGATGACCCCAATTAAC
RT-PCR
RT-PCR
RT-PCR
RT-PCR
RT-PCR
RT-PCR
Maspin
SerpinB5Fwd_877
ACACCAAACCAGTGCAGATG
RT-PCR
Maspin
SerpinB5Rev_877
CTGTGACAGTGACTCTGAGTTGAG
RT-PCR
MMP7
MMP7Fwd_822
TTGGGTATGGGACATTCCTC
RT-PCR
MMP7
MMP7Rev_822
GAATGGATGTTCTGCCTGAAG
RT-PCR
MRP2
MRP2_4453Fwd
AGCCTGCAACTTGGGTTATC
RT-PCR
MRP2
MRP2_4453Rev
TGGTCGTCTGAATGAGGTTG
RT-PCR
MRP4
MRP2_2655Fwd
CTGCCGCTGACGTTTTTAG
RT-PCR
MRP4
MRP2_2655Fwd
AATCCCAGCACTGAGGTTTG
RT-PCR
MRP5
MRP2_453Fwd
AGATGCCTTGGAAACAGCAG
RT-PCR
MRP5
MRP2_453Fwd
AAAAAGCCCAGCATTGTCC
RT-PCR
Musashi
Msi1Fwd_1174
CTTTGATTGCCACAGCCTTC
RT-PCR
Musashi
MsiRev_1174
GCTGGCTCACTCGTGGTC
RT-PCR
N-cadherin
N-Cad2973Fneu
GAATGGATGAAAGACCCATCC
RT-PCR
N-cadherin
Olfactomedin 4
Olfactomedin 4
N-Cad2973R
Olfm4_730Fa
Olfm4_730Ra
AGTCATATGGTGGAGCTGTGG
AGATCAAAACACCCCTGTCG
GAAAACCCTCTCCAGTTGAGC
RT-PCR
RT-PCR
RT-PCR
P21
p21Fwd_571
ATGTGGACCTGTCACTGTCTTG
RT-PCR
P21
p21Rev_571
GGATTAGGGCTTCCTCTTGG
RT-PCR
Snail1
SnailR_680
TCTTGACATCTGAGTGGGTCTG
RT-PCR
Snail1
SnailFb_680
TCTAGGCCCTGGCTGCTAC
RT-PCR
Survivin
Birc5Fwd_232
TTTTCATCGTCGTCCCTAGC
RT-PCR
Survivin
Birc5Rev_232
AGCCCGGATGATACAAACAG
RT-PCR
TP53
p53Fwd_1297
TGAATGAGGCCTTGGAACTC
RT-PCR
TP53
P53Rev_1297
TTTTATGGCGGGAGGTAGAC
RT-PCR
Twist1
Twist1
Vimentin
Vimentin
Twi1100Fb
Twi1100R
Vim1483F
Vim1483R
GGGCCGGAGACCTAGATG
CCACGCCCTGTTTCTTTG
TTTTCCTCCCTGAACCTGAG
CGTGATGCTGAGAAGTTTCG
RT-PCR
RT-PCR
RT-PCR
RT-PCR
ZEB1
ZEB1_2664Fwd
CTAGCTGCCAATAAGCAAACG
RT-PCR
ZEB1
ZEB1_2664Rev
TTGGGCGGTGTAGAATCAG
RT-PCR
Β-Actin
Β-Actin
b-Act_1404Fa
b-Act_1510Ra
ATGTGGCCGAGGACTTTGATT
AGTGGGGTGGCTTTTAGGATG
RT-PCR
RT-PCR
91
Material and methods
Preamplification
P53E5/6
P53E5/6
P53E7
P53E7
P53E8
P53E8
BRAF
BRAF
Kras
Kras
P53E5/6
P53E5/6
P53E7
P53E7
P53E8
P53E8
E5/6Fwdx
E5/6Revx
E789seqP53Fwd
E7 Revx
E8Fwdx
E8Revx
BRAF_Fwd
BRAF-Rev
KrasFwd
KrasRev
P53Exon5/6Fwd
P53 Exon5/6Rev
P53E7Fwd
P53E7Rev
P53E8_Fwd
P53E8 _Rev
TTCACTTGTGCCCTGACTTTCAAC
GCCACTGACAACCACCCTTAAC
CCTCATCTTGGGCCTGTGTTATC
TGGAAGAAATCGGTAAGAGGTGG
GGACCTGATTTCCTTACTGCCTC
CTGAGGCATAACTGCACCCTTG
GTGGGATTCCTGCTGTCAGTTAAAG
AGCCTCAATTCTTACCATCCAC
TCCTGCTGTCAGTTAACCTTATG
CTGTCAGTTAAAACAAGATTTACC
GCTGTCAGTTAATGTGTGATCTC
GTCAGTTAACCCCTCCTCCCAG
GCAGTGGCTCATGCCTGTAATC
GAGTGGGAGCAGTAAGGAGA
GGGATTCCTGCTGTCAGTTAAATGG
CATTGTCTTTGAGGCATC
Tissue
Tissue
Tissue
Tissue
Tissue
Tissue
CTC
CTC
CTC
CTC
CTC
CTC
CTC
CTC
CTC
CTC
Name
BRAF fwd
BRAF rev
Kras Fwd
Kras Rev1
p53E5/E6Fwd
p53 rev1
p53E7rev
p53E8-Fwd
p53E8-Rev
E5p53gDNAFwd
E5p53gDNARev
E7p53gDNAFwd
E7p53gDNARev
E8p53gDNaFwd
E8p53gDNARev
KrasFwd
KrasRev
sequence
CTCTTCATAATGCTTGCTC
CCATCCACAAAATGGATCC
TCCTGCTGTCAGTTAACC
ACCTCTATTGTTGGATC
GTGATCTCTGACTCCTGTC
AGAGACCCCAGTTGCAAAC
GGGAGCAGTAAGGAGATTC
AAATGGGACAGGTAGGAC
CATTGTCTTTGAGGCATC
TTCACTTGTGCCCTGAC
ACTGACAACCACCCTTAAC
ATCTTGGGCCTGTGTTATC
TGGAAGAAATCGGTAAGAG
GGACCTGATTTCCTTACTG
CTGAGGCATAACTGCAC
AAGGCCTGCTGAAAATGACTG
AGAATGGTCCTGCACCAGTAA
Purpose
CTC + Tissue
CTC + Tissue
CTC
CTC
CTC
CTC
CTC
CTC
CTC
Tissue
Tissue
Tissue
Tissue
Tissue
Tissue
Tissue
Tissue
Name
NR-21_fwd
NR-21_rev
NR-24_fwd
NR-24_rev
Bat25F
Bat25R
sequence
TAAATGTATGTCTCCCCTGG
ATTCCTACTCCGCATTCACAA
CCATTGCTGAATTTTACCTC
ATTGTGCCATTGCATTCCAAA
TCGCCTCCAAGAATGTAAGT
TATGGCTCTAAAATGCTCTGTTC
Purpose
MSI analysis
MSI analysis
MSI analysis
MSI analysis
MSI analysis
MSI analysis
Sequencing
Target
BRAF
BRAF
Kras
Kras
P53E5/6
P53E5/6
P53 E7
P53 E8
P53 E8
P53E5
P53E5
P53E7
P53E7
P53E8
P53E8
Kras
Kras
MSI Analysis
Target
NR21
NR21
NR24
NR24
BAT25
BAT25
92
Material and methods
Multiplex Primer quality control CTC amplification
LAMC1For
TCTGCTTTGGGCATTCTTCT
LAMC1Rev
TTCTAACAGGTTGGGGGATG
CADPSFor
CCCCACCCTTCTTCACTACA
CADPSRev
GTGTGCACATACCACCGAAG
GRIK5For
CTAGCTCCCACCAACCTCAG
GRIKRev
CTCGATGATCCCGTTGATCT
NEK9For
GCAGGAGGGAACCTGTATGA
NEK9Rev
CAGGAAAGAAAGCCCACAGA
PICK1For
TCGTATGCTGGAGTCCTGTG
PICK1Rev
GGGATGGCTTTGTTGAGGTA
DNAH9For
GGGTCTCATCACCAGCATTT
DNAH9Rev
GCCATCTTCCACATGGTCTT
Cell lines
Cell lines were obtained from DSMZ and their identity is regularly confirmed.
- HCT116
- HT29
- SW480
- DLD1
- Colo205
93
Appendix
Appendix
List of Figures
Figure 1: UICC staging of tumor progression. ............................................................... 9
Figure 2: The Wnt signaling pathway. .......................................................................... 10
Figure 3: Multistep genetic model of CRC development. ............................................ 12
Figure 4: EMT and tumor cell dissemination ............................................................... 14
Figure 5: Early or late dissemination of cancer cells .................................................... 18
Figure 6: The CellSearch. ............................................................................................. 20
Figure 7: CTC count and prognosis in mCRC .............................................................. 22
Figure 8: Heterogeneity of DTCs ................................................................................. 23
Figure 9: CTC amount in central venous blood compartment (CVBC) and mesenteric
venous blood compartment (MVBC) of CRC patients. ................................................. 28
Figure 10: Surface antigen expression allows identification of CTCs .......................... 33
Figure 11: The micromanipulator .................................................................................. 35
Figure 12: Surface antigen expression enables identification of DTCs ......................... 36
Figure 13: Validation of the cDNA amplification strategy. .......................................... 37
Figure 14: Human CTCs ................................................................................................ 38
Figure 15: mRNA expression of CTCs.......................................................................... 41
Figure 16: Compared mRNA expression of CTCs and matched tumor (A) or liver
metastasis (B) ................................................................................................................. 43
Figure 17: mRNA expression of DTCs. ........................................................................ 45
Figure 18: Compared mRNA expression of DTCs and matched cancer tissue. ............ 45
Figure 19: Mutation analysis. ........................................................................................ 46
Figure 20: MSI in CTCs ................................................................................................ 48
Figure 21: Validation experiment. ................................................................................. 51
Figure 22: Array-CGH profiles of a liver vein - derived CTC and liver metastasis tissue
of patient HD 2095......................................................................................................... 54
Figure 23: Array-CGH profiles of a CTC and liver metastasis of patient HD 2288. .... 55
Figure 24: Accumulative penetrance plots of array-CGH data. .................................... 56
Figure 25: The orthotopic mouse model of CRC metastasis ......................................... 58
Figure 26: Histological staining. .................................................................................... 58
Figure 27: EpCAM expression of HCT116 after xenotransplantation. ......................... 59
Figure 28: Tumor development. .................................................................................... 60
Figure 29: Colony formation of CTCs isolated from murine blood .............................. 61
Figure 30: Fluorescence staining of AS-CTCs. ............................................................. 61
Figure 31: Fluorescence staining of CTCs..................................................................... 62
Figure 32: Comparison of mRNA expression. .............................................................. 65
Figure 33: Expression of epithelial markers in CTCs. ................................................... 66
Figure 34: mRNA expression of Survivin and Ki67. .................................................... 67
Figure 35: mRNA expression of CD44s and EGFR ...................................................... 68
Supplementary figure 1: Genomic analyses of CTCs……………………………….....96
Supplementary figure 2: Patients characteristics…………………………………..…..98
Supplementary figure 3: Array-CGH profiles of CTCs and cancer tissue samples…...99
94
Appendix
List of tables
Tabel 1: Clinicopathological characteristics of the study population. ........................... 28
Table 2: Clinicopathologic correlations. ........................................................................ 29
Table 3: Detection of CTCs in mesenteric and central venous blood compartments and
correlation to clinical staging. ........................................................................................ 31
Table 4: Patients characteristics. .................................................................................... 39
Table 5: Overview about detected point mutations in CTCs. ........................................ 47
Table 6: cMS analysis of MSI+-CTCs. .......................................................................... 49
Table 7: Genomic inter-CTC heterogeneity of CTCs from patient HD 2215. .............. 50
Table 8: Genomic inter-CTC heterogeneity of CTCs from patient HD 2328. .............. 50
List of abbreviations
AS-CTC
ATP
BMDC
CGH
CRC
CTC
CVBC
DNA
DTC
gDNA
FACS
mRNA
MS
MSI
MSS
MVBC
NCI
NSCLC
PBMC
PCR
ptci
RNA
Wt
Ascites-asscociated CTC
Adenosine triphosphate
Bone marrow derived cells
Comparative genomic hybridization
Colorectal cancer
Circulating tumor cell
Central venous blood compartment
Desoxyribonucleic acid
Disseminated tumor cell
genomic DNA
Fluorescence activated cell sorting
Messenger RNA
Microsatellite
Microsatellite instability
Microsatellite stable
Mesenterial venous blood compartment
National cancer institute
Non small cell lung cancer
Peripheral blood mononucleated cells
Polymerase chain reaction
post tumor cell injection
Ribonucleic acid
Wild type
95
Appendix
Supplementary figure 1: Genomic analyses of CTCs. MSS= microsatellite stable; MSI=microsatellite instablity; wt =wild type; mutations are indicated in red; number in
brackets indicate the number of analyzed cells
MSS / MSI
Kras
Braf
TP53 Exon 5/6
TP53 Exon 7
TP53 Exon8
Patient ID
CTC
tissue
CTC
tissue
CTC
tissue
CTC
tissue
CTC
tissue
CTC
tissue
HD1960
MSS (1)
MSS
X
wt
wt (1)
wt
X
wt
X
wt
X
wt
HD2026
MSS (1)
MSS
X
wt
wt (1)
wt
wt (1)
wt
wt (1)
wt
X
FS_C184
HD2030
MSS (5)
X
wt (8)
wt
wt (4)
wt
wt (5)
wt
wt (7)
wt
wt (6)
wt
HD2050
MSS (1)
X
wt (1)
X
wt (1)
X
wt (1)
X
wt (1)
X
wt (1)
X
HD2054
MSS (3)
MSS
wt (2)
wt
V600E
wt (2)
wt
wt (2)
wt
wt (2)
wt
HD2068
MSS (2)
MSS
wt (1)
wt
X
wt
R175H (2)
R175H
wt (1)
wt
wt (1)
wt
HD2077
MSS (1)
MSS
wt (1)
wt
X
wt
wt (1)
wt
wt (1)
wt
wt (1)
wt
HD2078
MSS (1)
MSS
wt (1)
wt
wt (1)
wt
wt (1)
wt
X
X
wt (1)
X
HD2091
MSS (1)
MSS
wt (1)
wt
X
wt
wt (1)
wt
wt (1)
wt
R273C (1)
R273C
HD2094
MSS (1)
MSS
X
wt
Wt
wt
wt (1)
X
wt (1)
X
wt (1)
X
HD2095
MSS (13)
X
wt (8)
G12D
wt (1)
wt
wt (9)
wt
wt (9)
wt
wt (7)
wt
HD2101
MSS (1)
MSS
wt (1)
G12D
X
wt
wt (1)
wt
wt (1)
wt
wt (1)
wt
HD2113
MSS (4)
MSS
wt (1)
wt
wt (2)
wt
wt (2)
wt
wt (4)
wt
R282W (4)
R282W
HD2126
MSS (1)
MSS
X
G12D
wt (1)
wt
A159V (1)
A159V
wt (1)
wt
wt (1)
wt
HD2131
MSS (1)
MSS
wt (1)
G12V
wt (1)
wt
X
wt
wt (1)
wt
wt (1)
wt
HD2165
MSS (14)
MSS
wt (8L; 5P)
wt
wt (8)
wt
wt (13)
wt
wt (14)
wt
wt (14)
wt
HD2179
MSS (1)
MSS
wt (1)
wt
wt (2)
V600E
R175H (1)
R175H
X
wt
wt (1)
wt
HD2201
MSS (1)
MSS
G12C (1)
G12C
X
wt
R175H (1)
R175H
X
wt
wt (2)
wt
HD2202
MSS (1)
MSS
G13D (1)
wt
wt (1)
wt
wt (1)
wt
wt (1)
wt
wt (1)
wt
HD2203
MSS (3)
X
wt (1)
X
wt (3)
wt
wt (2)
wt
wt (2)
wt
V600E(1)/
wt(1)
EE285,
286DA(3)
wt
96
Appendix
HD2215
MSS(9)/
MSI(2)
MSS
G12S(4)/
wt (2)
G12S
wt (6)
wt
wt (2)
wt
wt (4)
wt
wt (5)
wt
HD2224
MSS (7)
MSS
wt (7)
G12R
wt (6)
wt
wt (6)
wt
wt (5)
wt
wt (7)
wt
HD2231
MSS (4)
MSS
wt (4)
wt
wt (4)
X
wt (4)
wt
G245S (3)
G245S
wt (4)
wt
HD2246
MSS (3)
X
wt (3)
xx
wt (3)
X
wt (3)
X
wt (3)
X
wt (2)
X
HD2257
X
MSS
X
wt
wt (1)
X
X
wt
wt (1)
R248Q
wt (1)
wt
HD2288
MSS (7)
MSS
G12A
wt (5)
wt
wt (5)
wt
wt (6)
wt
HD2295
MSS (7)
MSS
wt
wt (6)
V600E
wt (7)
wt
wt (7)
wt
wt (8)
wt
wt (10)
wt
wt (11)
wt
G12A(4)/
wt(2)
wt (7)
F134C(3L)+
HD2328
MSS (14)
MSS
wt (12)
wt
wt (9)
wt
wt(3LV,4M
es)
HD2334
HD2341
HD2351
MSS (8)
MSI(2)
/MSS(1)
MSS (1)
MSS
X
MSS
G12S(7)/
wt(1)
G13D(2)/
wt(1)
wt(1)
F134C_LM;
wt_Tu?
R273C(4)/
wt(1)
R273C
G12S
wt (8)
X
wt (7)
X
wt (7)
wt
wt (8)
wt
X
wt (2)
X
wt (3)
X
wt (3)
X
wt (3)
X
X
wt (1)
X
wt (1)
X
wt (1)
wt
wt (1)
97
Appendix
Supplementary figure 2: Patients characteristics. Patients from whom CTCs and cancer tissue
specimens were obtained for genomic analyses are listed. f=female, m=male, Tu=tumor, LM=liver
metastasis, LuMet=lung metastasis, LN=lymph node metastasis, PB peripheral blood, LV= liver vein,
PV=portal vein
Detected genomic
disparity between
CTCs and cancer
tissue
Patient ID
Sex
Age
UICC Stage
Analyzed tissue
CTC obtained
from
HD1960
f
81
IV
Tu
PB
HD2026
M
66
IV
LM
PB
HD2030
m
54
IV
Tu
LV
HD2050
M
53
I
-
PV
HD2054
m
53
III
Tu
PB
HD2068
m
36
IV
LM+LuMet
LV
HD2077
f
72
I
Tu1+Tu2
LV
HD2078
m
68
IV
Tu
LV
HD2091
M
53
IV
Tu
LV
HD2094
f
74
IV
LM
LV
HD2095
f
69
IV
Tu
LV
x
HD2101
f
85
III
Tu
LV
x
HD2113
f
48
IV
Tu
LV
HD2126
f
64
IV
Tu
LV
HD2131
f
61
IV
Tu
PB
HD2165
m
50
III
LM
LV+PV
HD2179
f
60
IV
LM
PB
HD2201
m
65
IV
LM
LV+PV
HD2202
m
56
III
LN-Met
PV
HD2203
f
38
IV
-
PB
HD2215
f
85
IV
LM
LV
HD2224
m
62
III
LN-Met
LV
HD2231
m
63
IV
LM
LV
HD2246
m
64
I
-
LV
HD2257
f
55
IV
LM
PB
HD2288
F
59
IV
LM
LV
HD2295
f
80
III
Tu
LV
HD2328
f
64
IV
Tu
LV+PV
HD2334
m
74
III
LN-Met
PV
HD2341
m
51
IV
-
PV
HD2351
w
83
II
Tu
PV
x
x
x
x
x
x
98
Appendix
Supplementary figure 3: Array-CGH profiles of CTCs and cancer tissue samples. The human set of
chromosomes is shown (1-22, X, Y). Beside each chromosome the detected signal intensity of the arrayCGH analysis is depicted. Amplitudes to the left reflect chromosomal losses. Amplitudes to the right
reflect chromosomal gains. CTCs were obtained from blood samples as indicated: PB=peripheral or
central blood, LV=liver vein, PV=portal vein.
Matrix-CGH profile HD2026 CTC (PB)
Matrix-CGH profile HD2030 CTC (LV)
99
Appendix
Matrix-CGH profile HD2095 CTC (LV)
Matrix-CGH profile HD 2126 CTC (LV)
100
Appendix
Matrix-CGH profile HD 2224 CTC (LV)
Matrix-CGH profile HD2231 CTC (LV)
101
Appendix
Matrix-CGH profile HD2288 CTC (LV)
Matrix-CGH profile HD2295 CTC (LV)
102
Appendix
Matrix-CGH profile HD2328 CTC (LV)
Matrix-CGH profile HD2334 CTC (PV)
103
Appendix
Matrix CGH profile of HD2026 liver metastasis tissue
Matrix CGH profile of HD2095 liver metastasis tissue
104
Appendix
Matrix CGH profile of HD2126 liver metastasis tissue
Matrix CGH profile of HD2224 tumor tissue
105
Appendix
Matrix CGH profile of HD2231 liver metastasis tissue
Matrix CGH profile of HD2288 liver metastasis tissue
106
Appendix
Matrix CGH profile of HD2328 liver metastasis tissue
Matrix CGH profile of HD2334 tumor tissue
107
Literature
Literature
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
13.
14.
15.
16.
17.
18.
19.
20.
Weitz, J. et al. Colorectal cancer. Lancet 365, 153-165 (2005).
Boyle, P. & Ferlay, J. Cancer incidence and mortality in Europe, 2004. Ann
Oncol 16, 481-488 (2005).
Benson, A.B., 3rd Epidemiology, disease progression, and economic burden of
colorectal cancer. J Manag Care Pharm 13, S5-18 (2007).
Benson, A.B., 3rd et al. American Society of Clinical Oncology
recommendations on adjuvant chemotherapy for stage II colon cancer. J Clin
Oncol 22, 3408-3419 (2004).
Chua, T.C., Liauw, W., Chu, F. & Morris, D.L. Viewing metastatic colorectal
cancer as a curable chronic disease. Am J Clin Oncol 35, 77-80 (2012).
Edge, S.B. & Compton, C.C. The American Joint Committee on Cancer: the 7th
edition of the AJCC cancer staging manual and the future of TNM. Ann Surg
Oncol 17, 1471-1474 (2010).
Saif, M.W. & Chu, E. Biology of colorectal cancer. Cancer J 16, 196-201
(2010).
Migliore, L., Migheli, F., Spisni, R. & Coppede, F. Genetics, cytogenetics, and
epigenetics of colorectal cancer. J Biomed Biotechnol 2011, 792362 (2011).
Sparks, A.B., Morin, P.J., Vogelstein, B. & Kinzler, K.W. Mutational analysis
of the APC/beta-catenin/Tcf pathway in colorectal cancer. Cancer Res 58,
1130-1134 (1998).
Fearon, E.R. Molecular genetics of colorectal cancer. Annu Rev Pathol 6, 479507 (2011).
Pino, M.S. & Chung, D.C. The chromosomal instability pathway in colon
cancer. Gastroenterology 138, 2059-2072 (2010).
Boland, C.R. & Goel, A. Microsatellite instability in colorectal cancer.
Gastroenterology 138, 2073-2087 e2073 (2010).
Markowitz, S. et al. Inactivation of the type II TGF-beta receptor in colon
cancer cells with microsatellite instability. Science 268, 1336-1338 (1995).
Massague, J., Blain, S.W. & Lo, R.S. TGFbeta signaling in growth control,
cancer, and heritable disorders. Cell 103, 295-309 (2000).
Markowitz, S.D. & Bertagnolli, M.M. Molecular origins of cancer: Molecular
basis of colorectal cancer. N Engl J Med 361, 2449-2460 (2009).
Roth, A.D. et al. Prognostic role of KRAS and BRAF in stage II and III resected
colon cancer: results of the translational study on the PETACC-3, EORTC
40993, SAKK 60-00 trial. J Clin Oncol 28, 466-474 (2010).
Baker, S.J. et al. p53 gene mutations occur in combination with 17p allelic
deletions as late events in colorectal tumorigenesis. Cancer Res 50, 7717-7722
(1990).
Greenblatt, M.S., Bennett, W.P., Hollstein, M. & Harris, C.C. Mutations in the
p53 tumor suppressor gene: clues to cancer etiology and molecular
pathogenesis. Cancer Res 54, 4855-4878 (1994).
Samuels, Y. & Velculescu, V.E. Oncogenic mutations of PIK3CA in human
cancers. Cell Cycle 3, 1221-1224 (2004).
Michl, P. & Downward, J. Mechanisms of disease: PI3K/AKT signaling in
gastrointestinal cancers. Z Gastroenterol 43, 1133-1139 (2005).
108
Literature
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
35.
36.
37.
38.
39.
40.
Li, R. et al. Prognostic value of Akt-1 in human prostate cancer: a computerized
quantitative assessment with quantum dot technology. Clin Cancer Res 15,
3568-3573 (2009).
Tomita, Y. et al. Prognostic significance of activated AKT expression in softtissue sarcoma. Clin Cancer Res 12, 3070-3077 (2006).
Dai, D.L., Martinka, M. & Li, G. Prognostic significance of activated Akt
expression in melanoma: a clinicopathologic study of 292 cases. J Clin Oncol
23, 1473-1482 (2005).
Baba, Y. et al. Phosphorylated AKT expression is associated with PIK3CA
mutation, low stage, and favorable outcome in 717 colorectal cancers. Cancer
117, 1399-1408 (2011).
Liu, M. & Chen, H. The role of microRNAs in colorectal cancer. J Genet
Genomics 37, 347-358 (2010).
Christiansen, J.J. & Rajasekaran, A.K. Reassessing epithelial to mesenchymal
transition as a prerequisite for carcinoma invasion and metastasis. Cancer Res
66, 8319-8326 (2006).
Tarin, D., Thompson, E.W. & Newgreen, D.F. The fallacy of epithelial
mesenchymal transition in neoplasia. Cancer Res 65, 5996-6000; discussion
6000-5991 (2005).
Polyak, K. & Weinberg, R.A. Transitions between epithelial and mesenchymal
states: acquisition of malignant and stem cell traits. Nat Rev Cancer 9, 265-273
(2009).
Huber, M.A., Kraut, N. & Beug, H. Molecular requirements for epithelialmesenchymal transition during tumor progression. Curr Opin Cell Biol 17, 548558 (2005).
Boyer, B., Valles, A.M. & Edme, N. Induction and regulation of epithelialmesenchymal transitions. Biochem Pharmacol 60, 1091-1099 (2000).
Comijn, J. et al. The two-handed E box binding zinc finger protein SIP1
downregulates E-cadherin and induces invasion. Mol Cell 7, 1267-1278 (2001).
Batlle, E. et al. The transcription factor snail is a repressor of E-cadherin gene
expression in epithelial tumour cells. Nat Cell Biol 2, 84-89 (2000).
Birchmeier, W. & Behrens, J. Cadherin expression in carcinomas: role in the
formation of cell junctions and the prevention of invasiveness. Biochim Biophys
Acta 1198, 11-26 (1994).
Bohm, M., Totzeck, B., Birchmeier, W. & Wieland, I. Differences of Ecadherin expression levels and patterns in primary and metastatic human lung
cancer. Clin Exp Metastasis 12, 55-62 (1994).
Thiery, J.P. Epithelial-mesenchymal transitions in tumour progression. Nat Rev
Cancer 2, 442-454 (2002).
Peinado, H., Olmeda, D. & Cano, A. Snail, Zeb and bHLH factors in tumour
progression: an alliance against the epithelial phenotype? Nat Rev Cancer 7,
415-428 (2007).
Wu, Y. & Zhou, B.P. Snail: More than EMT. Cell Adh Migr 4, 199-203 (2010).
Alves, C.C., Carneiro, F., Hoefler, H. & Becker, K.F. Role of the epithelialmesenchymal transition regulator Slug in primary human cancers. Front Biosci
14, 3035-3050 (2009).
Micalizzi, D.S., Farabaugh, S.M. & Ford, H.L. Epithelial-mesenchymal
transition in cancer: parallels between normal development and tumor
progression. J Mammary Gland Biol Neoplasia 15, 117-134 (2010).
Yang, J. & Weinberg, R.A. Epithelial-mesenchymal transition: at the crossroads
of development and tumor metastasis. Dev Cell 14, 818-829 (2008).
109
Literature
41.
42.
43.
44.
45.
46.
47.
48.
49.
50.
51.
52.
53.
54.
55.
56.
57.
58.
59.
60.
61.
62.
Tarin, D. et al. Mechanisms of human tumor metastasis studied in patients with
peritoneovenous shunts. Cancer Res 44, 3584-3592 (1984).
Luzzi, K.J. et al. Multistep nature of metastatic inefficiency: dormancy of
solitary cells after successful extravasation and limited survival of early
micrometastases. Am J Pathol 153, 865-873 (1998).
Brooks, D.E. The biorheology of tumor cells. Biorheology 21, 85-91 (1984).
Taddei, M.L., Giannoni, E., Fiaschi, T. & Chiarugi, P. Anoikis: an emerging
hallmark in health and diseases. J Pathol 226, 380-393 (2012).
Hou, J.M. et al. Circulating tumor cells as a window on metastasis biology in
lung cancer. Am J Pathol 178, 989-996 (2011).
Andratschke, M., Pauli, C., Stein, M., Chaubal, S. & Wollenberg, B. MHCclass I antigen expression on micrometastases in bone marrow of patients with
head and neck squamous cell cancer. Anticancer Res 23, 1467-1471 (2003).
Nieswandt, B., Hafner, M., Echtenacher, B. & Mannel, D.N. Lysis of tumor
cells by natural killer cells in mice is impeded by platelets. Cancer Res 59,
1295-1300 (1999).
Placke, T. et al. Platelet-derived MHC class I confers a pseudonormal
phenotype to cancer cells that subverts the antitumor reactivity of natural killer
immune cells. Cancer Res 72, 440-448 (2012).
Pawelek, J.M. & Chakraborty, A.K. The cancer cell--leukocyte fusion theory of
metastasis. Adv Cancer Res 101, 397-444 (2008).
Colombano, S.P. & Reese, P.A. The cascade theory of metastatic spread: are
there generalizing sites? Cancer 46, 2312-2314 (1980).
Makrilia, N., Kollias, A., Manolopoulos, L. & Syrigos, K. Cell adhesion
molecules: role and clinical significance in cancer. Cancer Invest 27, 1023-1037
(2009).
Muller, A. et al. Involvement of chemokine receptors in breast cancer
metastasis. Nature 410, 50-56 (2001).
Sastre, J. et al. Circulating tumor cells in colorectal cancer: correlation with
clinical and pathological variables. Ann Oncol 19, 935-938 (2008).
Fidler, I.J. The pathogenesis of cancer metastasis: the 'seed and soil' hypothesis
revisited. Nat Rev Cancer 3, 453-458 (2003).
Psaila, B., Kaplan, R.N., Port, E.R. & Lyden, D. Priming the 'soil' for breast
cancer metastasis: the pre-metastatic niche. Breast Dis 26, 65-74 (2006).
Kaplan, R.N. et al. VEGFR1-positive haematopoietic bone marrow progenitors
initiate the pre-metastatic niche. Nature 438, 820-827 (2005).
Tsuji, T. et al. Epithelial-mesenchymal transition induced by growth suppressor
p12CDK2-AP1 promotes tumor cell local invasion but suppresses distant
colony growth. Cancer Res 68, 10377-10386 (2008).
Galandiuk, S. et al. Patterns of recurrence after curative resection of carcinoma
of the colon and rectum. Surg Gynecol Obstet 174, 27-32 (1992).
Li, J. et al. Tumor heterogeneity in neoplasms of breast, colon, and skin. BMC
Res Notes 3, 321 (2010).
Grady, W.M. & Markowitz, S. Genomic instability and colorectal cancer. Curr
Opin Gastroenterol 16, 62-67 (2000).
Visvader, J.E. & Lindeman, G.J. Cancer stem cells in solid tumours:
accumulating evidence and unresolved questions. Nat Rev Cancer 8, 755-768
(2008).
Pang, R. et al. A subpopulation of CD26+ cancer stem cells with metastatic
capacity in human colorectal cancer. Cell Stem Cell 6, 603-615 (2010).
110
Literature
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
75.
76.
77.
78.
79.
80.
81.
82.
Ricci-Vitiani, L. et al. Identification and expansion of human colon-cancerinitiating cells. Nature 445, 111-115 (2007).
Kelly, P.N., Dakic, A., Adams, J.M., Nutt, S.L. & Strasser, A. Tumor growth
need not be driven by rare cancer stem cells. Science 317, 337 (2007).
Hill, R.P. Identifying cancer stem cells in solid tumors: case not proven. Cancer
Res 66, 1891-1895; discussion 1890 (2006).
Quintana, E. et al. Efficient tumour formation by single human melanoma cells.
Nature 456, 593-598 (2008).
Brabletz, T., Jung, A., Spaderna, S., Hlubek, F. & Kirchner, T. Opinion:
migrating cancer stem cells - an integrated concept of malignant tumour
progression. Nat Rev Cancer 5, 744-749 (2005).
Ovaska, J., Jarvinen, H., Kujari, H., Perttila, I. & Mecklin, J.P. Follow-up of
patients operated on for colorectal carcinoma. Am J Surg 159, 593-596 (1990).
Wolpin, B.M. & Mayer, R.J. Systemic treatment of colorectal cancer.
Gastroenterology 134, 1296-1310 (2008).
Klein, C.A. Parallel progression of primary tumours and metastases. Nat Rev
Cancer 9, 302-312 (2009).
Lianidou, E.S. & Markou, A. Circulating Tumor Cells in Breast Cancer:
Detection Systems, Molecular Characterization, and Future Challenges. Clin
Chem (2011).
Yu, M., Stott, S., Toner, M., Maheswaran, S. & Haber, D.A. Circulating tumor
cells: approaches to isolation and characterization. J Cell Biol 192, 373-382
(2011).
Cristofanilli, M. et al. Circulating tumor cells, disease progression, and survival
in metastatic breast cancer. N Engl J Med 351, 781-791 (2004).
Sequist, L.V., Nagrath, S., Toner, M., Haber, D.A. & Lynch, T.J. The CTCchip: an exciting new tool to detect circulating tumor cells in lung cancer
patients. J Thorac Oncol 4, 281-283 (2009).
Talasaz, A.H. et al. Isolating highly enriched populations of circulating
epithelial cells and other rare cells from blood using a magnetic sweeper device.
Proc Natl Acad Sci U S A 106, 3970-3975 (2009).
Vona, G. et al. Isolation by size of epithelial tumor cells : a new method for the
immunomorphological and molecular characterization of circulatingtumor cells.
Am J Pathol 156, 57-63 (2000).
Marrinucci, D. et al. Case study of the morphologic variation of circulating
tumor cells. Hum Pathol 38, 514-519 (2007).
Pantel, K. et al. Circulating Epithelial Cells in Patients with Benign Colon
Diseases. Clin Chem (2011).
Schmitt, M. & Foekens, J.A. Circulating tumor cells in blood of primary breast
cancer patients assessed by a novel RT-PCR test kit and comparison with status
of bone marrow-disseminated tumor cells. Breast Cancer Res 11, 109 (2009).
Shimomura, M. et al. Adequate lymph node examination is essential to ensure
the prognostic value of the lymph node ratio in patients with stage III colorectal
cancer. Surg Today 41, 1370-1379 (2011).
Weitz, J. et al. Detection of disseminated colorectal cancer cells in lymph
nodes, blood and bone marrow. Clin Cancer Res 5, 1830-1836 (1999).
Wolfrum, F., Vogel, I., Fandrich, F. & Kalthoff, H. Detection and clinical
implications of minimal residual disease in gastro-intestinal cancer.
Langenbecks Arch Surg 390, 430-441 (2005).
111
Literature
83.
84.
85.
86.
87.
88.
89.
90.
91.
92.
93.
94.
95.
96.
97.
98.
99.
100.
Nicastri, D.G., Doucette, J.T., Godfrey, T.E. & Hughes, S.J. Is occult lymph
node disease in colorectal cancer patients clinically significant? A review of the
relevant literature. J Mol Diagn 9, 563-571 (2007).
Rahbari, N.N. et al. Meta-analysis shows that detection of circulating tumor
cells indicates poor prognosis in patients with colorectal cancer.
Gastroenterology 138, 1714-1726 (2010).
Allard, W.J. et al. Tumor cells circulate in the peripheral blood of all major
carcinomas but not in healthy subjects or patients with nonmalignant diseases.
Clin Cancer Res 10, 6897-6904 (2004).
Cohen, S.J. et al. Prognostic significance of circulating tumor cells in patients
with metastatic colorectal cancer. Ann Oncol 20, 1223-1229 (2009).
Miller, M.C., Doyle, G.V. & Terstappen, L.W. Significance of Circulating
Tumor Cells Detected by the CellSearch System in Patients with Metastatic
Breast Colorectal and Prostate Cancer. J Oncol 2010, 617421 (2010).
Chambers, A.F., Groom, A.C. & MacDonald, I.C. Dissemination and growth of
cancer cells in metastatic sites. Nat Rev Cancer 2, 563-572 (2002).
Peach, G., Kim, C., Zacharakis, E., Purkayastha, S. & Ziprin, P. Prognostic
significance of circulating tumour cells following surgical resection of
colorectal cancers: a systematic review. Br J Cancer 102, 1327-1334 (2010).
Riethdorf, S., Wikman, H. & Pantel, K. Review: Biological relevance of
disseminated tumor cells in cancer patients. Int J Cancer 123, 1991-2006
(2008).
Pantel, K. et al. Detection and clinical implications of early systemic tumor cell
dissemination in breast cancer. Clin Cancer Res 9, 6326-6334 (2003).
Zia, A., Schildberg, F.W. & Funke, I. MHC class I negative phenotype of
disseminated tumor cells in bone marrow is associated with poor survival in
R0M0 breast cancer patients. Int J Cancer 93, 566-570 (2001).
Wulfing, P. et al. HER2-positive circulating tumor cells indicate poor clinical
outcome in stage I to III breast cancer patients. Clin Cancer Res 12, 1715-1720
(2006).
Pierga, J.Y. et al. Real-time quantitative PCR determination of urokinase-type
plasminogen activator receptor (uPAR) expression of isolated micrometastatic
cells from bone marrow of breast cancer patients. Int J Cancer 114, 291-298
(2005).
Theodoropoulos, P.A. et al. Circulating tumor cells with a putative stem cell
phenotype in peripheral blood of patients with breast cancer. Cancer Lett 288,
99-106 (2010).
Muller, P., Weckermann, D., Riethmuller, G. & Schlimok, G. Detection of
genetic alterations in micrometastatic cells in bone marrow of cancer patients by
fluorescence in situ hybridization. Cancer Genet Cytogenet 88, 8-16 (1996).
Klein, C.A. et al. Genetic heterogeneity of single disseminated tumour cells in
minimal residual cancer. Lancet 360, 683-689 (2002).
Al-Mulla, F. et al. Heterogeneity of mutant versus wild-type Ki-ras in primary
and metastatic colorectal carcinomas, and association of codon-12 valine with
early mortality. J Pathol 185, 130-138 (1998).
Stoecklein, N.H. & Klein, C.A. Genetic disparity between primary tumours,
disseminated tumour cells, and manifest metastasis. Int J Cancer 126, 589-598
(2010).
Smirnov, D.A. et al. Global gene expression profiling of circulating tumor cells.
Cancer Res 65, 4993-4997 (2005).
112
Literature
101.
102.
103.
104.
105.
106.
107.
108.
109.
110.
111.
112.
113.
114.
115.
116.
117.
118.
119.
Sieuwerts, A.M. et al. mRNA and microRNA expression profiles in circulating
tumor cells and primary tumors of metastatic breast cancer patients. Clin
Cancer Res 17, 3600-3618 (2011).
Watson, M.A. et al. Isolation and molecular profiling of bone marrow
micrometastases identifies TWIST1 as a marker of early tumor relapse in breast
cancer patients. Clin Cancer Res 13, 5001-5009 (2007).
Klein, C.A. et al. Combined transcriptome and genome analysis of single
micrometastatic cells. Nat Biotechnol 20, 387-392 (2002).
Rahbari, N.N. et al. Compartmental Differences of Circulating Tumor Cells in
Colorectal Cancer. Ann Surg Oncol (2012).
Kim, M.J. et al. Association of CD47 with natural killer cell-mediated
cytotoxicity of head-and-neck squamous cell carcinoma lines. Tumour Biol 29,
28-34 (2008).
Losi, L., Baisse, B., Bouzourene, H. & Benhattar, J. Evolution of intratumoral
genetic heterogeneity during colorectal cancer progression. Carcinogenesis 26,
916-922 (2005).
Heng, H.H. et al. Genetic and epigenetic heterogeneity in cancer: the ultimate
challenge for drug therapy. Curr Drug Targets 11, 1304-1316 (2010).
Klein, C.A. et al. Comparative genomic hybridization, loss of heterozygosity,
and DNA sequence analysis of single cells. Proc Natl Acad Sci U S A 96, 44944499 (1999).
Sargent, D.J. et al. Defective mismatch repair as a predictive marker for lack of
efficacy of fluorouracil-based adjuvant therapy in colon cancer. J Clin Oncol
28, 3219-3226 (2010).
Oki, E. et al. Chemosensitivity and survival in gastric cancer patients with
microsatellite instability. Ann Surg Oncol 16, 2510-2515 (2009).
Umar, A. et al. Revised Bethesda Guidelines for hereditary nonpolyposis
colorectal cancer (Lynch syndrome) and microsatellite instability. J Natl
Cancer Inst 96, 261-268 (2004).
Shemirani, A.I. et al. Simplified MSI Marker Panel for Diagnosis of Colorectal
Cancer. Asian Pac J Cancer Prev 12, 2101-2104 (2011).
Woerner, S.M. et al. Microsatellite instability of selective target genes in
HNPCC-associated colon adenomas. Oncogene 24, 2525-2535 (2005).
Halford, S.E. et al. MSI-low, a real phenomenon which varies in frequency
among cancer types. J Pathol 201, 389-394 (2003).
Halford, S. et al. Low-level microsatellite instability occurs in most colorectal
cancers and is a nonrandomly distributed quantitative trait. Cancer Res 62, 5357 (2002).
Tanaka, T. et al. Chromosome 18q deletion as a novel molecular predictor for
colorectal cancer with simultaneous hepatic metastasis. Diagn Mol Pathol 18,
219-225 (2009).
Knosel, T. et al. Genetic imbalances with impact on survival in colorectal
cancer patients. Histopathology 43, 323-331 (2003).
Catenacci, D.V., Kozloff, M., Kindler, H.L. & Polite, B. Personalized colon
cancer care in 2010. Semin Oncol 38, 284-308 (2011).
Naomoto, Y., Kondo, H., Tanaka, N. & Orita, K. Novel experimental models of
human cancer metastasis in nude mice: lung metastasis, intraabdominal
carcinomatosis with ascites, and liver metastasis. J Cancer Res Clin Oncol 113,
544-549 (1987).
113
Literature
120.
121.
122.
123.
124.
125.
126.
127.
128.
129.
130.
131.
132.
133.
134.
135.
136.
137.
Thalheimer, A. et al. The intraportal injection model: a practical animal model
for hepatic metastases and tumor cell dissemination in human colon cancer.
BMC Cancer 9, 29 (2009).
Flatmark, K., Maelandsmo, G.M., Martinsen, M., Rasmussen, H. & Fodstad, O.
Twelve colorectal cancer cell lines exhibit highly variable growth and
metastatic capacities in an orthotopic model in nude mice. Eur J Cancer 40,
1593-1598 (2004).
Unemori, E.N., Ways, N. & Pitelka, D.R. Metastasis of murine mammary
tumour lines from the mammary gland and ectopic sites. Br J Cancer 49, 603614 (1984).
Kubota, T. Metastatic models of human cancer xenografted in the nude mouse:
the importance of orthotopic transplantation. J Cell Biochem 56, 4-8 (1994).
Glinskii, A.B. et al. Viable circulating metastatic cells produced in orthotopic
but not ectopic prostate cancer models. Cancer Res 63, 4239-4243 (2003).
Pantelouris, E.M. Observations on the immunobiology of 'nude' mice.
Immunology 20, 247-252 (1971).
Cespedes, M.V. et al. Orthotopic microinjection of human colon cancer cells in
nude mice induces tumor foci in all clinically relevant metastatic sites. Am J
Pathol 170, 1077-1085 (2007).
Sasaki, H. et al. Orthotopic implantation mouse model and cDNA microarray
analysis indicates several genes potentially involved in lymph node metastasis
of colorectal cancer. Cancer Sci 99, 711-719 (2008).
Shultz, L.D., Ishikawa, F. & Greiner, D.L. Humanized mice in translational
biomedical research. Nat Rev Immunol 7, 118-130 (2007).
Morikawa, K., Walker, S.M., Jessup, J.M. & Fidler, I.J. In vivo selection of
highly metastatic cells from surgical specimens of different primary human
colon carcinomas implanted into nude mice. Cancer Res 48, 1943-1948 (1988).
Kuhn, S. et al. A complex of EpCAM, claudin-7, CD44 variant isoforms, and
tetraspanins promotes colorectal cancer progression. Mol Cancer Res 5, 553567 (2007).
Koch, M. et al. Comparative analysis of tumor cell dissemination in mesenteric,
central, and peripheral venous blood in patients with colorectal cancer. Arch
Surg 136, 85-89 (2001).
Taniguchi, T., Makino, M., Suzuki, K. & Kaibara, N. Prognostic significance of
reverse transcriptase-polymerase chain reaction measurement of
carcinoembryonic antigen mRNA levels in tumor drainage blood and peripheral
blood of patients with colorectal carcinoma. Cancer 89, 970-976 (2000).
Fujita, S., Kudo, N., Akasu, T. & Moriya, Y. Detection of cytokeratin 19 and 20
mRNA in peripheral and mesenteric blood from colorectal cancer patients and
their prognosis. Int J Colorectal Dis 16, 141-146 (2001).
Iinuma, H. et al. Usefulness and clinical significance of quantitative real-time
RT-PCR to detect isolated tumor cells in the peripheral blood and tumor
drainage blood of patients with colorectal cancer. Int J Oncol 28, 297-306
(2006).
Sato, H. et al. Significance of serum concentrations of E-selectin and CA19-9 in
the prognosis of colorectal cancer. Jpn J Clin Oncol 40, 1073-1080 (2010).
Dabelsteen, E. Cell surface carbohydrates as prognostic markers in human
carcinomas. J Pathol 179, 358-369 (1996).
Park, I.J., Choi, G.S. & Jun, S.H. Prognostic value of serum tumor antigen
CA19-9 after curative resection of colorectal cancer. Anticancer Res 29, 43034308 (2009).
114
Literature
138.
139.
140.
141.
142.
143.
144.
145.
146.
147.
148.
149.
150.
151.
152.
153.
154.
155.
Sieuwerts, A.M. et al. Anti-epithelial cell adhesion molecule antibodies and the
detection of circulating normal-like breast tumor cells. J Natl Cancer Inst 101,
61-66 (2009).
Krebs, M.G. et al. Analysis of Circulating Tumor Cells in Patients with Nonsmall Cell Lung Cancer Using Epithelial Marker-Dependent and -Independent
Approaches. J Thorac Oncol (2011).
van de Stolpe, A., Pantel, K., Sleijfer, S., Terstappen, L.W. & den Toonder,
J.M. Circulating tumor cell isolation and diagnostics: toward routine clinical
use. Cancer Res 71, 5955-5960 (2011).
Raj, A. & van Oudenaarden, A. Nature, nurture, or chance: stochastic gene
expression and its consequences. Cell 135, 216-226 (2008).
Elzagheid, A. et al. E-cadherin expression pattern in primary colorectal
carcinomas and their metastases reflects disease outcome. World J
Gastroenterol 12, 4304-4309 (2006).
Tsuji, T., Ibaragi, S. & Hu, G.F. Epithelial-mesenchymal transition and cell
cooperativity in metastasis. Cancer Res 69, 7135-7139 (2009).
Mehes, G., Witt, A., Kubista, E. & Ambros, P.F. Circulating breast cancer cells
are frequently apoptotic. Am J Pathol 159, 17-20 (2001).
Nubel, T. et al. Claudin-7 regulates EpCAM-mediated functions in tumor
progression. Mol Cancer Res 7, 285-299 (2009).
Muller, V. et al. Circulating tumor cells in breast cancer: correlation to bone
marrow micrometastases, heterogeneous response to systemic therapy and low
proliferative activity. Clin Cancer Res 11, 3678-3685 (2005).
Fuchs, S.Y., Ougolkov, A.V., Spiegelman, V.S. & Minamoto, T. Oncogenic
beta-catenin signaling networks in colorectal cancer. Cell Cycle 4, 1522-1539
(2005).
Kikuchi, A., Kishida, S. & Yamamoto, H. Regulation of Wnt signaling by
protein-protein interaction and post-translational modifications. Exp Mol Med
38, 1-10 (2006).
Pantel, K. et al. Differential expression of proliferation-associated molecules in
individual micrometastatic carcinoma cells. J Natl Cancer Inst 85, 1419-1424
(1993).
Solakoglu, O. et al. Heterogeneous proliferative potential of occult metastatic
cells in bone marrow of patients with solid epithelial tumors. Proc Natl Acad
Sci U S A 99, 2246-2251 (2002).
Di Fiore, F. et al. Clinical relevance of KRAS mutation detection in metastatic
colorectal cancer treated by Cetuximab plus chemotherapy. Br J Cancer 96,
1166-1169 (2007).
Lievre, A. et al. KRAS mutation status is predictive of response to cetuximab
therapy in colorectal cancer. Cancer Res 66, 3992-3995 (2006).
Di Nicolantonio, F. et al. Wild-type BRAF is required for response to
panitumumab or cetuximab in metastatic colorectal cancer. J Clin Oncol 26,
5705-5712 (2008).
Siena, S., Sartore-Bianchi, A., Di Nicolantonio, F., Balfour, J. & Bardelli, A.
Biomarkers predicting clinical outcome of epidermal growth factor receptortargeted therapy in metastatic colorectal cancer. J Natl Cancer Inst 101, 13081324 (2009).
Vidaurreta, M. et al. Colorectal carcinoma prognosis can be predicted by
alterations in gene p53 exons 5 and 8. Int J Colorectal Dis 23, 581-586 (2008).
115
Literature
156.
157.
158.
159.
160.
161.
162.
163.
164.
165.
166.
167.
168.
169.
170.
171.
172.
173.
174.
Mollevi, D.G. et al. Mutations in TP53 are a prognostic factor in colorectal
hepatic metastases undergoing surgical resection. Carcinogenesis 28, 12411246 (2007).
Elsaleh, H. et al. P53 alteration and microsatellite instability have predictive
value for survival benefit from chemotherapy in stage III colorectal carcinoma.
Clin Cancer Res 7, 1343-1349 (2001).
Samowitz, W.S., Curtin, K., Neuhausen, S., Schaffer, D. & Slattery, M.L.
Prognostic implications of BAX and TGFBRII mutations in colon cancers with
microsatellite instability. Genes Chromosomes Cancer 35, 368-371 (2002).
Benatti, P. et al. Microsatellite instability and colorectal cancer prognosis. Clin
Cancer Res 11, 8332-8340 (2005).
Brosens, R.P. et al. Deletion of chromosome 4q predicts outcome in stage II
colon cancer patients. Cell Oncol (Dordr) 34, 215-223 (2011).
Watanabe, T. et al. Molecular predictors of survival after adjuvant
chemotherapy for colon cancer. N Engl J Med 344, 1196-1206 (2001).
Chen, Z. et al. Chromosomal copy number alterations are associated with tumor
response to chemoradiation in locally advanced rectal cancer. Genes
Chromosomes Cancer 50, 689-699 (2011).
Fehm, T. et al. Cytogenetic evidence that circulating epithelial cells in patients
with carcinoma are malignant. Clin Cancer Res 8, 2073-2084 (2002).
Baker, S.J. et al. Chromosome 17 deletions and p53 gene mutations in
colorectal carcinomas. Science 244, 217-221 (1989).
Vogelstein, B. & Kinzler, K.W. Cancer genes and the pathways they control.
Nat Med 10, 789-799 (2004).
Schardt, J.A. et al. Genomic analysis of single cytokeratin-positive cells from
bone marrow reveals early mutational events in breast cancer. Cancer Cell 8,
227-239 (2005).
Albanese, I. et al. Heterogeneity within and between primary colorectal
carcinomas and matched metastases as revealed by analysis of Ki-ras and p53
mutations. Biochem Biophys Res Commun 325, 784-791 (2004).
Artale, S. et al. Mutations of KRAS and BRAF in primary and matched
metastatic sites of colorectal cancer. J Clin Oncol 26, 4217-4219 (2008).
Santini, D. et al. High concordance of KRAS status between primary colorectal
tumors and related metastatic sites: implications for clinical practice. Oncologist
13, 1270-1275 (2008).
Baldus, S.E. et al. Prevalence and heterogeneity of KRAS, BRAF, and PIK3CA
mutations in primary colorectal adenocarcinomas and their corresponding
metastases. Clin Cancer Res 16, 790-799 (2010).
Yao, J., Goh, H.S. & Smith, D.R. p53 mutations in primary colorectal
adenocarcinomas and liver metastases. Br J Surg 83, 1245-1246 (1996).
Zhang, J.S., Caplin, S., Bosman, F.T. & Benhattar, J. Genetic diversity at the
p53 locus between primary human colorectal adenocarcinomas and their lymphnode metastases. Int J Cancer 70, 674-678 (1997).
Schmidt-Kittler, O. et al. From latent disseminated cells to overt metastasis:
genetic analysis of systemic breast cancer progression. Proc Natl Acad Sci U S
A 100, 7737-7742 (2003).
Kraus, J., Pantel, K., Pinkel, D., Albertson, D.G. & Speicher, M.R. Highresolution genomic profiling of occult micrometastatic tumor cells. Genes
Chromosomes Cancer 36, 159-166 (2003).
116
Literature
175.
176.
177.
178.
179.
180.
181.
182.
183.
184.
185.
186.
Gangnus, R., Langer, S., Breit, E., Pantel, K. & Speicher, M.R. Genomic
profiling of viable and proliferative micrometastatic cells from early-stage
breast cancer patients. Clin Cancer Res 10, 3457-3464 (2004).
Gorelik, E., Gunji, Y. & Goldfarb, R.H. Interaction of tumor cells and immune
system in the metastatic process. Biochem Cell Biol 66, 617-625 (1988).
Khan, Z.A. et al. P53 mutations in primary and metastatic tumors and
circulating tumor cells from colorectal carcinoma patients. Clin Cancer Res 6,
3499-3504 (2000).
Meng, S. et al. HER-2 gene amplification can be acquired as breast cancer
progresses. Proc Natl Acad Sci U S A 101, 9393-9398 (2004).
Hayes, D.F. et al. Monitoring expression of HER-2 on circulating epithelial
cells in patients with advanced breast cancer. Int J Oncol 21, 1111-1117 (2002).
Eliane, J.P. et al. Monitoring serial changes in circulating human breast cancer
cells in murine xenograft models. Cancer Res 68, 5529-5532 (2008).
Helzer, K.T. et al. Circulating tumor cells are transcriptionally similar to the
primary tumor in a murine prostate model. Cancer Res 69, 7860-7866 (2009).
Howard, E.W. et al. Decreased adhesiveness, resistance to anoikis and
suppression of GRP94 are integral to the survival of circulating tumor cells in
prostate cancer. Clin Exp Metastasis 25, 497-508 (2008).
Wang, D. et al. Downregulation of E-Cadherin enhances proliferation of head
and neck cancer through transcriptional regulation of EGFR. Mol Cancer 10,
116 (2011).
Misra, S. et al. Hyaluronan-CD44 interactions as potential targets for cancer
therapy. FEBS J 278, 1429-1443 (2011).
Boland, C.R. et al. A National Cancer Institute Workshop on Microsatellite
Instability for cancer detection and familial predisposition: development of
international criteria for the determination of microsatellite instability in
colorectal cancer. Cancer Res 58, 5248-5257 (1998).
Findeisen, P. et al. T25 repeat in the 3' untranslated region of the CASP2 gene:
a sensitive and specific marker for microsatellite instability in colorectal cancer.
Cancer Res 65, 8072-8078 (2005).
117
Acknowledgement
Acknowledgment
I am very grateful to everyone who supported my work on this thesis. In particular I
would like to thank Prof. Dr. Jürgen Weitz for giving me the chance to work on a very
challenging field of research in his lab. Furthermore, I have to thank Prof. Dr. Philipp
Beckhove for his supervision.
Thank you Prof. Dr. Moritz Koch for your help and advises.
Thank you, Dr. Gunnar Steinert for your help with the project as well as for reading the
manuscript and your suggestions for improvement.
Thanks to the whole lab team for the pleasant working atmosphere. I really learned a
lot about playing Schnauz.
Thanks to the external collaboration partners Dr. Matthias Kloor and Anita Voigt as
well as Prof. Dr. N. Stöcklein and Bianca Behrens for their contributions and support.
Furthermore, I will not omit to mention Martin Mollenhauer and Markus Stauch who
became good friends and helped me to overcome some times of disappointment and
despair.
Finally, I am very grateful to my whole family and in particular to my girl friend Janine
Kästner for her support and comfort during a hard time.
118
Declaration
I have written this thesis independently, solely based on the literature, methods and
devices mentioned in the chapters and the appendix.
119
Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement