Abstract |
Breast cancer, the most common type of cancer among women, is frequently
diagnosed at an early stage and successfully treated. Nevertheless, distant metastases
ensue in a significant proportion of patients, thus representing the main cause of death
for these patients. Since the exact mechanism of metastases formation has not yet
been elucidated, there is a great need for a more extensive study of the metastatic
process. Numerous studies in cancer cell lines and animal models have demonstrated
the complexity of the metastatic process. The main stages of metastasis development
include the proliferation of cancer cells in the primary site, invasion in surrounding
tissues and intravasation into the circulation, where they can migrate either as single
cells or clusters. Afterwards, cancer cells can extravasaste at distant sites and invade
new organs to form micrometastases, where their subsequent proliferation leads to the
formation of overt metastases. These cells, identified either as disseminated tumor
cells (DTCs) in secondary sites or as circulating tumor cells (CTCs) in the peripheral
blood of patients, consist the population of micrometastatic cells.
Several studies have shown that the detection of CTCs in peripheral blood of
patients with breast cancer is a strong and independent marker for increased risk of
relapse and reduced overall survival. Moreover, the presence of CTCs in breast cancer
has been strongly correlated with lower response rates to conventional
chemotherapies. However, CTCs are a highly heterogeneous population exhibiting
differential metastatic potential, therefore their further molecular and phenotypic
characterization is of outmost importance. This could highlight those characteristics
that prosper metastatic process, contributing to a better understanding of the
mechanism of metastasis. Moreover, CTCs are considered as a real time liquid
biopsy, which allows the molecular characterization of the tumor at different time
points. The identification of CTCs bearing characteristics associated with aggressive
behavior or resistance to conventional treatment might also help to identify subgroups
of patients with poor prognosis. These patients could be offered more aggressive
therapeutic approaches and/or targeted therapies against molecules selected according
to their expression on CTCs. Thus, during the last years, a growing number of
techniques are being developed for the detection and characterization of CTCs.
20
Epithelial-to-mesenchymal transition (EMT) of cancer cells is a process holding
an important role in metastasis formation, during which the epithelial phenotype
dynamically converts to a mesenchymal one. These mesenchymal characteristics
enable the initial invasion and intravasation of cancer cells, their migration through
the systemic circulation and the subsequent extravasation and invasion at distant sites.
Nevertheless, it has been shown that during the colonization of distant organs, they
regain an epithelial phenotype through the reverse process of mesenchymal-toepithelial
transition (MET). One of the major changes during EMT is related to the
expression pattern of the cytoskeleton intermediate filaments. Specifically, a
significant reduction of cytokeratin expression along with a simultaneous increase in
vimentin expression is usually observed. Moreover, the expression of molecules
involved in cell to cell contacts, induced by a series of transcription factors, such as
TWIST, SNAIL and SLUG is also modulated. The expression of the putative EMT
markers, Vimentin and TWIST is considered essential for the invasion and
intravasation of cancer cells in the circulation. In addition, numerous studies have
shown that characteristics suggestive of the presence of EMT in primary tumors, are
strongly associated with resistance to chemotherapy, high risk of relapse and
decreased survival.
A growing body of evidence suggests a correlation between the EMT process and
the generation of cancer stem cells (CSCs). CSCs have been identified as a small
subpopulation in several types of cancers, including breast cancer, and are considered
to bear properties of normal stem cells. The CSC model proposes that these cells only,
are capable of self-renewal, unlimited proliferation and differentiation, whereas they
have also been shown to participate in the metastatic process, giving rise to all the
subpopulations that constitute the primary tumor. In breast cancer, CSCs have been
effectively identified and isolated from cell lines and tumors on the basis of the
CD44high/CD24-/low phenotype or according to the high enzymatic activity of aldehyde
dehydrogenase (ALDH). The expression of the ALDH1A1 isoenzyme is also
considered as a putative stem cell marker, however, contradictory data appear in the
literature concerning the activity of the 19 ALDH isoenzymes found in cell lines and
breast tumors. Numerous in vitro and in vivo studies have shown that CSCs are highly
resistant to radiotherapy, chemotherapy and hormone therapy, whereas the detection
21
of putative CSCs in primary tumors has been strongly associated with worse clinical
outcome.
Based on the above, the detection of putative EMT and CSC markers on CTCs
could distinguish a subpopulation of CTCs demonstrating chemoresistance and
enhanced metastatic potential. In this context, in 2010 our research group showed that
CTCs of patients with metastatic breast cancer expressed putative CSC markers, such
as CD44high/CD24-/low and ALDH1high. Furthermore, we have recently shown that
CTCs of early and metastatic breast cancer patients acquired a partial EMT
phenotype, as defined by the co-expression of epithelial and mesenchymal markers,
such as Vimentin and TWIST. It was further shown that the frequency of CTCs
bearing this partial EMT phenotype was significantly increased in the metastatic
setting, suggesting their selection during disease progression. Several studies during
the last years have also reported the presence of EMT and CSC characteristics on
CTCs by the use of different techniques, however there are no clinical studies
including well defined groups with large number of patients. Moreover, the coexpression
of CSC and EMT markers at the single CTC level has not yet been
reported. On this basis, we investigated whether CTCs of patients with breast cancer,
co-express putative stem cell and intermediate EMT phenotypes. Furthermore, we
interrogated the frequency of these CTC subpopulations among patients with early
and metastatic disease, evaluated their potential prognostic value and assessed the
effect of chemotherapy on their distribution.
For this purpose, a novel triple immunofluorescence methodology was developed
on isolated peripheral blood mononuclear cells (PBMCs) of patients, for the detection
of cytokeratins 8, 18 and 19 as a CTC marker, along with ALDH1 and TWIST, as
putative CSC and EMT markers, respectively. The co-expression of these molecules
was evaluated at the single cell level by the use of the semi-automated fluorescence
microscopy system ARIOL. For the development of the method, the different
expression patterns of ALDH1 and TWIST was initially evaluated in control cell lines
and three breast cancer cell lines, SKBR3, MCF7 and MDA.MB.231, representative
of the main molecular breast cancer subtypes, HER2-positive, hormone-positive and
triple-negative, respectively. Specifically, the expression levels of ALDH1 were
characterized as high (ALDH1high) and low or absent (ALDH1low/neg), according to the
quantification of the fluorescence levels by the use of the ARIOL system, while the
22
expression of TWIST was defined according to its subcellular localization as nuclear
(TWISTnuc) and cytoplasmic or absent (TWISTcyt/neg). According to the current
literature, CSC and EMT phenotypes were defined as those bearing high ALDH1
expression levels and nuclear localization of TWIST, respectively. A series of control
experiments were also performed in all cell lines above, as well as in normal blood
donor samples, to verify that the methodology allows the detection of these markers
on CTCs with high sensitivity and specificity.
Subsequently, the current methodology was applied to CTCs from 80 patients
with early and 50 with metastatic breast cancer. Both markers could be detected on
CTCs in almost all patients evaluated, however their expression pattern was different
between the two clinical stages. Specifically, CTCs bearing the CSC phenotype
(ALDH1high) were more frequently detected in metastatic disease, in contrast to the
early disease setting wherein CTCs with a non-stem cell phenotype were mainly
detected (ALDH1low/neg). Furthermore, CTCs expressing an intermediate EMT
phenotype (TWISTnuc) were more frequent in metastatic disease, whereas in the early
stage, CTCs bearing an epithelial phenotype were mainly identified (TWISTcyt/neg).
Assessing the co-expression of the two markers at the single cell level, it was shown
that the majority of metastatic patients had detectable CTCs co-expressing the two
phenotypes (ALDH1high/TWISTnuc), in contrast to the patients with early disease
which harboured mainly CTCs expressing a non-stem, epithelial phenotype
(ALDH1low/neg /TWISTcyt/neg). The correlation between the two phenotypes on CTCs
of metastatic breast cancer patients was further statistically confirmed. This study has
provided a new, sensitive and specific method for the characterization of CTCs
according to the co-expression of two putative CSC and EMT markers. The finding
that the expression pattern of the two markers differed between early and metastatic
disease reflects the dynamic evolution of the CSC and EMT states. Furthermore, the
finding that the frequency of CTCs co-expressing the two phenotypes was
significantly increased in metastatic disease, indicates that this subpopulation might
be involved in the metastatic process and that it could be selected during disease
progression. Finally, this study showed for the first time that two putative CSC and
EMT markers are frequently co-expressed on single CTCs, reinforcing their
association, as well as that that the two phenotypes of CTCs changed in parallel
during metastatic progression.
23
We next evaluated the effect of chemotherapy on the frequency of CTCs bearing
CSC and intermediate EMT characteristics. For this purpose, the current methodology
was applied to a larger and well defined group of 154 metastatic breast cancer patients
before the initiation of first-line chemotherapy. Patients with detectable CTCs, were
also evaluated after the completion of treatment. The majority of CTCs identified,
both before and after chemotherapy, were shown to express CSC and partial EMT
characteristics. Even though treatment resulted in a significant reduction or
elimination of CTCs, cells bearing CSC or partial EMT phenotypes were enriched
after treatment. Moreover, a statistically significant increase in CTCs co-expressing
the two phenotypes was further confirmed. These findings provide important evidence
that CTCs with CSC and intermediate EMT features exhibit chemo-resistance and
therefore are selected after chemotherapy administration. The current study showed
for the first time that the resistance of CTCs to conventional chemotherapy regiments
is enhanced by the simultaneous presence of CSC and intermediate EMT
characteristics, further supporting the participation of this CTC subpopulation in the
metastatic process.
The current study also showed for the first time the prognostic value of CTCs
detected by the ARIOL system. Specifically, the total CTC number both before and
after chemotherapy was correlated with disease progression at the end of
chemotherapy, while CTC detection prior to chemotherapy was further associated
with lower rates of 2-year survival. Moreover, the presence of CTCs bearing CSC and
EMT phenotypes was also correlated to metastases in specific organs such as bones
and lung, however, contradictory data emerged regarding their prognostic value in
terms of response to treatment and survival. Even though chemotherapy led to a
significant increase of CTCs expressing CSC and partial EMT phenotypes exclusively
among patients who progressed at the end of treatment, the detection of these
subpopulations was associated with better response to chemotherapy and increased
progression-free survival.
We further investigated whether CTCs bearing the stem cell phenotype, as
defined by the high expression of the isoenzyme ALDH1A1, also acquire functional
stem cell characteristics. The enzymatic activity of ALDH was used as a functional
marker of CSCs, which was evaluated by ALDEFLUOR assay and the use of flow
cytometry. First, the correlation between ALDH enzymatic activity and high ALDH1
24
protein expression was interrogated via a series of immunofluorescence experiments,
immunoblotting and ALDEFLUOR in HepG control cells and three breast cancer cell
lines, SKBR3, MCF7 and MDA.MB.231. A positive correlation between activity and
expression was shown in HepG2 cells, confirming that the methodology developed by
ARIOL allows the identification of the functional cell population. However, the
percentage of cells expressing high ALDH1 levels was higher than the percentage of
those bearing high ALDH activity in all breast cancer cell lines evaluated. These
findings indicate that the ALDH1A1 isoenzyme which is expressed in breast cancer
cells is not always enzymatically active. Therefore, it is suggested that the
methodology developed by ARIOL system allows the identification of cell
populations that express an enzymatically active ALDH1 protein, however cells that
express an inactive protein are also being detected. Subsequently, we investigated the
correlation between high ALDH1 expression and ALDH activity in CTCs from
metastatic breast cancer patients. Τhe methodology for the detection of ALDH
activity on CTCs was developed through a series of experiments using serial dilutions
of tumor cells from each cell line into isolated PBMCs from normal blood donors
(dilutions of 10, 100, 1.000 and 10.000 cells per 106 PBMCs). Tumor cells were
detected by the expression of EpCAM and CD45 surface molecules instead of
cytokeratin, since the evaluation of intracellular proteins cannot be combined with the
ALDEFLUOR assay. CTCs could be identified in all the above dilutions, however an
overestimation of cell counts was observed in low concentrations of 10 and 100 tumor
cells per 106 PBMCs. ALDH activity could also be detected in all dilutions, however
the recovery rate was reduced in low cell concentrations, suggesting that the
developed methodology is appropriate for patients bearing high CTC counts only.
Subsequently, the enzymatic activity and expression of ALDH was investigated
by the use of flow cytometry and the ARIOL system, respectively, on CTCs from nine
breast cancer patients, who had multiple metastases and had received at least two lines
of treatment. A statistically significant correlation was confirmed regarding the CTC
counts detected by the two methods, as well as between the number of CTCs bearing
high ALDH activity and high ALDH1 expression. This study showed for the first time
that the expression of ALDH1 on CTCs of metastatic breast cancer patients was
correlated to ALDH enzymatic activity, further confirming that the methodology
developed by the ARIOL system allows the identification of a CTC population with
25
CSC activity. Finally, ALDH enzymatic activity was evaluated on CD45-negative
cells, which include the subpopulation of CTCs that have completely lost the
expression of epithelial markers in the context of full EMT. It was shown that ALDH
activity in these cells was increased, compared to the population of EpCAMpositive/
CD45-negative cells, suggesting that the current approach allows the
detection of CTCs bearing a fully EMT phenotype and also functional characteristics
of CSCs.
To summarize, in the context of the present thesis, we performed a phenotypic
analysis of CTCs from breast cancer patients, regarding the presence of stem cell and
partial EMT characteristics. The present study provides a new methodology that
allows the sensitive and specific detection of CSC and EMT markers and their coexpression
at the single CTC level. The expression pattern of both markers was
differentiated between the early and metastatic disease stage, highlighting the
dynamic evolution of these characteristics on CTCs. The observation that the
frequency of CTCs bearing a stem cell and partial EMT phenotype was significantly
higher in metastatic disease, provides evidence for their involvement in the metastatic
process. It was also described for the first time that CSC and EMT markers are
frequently co-expressed on the same CTC, further confirming the correlation between
the two states in CTCs. Furthermore, it was found that CTCs co-expressing these
phenotypes were highly resistant to conventional chemotherapy and that the
enrichment of this CTC subpopulation was associated with disease progression at the
end of treatment. The detection of CTCs bearing these phenotypes was also associated
with metastases in specific organs, while their prognostic value for patients' survival
remains to be further investigated. Finally, the simultaneous evaluation of the
functional stem cell activity of CTCs in a subset of metastatic breast cancer patients
further confirmed that the methodology developed by the ARIOL, identified an active
CTC subpopulation with CSC characteristics. To conclude, CTCs acquiring CSC and
partial EMT characteristics consist an aggressive CTC population, which might be
involved in the metastatic progression of breast cancer and exhibit increased chemoresistance.
The present study highlights the significance of the extensive phenotypic
analysis of CTCs for the determination of their role in the metastatic process, as well
as the need for alternative targeted therapies against molecules associated with stem
26
cell and EMT features. This approach could contribute to a more effective
management of breast cancer patients.
|