tumor initiating cells in mesenchymal neoplasms · population cells to initiate tumor formation in...
TRANSCRIPT
Tumor Initiating Cells in Mesenchymal Neoplasms
by
Colleen Wu
A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy
Institute of Medical Sciences University of Toronto
© Copyright by Colleen Wu 2010
ii
Tumor Initiating Cells in Mesenchymal Neoplasms
Colleen Wu
Degree of Doctor of Philosophy
Institute of Medical Sciences University of Toronto
2010
Abstract
Despite the clonal origins of tumors, the majority of neoplasms are composed of a heterogeneous
population of cells. The origins of this phenotype these cells have the potential to get can be
associated with cancer stem cells or tumor initiating cells have the potential to self-renew and to
differentiate giving rise to all cell types compromising a heterogeneous malignancy. These cells
are clinically important as they preferentially give rise to tumors and are therefore hypothesized
to account for the longevity and recurrence of neoplastic lesions. Cancer stem cells have been
identified from a broad range of hematopoietic, neural and epithelia tumors; however, their
function in mesenchymal neoplasms is less well defined. Using the side population assay, we
identified a subpopulation of cells within mesenchymal neoplasms, referred to as side population
cells, which are enhanced for tumor initiating potential. Importantly, we show a correlation
between the percentage of side population cells and tumor grade suggesting clinical prognostic
value as the proportion of side population cells may be a predictor of patient outcome.
Interestingly side population cells show distinct molecular features when compared to non-side
population cells and manipulation of these molecular mechanisms reduces the ability of side
population cells to initiate tumor formation in osteosarcoma cell lines. In conjunction with these
experiments, we also sought to determine the cellular origins of the mesenchymal neoplasm,
aggressive fibromatosis. Using mouse models we show the influence of a mesenchymal
precursor cells in the development of this malignancy. These results identify important biological
features of mesenchymal neoplasms from which the development of targeted treatment strategies
can begin.
iii
Table of Contents
TableofContents ................................................................................................................................ iii
ListofFigures.......................................................................................................................................vii
ListofTables ...................................................................................................................................... viii
ListofAbbreviations .......................................................................................................................... ix
Chapter1 .................................................................................................................................................2
1 Theroleofcancerstemcellsintheinitiationandprogressionofmesenchymal
neoplasms ...............................................................................................................................................2
1.1 Abstract....................................................................................................................................................21.2 MesenchymalNeoplasms...................................................................................................................31.3 CancerStemCells..................................................................................................................................61.3.1 Overview..............................................................................................................................................................61.3.2 CSCsinsolidtumors........................................................................................................................................71.3.3 CharacteristicsofCSCs...................................................................................................................................81.3.4 ClinicalsignificanceandtherapeutictargetingofCSCs................................................................101.3.5 CSCs:ongoingcontroversies ....................................................................................................................121.3.6 Isolationtechniques.....................................................................................................................................14
1.4 SidePopulationCells ........................................................................................................................ 141.4.1 Sidepopulationcellsinhumanneoplasms........................................................................................151.4.2 Cellularphenotype .......................................................................................................................................161.4.3 Tumorigenicpotential.................................................................................................................................171.4.4 Expressionofstem‐likegenes .................................................................................................................171.4.5 Drugefflux........................................................................................................................................................18
1.5 TheOriginsofCSCs............................................................................................................................ 191.6 MesenchymalProgenitorCells ..................................................................................................... 201.6.1 Overview...........................................................................................................................................................201.6.2 Definingthemesenchymalstem/progenitorcell ...........................................................................211.6.3 Locationofmesenchymalstemcells ....................................................................................................221.6.4 Arepericytesmesenchymalstemcells?..............................................................................................22
1.7 MesenchymalProgenitorCellsastheSarcomaCellofOrigin ............................................ 23
iv
1.8 AggressiveFibromatosis................................................................................................................. 231.8.1 Overview...........................................................................................................................................................231.8.2 Molecularetiologyofaggressivefibromatosis.................................................................................241.8.3 Mousemodelsofaggressivefibromatosis .........................................................................................25
1.9 Summaryandconclusions.............................................................................................................. 251.10 ThesisSummaryandRationale.................................................................................................. 261.11 References ......................................................................................................................................... 30
Chapter2 .............................................................................................................................................. 42
2 Sidepopulationcellsisolatedfrommesenchymalneoplasmscontaintumor
initiatingcells ..................................................................................................................................... 42
2.1 Abstract................................................................................................................................................. 422.2 Introduction ........................................................................................................................................ 432.3 Results ................................................................................................................................................... 452.3.1 Mesenchymaltumorscontainsidepopulation(SP)cells............................................................452.3.2 Theproportionofsidepopulationcellscorrelateswithaggressivenessofthetumor ..452.3.3 SPcellshavethecapacitytoformtumorsuponserialtransplantationinNOD/SCID
mice 452.3.4 SPCellsEffluxRhodamine‐123...............................................................................................................47
2.4 Discussion ............................................................................................................................................ 472.5 MaterialsandMethods .................................................................................................................... 632.5.1 PrimaryTumors ............................................................................................................................................632.5.2 Pathology ..........................................................................................................................................................632.5.3 FlowCytometry..............................................................................................................................................632.5.4 CellTransplantationintoNOD/SCIDmice .........................................................................................64
2.6 References ........................................................................................................................................... 65
Chapter3 .............................................................................................................................................. 71
3 Blockadeofhedgehogsignalinginhibitstheformationoftumorsderivedfrom
osteosarcomasidepopulationcells ............................................................................................ 71
3.1 Abstract................................................................................................................................................. 713.2 Introduction ........................................................................................................................................ 723.3 Results ................................................................................................................................................... 753.3.1 Sidepopulationcellsarepresentinosteosarcomacelllines.....................................................75
v
3.3.2 Hedgehogsignalinginosteosarcomasidepopulationcells .......................................................753.3.3 Invivotumorigenicpotentialofosteosarcomasidepopulationandnon‐sidepopulation
cells 763.3.4 Blockadeofhedgehogsignalinginhibitsinvivotumorformationofsidepopulation
cells 763.3.5 BlockadeofHedgehogSignalingdecreasesthecellularityofsidepopulationderived
tumors773.4 Discussion ............................................................................................................................................ 773.5 Figures................................................................................................................................................... 813.6 MaterialsandMethods .................................................................................................................... 913.6.1 Celllines ............................................................................................................................................................913.6.2 Flowcytometry ..............................................................................................................................................913.6.3 RNAextractionandrealtime‐RT‐PCR.................................................................................................913.6.4 Xenograftmodels ..........................................................................................................................................923.6.5 Dissociationofxenograftedtumors ......................................................................................................923.6.6 Invitroandinvivoblockadeofhedgehogsignaling.......................................................................92
3.7 References ........................................................................................................................................... 94
Chapter4 .............................................................................................................................................. 98
4 Thedevelopmentofaggressivefibromatosis(desmoidtumor)isinfluencedby
mesenchymalprogenitorcells...................................................................................................... 98
4.1 Abstract................................................................................................................................................. 984.2 Introduction ........................................................................................................................................ 994.3 Results .................................................................................................................................................1014.3.1 Aggressivefibromatosiscontainasubpopulationofcellswithprogenitorproperties
1014.3.2 PositivecorrelationbetweennumbersofaggressivefibromatosisandCFU‐Fsin
Apcwt/1638nmice............................................................................................................................................................ 1014.3.3 Mesenchymalprogenitorsareinvolvedinthedevelopmentofaggressivefibromatosis
1024.3.4 MesenchymalbutnotepithelialderivedtumorsareimpactedbyalterationofMPCs 1024.3.5 MesenchymalprecursorsfromApcwt/1638Nhavethecapacitytoinitiatetumorformation
1034.4 Discussion ..........................................................................................................................................103
vi
4.5 MaterialsandMethods ..................................................................................................................1184.5.1 Primarytumors........................................................................................................................................... 1184.5.2 Flowcytometry ........................................................................................................................................... 1184.5.3 Generationofgeneticallyengineeredmice..................................................................................... 1184.5.4 Scoringoftumors ....................................................................................................................................... 1194.5.5 CellCulture.................................................................................................................................................... 1194.5.6 Xenograftmodels ....................................................................................................................................... 1194.5.7 Geneprofiling .............................................................................................................................................. 120
4.6 References .........................................................................................................................................121
Chapter5 ............................................................................................................................................125
5 Summary,conclusions,andfuturedirections ................................................................125
5.1 Summary ............................................................................................................................................1265.2 Conclusions........................................................................................................................................1265.3 FutureDirections.............................................................................................................................1275.3.1 Thesidepopulationassay:considerationsforitsuseintheisolationofcancerstem
cells/tumorinitiatingcells .................................................................................................................................... 1275.3.2 Characterizationofsidepopulationcellsinmesenchymalneoplasms .............................. 1305.3.3 Clinicalsignificanceofsarcomasidepopulationcell ................................................................. 1335.3.4 Tumormicroenvironment ..................................................................................................................... 1345.3.5 Mesenchymalprogenitorcellsandtheirinvolvementinthedevelopmentofaggressive
fibromatosis ................................................................................................................................................................. 1375.4 ConcludingRemarks.......................................................................................................................1395.5 References .........................................................................................................................................140
vii
List of Figures
Figure 1.1 Cellular origins of cancer stem cells ........................................................................... 28
Figure 2.1 Mesenchymal tumors contain side population cells .................................................... 51
Figure 2.2 High grade sarcomas have an increased prevalence of SP cells when compared to
lower grade lesions. ...................................................................................................................... 53
Figure 2.3 Histopathologic features of SP tumors ........................................................................ 57
Figure 2.4 Characteristics of tumors derived from SP and non-SP cells...................................... 59
Figure 2.5 SP cells efflux Rhodamine-123 ................................................................................... 61
Figure 3.1 Side population cells are present in osteosarcoma cell lines ....................................... 81
Figure 3.2 Hedgehog signaling in osteosarcoma side population cells ........................................ 83
Figure 3.3 Blockade of hedgehog signaling inhibits in vivo tumor formation of osteosarcoma side
population cells ............................................................................................................................. 87
Figure 3.4 Side population tumors treated with triparanol have decreased cellularity................. 89
Figure 4.1 Human AF tumors contain progenitor cells .............................................................. 106
Figure 4.2 Correlation of CFU-F with numbers of AF tumors................................................... 108
Figure 4.3 Modulation of MPCs impacts tumor development ................................................... 110
Figure 4.4 Progenitor statuses in Apcwt/1638N/Sca-1-/- Mice ......................................................... 112
Figure 4.5 Loss of Sca-1 does not impact the formation of epithelial lesions............................ 114
Figure 4.6 Stromal cells with oncogenic mutations have tumor initiating potential .................. 116
viii
List of Tables
Table 2.1 The proportion of tumors that formed from injection of various numbers of cells from
each subpopulation into NOD/SCID mice.................................................................................... 55
Table 3.1 Proportion of tumors formed in NOD/SCID mice from injection of various numbers of
side population and non-side population cells. ............................................................................. 85
ix
List of Abbreviations
ABC ATP binding cassette
APC adenomatous polyposis coli
AF aggressive fibromatosis
AML acute myeloid leukemia
CFU-F colony forming unit-fibroblastic
CML chronic myelogenous leukemia
CSC cancer stem cell
DMSO dimethyl sulfoxide
FAP familial adenomatous polyposis
FBS fetal bovine serum
FIF familial infiltrative fibromatosis
HSC hematopoietic stem cell
LSC leukemic stem cell
MPC mesenchymal progenitor cells
MSC mesenchymal stem cell
OS osteosarcoma
PTCH1 Patched 1
SP side population
SMO smoothened
TIC tumor initiating cell
1
CHATPER 1
INTRODUCTION
The role of cancer stem cells in the initiation and progression of
mesenchymal neoplasms
2
Chapter 1
Introduction
1 The role of cancer stem cells in the initiation and progression of mesenchymal neoplasms
1.1 Abstract
Cancer stem cells (CSCs) are found in multiple tumor types. While the presence of surface
markers selectively expressed on CSCs is used to isolate these cells, no marker or pattern of
makers is known to prospectively identify CSCs in many tumor types, including mesenchymal
neoplasms. In such cases exploitation of stem cell characteristics can be used to identify CSCs,
and one such characteristic is the capacity to extrude dyes such as Hoechst 33342. Cell that
exclude this dye are referred to as side population (SP) cells. These cells share characteristics of
CSCs, specifically, they are enriched for tumor initiating capacity, they express stem-like genes,
and they are resistant to chemotherapeutic drugs. Dye exclusion is a valuable technique as it
identifies a unique population of cells with stem-like characteristics.
3
1.2 Mesenchymal Neoplasms
Mesenchymal neoplasms or sarcomas, unlike epithelial tumors, do not originate from a specific
organ, but rather develop from supporting tissues, such as muscle, fat, fibrous tissue and bone.
Mesenchymal neoplasms can be grouped into two general categories, soft tissue sarcomas and
primary bone sarcomas. In each class, based on histological lines of differentiation, there are
approximately 50 subgroups demonstrating the heterogeneity of this tumor type. In addition to
location of origin and histological appearance, sarcomas can also be broadly classified based on
molecular features into two categories, those with defined diagnostic molecular events and those
with variable complex histological genetic changes[1-3].
The defined molecular event can be a specific point mutation or a translocation event.
Chromosomal translocations constitute the majority of specific genetic alterations associated
with sarcomas. Often, they result in the expression of an oncogenic fusion protein that act as an
abnormal transcription factor deregulating the transcription of multiple downstream genes and
pathways[4]. Fusion gene related sarcomas may account for a third of all sarcomas[5, 6]. Many
specific recurrent chromosomal translocations have been cloned and the resulting fusion genes
identified. Sarcomas affiliated with specific translocation events include alveolar
rabdomyosarcoma, synovial sarcoma, myxoid lipsarcoma, and Ewing’s sarcoma.
The contribution of these fusion proteins in sarcomagenesis supported with the generation of
mouse models for this class of mesenchymal neoplasms. These models also reveal important
information regarding the cell of origin for certain sarcomas. For example, approximately eighty
five percent of alveolar rabdomyosarcoma are associated with a translocation that fuses the Pax
transcription factors, most often Pax3, to a Fkhr head transcription factor[7]. The resultant
protein is believed to act as an oncogene, forcing somatic cells down aberrant embryonic
differentiation. Generation of a Cre/loxP-mediated conditional “knock in” system to insert a
silenced portion of the fkhr gene at the Pax3 locus whereby Pax3 can be normally transcribed
until the Cre recombinase converts Pax3 into the fusion protein. Spatial and temporal regulating
the expression of Pax3:Fkhr reveals expression in late muscle progenitors and not embryonic
cells results in the formation of tumors in small numbers of mice. Tumor frequency could be
increased by the conditional ablation of either p53 or CDKN2A in these mice and produced
neoplasms that phenotypically recapitulated the human disease[8]. Synovial sarcomas are often
4
marked by a translocation leading to the production of the chimeric fusion protein SYT-SSX[9].
Generation of a Cre/loxP-mediated conditional “knock in” whereby the ubiquitous Rosa 26
promoter regulates SYT-SSX expression and Myf5, a promoter specific to myoblasts, drives the
conditional expression of Cre; revealed that expression of the oncogene in committed muscle cell
progenitors results in the formation of synovial sarcoma with 100% penetrance. In contrast,
expression driven by a promoter specific to differentiated muscle cells induces myopathy without
tumor induction[10]. It should be noted that not all mouse models expressing sarcoma associated
translocations result in mesenchymal neoplasms that recapitulate the human disease. For
example, Ewing’s sarcomas are characterized by the presence of the fusion of a portion of the
EWS gene to a segment of one of the ets family of genes. The most common translocation
generating the EWS-FLI-1 fusion protein, believed to act as an aberrant transcriptional activator
contributing to the development of Ewing’s sarcoma[11-13]. Conditional expression of this
protein in the bone marrow and mesenchymal tissue results in the development of leukemia.
While the use of mouse models help in understanding the biologic significance of these fusion
proteins, it is important to note that, not only do these translocations contribute to the molecular
pathogenesis of the disease, they also serve as powerful diagnostic markers for classification[14].
The second broad category of sarcomas, characterized by complex karyotypes, can include
osteosarcomas, leiomyosarcomas, malignant fibrous histocytoma, embryonal
rhabdomyosarcoma, and chondrosarcomas. Inactivation of the p53 pathways appears to be a key
differentiating factor between the two sarcoma classes[15]. Despite their low prevalence,
secondary alterations in the p53 pathway impact the clinical behavior of sarcomas with specific
gene mutations. In contrast, sarcomas with unbalanced karyotypes, mutations in p53 have a
weaker prognostic value, despite a higher prevalence of mutations[16, 17]. It can be
hypothesized, that the p53 pathway is at least partially functional in most sarcomas with specific
translations, possibly acting in concert with the cellular effects of fusion oncoproteins. However,
in sarcomas with nonspecific genetic alterations, p53 pathway inactivation may be a common
early event needed to overcome checkpoints triggered by senescence, telomere erosion, or
double stranded DNA breaks in their progression[3]. The relative timing of common oncogenic
events remains unclear and cannot be readily extrapolated from mouse models.
Despite this, mouse models confirm the importance the p53 pathway in contributing to
sarcomagenesis with p53 knockout mice developing both sarcomas and lymphomas[18, 19].
5
Also, mice harboring specific germline point mutations in p53 develop a spectra of tumors,
including osteosarcomas[20, 21]. Other signaling pathways can act in conjunction with an
aberrant p53 pathway to induce sarcoma formation. For example, expression of the Kras allele in
mice has the potential to induce lung cancer with a high penetrance; however this is not
sufficient to induce sarcoma formation[22]. In contrast, Kras expression in concert with loss of
p53 in muscle tissue results in the formation of poorly differentiated primary soft tissue
sarcomas[23]. Furthermore, in mice, over expression of the Gli2 transcription factor in
chondrocytes results in the development of benign cartilage lesions; however, these lesions
develop into neoplastic entities that resemble chondrosarcomas when combined with a p53
deficiency[24]. This data highlights the importance of the p53 inactivation in sarcomagenesis;
however, given the wide spectrum of sarcoma tumor types that can arise from p53 mutations,
elucidation of both temporal and spatial mechanisms that regulate this process will be required to
provide further insight into this neoplastic process. For example, it has been postulated that
mesenchymal precursors may strongly influence the development of certain sarcomas.
Supporting this notion is the observation in mouse models where the loss of p53 in osteoblast
precursors results in the formation of highly metastatic osteosarcomas with 100%
penetrance[25]. 63% of mice harboring heterozygous germline deletions in NF2 develop highly
differentiated osteosarcomas with high metastatic potential[26]. However, the role of NF2 in the
formation of mouse osteosarcomas is not clear and it should be noted that humans carrying
germline or somatic mutations in this gene do not acquire osteosarcomas, but rather develop
benign Schwann cell neoplasms[27].Also, deletion of PTEN in smooth muscle induced the
formation of leiomyosarcomas in 80% of mice found, in the abdominal wall however no tumors
were found in cardiac muscle despite loss of PTEN in their precursor cells[28]. It is clear given
the broad class and heterogeneity of these tumors that a common cell of origin may not be likely.
Interestingly, ectopic expression of these translocations in mesenchymal stem cells results in the
malignant transformation of these cells and they are able to generate tumors in immuno-deficient
mice suggesting some influence of mesenchymal precursors in the development of sarcomas.
Given the heterogeneity of sarcomas, treatment for those afflicted with these tumors is complex.
Primary presentation with metastasis of soft tissue sarcomas to the lymph nodes is uncommon,
however, when this occurs, prognosis is poor[29, 30]. With the exception of a few, such as
osteosarcoma and Ewing’s sarcoma, generally these tumors do not respond well to conventional
6
chemotherapy. Small, low grade sarcomas, can be treated with surgery and larger or higher grade
lesions can be treated with surgery in conjunction with radiation therapy. While this approach
decreases local recurrence it does not effect overall survival[2]. Furthermore, fibrosis or
impairment in wound healing often result as an effect of radiation therapy[31, 32]. Customized
therapy for some sarcomas for has shown some experimental promise. For example, blockade of
the hedgehog signaling pathway reduces the size of tumor formation in xenografted
chondrosarcomas [33]. Given the cellular diversity within mesenchymal neoplasms,
identification a common modality within this tumor type may and help in developing effective
targeted treatments. As such, the identification of cancer stem cells (CSC) or tumor initiating
cells (TIC) within mesenchymal may be a promising avenue of investigation.
1.3 Cancer Stem Cells
1.3.1 Overview
Despite their monoclonal origins, tumors demonstrate remarkable heterogeneity as evidenced by
marked differences in cellular morphology, kinetic growth properties and expression of cell
surface markers. Perhaps the most striking demonstration of tumor heterogeneity is the differing
capacity of neoplastic cells to initiate de novo tumor formation. In one prominent study
performed in the 1960’s, 35 patients were injected with 1 billion of their own cancer cells.
Remarkably, only seven of the thirty-five patients developed tumors from these autotransplants
[34]. Since then the development of immuno-deficient mice that can tolerate the growth of
human cells has led to a series of experiments validating this initial observation. While
differences in in vivo tumor initiating potential is only one measure of tumor heterogeneity, it
effectively establishes that all cells within a pre-existing tumor are not functionally equivalent.
The cellular mechanisms that underlie tumor heterogeneity are a widely debated issue. The
stochastic model proposes each cell within a tumor is influenced by events, either intrinsic (i.e.
cellular mutations) or extrinsic (i.e. influence of the microenvironment) that confer a
deterministic phenotype to individual cells. In this model every cell within a neoplastic lesion is
biologically equivalent and is influenced by events in a completely random manner. Importantly,
within a tumor each cell has an equal opportunity to be impacted by these events to influence
their cellular phenotype. This random, non-static process results in tumor heterogeneity that can
be assessed in cell surface marker expression, growth kinetics, or the capacity to initiate tumor
7
formation. In contrast, the cancer stem cell model postulates that tumors are similar to normal
tissues in that cells are organized into a cellular hierarchy at the top of which resides the CSC or
tumor initiating cell. These cells, like their normal counterparts, have the potential to self-renew
and to differentiate into all the cell types within heterogeneous a neoplastic lesion. In this model,
cells are not influenced by random events, but rather heterogeneity is conferred by the behavioral
similarities of CSCs to their normal counterparts. The cancer stem cell model proposes that cells
within a lesion are not functionally equivalent, but rather, only a specific subset of cells, the
CSCs, are capable of establishing the heterogeneous phenotype of a tumors. As CSC have the
potential to both self-renew and to differentiate, these cells become clinically important as it is
postulated that similar to normal stem cells that maintain tissue, CSCs are responsible for tumor
growth and maintenance.
1.3.2 CSCs in solid tumors
Human CSCs were originally characterized by their immunophenotype and their ability to
reconstitute whole tumors after serial xenotransplantaion into sub-lethally irradiated immuno-
deficient mice. The presence of CSCs was originally identified in acute myeloid leukemias
(AML). It was demonstrated that only the phenotypically distinct population of AML cells
marked by the expression of CD34+ /CD38- were able to transplant human disease into immuno-
deficient mice. The remaining AML cells were in various stages of differentiation and contained
multiple mature blood cell types with limited proliferative potential. Importantly the transplanted
leukemias closely resembled the disease of the original patients demonstrating the stem like
behavior of these cells [35, 36]. Since this original finding, CSCs have been prospectively
isolated from a broad range of solid tumors. For example, in breast cancer, CSCs have been
identified as a subpopulation of cells that selectively express CD 44high/CD24 low// Lin-. As few as
100 of these cells injected into the mammary gland of mice had the potential to initiate tumor
formation whereas tens of thousands of the negative fraction failed to instigate a malignant
growth when injected into the mammary gland. Importantly these cells had the capacity to form
tumors after secondary and tertiary transplants and the new tumors recapitulated the original
caner [37]. Following this work, CSCs in brain and colon cancers were identified and isolated
based on the selective expression of CD133. In both these malignancies, CD133+ cells have the
ability to initiate tumorigenesis, whereas CD133- cells are incapable of triggering neoplastic
lesions[38-41]. What has followed has been an explosion of similar data in melanoma[42],
8
prostate[43-45], sarcomas[46, 47], head and neck[48], pancreas[49, 50], liver[51] and various
other malignancies thereby highlighting the importance and universality of CSCs in the
neoplastic process.
1.3.3 Characteristics of CSCs
As CSCs are postulated to share characteristics with their normal counterparts, it can be
hypothesized that the two cell types share common cellular and molecular features critical for
regulating self-renewal, proliferation, and differentiation. For example, one family of proteins
that have been shown to be involved in regulating the self-renewal of stem cells is the Polycomb
genes, which repress the expression of their target genes through chromatin modifications. Bmi1
is a member of the Polycomb group protein family and is crucial for the self-renewal of
hematopoietic stem cells and mouse leukemic stem cells[52-54]. Firstly, Bmi1 is highly
expressed in purified mouse and human HSCs[55]. Secondly, although hematopoietic stem cells
are present in normal numbers in the fetal liver of Bmi1-/- mice, they are depleted in the postnatal
bone marrow. Reconstitution experiments indicate the Bmi1 fetal liver cells were only able to
reconstitute primary recipient mice and not secondary recipient mice indicating that the
hematopoietic stem cells are impaired in the potential to self-renew[55]. Thirdly, Bmi-1-/- HSCs
transfected with genes known to induce AML in normal HSCs resulted in leukemias that could
not be serially transplanted suggesting that the polycomb gene is also important in leukemic stem
cell self-renewal[52]. Bmi1 also promotes the proliferation of leukemic stem cells in a mouse
model of AML, as Bmi1 expressing leukemic cells are able to induce leukemia when
transplanted into irradiated mice. Furthermore, in patients with AML, expression of Bmi1 is
higher in AML cells that in normal bone marrow[56]. Bmi1 has also been implicated in self-
renewal so mammary CSCs as down regulation of Bmi1 in mammosphere initiating cells results
in an impairment of these cells to form secondary mammospheres structures by 60% when
compared to control cells[57].
Signaling cascades, such as the Hedgehog and the canonical Wnt pathways, are utilized by both
normal stem cells and CSCs to modulate their behavior. For example, the Wnt/β-catenin
pathway is implicated in regulation of stem cell self-renewal for a variety of tissue systems.
Perhaps the best-characterized function of β-catenin is in the regulation of tissue specific stem
cells within the intestine. Deregulation of this pathway can result in aberrant stem/progenitor cell
9
proliferation and differentiation, often leading to carcinoma formation. Mutations that activate
the Wnt signaling cascade can induce the hyper-proliferation of crypt progenitor cells, generating
benign polyps in which multi-lineage differentiation is evident[58, 59]. Furthermore inhibition of
the pathway induces cell cycle arrest and the expression of differentiation markers in colorectal
cancer cells. A dominant negative form of TCF4 induced cell cycle arrest and the expression of
differentiation markers in colorectal cancer cells in vitro. Taken together, this suggests that β-
catenin may play a role in colon carcinogenesis by initially by increasing the self-renewal of
intestinal crypt stem cells[60, 61]. Expression of stabilized β-catenin also promotes the self-
renewal of central nervous system stem cells keratinocyte stem cells and leads to the
tumorigenesis of the central nervous system and skin. Furthermore, CSCs isolated from
cutaneous cancers are dependent on β-catenin signaling. Using mouse models, Malanchi et al
identified a population of CSC in the mouse skin and ablation of β-catenin in this population of
cells resulted in complete tumor regression[62]. In the hematopoietic system, over-expression of
β-catenin in cultured bone-marrow from mice increase the numbers of stem cells as measured by
the enhanced ability to reconstitute the hematopoietic systems of irradiated mice[63]. Also,
purified WNT3a promotes HSC self-renewal and inhibits the differentiation of HSCs in
culture[64]. In the neoplastic process, In addition, in CML β-catenin has the capacity to
transform committed progenitors into leukemic stem cells[65]. As such it is possible that
deregulation of WNT signaling causes the neoplastic proliferation of normal stem cells by over
activation of their self-renewal program.
Hedgehog signaling also plays an important role in the regulation of both normal and malignant
stem cells. Initial in vitro studies demonstrated that in the presence of Hedgehog antibodies, the
cytokine induced proliferation of HSCs was inhibited. Conversely, activation of the hedgehog
pathway via addition of Sonic hedgehog resulted in the expansion of primitive hematopoeitc
repopulating cells[66]. These in vitro observations are supported by data generated using in vivo
mouse models of this signaling pathway. Ptch+/- mice, which exhibit elevated hedgehog
signaling, show an alteration the proliferation and self-renewal homeostasis of HSCs resulting in
the expansion of hematopoietic stem cells ultimately leading to the exhaustion of HSC numbers
as measured by a diminished long-term engraftment potential after bone marrow
transplantation[66, 67]. Furthermore, mice with the condition deletion of Smo in the
hematopoietic system demonstrate a defect in long term HSC function in primary and secondary
10
transplants. This was not due to impairment in either homing or the loss a specific blood
lineage[68]. However, it should be noted, that the requirement of hedgehog signaling for
hematopoiesis remains controversial, as recent papers demonstrate that contrary to previous
reports, the signaling pathway is dispensable for HSC function. HSC isolated from wild type and
Smo null fetal livers show no differences in HSC function suggesting hedgehog signaling may
have different roles in embryonic HSC function in comparison to adult HSC function. In
hematopoietic malignancies hedgehog signaling was found to be required for the development of
BCR-ABL induced leukemias. Transduction of Smo-/- hematopoietic progenitors with BCR-
ABL1 resulted in a diminished capacity of these cells to induce leukemia. Importantly, this was
not due to effects in homing or engraftments but rather a subsequent loss in the number of
leukemic stem population was detected suggesting hedgehog activity is required for the initiation
of BCR-ABL1 induced leukemias[68, 69].
1.3.4 Clinical significance and therapeutic targeting of CSCs
The presence of embryonic stem cell-like gene expression signatures in human cancers is
associated with aggressive histopathology confirming the clinical-prognostic significance of the
suggests that these cells need to be eliminated in order for disease free survival to occur. As
such, developing treatment strategies that exploit pathways thought to be involved in CSCs self-
renewal and/or differentiation are beginning to emerge. For example, glioblastoma cells express
bone morphogenic proteins (BMP) and their cell surface receptors. BMP treatment of
undifferentiated glioblastoma cells results in reduced cell proliferation and induced
differentiation of cells into mature astrocytes. Moreover, glioma cancer stem cells can be
identified by the expression of the cell surface marker CD133. Importantly, the treatment of
CD133+ glioma cells with BMP reduces the size of tumors implanted into immuno-deficient
mice increasing animal survival. This suggests that BMPs can induce differentiation of CD 133+
glioma stem cells to astrocytes, markedly attenuating their tumor-forming ability[70]. L1CAM
has higher expression in CD133+ glioma cells than in CD133+ normal neural stem cells.
Inhibition of L1CAM exression CD133+ glioma cells resulted in the disruption of neurosphere
formation and the disruption of the growth of the glioma stem cells. Importantly, using shRNA
to reduce L1CAM expression in glioma cells prior to injection into immuno-deficient mice
resulted in decreased tumor growth and increased survival of tumor bearing animals[71].
11
Targeting CSCs with monoclonal antibodies against cell surface markers differentially expressed
on cancer stem cells also shows clinical promise as a treatment strategy. In AML leukemic, but
not normal stem cells exhibit high levels of the IL3 receptor (CD123). Administration of a
CD123 antibody inhibits the engraftment of AML leukemic stem cells in NOD/SCID mice and
also diminishes engraftment of secondary leukemias; suggesting an impact on the self-renewal
capacity of the leukemic stem cells. Importantly, this did not confer cytotoxic effects to normal
stem cells demonstrating specificity to the cancer stem cell population[72]. The use of a
monoclonal antibody specific for the adhesion molecule CD44 in AML led to a marked delay in
the progression of leukemia in mouse models by altering the ability of leukemia stem cells to
traffic to their supportive microenvironments and by altering lineage commitment. Interestingly,
normal hematopoietic stem cells do not appear to rely on this adhesion molecule for their
function to the same extent as leukemais, once again suggesting tumor cell specificity[73, 74]. In
gliomas, the development of a neutralizing monoclonal antibody against the Delta-like 4 ligand
(DLL), which is a member of the notch pathway, inhibits the ability to colon cancer cells to
serial transplant into immuno-deficient mice suggesting that this targets the self-renewing cancer
stem cells[75]. Importantly, treatment with notch inhibitor decreased the frequency of colon
cancer stem cells and as such is believed to impair the capacity for these cells to self- renew.
Not surprisingly, CSCs seem particular well equipped to tolerate external insults making them
relatively resistant to conventional treatments that seemingly target bulk tumors cells. Glioma,
CSCs have an enhanced ability to activate DNA damage response pathways allowing for the
rapid repair of DNA damage caused by radiation which is a major therapeutic modality for this
malignancy. Exposure to non-ionizing radiation of glioma cell cultures and xenografted gliomas
tumors resulted in an increased proportion of CD133 expressing cells. Importantly serial
transplantation of irradiated tumors demonstrated decreased latency in tumor formation, showing
the importance of enrichment of the CD133+ cells. The decreased sensitivity to radiation could
be attributed in part to the increased activation of DNA checkpoint proteins in CD133+ cells in
response to radiation when compared to CD 133- cells. Importantly, in vitro inhibition of these
proteins resulted in a loss of radiation sensitivity in the CD133+ fraction suggesting possible
clinical importance of these findings but this observation was not shown in vivo[76]. However, in
MMTV mice that develop breast cancers, exposure to ionizing radiation results in the enrichment
of breast cancer stem cells as detected by flow cytometry against specific CSC markers[77]. It is
12
important to consider that radioresistance of cancer cells is strongly influenced by the tumor
microenvironment. For example, hypoxic tumors cells are more radioresistant than well-
oxygenated ones[78]. Interestingly, brain cancer stem cells may be enriched in areas of hypoxia
as it has been demonstrated that the fraction of CD133+ medulloblastoma cell lines and primary
tumors increases when cultured at reduced oxygen levels[79, 80]. As hypoxia promotes stem cell
maintenance and blocks differentiation it may play a critical role in by defining both stem cell
maintenance and resistance to radiotherapy. However, other factors may also contribute to the
increased resistance of CSCs to radiation treatment. For example, reactive oxygen species (ROS)
are critical mediators of ionizing radiation induced killing. Examination of both primary human
and mouse breast cancer stem cells reveal that these cells contain lower levels of ROS when
compared to their non CSC counterparts. Treatment of cells to ionizing radiation revealed fewer
DNA strand breaks in CSCs when compared to bulk tumor cells. This observation was attributed
to lower levels of ROS in the CSCs population as depletion of ROS scavengers in the non CSCs
population resulted in increased survival of the cells when treated with ionizing radiation[77].
The impact of targeting CSCs as a means of treatment is in the initial stages of development.
Correlations between CSC activity and clinical outcome will be of critical importance in
determining the relevance of this treatment strategy. Furthermore, most studies have been
preformed by isolating of CSCs from untreated malignancies, however, the main clinical
problem is disease relapse after initial response to therapy, as such, testing of cells after
treatment failure may provide insight into the strength of this treatment strategy.
1.3.5 CSCs: ongoing controversies
The assumption that cancer stem cells represent only a small fraction of tumor cells is
controversial as some studies demonstrate that greater than 10% of tumor cells in transgenic
mouse models of leukemia and lymphoma are capable of initiating tumors in histocompatible
mice[81] . It should be noted that the majority of in vivo data previously generated has taken
advantage of xenograft models in which human tumor cells are injected into immuno-deficient
mice[36-38, 41]. The recent data generated by Kelly et al. suggests that previous work fails to
capture the true frequency of tumor initiating cells as the mouse microenvironment may only be
conducive for the growth of a very small fraction of cells derived from human neoplasms. For
example, using a melanoma model, the frequency of tumor initiating cells was estimated to be
13
0.00083% within metastatic melanomas based on the selection of the ABCG5 transporter
protein. However, Quintana et al, demonstrate that manipulation of the host environment has a
dramatic impact on the capacity of melanoma tumor cells to engraft. Prolonged incubation time,
co-culture of cells with extra-cellular matrix components, and using strains of mice that have
higher degree of immunodeficiency dramatically increased the frequency of melanoma CSCs
[42, 82]. In this work, 27% of implanted, single, unselected melanoma cells had the capacity to
engraft and to form tumors in vivo. Furthermore, expression of cell surface markers that had
previously been shown to enrich for tumorigenic potential failed to do so within this modified
microenvironment, raising the possibility that markers that enrich for rare cells with tumorigenic
potential in NOD/SCID mice may fail to distinguish tumorigenic cells from non tumroigenic
cells in assays that detect higher frequencies of tumorigenic cells. This work suggests that only a
very small proportion of tumor initiating stem cells may be represented in current NOD/SCID
xenograft models. Thus, in some neoplasms, cancer stem cells may not be a small population of
the bulk tumor and these numbers may vary widely depending on the tumor type [83]. This work
highlights the importance of establishing assays that are optimized to capture the stem cell
potential within a neoplasm.
Another source of controversy surrounding the CSC field arises from the assumptions of the
behavior of cancer stem cells in relation to their normal counterparts. It is important to define a
CSC based not on their cell of origin, but rather, on their functional properties, as such, the term
CSC does not reflect the derivation of the cell or imply a normal cell of origin. Furthermore, as
CSCs arise from tumorigenic processes, these cells may not behave in a similar fashion to
normal stem cells. For example, a normal stem cell may respond in to normal external/internal
stimuli maintaining features such as immuophenotype and frequency thereby maintaining a
steady state of a particular stem cell component and their subsequent offspring. However, this
same stability may not hold true for CSCs during the course of a disease. For example, there is a
large variance in the prevalence of leukemia stem cells isolated from multiple independent
samples indicating that the size of the CSC compartment can be variable. In addition, analysis of
leukemia populations with respect to cell surface markers associated with a primitive phenotype
showed variability from patient to patient. As such it is important to define stem cells based on
their functionality and to develop assays that properly capture and read out these stem cell
functions as opposed to the expression of a relatively small set of surface markers which may or
14
may not remain static. In addition, CSCs may themselves represent a heterogeneous population
with varying intrinsic capabilities for self-renewal and differentiation. As such frequency of
CSCs may not be an important feature, but rather, the ability for CSCs to be targeted for
effective treatment may be a more effective endpoint.
However, regardless of the absolute frequency of cancer stem cells present in a given tumor, the
capacity to prospectively identify a subpopulation of cells from a heterogeneous tumor that are
enriched for the capacity to initiate tumor growth, is important. Therefore, regardless of the rarity
of the stem cell population, these CSCs represent a biologically distinct population of cells as
such, their prospective identification may prove to be a means of successful target treatment in
specific tumor types.
1.3.6 Isolation techniques
Prospective identification of CSCs in solid tumors has been propelled by the discovery of
markers that are selectively expressed on CSCs and not on the bulk of the tumor cells. Although
the use of makers does confer specificity to the identification and the isolation of CSCs, this
strategy fails in the absence of known markers. Furthermore, different CSC markers are used for
different tumor types and markers of normal stem cells may not be necessarily successful in the
identification of cancer stem cells. Hence, the discovery of markers that will universally identify
a TIC seems unlikely, and as such, the identification of CSCs using this strategy may prove to be
elusive. An alternative means of CSC isolation exploits manipulation of the differing
characteristics between stem like cells and non-stem cells offers. Furthermore, the universality of
these assays makes them an alluring alternative means of isolation. One such assay, Hoechst dye
exclusion, has proven to be successful on this front.
1.4 Side Population Cells
Hoechst 33342 dye binds to the AT-rich regions of the minor groove of DNA. Fluorescence
intensity is dependant on many factors involved in DNA structure such as chromatin structure,
DNA content and position of the cell within the cell cycle[84, 85]. While uptake occurs
universally in all cells, efflux is less permissive. Cells with the capacity to efflux the dye were
first identified in the mouse bone marrow and they were referred to as side population cells as
they fell to the “side” of the bulk of the positively stained cells in FACS analysis plots[86].
15
Mouse bone marrow SP cells are highly enriched for long term repopulating cells. Since this
original discovery, SP cells have been identified in a variety of tissues[87-89], including skin[90,
91], lung[92, 93], liver[94], heart[95], brain[96], mammary gland[97] and skeletal muscle[98]. In
normal tissues they express high levels of stem-like genes and possess multi-potent
differentiation potential, and as such, they are thought to behave in a similar fashion to stem
cells. In addition, SP cells isolated from mouse bone marrow and muscle share transcriptome
signatures and they under express genes representing tissue specific functions[99].
The mechanism regulating the efflux of Hoechst dye is conferred in part, through the expression
of ATP binding cassette protein (ABC) transporters[100]. Forced expression of these membrane
transporters has direct effects on murine stem cells[101, 102]. However, it should be noted that
Mdr1a/b/Bcrp1 triple knockout mice are viable and still retain some SP cells in the bone
marrow[103]. This suggests that there is either a redundancy in transporter function and/or the
mechanism in which the SP phenotype is determined is not solely conferred through the
expression of ABC transporter proteins. While the exact mechanism of Hoechst dye exclusion
has yet to be fully elucidated, non the less, in adult tissues, SP cells appear to share similar
features to stem like cells.
1.4.1 Side population cells in human neoplasms
SP cells have been identified in a large variety of cancer cell lines with their presence ranging
from 0-20% of the total cell population[46, 104-111]. Given the cancer stem cell hypothesis, it
would be interesting to postulate that the percentage of SP cells within a given cancer cell line
would correlate to its tumorigenicity and/or its aggressiveness; however to date, there has been
no data demonstrating this relationship in cell lines. Also, given that some cell lines lack SP
cells, tumorigenicity, is therefore unlikely to be solely dependant on their presence. In any case,
SP cells contribute to the maintenance and the tumorigenic potential in those cell lines in which
they are present[104, 105, 107-109, 112, 113]. For example, in the C6 glioma cell line, only SP
cells have the capacity to form both SP and non-SP populations suggesting that only SP cells
have the capacity to self-renew and recapitulate the original phenotype of the cell line. In
addition, only SP cells had the ability to grow as neurospheres, a hallmark of neuronal stem cells
and differentiate down the different neuronal lineages[113]. Similar observations have also been
demonstrated in the MCF-7 breast cancer cell line[112]. Interestingly, many tumor derived cell
16
lines contain side population cells and it has been demonstrated that in these cell lines, SP cells
have an increased capacity for self-renewal as measured by the ability to form colonies from
single cells. This suggests that in cell lines, SP cells also have stem cell-like characteristics
similar to those of primary tumors. However, it should be noted, that not all cell lines contain SP
cells hence this population is not exclusively responsible for the prolonged in vitro lifespan of
cell lines. Taken together, this data suggests that in cell lines, SP cells do have characteristics
similar to stem cells.
While much of the work examining the presence of SP cells in neoplasms has been through the
use of existing cancer cell lines, they have been also discovered in primary tumors. Specifically,
they have been detected in primary neuroblastomas[111], the ascites of ovarian cancers[106],
and in a wide range of mesenchymal neoplasms[46]. Interestingly, in primary mesenchymal
tumors, there exists a correlation between tumor grade and the percentage of SP cells
present[46]. Hence, the SP percentage may potentially be a valuable prognostic indicator in this
tumor type.
Paradoxically, SP cells are present in primary mesenchymal neoplasms, including primary
osteosarcomas; however, both SaOS and U2OS osteosarcoma cell lines do not appear to contain
them[111]. This can be explained, in part, by the fact that cell lines only very crudely represent
the true in vivo nature of tumors. For example, culture conditions used in these assays may not
completely recapitulate the in vivo conditions required to detect SP cells. Notably, the presence
of the specific growth factors fibroblastic growth factor (FGF)and basic fibroblastic growth
factor (bFGF) and a lack of serum were required to enhance the detection SP cells in the C6
glioma cell line[113]. In addition, one hallmark of osteosarcoma tumors is their notoriously
heterogeneous karyotypes, both within cells of the same tumors and in comparison to cells from
different osteosarcoma tumors[114]. In contrast, cell lines lack this feature and as such, this may
account for the differences seen in the presence of SP cells between the primary tumors and that
of the cell lines.
1.4.2 Cellular phenotype
The cellular phenotype of SP cells in mice has been characterized in a variety of tissues,
including, but not exclusive to the bone marrow, skeletal muscle, mammary gland, testis, and
skin. SP cells from these tissues highly express stem cells markers such as Sca-1, and CD
17
34[86]. Unfortunately, to date, data on the cellular phenotype of SP cells isolated from tumors
has not been extensively studied. In human neuroblastomas, SP cells were shown to be negative
for the hematopoietic marker, CD45. Also, compared to non-SP cells, SP cells from
neuroblatsomas had increased expression of C-kit/CD117 and had lowered expression of
AC133/CD71 and CD56[111]. This staining pattern is similar to the cellular phenotype of neural
crest progenitor cells, suggesting that SP cells from these tumors may have stem cell
characteristics[111].
1.4.3 Tumorigenic potential
Data from several independent laboratories have demonstrated that when compared to both the
bulk tumor cell populations and to the non-SP population, SP cells isolated from
hepatocellular[104], lung[107], gastric[109], and nasopharyngeal carcinoma[108] cell lines are
highly enriched for the capacity to initiate tumor formation when xenografted into NOD/SCID
mice. Importantly, this observation has also been demonstrated in primary mesenchymal tumors.
However, in this neoplasm non-SP cells initially do have the capacity to form tumors in
NOD/SCID mice, but only SP cells have the capacity to initiate tumors upon serial
transplantation. Interestingly, in primary mesenchymal tumors, cells derived from non-SP tumors
have increased DNA content when compared to cells derived from SP tumors. This observation
has also been noted in primary ovarian neoplasms[46]. Taken together, this data suggests that the
non-SP fraction may contain a population of transiently amplifying cells that have the ability to
initially form tumors through rapid proliferation. However, these cells do not have the capacity
to self-renew and therefore cannot sustain tumor initiation upon serial transplantation.
Importantly, in primary mesenchymal tumors cells from non-SP tumors only gave rise to non-SP
cells and were therefore unable to recapitulate the original tumor phenotype[46]. Thus, similar to
CSCs, only SP cells can self renew and differentiate.
1.4.4 Expression of stem-like genes
In comparison to non-SP cells, SP cells have increased expression of genes that that are believed
to be involved in the regulation of stem cell function. Using microarray analysis and validation
with RT-PCR analysis, SP cells from MCF-7 breast cancer[112], hepatocellular[104],
gastrointestinal[109], and thyroid[105] cancer cell lines have shown to be up-regulated in the
expression of the ABCG2 transporter when compared to the non-SP cells. In addition, SP cells
18
from both colon carcinoma and breast carcinoma cell lines have increased expression of genes
involved in the WNT/beta-catenin signaling pathway when compared to non-SP cells[109]. This
pathway has been shown to be involved the self-renewal of hematopoietic stem cells, and as
such, may play a similar role in SP cells[63, 64]. Recent data from the MCF-7 breast carcinoma
cell line demonstrates that in comparison to the non-SP cells, SP cells show an increase in the
expression of genes involved in cell cycle regulation, including EXT1, INHBA and CCNT2.
Furthermore, SP cells were shown to have more cells in G1/G0 than non-SP cells. In MCF-7
cells, SP cells are also up regulated in genes belonging to PI3K/AKT pathway. Interestingly, in
the presence of the PI3K inhibitor LY4498 there was a decrease in the percentage of SP cells.
Furthermore, inhibition resulted in a decreased ability of in vitro colony formation and in vivo
tumor formation[115]. This suggests that identifying signaling pathways upregulated in SP cells
may prove to be useful strategies for therapy. However, it should be noted that these array
studies were based on cancer cell lines that have been maintained in culture over long periods of
time and as such, they may not recapitulate the true tumor phenotype. Microarray data stemming
from primary tumors should shed some insight into the expression of “stemness” genes in SP
cells.
Interestingly, recent work has shown that in breast and colon cancers a relatively small number
of signaling pathways are disrupted at a high frequency while a large number of pathways are
disrupted at a low frequency [116]. Given these findings, it would be interesting to determine if
signaling pathways disrupted in the SP fraction differ from those in the bulk of the tumor cells
and as such, this may account for the differences in tumor initiating potential between the two
populations of cells.
1.4.5 Drug efflux
Even when tumors appear to be eradicated by chemotherapy, relapse often occurs. One
hypothesis is that cancer stem cells have the capacity to elude such treatments and as such,
remain viable and are therefore responsible for disease reoccurrence. One mechanism by which
this may arise is through the expression of ABC transporter proteins, as they efflux lipophillic
chemotherapeutic agents such as doxorubicin[117]. SP cells from a variety of mouse tissues have
increased expression of ABC transporters in comparison to non-SP cells, as do cancer cell lines
and cells derived from primary tumors[104, 105, 109, 112]. Expression of these proteins may be
19
one mechanism by which tumorigenic potential is conferred upon SP cells, as such; these
proteins would make ideal targets for cancer therapy. Unfortunately, to date, a correlation
between the tumor initiating potential of neoplasms and the expression of ABC transporters has
not been demonstrated. In fact, ABCG2 positive MCF-7 cells showed no more tumorigenic
potential than ABCG2 negative cells[112]. Also, in microarray analysis of over 500 soft tissue
sarcomas no correlation between the expression of ABCG1 transporter and tumor grade was
observed[118]. However, this data does not rule out the possibility of the existence of a
relationship between the two. As in the case with CSC markers, transporter expression and
function may be diverse amongst different tumor types. For example, in primary mesenchymal
tumors, while SP cells are capable of forming tumors in immunodeficient mice, Rhodaminelow
cells are not[46]. Efflux of Rhodamine-123 dye is largely conferred by the expression of the
ABCG1 transporter, and as such, ABCG2 may play the critical role in tumor initiation in soft
tissue sarcomas but not in breast carcinomas. Furthermore, Mdr1a/b/Bcrp1 triple knockout mice
are viable and although diminished, SP cells are still present in the bone marrow[100]. It is
therefore likely that multiple transporters are expressed and responsible for the SP phenotype in
any given tumor. In those tumors that are particularly resistant to chemotherapeutic agents,
transporter expression may strongly correlate to the percentage of SP cells present and these
tumors may be ideal candidates for targeted treatment of such transporters.
1.5 The Origins of CSCs
The prospective isolation of cancer stem cells demonstrates that within tumors there resides a
subpopulation of cells with stem-like characteristics; however, their cellular origins cannot be
extrapolated from these experiments. It is not clear whether cancer stem cells arise from
mutations in either normal stem cells or more differentiated progenitors. The cellular longevity
allows stem cells to be subjected to the acquisition of multiple genetic abnormalities required for
tumorigenesis; however, they are relatively quiescent they may not undergo sufficient number of
cell divisions to become a neoplastic entity. Alternatively, differentiated cells can acquire
mutations giving them the capacity to self-renew and differentiate, thereby developing stem-like
features (Figure 1.1).
Evidence from both leukemia and brain tumors suggests that mutations in both normal stem cells
and differentiated progenitors can result in tumorigenesis indicating both sources can act as
20
potential candidates from which cancer stem cells can arise. In myoproliferative disorders,
overexpression of BCR-ABL will only induce tumorigenesis in stem-like cells and not
progenitors. As expression of oncogenes did not confer self-renewal properties to committed
progenitors it can be concluded that these mutations must impact a population of cells with the
pre-existing capacity to self renew[119]. However, the expression of MLL fusions in committed
myeloid progenitors can generate AML upon transplantation into immune-deficient mice[120].
Mouse models of neurofibromas indicate that loss of the tumor suppressor neurofibromin (NF1)
transforms committed progenitors rather than stem cells. Furthermore, conditional deletion of
NF1 at different stages of development demonstrates that committed Schwann cell progenitors,
proliferate and predominately contribute to neurofibroma formation [121, 122]. However,
another study demonstrates that NF-1 deficient skin derived precursors can give rise to
neurofibromas[123]. In prostate cancer, genetic lineage marking demonstrates that rare luminal
epithelial can act as stem cells and deletion of PTEN in these cells results in carcinoma
formation[119]. In the gastrointestinal tract, Lgr+ cells mark normal stem cell in both the
proximal and distal sites[124]. Conditional ablation of the APC tumor suppressor gene, which
mediates the development of colorectal cancer, from Lgr+ intestinal stem cells results in the
formation of adenomas in the small intestine and colon. In contrast ablation of APC in transiently
amplifying cells did not result in neoplastic transformation[125].
The cellular origin of mesenchymal neoplasms has not been identified; however, it has been
postulated that they may arise from mutations in mesenchymal progenitor cells.
1.6 Mesenchymal Progenitor Cells
1.6.1 Overview
The identification of cells with the potential to differentiate down the various lineages of
mesenchymal tissue was first described in the bone marrow. Originally these cells were used as
feeder layers to promote the growth of hematopoietic stem cells, however it became apparent
that these stromal cells had stem like capacities[126]. Specifically, these cells had capacity to
form adherent colonies in culture and were able to differentiate into cells of bone, cartilage, and
fat demonstrating multipotent potential[127]. Since the discovery of multi-potent mesenchymal
progenitors in the bone marrow, cells with similar characteristics have been identified from
various adult and fetal tissues. However, it has yet to be determined whether these tissues arise
21
from a common precursor such as a mesenchymal stem/progenitor cell as the isolation of a single
clonogenic self-renewing cell that can generate on one or more of the specialized cell types that
constitutes mesenchymal tissue has yet to be definitively isolated. This work is largely hampered
by the lack of cell surface markers or pattern of markers known to identify a mesenchymal stem
cell (MSC).
1.6.2 Defining the mesenchymal stem/progenitor cell
The ambiguity regarding the existence of the MSC stems from the lack of identifying markers,
various locations of isolation, and disparity in methods used to culture and expand cells. This has
resulted in the isolation of a heterogeneous population of “MSC” that exhibit variable
phenotypes, and as such, hampers the precision of defining such a population of cells. Currently,
the commonly accepted tests for the identification of MSCs include the capacity to form colony
forming unit fibroblastic (CFU-F) in culture, analysis of surface marker profiles, and multi-
lineage potential, particularly oteogenesis, chondrogenesis and adipogeneiss[128]. As mentioned,
to date no single marker or pattern of markers is known to isolate a clonogenic MSC. Profiling of
MSC surface antigen expression demonstrates some consistency amongst the differing
populations of MSCs. To date, the consensus for marker profile by the International Society for
Cellular Therapy establishes that by flow cytometry, MSC express CD73, CD90 (Thy-1) and
CD105. Other markers know to be expressed on MSCS are CD49a and STRO-1. While STRO-1
is the best known MSC marker, it is not exclusive to these cells and is lost during culture[129].
CD146 in conjunction with NG2 and PDGF2, identifies pericytes, a population of sub-
endothelial cells ability to differentiate down mesodermal lineages and is another putative
marker that shows promise in the identification of MSCs.
The majority of work identifying MSCs has focuses on the multi-potent abilities of isolated cells.
However, by definition, a stem cell is required to self-renew to create more stem cells. While
CFU-F are capable of forming colonies, clonal expansion does not equate to self-renewing
potential. In vitro work demonstrating the self-renewal potential of clonogenic MSC reveal that
these cell are organized into a cellular hierarchy in which differentiation down multi lineages
occurs in a regulated fashion. These clones had the capacity to self-renew, demonstrating in
vitro, isolation of a single clonogenic MSC[130].
22
One challenge of identifying MSC such cells is that by definition these cells must have the
capacity to regenerate their tissue of origin. This has best been exemplified in the hematopoietic
system in which a single HSC has the capacity to reconstitute the entire blood system. However,
in other tissue types this becomes a more challenging question. Currently, the strongest evidence
for the existence of a mesenchymal progenitor cells is provided by experiments demonstrating
muscle satellite cells, which have the capacity to differentiate into cells of various mesenchymal
lineages, also have the capacity to regenerate in vivo muscle tissue in mouse models of
Duchanne’s muscular dystrophy [131].
1.6.3 Location of mesenchymal stem cells
Adult, tissue specific stem cells are found in specialized niches in their corresponding tissues of
origin. For example, hematopoietic stem cells can be found in the bone marrow, epidermal stem
cells in mammalian hair follicles, intestinal stem cell in the intestinal crypts[132]. However, by
current definitions, MSCs can be found throughout an entire organism[133]. This phenomenon
can be occur by several mechanisms. For example, MSCs may reach all these areas by
circulating through the blood system from a primary source; however, the difficulty in isolating
of MSCs from peripheral blood argues against this possibility[134, 135]. A second possibility,
based on the fact that postnatal MSC have been isolated from different tissues, is that each tissue
type possesses tissue specific MSC with intrinsic stem cell like properties when characterized in
vitro. Thirdly, there has been increasing evidence that an intimate relationship between MSCs
and perivascular cells[136]. Furthermore, this would explain the ability to isolate mesenchymal
progenitor cells from a wide range of tissues and the establishment of MSC-like cultures from
blood vessels supports this hypothesis[133]. As such the pericyte may be an in vivo source of
MSCs.
1.6.4 Are pericytes mesenchymal stem cells?
Pericytes are defined morphologically on their basis of their location in relation to endothelial
cells. They are located on the abluminal side of blood vessels, immediately opposed to
endothelial cells[137]. Tissue sections stained with an antibody specific to periciytes
demonstrates their presence in both small and large blood vessels, suggesting that they may form
a subendothelial network spanning the vasculature[138]. Both pericytes and MSC express similar
cell surface markers, and interestingly, there is also evidence that suggesting that pericytes have
23
the in vivo potential to act as progenitors for adipocytes, cartilage and bone after injury. Further
supporting this relationship between MSCs and pericytes, is the identification of markers that
allows for the prospective isolation of pericyets from a variety of tissue types. Using these
markers, Crisan et al were able to identify clones that were multi-potent for osteogenic,
chondrogenic, and myogenic lineages in vitro from multiple organs including skeletal muscle,
pancreas, and adipose tissue. Thus blood vessel walls may harbor a reserve of progenitor cells
that may be central to the origins of MSC[139].
1.7 Mesenchymal Progenitor Cells as the Sarcoma Cell of Origin
Given the cellular diversity of mesenchymal neoplasms, it has been postulated that multi-potent
MSCs are their cell of origin. The demonstration that some neoplasms can arise from oncogenic
mutations in normal stem cells, aids in supporting this hypothesis. Recent publications have
begun to address this question in Ewing’s sarcoma and in malignant fibrous histiocytoma
(MFH), both of which are mesenchymal tumors. Specifically, it has been shown that over-
expression of the EWS-FL-1 fusion protein, the transforming event in Ewing’s sarcoma, results
in the transformation of primary bone marrow derived mesenchymal progenitor cells.
Furthermore, these cells generate tumors that display the hallmarks of Ewing’s sarcoma [140].
Tirode et al. demonstrate shRNA silencing of the EWS-FL1 gene in Ewing’s sarcoma cell lines
results in the generation of a transcriptional profile similar to mesenchymal stem cells. These
silenced cells also recovered the phenotype of mesenchymal stem cells in that they had the
potential to differentiate into both adipocytes and osteoblasts[141]. Finally, it has been shown
that human mesenchymal stem cells can be transformed, via inhibition of the Wnt/β-catenin
signaling pathway, to form MFHs[142]. Taken together this suggests that mesenchymal
neoplasms may indeed originate from mesenchymal progenitor cells.
1.8 Aggressive Fibromatosis
1.8.1 Overview
Aggressive fibromatosis (AF: also called desmoids tumor) is a rare, benign neoplasm
representing 0.03 to 0.1% of all tumors and 3.5% of fibrous tissue tumors with an occurrence
frequency of approximately 2-4 cases per million population[143]. Tumors are characteristically
24
slow growing, although a wide variation exists, and is often associated with local invasion of fat
and muscle with infiltration of neurovascular structures. Metastasis is exceedingly rare[144].
Aggressive fibromatosis are benign lesions of fibrous growth that are believed to arise from
deregulation of connective tissue growth[145]. The lesions lack a capsule and are characterized
by poorly demarcated basses that blend into the surrounding tissue[146, 147]. Histopatholocially,
these tumors are composed of bipolar cells that stain positively for the intermediate filament
vimentin but do not express the epithelial marker keratin; strongly resembling well differentiated
fibroblasts[148]. The anatomical location, cellular morphology, and histological profile of these
lesions suggest that they are of mesenchymal origins.
There are several therapeutic options for the management of AF including surgical excision,
radiation and pharmacological therapy. Overall, however, the effectiveness of these approaches
is limited and recurrence rates are high[149]. The primary therapy of choice is wide local
surgical excision to a normal tissue margins, but his approach is limited to mainly extremity and
small mesenteric AF[150]. Complete excision of large intra-abdominal AF is more difficult as it
may involve sacrificing critical structures. Local recurrence rates are high, despite
microscopically clear surgical margins[151].
1.8.2 Molecular etiology of aggressive fibromatosis
In children, AF is found at the highest frequency in the extremities, but can also be located in
intro abdominal regions. Most AF in children occurs sporadically. The molecular etiology of
aggressive fibromatosis is well characterized. In adults, patients with autosomal dominant
hereditary condition know as familial adenomatous polyposis (FAP) have a 1000-fold higher
susceptibility to developing AF[152, 153]. The adenomatous polyposis coli (APC) tumor
suppressor gene is mutated in FAP patients leading to the truncation and loss of function of the
APC protein. The frequency of occurrence of AF in FAP patients varies in adults between
different studies ranging from 3.6-34%[143, 152, 154]. Mutations in β-catenin protein that result
in the up-regulation of the β-catenin/WNT signaling pathway is a consistent feature of both
sporadic and FAP-associated AF tumors. In sporadic cases, most tumors contain a mutation in
CTNNB1, the gene that codes for β-catenin, resulting in protein elevation are present in 75% of
cases[145, 155, 156]. Regardless of the causative mutation, one unifying factor in AF
pathogenesis is the elevation of the β-catenin/WNT signaling pathway.
25
1.8.3 Mouse models of aggressive fibromatosis
The Apc1638N mouse is a well-characterized mouse that closely approximates human FAP. The
APC1683N mice carries a targeted mutation at codon 1638 of the mouse Apc gene. This
mutation, a targeted frameshift at codon 1638 of the mouse Apc gene, represents a null allele and
leads to haploinsufficiency in heterozygous animals. These mice begin to develop
gastrointestinal and desmoids tumors by two months of age[157, 158]. As such this mouse model
provides an excellent tool from which we can investigate the role of mesenchymal progenitor
cells in the development and progression of aggressive fibromatosis.
1.9 Summary and conclusions
Cancer stem cells have been identified in many human malignancies. Traditionally, isolation is
based on the selective expression of cell surface markers; however, alternative methods founded
on functional characteristics of stem cells can also be used. For example, the side population
assay, based on the active efflux of the fluorescent dye Hoechst 33342, allows for the
prospective isolation of CSCs from many neoplasms. Importantly, identification of cells with
stem like characteristics within neoplastic lesions does not suggest that they arise from normal
stem cells harboring oncogenic mutations. The use of mouse models have begun to address the
cellular origins of neoplasms with recent evidence suggests both stem cells and committed
progenitors have the potential to induce tumorigenesis.
The mechanisms underlying the pathogenesis of mesenchymal neoplasm is poorly understood.
These malignancies may model the characteristics described in hematological, neuronal, and
epithelial tumors. The identification of the biological processes that give rise to mesenchymal
lesions can potentially aid in the development of new therapies.
26
1.10 Thesis Summary and Rationale
The goal of this research is to investigate the development and progression of mesenchymal
neoplasms. We hypothesize that, similar to other solid malignancies, mesenchymal neoplasms
contain a subpopulation of cells that are enriched for tumor initiating potential. Furthermore, as
these tumors may be of mesodermal origins, we postulate that normal mesenchymal progenitor
cells can influence the development of certain mesenchymal neoplasms.
To examine these hypotheses, the following questions will be addressed:
1) Do mesenchymal neoplasms contain a subpopulation of cells with enhanced tumor
initiating potential? As no known markers, or pattern of markers are known to
distinguish a mesenchymal progenitor cell, we use the side population assay to identify
and prospectively isolate tumor-initiating cells within mesenchymal neoplasms (Chapter
2).
2) Can we identify a biological feature that can be exploited and used as a therapeutic
target to diminish the tumorigenic potential of mesenchymal side population cells?
Deregulation of the hedgehog signaling pathway is associated with a broad range of
cancers, including mesenchymal neoplasms. Furthermore, blockade of this pathway is an
effective therapeutic treatment modality for a variety of tumor types. Given these
observations, we investigated how chemical modulation of hedgehog signaling would
impact the behavior of tumors derived from SP cells isolated from osteosarcoma cell
lines (Chapter 3).
3) Can normal mesenchymal progenitor cells influence the development of the
mesenchymal neoplasm, aggressive fibromatosis? The cellular origins of tumors are
poorly defined. Through the use of mouse models we sought to determine if altering the
numbers of mesenchymal progenitor cells would impact the development of aggressive
fibromatosis in mice predisposed to forming these tumors. Furthermore we questioned
whether mesenchymal progenitor cells harboring a mutation associated with aggressive
fibromatosis would have tumorigenic potential. (Chapter 4).
27
The identification of a subpopulation of cells within mesenchymal neoplasms enriched for tumor
initiating potential, reveals a previously unknown aspect of this tumor type and provides a
powerful tool for further investigation into mesenchymal tumorigenesis. Further elucidation of
the molecular mechanisms, conferring this phenotype will uncover important information and
will aid in the development of effective treatment modalities targeted against what may be the
most potent malignant cells. Finally, the identification of progenitor cells with the capacity to
influence the development of mesenchymal neoplasms such as aggressive fibromatosis raises the
intriguing possibility that protecting these cells in patients with known genetic predispositions to
tumor development, such as familial adenomatous polyposis patients, can prevent the
development of these neoplastic lesions. These studies highlight the importance of understanding
mesenchymal progenitor cell biology in the development of potential new treatments.
28
Figure 1.1 Cellular origins of cancer stem cells
Tumors are composed of a heterogeneous population of cells from which cancer stem cells can
be prospectively isolated; however, the origins of these cells remain unclear. It has been
postulated that cancer stem cells can arise from normal stem cells harboring oncogenic
mutations. Conversely, it is hypothesized that arise from committed progenitor cells that that
acquire stem like characteristics such as the ability to self-renew.
29
!"#$%&'()*+*(,&-.//&
0)1.%.(*+2.3&-.//&4$%#+/&52.#&&
-.//&
30
1.11 References
1. Pollock, R.E., The seminal role of Cancer in our understanding of sarcoma. Cancer, 2008. 113(7 Suppl): p. 1969-79.
2. Skubitz, K.M. and D.R. D'Adamo, Sarcoma. Mayo Clin Proc, 2007. 82(11): p. 1409-32.
3. Borden, E.C., et al., Soft tissue sarcomas of adults: state of the translational science. Clin Cancer Res, 2003. 9(6): p. 1941-56.
4. Mitelman, F., B. Johansson, and F. Mertens, The impact of translocations and gene fusions on cancer causation. Nat Rev Cancer, 2007. 7(4): p. 233-45.
5. Mitelman, F., Recurrent chromosome aberrations in cancer. Mutat Res, 2000. 462(2-3): p. 247-53.
6. Bennicelli, J.L. and F.G. Barr, Genetics and the biologic basis of sarcomas. Curr Opin Oncol, 1999. 11(4): p. 267-74.
7. Sorensen, P.H., et al., PAX3-FKHR and PAX7-FKHR gene fusions are prognostic indicators in alveolar rhabdomyosarcoma: a report from the children's oncology group. J Clin Oncol, 2002. 20(11): p. 2672-9.
8. Keller, C., et al., Pax3:Fkhr interferes with embryonic Pax3 and Pax7 function: implications for alveolar rhabdomyosarcoma cell of origin. Genes Dev, 2004. 18(21): p. 2608-13.
9. Clark, J., et al., Identification of novel genes, SYT and SSX, involved in the t(X;18)(p11.2;q11.2) translocation found in human synovial sarcoma. Nat Genet, 1994. 7(4): p. 502-8.
10. Haldar, M., et al., A conditional mouse model of synovial sarcoma: insights into a myogenic origin. Cancer Cell, 2007. 11(4): p. 375-88.
11. Arvand, A. and C.T. Denny, Biology of EWS/ETS fusions in Ewing's family tumors. Oncogene, 2001. 20(40): p. 5747-54.
12. Delattre, O., et al., Gene fusion with an ETS DNA-binding domain caused by chromosome translocation in human tumours. Nature, 1992. 359(6391): p. 162-5.
13. May, W.A., et al., The Ewing's sarcoma EWS/FLI-1 fusion gene encodes a more potent transcriptional activator and is a more powerful transforming gene than FLI-1. Mol Cell Biol, 1993. 13(12): p. 7393-8.
14. Ladanyi, M. and J.A. Bridge, Contribution of molecular genetic data to the classification of sarcomas. Hum Pathol, 2000. 31(5): p. 532-8.
31
15. Marina, N., et al., Biology and therapeutic advances for pediatric osteosarcoma. Oncologist, 2004. 9(4): p. 422-41.
16. Drobnjak, M., et al., Prognostic implications of p53 nuclear overexpression and high proliferation index of Ki-67 in adult soft-tissue sarcomas. J Natl Cancer Inst, 1994. 86(7): p. 549-54.
17. Wurl, P., et al., Clinical relevance of pRb and p53 co-overexpression in soft tissue sarcomas. Cancer Lett, 1999. 139(2): p. 159-65.
18. Williams, B.O., et al., Cooperative tumorigenic effects of germline mutations in Rb and p53. Nat Genet, 1994. 7(4): p. 480-4.
19. Purdie, C.A., et al., Tumour incidence, spectrum and ploidy in mice with a large deletion in the p53 gene. Oncogene, 1994. 9(2): p. 603-9.
20. Olive, K.P., et al., Mutant p53 gain of function in two mouse models of Li-Fraumeni syndrome. Cell, 2004. 119(6): p. 847-60.
21. Lang, G.A., et al., Gain of function of a p53 hot spot mutation in a mouse model of Li-Fraumeni syndrome. Cell, 2004. 119(6): p. 861-72.
22. Tuveson, D.A., et al., Endogenous oncogenic K-ras(G12D) stimulates proliferation and widespread neoplastic and developmental defects. Cancer Cell, 2004. 5(4): p. 375-87.
23. Kirsch, D.G., et al., A spatially and temporally restricted mouse model of soft tissue sarcoma. Nat Med, 2007. 13(8): p. 992-7.
24. Ho, L., et al., Gli2 and p53 cooperate to regulate IGFBP-3- mediated chondrocyte apoptosis in the progression from benign to malignant cartilage tumors. Cancer Cell, 2009. 16(2): p. 126-36.
25. Berman, S.D., et al., Metastatic osteosarcoma induced by inactivation of Rb and p53 in the osteoblast lineage. Proc Natl Acad Sci U S A, 2008. 105(33): p. 11851-6.
26. McClatchey, A.I., et al., Mice heterozygous for a mutation at the Nf2 tumor suppressor locus develop a range of highly metastatic tumors. Genes Dev, 1998. 12(8): p. 1121-33.
27. Gutmann, D.H. and M. Giovannini, Mouse models of neurofibromatosis 1 and 2. Neoplasia, 2002. 4(4): p. 279-90.
28. Hernando, E., et al., The AKT-mTOR pathway plays a critical role in the development of leiomyosarcomas. Nat Med, 2007. 13(6): p. 748-53.
29. Fong, Y., et al., Lymph node metastasis from soft tissue sarcoma in adults. Analysis of data from a prospective database of 1772 sarcoma patients. Ann Surg, 1993. 217(1): p. 72-7.
32
30. Wunder, J.S., et al., Opportunities for improving the therapeutic ratio for patients with sarcoma. Lancet Oncol, 2007. 8(6): p. 513-24.
31. Davis, A.M., et al., Late radiation morbidity following randomization to preoperative versus postoperative radiotherapy in extremity soft tissue sarcoma. Radiother Oncol, 2005. 75(1): p. 48-53.
32. Akudugu, J.M., et al., Wound healing morbidity in STS patients treated with preoperative radiotherapy in relation to in vitro skin fibroblast radiosensitivity, proliferative capacity and TGF-beta activity. Radiother Oncol, 2006. 78(1): p. 17-26.
33. Tiet, T.D., et al., Constitutive hedgehog signaling in chondrosarcoma up-regulates tumor cell proliferation. Am J Pathol, 2006. 168(1): p. 321-30.
34. Southam, C.M., et al., Effect of leukocytes on transplantability of human cancer. Cancer, 1966. 19(11): p. 1743-53.
35. Lapidot, T., et al., A cell initiating human acute myeloid leukaemia after transplantation into SCID mice. Nature, 1994. 367(6464): p. 645-8.
36. Bonnet, D. and J.E. Dick, Human acute myeloid leukemia is organized as a hierarchy that originates from a primitive hematopoietic cell. Nat Med, 1997. 3(7): p. 730-7.
37. Al-Hajj, M., et al., Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S A, 2003. 100(7): p. 3983-8.
38. O'Brien, C.A., et al., A human colon cancer cell capable of initiating tumour growth in immunodeficient mice. Nature, 2007. 445(7123): p. 106-10.
39. Ricci-Vitiani, L., et al., Identification and expansion of human colon-cancer-initiating cells. Nature, 2007. 445(7123): p. 111-5.
40. Singh, S.K., et al., Identification of a cancer stem cell in human brain tumors. Cancer Res, 2003. 63(18): p. 5821-8.
41. Singh, S.K., et al., Identification of human brain tumour initiating cells. Nature, 2004. 432(7015): p. 396-401.
42. Schatton, T., et al., Identification of cells initiating human melanomas. Nature, 2008. 451(7176): p. 345-9.
43. Patrawala, L., et al., Highly purified CD44+ prostate cancer cells from xenograft human tumors are enriched in tumorigenic and metastatic progenitor cells. Oncogene, 2006. 25(12): p. 1696-708.
44. Patrawala, L., et al., Hierarchical organization of prostate cancer cells in xenograft tumors: the CD44+alpha2beta1+ cell population is enriched in tumor-initiating cells. Cancer Res, 2007. 67(14): p. 6796-805.
33
45. Collins, A.T., et al., Prospective identification of tumorigenic prostate cancer stem cells. Cancer Res, 2005. 65(23): p. 10946-51.
46. Wu, C., et al., Side population cells isolated from mesenchymal neoplasms have tumor initiating potential. Cancer Res, 2007. 67(17): p. 8216-22.
47. Levings, P.P., et al., Expression of an exogenous human Oct-4 promoter identifies tumor-initiating cells in osteosarcoma. Cancer Res, 2009. 69(14): p. 5648-55.
48. Prince, M.E., et al., Identification of a subpopulation of cells with cancer stem cell properties in head and neck squamous cell carcinoma. Proc Natl Acad Sci U S A, 2007. 104(3): p. 973-8.
49. Li, C., et al., Identification of pancreatic cancer stem cells. Cancer Res, 2007. 67(3): p. 1030-7.
50. Hermann, P.C., et al., Distinct populations of cancer stem cells determine tumor growth and metastatic activity in human pancreatic cancer. Cell Stem Cell, 2007. 1(3): p. 313-23.
51. Yang, Z.F., et al., Significance of CD90+ cancer stem cells in human liver cancer. Cancer Cell, 2008. 13(2): p. 153-66.
52. Lessard, J. and G. Sauvageau, Bmi-1 determines the proliferative capacity of normal and leukaemic stem cells. Nature, 2003. 423(6937): p. 255-60.
53. Lessard, J. and G. Sauvageau, Polycomb group genes as epigenetic regulators of normal and leukemic hemopoiesis. Exp Hematol, 2003. 31(7): p. 567-85.
54. Leung, C., et al., Bmi1 is essential for cerebellar development and is overexpressed in human medulloblastomas. Nature, 2004. 428(6980): p. 337-41.
55. Park, I.K., et al., Bmi-1 is required for maintenance of adult self-renewing haematopoietic stem cells. Nature, 2003. 423(6937): p. 302-5.
56. Sawa, M., et al., BMI-1 is highly expressed in M0-subtype acute myeloid leukemia. Int J Hematol, 2005. 82(1): p. 42-7.
57. Liu, S., et al., Hedgehog signaling and Bmi-1 regulate self-renewal of normal and malignant human mammary stem cells. Cancer Res, 2006. 66(12): p. 6063-71.
58. Powell, S.M., et al., APC mutations occur early during colorectal tumorigenesis. Nature, 1992. 359(6392): p. 235-7.
59. Moser, A.R., et al., The Min (multiple intestinal neoplasia) mutation: its effect on gut epithelial cell differentiation and interaction with a modifier system. J Cell Biol, 1992. 116(6): p. 1517-26.
34
60. Kinzler, K.W. and B. Vogelstein, Lessons from hereditary colorectal cancer. Cell, 1996. 87(2): p. 159-70.
61. Bienz, M. and H. Clevers, Linking colorectal cancer to Wnt signaling. Cell, 2000. 103(2): p. 311-20.
62. Malanchi, I., et al., Cutaneous cancer stem cell maintenance is dependent on beta-catenin signalling. Nature, 2008. 452(7187): p. 650-3.
63. Reya, T., et al., A role for Wnt signalling in self-renewal of haematopoietic stem cells. Nature, 2003. 423(6938): p. 409-14.
64. Willert, K., et al., Wnt proteins are lipid-modified and can act as stem cell growth factors. Nature, 2003. 423(6938): p. 448-52.
65. Jamieson, C.H., et al., Granulocyte-macrophage progenitors as candidate leukemic stem cells in blast-crisis CML. N Engl J Med, 2004. 351(7): p. 657-67.
66. Bhardwaj, G., et al., Sonic hedgehog induces the proliferation of primitive human hematopoietic cells via BMP regulation. Nat Immunol, 2001. 2(2): p. 172-80.
67. Trowbridge, J.J., M.P. Scott, and M. Bhatia, Hedgehog modulates cell cycle regulators in stem cells to control hematopoietic regeneration. Proc Natl Acad Sci U S A, 2006. 103(38): p. 14134-9.
68. Zhao, C., et al., Hedgehog signalling is essential for maintenance of cancer stem cells in myeloid leukaemia. Nature, 2009. 458(7239): p. 776-9.
69. Dierks, C., et al., Expansion of Bcr-Abl-positive leukemic stem cells is dependent on Hedgehog pathway activation. Cancer Cell, 2008. 14(3): p. 238-49.
70. Piccirillo, S.G., et al., Bone morphogenetic proteins inhibit the tumorigenic potential of human brain tumour-initiating cells. Nature, 2006. 444(7120): p. 761-5.
71. Bao, S., et al., Targeting cancer stem cells through L1CAM suppresses glioma growth. Cancer Res, 2008. 68(15): p. 6043-8.
72. Jin, L., et al., Monoclonal antibody-mediated targeting of CD123, IL-3 receptor alpha chain, eliminates human acute myeloid leukemic stem cells. Cell Stem Cell, 2009. 5(1): p. 31-42.
73. Krause, D.S., et al., Requirement for CD44 in homing and engraftment of BCR-ABL-expressing leukemic stem cells. Nat Med, 2006. 12(10): p. 1175-80.
74. Jin, L., et al., Targeting of CD44 eradicates human acute myeloid leukemic stem cells. Nat Med, 2006. 12(10): p. 1167-74.
75. Hoey, T., et al., DLL4 blockade inhibits tumor growth and reduces tumor-initiating cell frequency. Cell Stem Cell, 2009. 5(2): p. 168-77.
35
76. Bao, S., et al., Glioma stem cells promote radioresistance by preferential activation of the DNA damage response. Nature, 2006. 444(7120): p. 756-60.
77. Diehn, M., et al., Association of reactive oxygen species levels and radioresistance in cancer stem cells. Nature, 2009. 458(7239): p. 780-3.
78. Wright, E.A. and P. Howard-Flanders, The influence of oxygen on the radiosensitivity of mammalian tissues. Acta radiol, 1957. 48(1): p. 26-32.
79. Platet, N., et al., Influence of oxygen tension on CD133 phenotype in human glioma cell cultures. Cancer Lett, 2007. 258(2): p. 286-90.
80. Blazek, E.R., J.L. Foutch, and G. Maki, Daoy medulloblastoma cells that express CD133 are radioresistant relative to CD133- cells, and the CD133+ sector is enlarged by hypoxia. Int J Radiat Oncol Biol Phys, 2007. 67(1): p. 1-5.
81. Kelly, P.N., et al., Tumor growth need not be driven by rare cancer stem cells. Science, 2007. 317(5836): p. 337.
82. Quintana, E., et al., Efficient tumour formation by single human melanoma cells. Nature, 2008. 456(7222): p. 593-8.
83. Kennedy, J.A., et al., Comment on "Tumor growth need not be driven by rare cancer stem cells". Science, 2007. 318(5857): p. 1722; author reply 1722.
84. Arndt-Jovin, D.J. and T.M. Jovin, Analysis and sorting of living cells according to deoxyribonucleic acid content. J Histochem Cytochem, 1977. 25(7): p. 585-9.
85. Lalande, M.E. and R.G. Miller, Fluorescence flow analysis of lymphocyte activation using Hoechst 33342 dye. J Histochem Cytochem, 1979. 27(1): p. 394-7.
86. Challen, G.A. and M.H. Little, A side order of stem cells: the SP phenotype. Stem Cells, 2006. 24(1): p. 3-12.
87. Goodell, M.A., et al., Isolation and functional properties of murine hematopoietic stem cells that are replicating in vivo. J Exp Med, 1996. 183(4): p. 1797-806.
88. Goodell, M.A., et al., Dye efflux studies suggest that hematopoietic stem cells expressing low or undetectable levels of CD34 antigen exist in multiple species. Nat Med, 1997. 3(12): p. 1337-45.
89. Pearce, D.J., et al., Multiparameter analysis of murine bone marrow side population cells. Blood, 2004. 103(7): p. 2541-6.
90. Larderet, G., et al., Human side population keratinocytes exhibit long-term proliferative potential and a specific gene expression profile and can form a pluristratified epidermis. Stem Cells, 2006. 24(4): p. 965-74.
36
91. Yano, S., et al., Characterization and localization of side population cells in mouse skin. Stem Cells, 2005. 23(6): p. 834-41.
92. Summer, R., et al., Side population cells and Bcrp1 expression in lung. Am J Physiol Lung Cell Mol Physiol, 2003. 285(1): p. L97-104.
93. Majka, S.M., et al., Identification of novel resident pulmonary stem cells: form and function of the lung side population. Stem Cells, 2005. 23(8): p. 1073-81.
94. Shimano, K., et al., Hepatic oval cells have the side population phenotype defined by expression of ATP-binding cassette transporter ABCG2/BCRP1. Am J Pathol, 2003. 163(1): p. 3-9.
95. Martin, C.M., et al., Persistent expression of the ATP-binding cassette transporter, Abcg2, identifies cardiac SP cells in the developing and adult heart. Dev Biol, 2004. 265(1): p. 262-75.
96. Kim, M. and C.M. Morshead, Distinct populations of forebrain neural stem and progenitor cells can be isolated using side-population analysis. J Neurosci, 2003. 23(33): p. 10703-9.
97. Behbod, F., et al., Transcriptional profiling of mammary gland side population cells. Stem Cells, 2006. 24(4): p. 1065-74.
98. Meeson, A.P., et al., Cellular and molecular regulation of skeletal muscle side population cells. Stem Cells, 2004. 22(7): p. 1305-20.
99. Liadaki, K., et al., Side population cells isolated from different tissues share transcriptome signatures and express tissue-specific markers. Exp Cell Res, 2005. 303(2): p. 360-74.
100. Zhou, S., et al., The ABC transporter Bcrp1/ABCG2 is expressed in a wide variety of stem cells and is a molecular determinant of the side-population phenotype. Nat Med, 2001. 7(9): p. 1028-34.
101. Bunting, K.D., et al., Transduction of murine bone marrow cells with an MDR1 vector enables ex vivo stem cell expansion, but these expanded grafts cause a myeloproliferative syndrome in transplanted mice. Blood, 1998. 92(7): p. 2269-79.
102. Bunting, K.D., et al., Enforced P-glycoprotein pump function in murine bone marrow cells results in expansion of side population stem cells in vitro and repopulating cells in vivo. Blood, 2000. 96(3): p. 902-9.
103. Jonker, J.W., et al., Contribution of the ABC transporters Bcrp1 and Mdr1a/1b to the side population phenotype in mammary gland and bone marrow of mice. Stem Cells, 2005. 23(8): p. 1059-65.
104. Chiba, T., et al., Side population purified from hepatocellular carcinoma cells harbors cancer stem cell-like properties. Hepatology, 2006. 44(1): p. 240-51.
37
105. Mitsutake, N., et al., Characterization of side population in thyroid cancer cell lines: cancer stem-like cells are enriched partly but not exclusively. Endocrinology, 2007.
106. Szotek, P.P., et al., Ovarian cancer side population defines cells with stem cell-like characteristics and Mullerian Inhibiting Substance responsiveness. Proc Natl Acad Sci U S A, 2006. 103(30): p. 11154-9.
107. Ho, M.M., et al., Side population in human lung cancer cell lines and tumors is enriched with stem-like cancer cells. Cancer Res, 2007. 67(10): p. 4827-33.
108. Wang, J., et al., Identification of cancer stem cell-like side population cells in human nasopharyngeal carcinoma cell line. Cancer Res, 2007. 67(8): p. 3716-24.
109. Haraguchi, N., et al., Characterization of a side population of cancer cells from human gastrointestinal system. Stem Cells, 2006. 24(3): p. 506-13.
110. Hadnagy, A., et al., SP analysis may be used to identify cancer stem cell populations. Exp Cell Res, 2006. 312(19): p. 3701-10.
111. Hirschmann-Jax, C., et al., A distinct "side population" of cells with high drug efflux capacity in human tumor cells. Proc Natl Acad Sci U S A, 2004. 101(39): p. 14228-33.
112. Patrawala, L., et al., Side population is enriched in tumorigenic, stem-like cancer cells, whereas ABCG2+ and ABCG2- cancer cells are similarly tumorigenic. Cancer Res, 2005. 65(14): p. 6207-19.
113. Kondo, T., T. Setoguchi, and T. Taga, Persistence of a small subpopulation of cancer stem-like cells in the C6 glioma cell line. Proc Natl Acad Sci U S A, 2004. 101(3): p. 781-6.
114. Hiddemann, W., et al., Tumor heterogeneity in osteosarcoma as identified by flow cytometry. Cancer, 1987. 59(2): p. 324-8.
115. Zhou, J., et al., Activation of the PTEN/mTOR/STAT3 pathway in breast cancer stem-like cells is required for viability and maintenance. Proc Natl Acad Sci U S A, 2007. 104(41): p. 16158-63.
116. Wood, L.D., et al., The genomic landscapes of human breast and colorectal cancers. Science, 2007. 318(5853): p. 1108-13.
117. Doyle, L.A. and D.D. Ross, Multidrug resistance mediated by the breast cancer resistance protein BCRP (ABCG2). Oncogene, 2003. 22(47): p. 7340-58.
118. Wunder, J.S., et al., A comparison of staging systems for localized extremity soft tissue sarcoma. Cancer, 2000. 88(12): p. 2721-30.
119. Wang, X., et al., A luminal epithelial stem cell that is a cell of origin for prostate cancer. Nature, 2009. 461(7263): p. 495-500.
38
120. Huntly, B.J., et al., MOZ-TIF2, but not BCR-ABL, confers properties of leukemic stem cells to committed murine hematopoietic progenitors. Cancer Cell, 2004. 6(6): p. 587-96.
121. Joseph, N.M., et al., The loss of Nf1 transiently promotes self-renewal but not tumorigenesis by neural crest stem cells. Cancer Cell, 2008. 13(2): p. 129-40.
122. Zheng, H., et al., Induction of abnormal proliferation by nonmyelinating schwann cells triggers neurofibroma formation. Cancer Cell, 2008. 13(2): p. 117-28.
123. Le, L.Q., et al., Cell of origin and microenvironment contribution for NF1-associated dermal neurofibromas. Cell Stem Cell, 2009. 4(5): p. 453-63.
124. Barker, N., et al., Identification of stem cells in small intestine and colon by marker gene Lgr5. Nature, 2007. 449(7165): p. 1003-7.
125. Barker, N., et al., Crypt stem cells as the cells-of-origin of intestinal cancer. Nature, 2009. 457(7229): p. 608-11.
126. Dexter, T.M., T.D. Allen, and L.G. Lajtha, Conditions controlling the proliferation of haemopoietic stem cells in vitro. J Cell Physiol, 1977. 91(3): p. 335-44.
127. Bianco, P. and A. Boyde, Confocal images of marrow stromal (Westen-Bainton) cells. Histochemistry, 1993. 100(2): p. 93-9.
128. Dominici, M., et al., Minimal criteria for defining multipotent mesenchymal stromal cells. The International Society for Cellular Therapy position statement. Cytotherapy, 2006. 8(4): p. 315-7.
129. Kolf, C.M., E. Cho, and R.S. Tuan, Mesenchymal stromal cells. Biology of adult mesenchymal stem cells: regulation of niche, self-renewal and differentiation. Arthritis Res Ther, 2007. 9(1): p. 204.
130. Sarugaser, R., et al., Human mesenchymal stem cells self-renew and differentiate according to a deterministic hierarchy. PLoS One, 2009. 4(8): p. e6498.
131. Dellavalle, A., et al., Pericytes of human skeletal muscle are myogenic precursors distinct from satellite cells. Nat Cell Biol, 2007. 9(3): p. 255-67.
132. Watt, F.M. and B.L. Hogan, Out of Eden: stem cells and their niches. Science, 2000. 287(5457): p. 1427-30.
133. da Silva Meirelles, L., P.C. Chagastelles, and N.B. Nardi, Mesenchymal stem cells reside in virtually all post-natal organs and tissues. J Cell Sci, 2006. 119(Pt 11): p. 2204-13.
134. Lazarus, H.M., et al., Human bone marrow-derived mesenchymal (stromal) progenitor cells (MPCs) cannot be recovered from peripheral blood progenitor cell collections. J Hematother, 1997. 6(5): p. 447-55.
39
135. Wexler, S.A., et al., Adult bone marrow is a rich source of human mesenchymal 'stem' cells but umbilical cord and mobilized adult blood are not. Br J Haematol, 2003. 121(2): p. 368-74.
136. Farrington-Rock, C., et al., Chondrogenic and adipogenic potential of microvascular pericytes. Circulation, 2004. 110(15): p. 2226-32.
137. Hirschi, K.K. and P.A. D'Amore, Pericytes in the microvasculature. Cardiovasc Res, 1996. 32(4): p. 687-98.
138. Andreeva, E.R., et al., Continuous subendothelial network formed by pericyte-like cells in human vascular bed. Tissue Cell, 1998. 30(1): p. 127-35.
139. Crisan, M., et al., A perivascular origin for mesenchymal stem cells in multiple human organs. Cell Stem Cell, 2008. 3(3): p. 301-13.
140. Riggi, N., et al., Development of Ewing's sarcoma from primary bone marrow-derived mesenchymal progenitor cells. Cancer Res, 2005. 65(24): p. 11459-68.
141. Tirode, F., et al., Mesenchymal stem cell features of Ewing tumors. Cancer Cell, 2007. 11(5): p. 421-9.
142. Matushansky, I., et al., Derivation of sarcomas from mesenchymal stem cells via inactivation of the Wnt pathway. J Clin Invest, 2007. 117(11): p. 3248-57.
143. McAdam, W.A. and J.C. Goligher, The occurrence of desmoids in patients with familial polyposis coli. Br J Surg, 1970. 57(8): p. 618-31.
144. Burke, A.P., et al., Intra-abdominal fibromatosis. A pathologic analysis of 130 tumors with comparison of clinical subgroups. Am J Surg Pathol, 1990. 14(4): p. 335-41.
145. Alman, B.A., et al., Aggressive fibromatosis (desmoid tumor) is a monoclonal disorder. Diagn Mol Pathol, 1997. 6(2): p. 98-101.
146. Lewis, J.J., et al., The enigma of desmoid tumors. Ann Surg, 1999. 229(6): p. 866-72; discussion 872-3.
147. Shields, C.J., et al., Desmoid tumours. Eur J Surg Oncol, 2001. 27(8): p. 701-6.
148. Alman, B.A., et al., Aggressive fibromatosis. J Pediatr Orthop, 1992. 12(1): p. 1-10.
149. Reitamo, J.J., T.M. Scheinin, and P. Hayry, The desmoid syndrome. New aspects in the cause, pathogenesis and treatment of the desmoid tumor. Am J Surg, 1986. 151(2): p. 230-7.
150. Khorsand, J. and C.P. Karakousis, Desmoid tumors and their management. Am J Surg, 1985. 149(2): p. 215-8.
151. Faulkner, L.B., et al., Pediatric desmoid tumor: retrospective analysis of 63 cases. J Clin Oncol, 1995. 13(11): p. 2813-8.
40
152. Gurbuz, A.K., et al., Desmoid tumours in familial adenomatous polyposis. Gut, 1994. 35(3): p. 377-81.
153. Hizawa, K., et al., Desmoid tumors in familial adenomatous polyposis/Gardner's syndrome. J Clin Gastroenterol, 1997. 25(1): p. 334-7.
154. Rodriguez-Bigas, M.A., et al., Desmoid tumors in patients with familial adenomatous polyposis. Cancer, 1994. 74(4): p. 1270-4.
155. Alman, B.A., et al., Increased beta-catenin protein and somatic APC mutations in sporadic aggressive fibromatoses (desmoid tumors). Am J Pathol, 1997. 151(2): p. 329-34.
156. Tejpar, S., et al., Predominance of beta-catenin mutations and beta-catenin dysregulation in sporadic aggressive fibromatosis (desmoid tumor). Oncogene, 1999. 18(47): p. 6615-20.
157. Smits, R., et al., Apc1638N: a mouse model for familial adenomatous polyposis-associated desmoid tumors and cutaneous cysts. Gastroenterology, 1998. 114(2): p. 275-83.
158. Fodde, R., et al., A targeted chain-termination mutation in the mouse Apc gene results in multiple intestinal tumors. Proc Natl Acad Sci U S A, 1994. 91(19): p. 8969-73.
41
CHATPER 2
Side population cells isolated from mesenchymal neoplasms have
tumor initiating potential
I preformed all experiments describe in the paper with equal contribution from Qingxia Wei with the following exceptions:
• Velani Utomo helped in processing primary tumor samples • Puviindran Nadesan helped in managing the mouse colony • Heather Whetstone preformed the hemotoxylin and eosin staining (Figure 2.3) • Rita Kandel preformed pathological analysis of the tumor samples (Figure 2.3) • Jay Wunder provided human tumor samples for the side population experiments
42
Chapter 2
2 Side population cells isolated from mesenchymal neoplasms contain tumor initiating cells
2.1 Abstract
Although many cancers are maintained by tumor initiating cells, this has not been demonstrated
for mesenchymal tumors, in part due to the lack of unique surface markers that identify
mesenchymal progenitors. An alternative technique to isolate stem-like cells is to isolate side
population cells based on efflux of Hoechst 33342 dye. We examined 29 mesenchymal tumors
ranging from benign to high grade sarcomas, and identified a side population of cells in all but
six samples. There was a positive correlation between the percentage of side population cells and
the grade of the tumor. Side population cells preferentially formed tumors when grafted into
immunodeficient mice, and only cells from tumors that developed from the side population cells
had the ability to initiate tumor formation upon serial transplantation. Although the side
population cells are able to efflux Rhodamine dye in addition to Hoechst 33342, we found that
the ability to efflux Rhodamine dye did not identify a population of cells enriched for tumor
initiating capacity. Here we identify a subpopulation of cells within a broad range of benign and
malignant mesenchymal tumors with tumor initiating capacity. In addition, our data suggests that
the proportion of side population cells could be used as a prognostic factor, and that
therapeutically targeting this subpopulation of cells could be used to improve patient outcome.
43
2.2 Introduction
Solid tumors are composed of a heterogeneous population of cells. These cells have different in
vitro proliferative capacities; only a minority have the ability to initiate tumor formation in
immuno-deficient mice. This observation led to the concept of cancer stem cells (CSC) which
have the ability to self-renew and differentiate. By manipulating these characteristics, CSCs
have been postulated to be responsible for driving the growth of tumors and the recurrence of
neoplasms after therapy (1, 2). Although these cells have been identified in a variety of
malignancies, such as hematologic, neural and epithelial cancers, they have not been identified in
neoplasms of mesenchymal origin. The solid tumors from which CSCs have been isolated are
not of mesenchymal origin. As such, cell surface markers used in these tumors may not be
successful for the isolation of these CSCs from mesenchymal neoplasms (3-11). Furthermore,
there are no unique markers for mesenchymal stem cells. Fortunately, other properties of stem
cells can be used to isolate cells with progenitor characteristics.
One such property is the ability of stem cells to efflux chemotherapeutic drugs and certain dyes
such as Hoechst 33342 (12, 13). This feature is conferred partly by high levels of expression of
ABC transporter proteins on primitive progenitor cells (14, 15). During flow cytometry analysis,
negatively stained cells fall to the “side” of the majority, hence they are commonly referred to as
side population (SP) cells (16). They enrich for progenitor cells from a variety of tissue types
(17-20); furthermore, data has proven that some cancer cell lines and primary tumors contain SP
cells. It is the SP cells that confer tumorgenicity of the tumor or cancer cell line (21, 22). This
technique provides an alternative means to isolate progenitor cells other than through the use of
specific surface markers, and may be one way to identify putative tumor initiating cells.
Primary musculoskeletal tumors are of mesenchymal origins. They range from benign lesions,
such as the soft tissue aggressive fibromatosis (desmoid), to very aggressive malignancies such
as synovial sarcomas. Mesenchymal progenitors express a variety of markers, such as Stro-1,
CD105, and CD44; however, to date no marker or pattern of markers universally defines a
consensus phenotype for a mesenchymal stem cell. This has hampered research into the cell of
origin of these lesions (23-26). Recent in vitro evidence suggests that mesenchymal tumors may
contain cells with stem-like characteristics. Cells within these tumors had the capacity to form
spheres similar to those formed by neural stem cells. Furthermore, these spheres preferentially
44
expressed genes that are involved in regulating stem cell fate(27). Despite this, the ability to
identify and isolate tumor initiating cells (TICs) within mesenchymal tumors has proved to be
elusive.
The initiation and progression of these tumor types is poorly understood. These lesions can have
a high rate of recurrence suggesting that a small population of cells have the potential to escape
treatment and may have tumor initiating cell characteristics (28). SP cells, which are enriched for
TICs, efflux dye through a mechanism that is similar to that used by chemotherapeutic resistant
cells. Therefore, TICs may be relatively resistant to chemotherapeutic drugs and may confer the
malignant phenotype to tumors (13, 29, 30). As such, these cells could be responsible for
sarcoma recurrence after chemotherapy. Here we report, for the first time, the identification of
SP cells from mesenchymal tumors. These cells have the capacity to initiate tumor formation.
Further investigation of these TICs cells may lead to the development of more effective and
efficient targeted treatment modalities.
45
2.3 Results
2.3.1 Mesenchymal tumors contain side population (SP) cells
We sought to determine whether we could isolate tumor initiating cells from primary
mesenchymal tumors based on the observation that stem-like cells have the ability to efflux
Hoecsht 33342 dye (16). Cells from 29 primary mesenchymal tumors, ranging from locally
invasive lesions to high grade sarcomas were incubated with the fluorescent dye, Hoechst 33342.
Cells were also incubated in the presence of verapamil, a chemical inhibitor of the ABC protein
family of transporters, which inhibits the efflux of Hoechst (15). A distinct SP was found in all
of the 29 tumors examined except for one dermatofibrosarcoma protuberans, one myxoid
chondrosarcoma, one of three chondrosarcomas, and one of two leiomyosarcomas. With the
addition of verapamil, the presence of SP cells was abolished, indicating that dye efflux
occurred, in part, through an ABC transporter regulated mechanism (Figure 2.1). The presence of
a SP within these tumors, raises the possibility this represents a subpopulation of cells with
tumor initiating characteristics.
2.3.2 The proportion of side population cells correlates with aggressiveness of the tumor
Mesenchymal tumors are highly heterogeneous groups of neoplasms with varying levels of
histologic and clinical aggressiveness. (28). We observed a trend between the proportion of SP
cells present in a given tumor and the relative aggressiveness of the tumor (Figure. 2.2). In
general, higher grade tumors had an increased prevalence of SP cells. This suggests that the
proportion of SP cells may be a predictor of patient outcome. Furthermore, cells in the side
population may be responsible for the maintenance and propagation of mesenchymal tumors
giving the high grade tumors a more aggressive behavior. However numbers of individual tumor
types were too small to determine if within a particular tumor type the percentage of cells sorting
to the SP correlated with grade, aggressiveness, or clinical outcome.
2.3.3 SP cells have the capacity to form tumors upon serial transplantation in NOD/SCID mice
We tested the ability of SP cells to initiate tumor formation when grafted into NOD/SCID mice.
Cells from one osteosarcoma, two malignant fibrous histocytomas, and one synovial sarcoma
46
were stained with Hoechst 33342 dye and sorted into SP and non-SP fractions. Both fractions
were then subcutaneously injected into NOD/SCID mice. Initially, both SP and non-SP fractions
had the capacity to form tumors, however, significantly smaller numbers of SP cells formed
tumors at a significantly higher frequency when compared to non-SP cells (Table 1.1). The SP
from each of the individual tumors examined showed the same enhanced ability to form tumors
in NOD/SCID mice. Furthermore, tumors that did form from the non-SP fraction, even though
they formed from larger numbers of injected cells, were significantly smaller when compared to
tumors from the SP fraction (0.30 –vs- 1.08 gm, p<0.01). Tumors expressed human GAPDH
and had a nearly identical cytology to the primary sarcoma (Figure 2.3).
To test for self-renewal (34), SP and non-SP tumors were dissociated and cells from both tumors
were re-stained with Hoechst 33342 dye. SP re-analysis demonstrated that only cells derived
from SP tumors were able to repopulate both SP and NSP fractions (Figure 2.4A). The
percentage of SP present from xenografted tumors was similar to that of the parental tumor cells
suggesting that SP cells had the ability to recapitulate the pheonotypic distribution of the original
tumor. When labeled with propidium iodide, there was significant decrease in viable cells in the
tumors derived from the non-SP compared to the cells derived from SP tumors (Figure 2.4B).
Viable cells from non-SP tumors had an increase in DNA content when compared to the SP
tumor cells suggesting a higher mitotic rate (Figure 2.4 C). Furthermore even after 32 weeks,
non-SP tumor cells injected into NOD/SCID failed to initiate tumor growth. Taken together, this
suggests non-SP tumors represent a population of cells with only transient amplifying potential.
To determine the tumorigenic potential of the cells from the xenograft tumors, we next serially
transplanted SP and non-SP cells from the SP tumors and non-SP tumors. In the secondary
transplant, cells from non-SP tumors did not initiate tumors in the NOD/SCID mice. Only the
SP fraction from tumors from the initial SP fraction had the capacity to initiate tumor formation.
Furthermore, as few as 100 SP cells resulted in the formation of tumors whereas 10,000 non-SP
cells failed to form tumors (Table 1.1). These secondary xenografted tumors had an identically
histological appearance to both primary xenografted tumors and to original parental tumors
(Figure 2.3 C). Hence, only tumors derived from SP cells can self-renew and differentiate in a in
a manner that can be propagated through serial transplantation into NOD/SCID mice.
47
2.3.4 SP Cells Efflux Rhodamine-123
The phenomenon of dye efflux is not exclusive to Hoechst 33342. For instance, Rhodamine-123
efflux can be used to identify populations enriched for hematopoetic stem cells from mouse bone
marrow. We next examined the ability of Rhodamine-123 efflux to recapitulate the results
demonstrated with Hoechst efflux. Staining with Rhodamine-123 resulted in a broad range of
florescent uptake. Two populations were sorted based on gating of the upper and lower 10% of
live cells, hereafter referred to as Rholow and Rhohigh. Cells were also dual stained with both
Rhodamine-123 and Hoechst 33342. We noted that the majority of SP cells did not stain highly
for Rhodamine-123 while the staining pattern of non-sp cells was identical to the general
population of cells (Figure 2.5 A). Thus, SP cells have the capacity to efflux Rhodamine-123 and
the two populations were not mutually exclusive. Rholow, Rhohigh , Rhohigh /non-SP, Rholow/SP
cells were sorted and injected into NOD/SCID mice (Figure 2.5 B). While both Rhodamine-123
populations failed to form tumors after 12 weeks, as few as 375 Rholow/SP cells did initiate
tumor formation in 2/2 mice. This corresponded to the number of cells required to form tumors
from SP cells, suggesting that the capacity to efflux Rhodamine-123 does not enhance the
tumorigenic potential of SP cells.
2.4 Discussion
Here we demonstrate for the first time that mesenchymal tumors contain a subpopulation of cells
with tumor initiating capacity, which can be identified based on their exclusion of Hoescht
33342 dye. Intriguingly, there are higher proportions of SP cells in high grade mesenchymal
malignancies than in less aggressive benign lesions. While tumor initiating cells (TICs) have
been identified in hematologic, neural, and epithelial cancer (3-11), to date, this has not been
proven for mesenchymal tumors whose origins differ from other solid tumor. Our data suggests
that TICs are present in a broad range of benign and malignant neoplastic processes, and as such,
are a general phenomenon in tumorigenesis.
The identification of TICs in mesenchymal tumors has proven to be more elusive than in tumors
that originate from other tissue types, in part due to the lack of universal markers for
mesenchymal progenitor cells (23, 25, 26). Other techniques to isolate stem cells include in-vitro
functional techniques. Indeed, recent work has shown that in an in vitro culture system, cells
derived from mesenchymal tumors have the capacity to form spheres similar to those derived
48
from neural stem cells(27). While these “sarcospheres” preferentially express “stemness” genes
in comparison to their adherent counterparts, it has not been demonstrated that these cells have a
preferential ability to form tumors when grafted into immuno-deficient mice.
Given these obstacles, we chose to use Hoecsht dye exclusion as an alternative method of stem
cell isolation. While this method has been successfully utilized to isolate stem-like cells in a
variety of cell lines(21), it is infrequently used in primary tumors(22). We found that cells from
most primary mesenchymal tumors contained a side population and that these cells have the
capacity to initiate tumors when transplanted into immuno-deficient mice. Furthermore, upon
serial transplantation, only SP tumors had the ability to re-initiate the tumor in immuno-deficient,
suggesting that only this population is able to self-renew in vivo.
Large numbers of non-SP cells also formed tumors in NOD/SCID mice, and others demonstrated
that in breast, prostate, and thyroid cancer cell lines, large numbers of non-SP cells have the
capacity to initiate tumor formation. It has been suggested that this finding is due to
contamination of small numbers of SP cells in the non-SP fraction(35-37). However in our
experiments, while cells from the non-SP fraction were initially able to form tumors in
NOD/SCID mice, when re-stained with Hoechst dye they did not contain an SP fraction. They
also exhibited high levels of cell death as demonstrated by a high level of propridium iodide
staining. Furthermore, analysis of DNA content demonstrated that these cells have an increase in
DNA content when compared to cells derived from SP tumors suggesting that non-SP tumor
cells have an increased proliferation rate. Importantly, non-SP tumor cells were not able to
initiate tumors after serial transplantation even after 32 weeks post injection. This suggests non-
SP cells are fundamentally different from their SP counterparts. Given these observations, it is
possible that cells from the non-SP fraction represent a more differentiated subpopulation,
characterized by a short term proliferative potential, such as a transient amplifying cell. Taken
together, this strongly suggests that Hoescht dye exclusion will enrich for cells with the capacity
to divide asymmetrically and to self-renew, a key feature of stem cells.
As Hoescht dye exclusion selects for cells with the capacity to exclude dye via the expression of
protein transporters(15, 38), we cannot exclude the possibility that SP cells are more mature
tumor cells that have acquired the ability to increase the expression of genes responsible for this
pumping mechanism. Such cells might also be resistant to chemotherapy and may be responsible
49
for the relapse of disease (13). To address this issue, we isolated cells that effluxed the
fluorescent dye, Rhodamine-123. Unlike SP cells, neither Rholow nor Rhohigh cells had the
capacity to initiate tumor formation, however, we cannot exclude the possibility that our gating
strategy was too generous for enrichment of tumor initiating cells and that with increased
stringency there may be tumor initiating cells in the Rholow population. Despite this, our data
suggests that the capacity to efflux materials may not the sole determinant of tumor initiating
cells since SP but not Rholow cells had the capacity enrich for TICs. In any event, we have shown
that SP cells are distinct from their non-SP counterparts, and are enriched for TICs. A more in-
depth analysis of this population may provide important clues into sarcoma biology.
A striking result of our study is the correlation seen between the percentage of SP cells and the
aggressiveness of the mesenchymal tumor examined. We found that benign, locally invasive
low-grade tumors had a low SP population whereas high-grade malignant tumors had high SP
populations. This data implies that SP is a predictor of patient outcome, however, prospective
studies are needed to determine if this hypothesis will be proven to be correct.
The high proportion of SP cells in malignant tumors might also correlate with chemotherapeutic
resistance, as this resistance may be due in part to the over-expression of transporter proteins
which efflux drugs. As we have shown that SP cells have the capacity to efflux material and have
enhanced tumorigenic potential when compared to non-SP cells, it is interesting to hypothesize
that the efflux mechanism plays a role in chemotherapeutic resistance. Perhaps the expression of
transporter proteins would correlate with both outcome and the proportion of SP cells. Hoechst
dye exclusion is conferred in part by the expression of the ABCG2 transporter protein while
Rhodamine dye exclusion is mediated by ABCB1/P-gp protein (15). Previous data has shown no
correlation between the expression of ABCB1/P-gp in soft tissue sarcomas and tumor grade and
not surprisingly, Rholow cells failed to initiate tumor formation(39, 40). To date the role of
ABCG2 transporter protein in soft tissue sarcomas has yet to be elucidated. However, it should
be noted that Bcrp1/Mdr1a/b triple knock-out mice still exhibit SP and it is likely multiple
transporters are involved in this process(41).
Our studies demonstrate for the first time, the existence of SP cells in primary mesenchymal
tumors. Furthermore, we found that these cells are enriched for tumor initiating cells. There
appears to be a direct correlation between the number of SP cells present in a tumor and the
50
aggressiveness of the tumor. Targeting this side population has the potential to be developed into
an effective treatment modality for mesenchymal tumors.
51
Figures
Figure 2.1 Mesenchymal tumors contain side population cells
(A) Cells from a representative grade two primary sarcoma were stained with Hoechst 33342 and
analyzed by flow cytometry. The SP from a representative tumor is shown. The SP cells are
outlined and shown as a percentage of the total cell population. (B) This cell population
disappears in the presence of verapamil.
52
53
Figure 2.2 High grade sarcomas have an increased prevalence of SP cells when compared to lower grade lesions.
(A) SP analysis from a benign but locally aggressive aggressive fibromatosis, a grade two
liposarcoma and a grade three malignant fibrous histiocytoma are shown. SP cells are outlined
and shown as a percentage of the total cell population. (B) Mean and 95% confidence intervals
for the percent of SP cells from benign and various grade lesions. A 95% confidence interval that
does not cross the mean of a comparison is a statistically significant difference at p<0.05. There
are significantly higher proportions of SP cells in higher grade tumors.
54
55
Table 2.1 The proportion of tumors that formed from injection of various numbers of cells
from each subpopulation into NOD/SCID mice.
Primary mice were injected with cells sorted from the primary tumor, and secondary mice are
injected with cells sorted from the tumor that developed in primary mice. Side population cells
from each tumor type examined formed tumors from the side population fraction at each cell
number examined. In contrast, only in a few instances did non-side population cells formed
tumors in primary mice. Larger numbers of injected cells were required for the non-side
population cells to form tumors. The side population fraction from tumors that formed in the
primary mice were able to form tumors in secondary mice, while in only one case would cells
from the non-side population form a tumor. None of the cells from the primary tumors that
formed from non-side population cells were able to form tumors in secondary mice.
Abbreviation: nt not tested
56
Cell Type Cell Dose
(number of primary mice with tumors)/(total number injected)
(number of secondary mice with tumors)/(total number injected)
Total number of mice with tumors (percent)
Side Population 1x102 11/16 5/8 16/24 (67)
5 x 102 5/5 nt 5/5 (100)
1x103 8/10 9/12 17/24 (71)
1x104 8/10 4/4 12/14 (86)
1x105 nt nt
Non Side Population 1x102 0/8 0/8 0/16 (0)
5 x 102 0/4 nt 0/4 (0)
1x103 3/14 0/12 3/26 (12)
1x104 1/10 0/10 1/20 (5)
1x105 4/16 1/10 5/26 (19)
57
Figure 2.3 Histopathologic features of SP tumors
A to C) SP cells from (A) parental, (B) primary, and (C) secondary xenografted tumors were
injected into NOD/SCID mice. After 12 weeks tumors were collected formalin fixed, paraffin
embedded and stained with H&E. The grafted tumors (B and C) formed had nearly identical
cytological appearances to the (A) primary tumors. (D) A representative NSP tumor also showed
similar cytologic appearances to the primary tumor. A representative malignant fibrous
histiocytoma is shown, and the size bar is 200um.
58
59
Figure 2.4 Characteristics of tumors derived from SP and non-SP cells.
Cells from derived from both SP and non-SP (NSP) primary xenograft tumors were reanalyzed
for SP cells . (A) Only SP tumors contained cells that had the capacity to reform both SP and
non-SP fractions . (B) Cell viability was determined by propidium iodide staining. 23% of cells
derived from SP tumors were viable, in contrast, only 8% of cells derived from non-SP tumors
were viable. (C) Cells from non-SP tumors had an increase in DNA content when compared to
the SP tumor cells suggesting a higher mitotic rate.
60
0 1000 2000 3000 4000
FL9
0
1000
2000
3000
4000
FL6
1.79
0 1000 2000 3000 4000
FL9
0
1000
2000
3000
4000
FL6
0.16
0 1000 2000 3000 4000
FSC
100
101
102
103
104
FL4
23
0 1000 2000 3000 4000
FSC
100
101
102
103
104
FL4
7.95
SPN
SP
Hoechst 33342 Propidum Iodide DNA ContentA B C
0 1000 2000 3000 4000
PE-A
0
100
200
300
# Cells
0 1000 2000 3000 4000
PE-A
0
50
100
150
200
250
# Cells
61
Figure 2.5 SP cells efflux Rhodamine-123
Cells derived from a primary MFH sample were dual stained with both Hoechst 33342 and
Rhodamine-123. (A) Twenty percent cutoffs were used to sort Rhohigh and Rholow fractions.
Shared profiles indicate SP cells (blue) have decreased Rhodamine-123 staining when compared
to the total population of cells (green). The non-SP cells (red) have a staining patter similar to
that of the total population of cells. (B) Cells were sorted into SP, non-SP, Rhohigh, Rholow,
SP/Rholow, non-SP/Rhohigh fractions and varying cell numbers were injected into NOD/SCID
mice. Tumor formation was assessed 12 weeks post injection. Numbers indicate the number of
tumors that formed/number of injections.
62
0 1000 2000 3000 4000
FL9
0
1000
2000
3000
4000
FL6
0.11
68.3
100
101
102
103
104
DSP(Comp FL1)
0
1000
2000
3000
4000
# Cells
20.2 20.6
100
101
102
103
104
DSP(Comp FL1)
0
20
40
60
80
100
% of Max
Hoechst 33342 Rhodamine 123
Cells injected
Sample Population 100 375 1000 5000
T-1 MFH SP 3/8 6/8NSP 0/8 0/8
T-1 MFH Rholow 0/6 0/4Rhohigh 0/5 0/4
T-1 MFH SP Rholow 2/2NSPRhohigh 0/4
A
B
63
2.5 Materials and Methods
2.5.1 Primary Tumors
29 mesenchymal tumors were processed at the time of surgical excision. Local ethical approval
was obtained for all tissue samples collected. The samples included 7 aggressive fibromatosis, 5
osteosarcomas, 3 chondrosarcomas, 3 synovial sarcomas, 2 leiomyosarcoma, 4 malignant fibrous
histocytomas, 1 myxoid liposarcoma, 1 pleomorphic liposarcoma, 1 dermatofibrosarcoma
protuberans, 1 myxoid chondrosarcoma, and 1 chordoma. Primary tumor samples were manually
minced and all visible clumps removed. Enzymatic digestion followed at 37oC for 45 minutes
with constant rotation using 10mg/ml of collagenase IV (Worthington), 2.4 U/ml of Dispase
(Becton Dickinson), 0.05% trypsin (Wisent). Further manual dissociation was performed by
passing the cell slurry through an 18 gauge needle. Cells were then centrifuged at 1400 rpm for 5
minutes and washed three times in PBS. After washing, cells were strained through 70µm filters
to remove remaining clumps. Collected cells were plated in alpha-MEM supplemented with L-
glutamine and containing 18% fetal bovine serum (Wisent) and cultured at 37oC with 5% CO2 in
a humidified chamber until subjected to fluorescent activated cells sorting (FACS).
2.5.2 Pathology
Formalin fixed, paraffin embedded samples were stained with hematoxylin and eosin (H&E) and
observed in a blinded manner by an experienced pathologist. Tumors were graded into benign or
malignant groups. Each tumor was sub-classified according to the WHO classification and
graded using the standard 3 scale American Joint Commission on Staging (31-33).
2.5.3 Flow Cytometry
Cells were trypsinized and resuspended in PBS supplemented with 2% FBS at a concentration of
1X106 cells/ml. For SP assays, cells were treated either alone with 2.5 µg/ml of Hoechst 33342
dye (Sigma) for 90 minutes or with 0.1ug/ml at 37oC, or in combination with 50 µM verapamil
(Sigma), an inhibitor of ABC transporters. For Rhodamine-123 staining, 1X106 cells/ml were
incubated with 0.1µg/ml of Rhodamine-123 (Molecular Probes) for 30 minutes at 37oC. For
analysis of Rhodamine-123 and Hoechst 33342 efflux activity, cells were initially incubated with
0.1µg/ml of Rhodamine-123 for 30 minutes at 37oC, washed and then re-suspended in the same
cell concentration with 2.5µg/ml of Hoechst dye for 90 minutes at 37oC. Cells were
64
counterstained with 1µg/ml of propidium iodide (Molecular Probes), and non-viable cells were
excluded from both analysis and sorting assays. To detect for SP, cells were analyzed by using a
dual wavelength analysis (blue, 424-444nm; red, 675nM) after excitation with 350nm UV light
(MoFlow, Cytomation). SP and non-SP cells were collected and either grown in culture medium
for in vitro studies, or injected into NOD/SCID mice. For Rhodamine-123 staining, ten percent
cutoffs were used to sort Rhohigh and Rholow fractions.
2.5.4 Cell Transplantation into NOD/SCID mice
Sorted SP and non-SP cells were collected and cells were resuspended in PBS at concentrations
ranging from 200-20,000 cells/50µl. Cells were then mixed with 50µl of Matrigel (Becton
Dickinson). This cell:matrigel suspension was then subcutaneously injected in to eight to ten
week old NOD/SCID mice. Mice were observed for up to 12 weeks after which they were
euthanized and tumor formation was assessed. Tumors that formed were removed and samples
from each tumor were harvested for FACS and histology. For FACS, tumors were dissociated, as
for the primary tumors and resorted into SP and NSP fractions. These secondary sorted cells
were then re-injected into eight to ten week old NOD/SCID mice and tumor formation was
assessed after 12 weeks. Samples of tumors were formalin fixed and processed for histology.
65
2.6 References
1. Al-Hajj, M. and Clarke, M. F. Self-renewal and solid tumor stem cells. Oncogene, 23:
7274-7282, 2004.
2. Pardal, R., Clarke, M. F., and Morrison, S. J. Applying the principles of stem-cell biology
to cancer. Nat Rev Cancer, 3: 895-902, 2003.
3. Bonnet, D. and Dick, J. E. Human acute myeloid leukemia is organized as a hierarchy
that originates from a primitive hematopoietic cell. Nat Med, 3: 730-737, 1997.
4. Singh, S. K., Hawkins, C., Clarke, I. D., Squire, J. A., Bayani, J., Hide, T., Henkelman,
R. M., Cusimano, M. D., and Dirks, P. B. Identification of human brain tumour initiating
cells. Nature, 432: 396-401, 2004.
5. Singh, S. K., Clarke, I. D., Terasaki, M., Bonn, V. E., Hawkins, C., Squire, J., and Dirks,
P. B. Identification of a cancer stem cell in human brain tumors. Cancer Res, 63: 5821-
5828, 2003.
6. Al-Hajj, M., Wicha, M. S., Benito-Hernandez, A., Morrison, S. J., and Clarke, M. F.
Prospective identification of tumorigenic breast cancer cells. Proc Natl Acad Sci U S A,
100: 3983-3988, 2003.
7. Li, C., Heidt, D. G., Dalerba, P., Burant, C. F., Zhang, L., Adsay, V., Wicha, M., Clarke,
M. F., and Simeone, D. M. Identification of pancreatic cancer stem cells. Cancer Res, 67:
1030-1037, 2007.
8. O'Brien, C. A., Pollett, A., Gallinger, S., and Dick, J. E. A human colon cancer cell
capable of initiating tumour growth in immunodeficient mice. Nature, 445: 106-110,
2007.
9. Prince, M. E., Sivanandan, R., Kaczorowski, A., Wolf, G. T., Kaplan, M. J., Dalerba, P.,
Weissman, I. L., Clarke, M. F., and Ailles, L. E. Identification of a subpopulation of cells
with cancer stem cell properties in head and neck squamous cell carcinoma. Proc Natl
Acad Sci U S A, 104: 973-978, 2007.
66
10. Ricci-Vitiani, L., Lombardi, D. G., Pilozzi, E., Biffoni, M., Todaro, M., Peschle, C., and
De Maria, R. Identification and expansion of human colon-cancer-initiating cells. Nature,
445: 111-115, 2007.
11. Patrawala, L., Calhoun, T., Schneider-Broussard, R., Li, H., Bhatia, B., Tang, S., Reilly,
J. G., Chandra, D., Zhou, J., Claypool, K., Coghlan, L., and Tang, D. G. Highly purified
CD44+ prostate cancer cells from xenograft human tumors are enriched in tumorigenic
and metastatic progenitor cells. Oncogene, 25: 1696-1708, 2006.
12. Challen, G. A. and Little, M. H. A side order of stem cells: the SP phenotype. Stem Cells,
24: 3-12, 2006.
13. Dean, M., Fojo, T., and Bates, S. Tumour stem cells and drug resistance. Nat Rev Cancer,
5: 275-284, 2005.
14. Bunting, K. D. ABC transporters as phenotypic markers and functional regulators of stem
cells. Stem Cells, 20: 11-20, 2002.
15. Zhou, S., Schuetz, J. D., Bunting, K. D., Colapietro, A. M., Sampath, J., Morris, J. J.,
Lagutina, I., Grosveld, G. C., Osawa, M., Nakauchi, H., and Sorrentino, B. P. The ABC
transporter Bcrp1/ABCG2 is expressed in a wide variety of stem cells and is a molecular
determinant of the side-population phenotype. Nat Med, 7: 1028-1034, 2001.
16. Goodell, M. A., Brose, K., Paradis, G., Conner, A. S., and Mulligan, R. C. Isolation and
functional properties of murine hematopoietic stem cells that are replicating in vivo. J
Exp Med, 183: 1797-1806, 1996.
17. Chiba, T., Kita, K., Zheng, Y. W., Yokosuka, O., Saisho, H., Iwama, A., Nakauchi, H.,
and Taniguchi, H. Side population purified from hepatocellular carcinoma cells harbors
cancer stem cell-like properties. Hepatology, 44: 240-251, 2006.
18. Redvers, R. P., Li, A., and Kaur, P. Side population in adult murine epidermis exhibits
phenotypic and functional characteristics of keratinocyte stem cells. Proc Natl Acad Sci
U S A, 103: 13168-13173, 2006.
67
19. Jackson, K. A., Mi, T., and Goodell, M. A. Hematopoietic potential of stem cells isolated
from murine skeletal muscle. Proc Natl Acad Sci U S A, 96: 14482-14486, 1999.
20. Asakura, A. and Rudnicki, M. A. Side population cells from diverse adult tissues are
capable of in vitro hematopoietic differentiation. Exp Hematol, 30: 1339-1345, 2002.
21. Kondo, T., Setoguchi, T., and Taga, T. Persistence of a small subpopulation of cancer
stem-like cells in the C6 glioma cell line. Proc Natl Acad Sci U S A, 101: 781-786, 2004.
22. Hirschmann-Jax, C., Foster, A. E., Wulf, G. G., Nuchtern, J. G., Jax, T. W., Gobel, U.,
Goodell, M. A., and Brenner, M. K. A distinct "side population" of cells with high drug
efflux capacity in human tumor cells. Proc Natl Acad Sci U S A, 101: 14228-14233,
2004.
23. Zannettino, A. C., Harrison, K., Joyner, C. J., Triffitt, J. T., and Simmons, P. J. Molecular
cloning of the cell surface antigen identified by the osteoprogenitor-specific monoclonal
antibody, HOP-26. J Cell Biochem, 89: 56-66, 2003.
24. Haynesworth, S. E., Baber, M. A., and Caplan, A. I. Cell surface antigens on human
marrow-derived mesenchymal cells are detected by monoclonal antibodies. Bone, 13: 69-
80, 1992.
25. Tondreau, T., Meuleman, N., Delforge, A., Dejeneffe, M., Leroy, R., Massy, M., Mortier,
C., Bron, D., and Lagneaux, L. Mesenchymal stem cells derived from CD133-positive
cells in mobilized peripheral blood and cord blood: proliferation, Oct4 expression, and
plasticity. Stem Cells, 23: 1105-1112, 2005.
26. Simmons, P. J. and Torok-Storb, B. Identification of stromal cell precursors in human
bone marrow by a novel monoclonal antibody, STRO-1. Blood, 78: 55-62, 1991.
27. Gibbs, C. P., Kukekov, V. G., Reith, J. D., Tchigrinova, O., Suslov, O. N., Scott, E. W.,
Ghivizzani, S. C., Ignatova, T. N., and Steindler, D. A. Stem-like cells in bone sarcomas:
implications for tumorigenesis. Neoplasia, 7: 967-976, 2005.
28. Olstad, O. K., Gautvik, V. T., Reppe, S., Rian, E., Jemtland, R., Ohlsson, C., Bruland, O.
S., and Gautvik, K. M. Molecular heterogeneity in human osteosarcoma demonstrated by
68
enriched mRNAs isolated by directional tag PCR subtraction cloning. Anticancer Res,
23: 2201-2216, 2003.
29. Doyle, L. A., Yang, W., Abruzzo, L. V., Krogmann, T., Gao, Y., Rishi, A. K., and Ross,
D. D. A multidrug resistance transporter from human MCF-7 breast cancer cells. Proc
Natl Acad Sci U S A, 95: 15665-15670, 1998.
30. Kim, M., Turnquist, H., Jackson, J., Sgagias, M., Yan, Y., Gong, M., Dean, M., Sharp, J.
G., and Cowan, K. The multidrug resistance transporter ABCG2 (breast cancer resistance
protein 1) effluxes Hoechst 33342 and is overexpressed in hematopoietic stem cells. Clin
Cancer Res, 8: 22-28, 2002.
31. Fletcher, C. D. The evolving classification of soft tissue tumours: an update based on the
new WHO classification. Histopathology, 48: 3-12, 2006.
32. Levine, E. A. Prognostic factors in soft tissue sarcoma. Semin Surg Oncol, 17: 23-32,
1999.
33. Wunder, J. S., Healey, J. H., Davis, A. M., and Brennan, M. F. A comparison of staging
systems for localized extremity soft tissue sarcoma. Cancer, 88: 2721-2730, 2000.
34. Reya, T., Morrison, S. J., Clarke, M. F., and Weissman, I. L. Stem cells, cancer, and
cancer stem cells. Nature, 414: 105-111, 2001.
35. Patrawala, L., Calhoun, T., Schneider-Broussard, R., Zhou, J., Claypool, K., and Tang, D.
G. Side population is enriched in tumorigenic, stem-like cancer cells, whereas ABCG2+
and ABCG2- cancer cells are similarly tumorigenic. Cancer Res, 65: 6207-6219, 2005.
36. Mitsutake, N., Iwao, A., Nagai, K., Namba, H., Ohtsuru, A., Saenko, V., and Yamashita,
S. Characterization of side population in thyroid cancer cell lines: cancer stem-like cells
are enriched partly but not exclusively. Endocrinology, 2007.
37. Szotek, P. P., Pieretti-Vanmarcke, R., Masiakos, P. T., Dinulescu, D. M., Connolly, D.,
Foster, R., Dombkowski, D., Preffer, F., Maclaughlin, D. T., and Donahoe, P. K. Ovarian
cancer side population defines cells with stem cell-like characteristics and Mullerian
69
Inhibiting Substance responsiveness. Proc Natl Acad Sci U S A, 103: 11154-11159,
2006.
38. Zhou, S., Morris, J. J., Barnes, Y., Lan, L., Schuetz, J. D., and Sorrentino, B. P. Bcrp1
gene expression is required for normal numbers of side population stem cells in mice, and
confers relative protection to mitoxantrone in hematopoietic cells in vivo. Proc Natl Acad
Sci U S A, 99: 12339-12344, 2002.
39. Oda, Y., Saito, T., Tateishi, N., Ohishi, Y., Tamiya, S., Yamamoto, H., Yokoyama, R.,
Uchiumi, T., Iwamoto, Y., Kuwano, M., and Tsuneyoshi, M. ATP-binding cassette
superfamily transporter gene expression in human soft tissue sarcomas. Int J Cancer, 114:
854-862, 2005.
40. Noonan, K. E., Beck, C., Holzmayer, T. A., Chin, J. E., Wunder, J. S., Andrulis, I. L.,
Gazdar, A. F., Willman, C. L., Griffith, B., Von Hoff, D. D., and et al. Quantitative
analysis of MDR1 (multidrug resistance) gene expression in human tumors by
polymerase chain reaction. Proc Natl Acad Sci U S A, 87: 7160-7164, 1990.
41. Jonker, J. W., Freeman, J., Bolscher, E., Musters, S., Alvi, A. J., Titley, I., Schinkel, A.
H., and Dale, T. C. Contribution of the ABC transporters Bcrp1 and Mdr1a/1b to the side
population phenotype in mammary gland and bone marrow of mice. Stem Cells, 23:
1059-1065, 2005.
70
CHATPER 3
Blockade of hedgehog signaling inhibits the formation of tumors
derived from osteosarcoma side population cells
I preformed all experiments described in this chapter.
71
Chapter 3
3 Blockade of hedgehog signaling inhibits the formation of tumors derived from osteosarcoma side population cells
3.1 Abstract
Our previous work demonstrates that side population (SP) cells are enriched for tumorigenic
potential in a broad range of primary mesenchymal neoplasms. Identification and manipulation
of the biological features underlying side population tumorigenesis may result in the
development of more effective treatment modalities against these neoplasms. Using
osteosarcoma cell lines as a model system, we showed that osteosarcoma side population cells
were up-regulated in the expression of the downstream Hedgehog targets genes Gli-1 and Ptch
when compared to their non-side population counterparts. In vitro blockade of the hedgehog
pathway with the chemical agent triparanol diminished the proportion of side population cells
when compared to vehicle control. Importantly, in primary transplants, triparanol treatment
resulted in an impairment of growth in side population derived tumors but not non-side
population derived tumors. Furthermore, treatment with triparanol decreased tumor cellularity
and the percentage of side population cells within SP derived tumors. Taken together, we
indentified a biologically distinguishing property between osteosarcoma side population and
non-side population cells that has the potential to be exploited and used as a treatment modality
for this mesenchymal neoplasm.
72
3.2 Introduction
Osteosarcoma (OS) is the most common primary bone cancer in children and adolescence
representing 15% of all primary bone tumors and 0.2% of all malignant tumors in children[1, 2].
While survival rates have improved over the last decade with patients presenting with non-
metastatic disease having a 70% chance of long term survival, patients presenting with
metastases at time of diagnosis or with recurrent disease have only 20% chance of survival at 5
years[3, 4]. Osteosarcoma is a highly aggressive tumor with approximately 10% of patients
presenting with lung metastases at diagnosis[5]. The mechanisms underlying osteosarcoma
initiation and progression are poorly understood; however, mutations in both p53 and Rb are
associated with the disease. Patients with Li-freumeni syndrome, an autosomal dominant
disorder characterized by germline mutations in p53, develop a wide variety of tumor including
osteosarcomas[6, 7]. Also, genetic alterations of RB have been found in up to 70% of sporadic
cases of osteosarcoma[8, 9]. Conventional treatment for this malignancy includes a combination
of chemotherapy and surgery has significantly improved the survival rate of certain patients,
however, the overall long-term disease-free survival for patients diagnosed with osteosarcoma is
still low, indicating the need for exploration of new treatment options.
During normal fetal development various cell signaling pathways are regulated in a coordinated
manner that allows for stem cells to proliferate, differentiate, move and die to allow for normal
organ formation. Later in life cells in neoplastic process can utilize these same pathways to drive
tumorigenesis. Understanding the manipulation of these pathways by malignant cells may reveal
novel therapeutic approaches based on modulating the activities of these pathways in neoplastic
entities. One potential pathway for exploitation in the treatment of malignancies is the hedgehog
pathway, as it controls cell growth and differentiation during embryonic development and
deregulation of it is associated with the development of a variety of epithelial neoplasms[10, 11].
Briefly, the mechanism of signal transduction in this pathway involves the binding of secreted
hedgehog ligand to the Patched 1 (PTCH1) receptor on target cells. This interaction relieves the
inhibition of Smoothened (SMO) by PTCH1, and SMO sends signals through a series of
interacting proteins, including suppressor of fused, resulting in activation of the downstream Gli
family of transcription factors: GLI1, GLI2 and GLI3[10]. Patients with Gorlin’s syndrome have
germline mutations in the PTCH1 gene predisposing them to the development of basal cell
carcinoma and medulloblastoma[12, 13]. Also, both sporadically occurring basal cell carcinomas
73
and medulloblastoma show up-regulation of various components of the hedgehog signaling
pathway[14, 15]. Deregulation of this pathway is also implicated in cartilaginous lesions, with
inhibition leading to a decrease in tumor volume, cellularity, and proliferation rates of
xenografted chondrosarcomas [16-18]. Osteosarcoma cell lines show elevated expression of
hedgehog signaling components and chemical blockade of cell lines with high level of GLI1 and
PTCH1 leds to impaired in vitro cellular proliferation[19]. However, the means by which
hedgehog signaling hinders the advancement of both epithelial and mesenchymal tumorigenesis
has yet to be fully elucidated.
The cancer stem cell theory is based on the observation that tumors are composed of a
heterogeneous population of cells organized into a cellular hierarchy. At the top of this hierarchy,
resides the cancer stem cell (CSC) or tumor initiating cell (TIC) that has the capacity to self-
renew and to differentiate into all the cells composing the heterogeneous tumor. The prospective
identification of TICs from a broad range of hematopoietic malignancies and solid tumors
strongly supports this hypothesis. Notably, only CSCs have the potential to induce tumor
formation when injected into immuno-deficient mice and a such, the CSC theory has important
clinical implications as it suggests that these cells need to be eliminated in order for disease free
survival to occur. As CSCs are thought to behave in a fashion similar to normal stem cells, it can
be reasoned that CSCs utilize similar signaling pathways to regulate self-renewal and
differentiation. As such, development of treatment strategies that specifically target these
pathways have begun to emerge[20]. Hedgehog signaling plays a critical role in the development
of both normal and malignant stem cells. It can be postulated that inhibition of this signaling
pathway may reduce tumor burden by effectively targeting the most potent malignant cells for
destruction. Treatment of glioma CSC with the hedgehog inhibitor, cyclopamine inhibits
proliferation and self-renewal while increasing apoptosis in the CSC population[21]. In breast
CSCs hedgehog signaling is also active[22], and treatment with cyclopamine or anti-hedgehog
blockade antibody reduces the clonogenicity of CD19+CD37+ multiple myeloma CSCs[23].
We previously demonstrated that primary mesenchymal tumors contain a subpopulation of cells,
SP cells, that preferentially form tumors when subcutaneously injected into immuno-deficient
mice[24]. We sought to identify a biological feature of osteosarcoma side population cells that
could be exploited to reduce their tumor forming capacity. As hedgehog signaling is implicated
74
in osteosarcoma pathogenesis, we sought to determine if chemical inhibition of this pathway in
osteosarcoma side population cells would reduce their tumorigenic potential.
75
3.3 Results
3.3.1 Side population cells are present in osteosarcoma cell lines
We previously identified the presence of SP cells in a broad range of primary mesenchymal
neoplasms, including osteosarcomas. To determine if osteosarcoma cell lines would recapitulate
this phenotype, we screened four well-characterized osteosarcoma cell lines; MG63, HOS, HOS-
MN, and KHOS for the presence of side population cells. Cells from each line were stained with
the fluorescent dye, Hoechst 33342, and staining was quantified using a fluorescence activated
cytometer. With the exception of the MG63, all other cell lines contained side population cells
with the percentage ranging from 0.58%-0.73% of the total cell population (Figure 3.1A&B).
This percentage falls within the range found in primary mesenchymal neoplasms[34], suggesting
that osteosarcoma cell lines maintain some features of primary mesenchymal neoplasms.
Staining in the presence of verapamil, an inhibitor of the ATP-binding cassette (ABC) family of
transporters, resulted in the loss of the side population fraction, indicating that the presence of
side populations cells are due in part, to the expression of the ABC family of transporters (Figure
3.1A). Neither passage number nor confluency of the cell lines at the time of staining impacted
the percentage of side population cells (data not shown).
3.3.2 Hedgehog signaling in osteosarcoma side population cells
Components of the hedgehog pathway are activated in mesenchymal neoplasms[20, 21, 35]. In
light of these findings, we sought to determine if differential activation of the hedgehog pathway
would be detected between osteosarcoma side population and non side population cells. KHOS
and HOS-MN cells were sorted into side population and non-side population fractions, RNA was
isolated, and real- time RT-PCR was performed to assess the levels of the hedgehog downstream
targets GLI-1 and PTCH-1. When compared to non side population cells, side population cells
showed a 3 fold higher expression in the hedgehog downstream target GLI-1 and a two fold
higher expression of the downstream target PTCH1 (Figure 3.2 A, B). Furthermore, KHOS cells
treated with 11.25µM of triparanol, a chemical inhibitor of the hedgehog pathway[36], exhibited
a diminished percentage of side population cells (Figure 3.2 C), suggesting that hedgehog
signaling may be important in regulating the behavior of these cells in osteosarcomas.
76
3.3.3 In vivo tumorigenic potential of osteosarcoma side population and non-side population cells
As blockade of the hedgehog signaling pathway is an effective therapeutic treatment for a variety
of tumor types[14, 37, 38], we next investigated the potential of exploiting this pathway to
therapeutic target side population cells in osteosarcomas. Initially, we ascertained the
tumorigenic potential of osteosarcoma side population and non-side population cells. 5000 and
10,000 side population and non-side population cells isolated from HOS-MN cell lines were
injected subcutaneously into NOD/SCID mice. For the KHOS cell line 100, 1000, 5000, and
10,000 cells were injected into immuno-compromised mice. As with primary mesenchymal
tumors, both HOS and KHOS side population cells and non-side population cells had the
potential to form tumors upon primary transplantation. In the KHOS cell line, as cell dosage
increased, the frequency of tumor formation also increased (Table 3.1). However, unlike primary
mesenchymal neoplasms, no statistically significant differences were seen in tumor formation
frequency between the side population cells and the non-side population cells in either of the two
cell lines.
3.3.4 Blockade of hedgehog signaling inhibits in vivo tumor formation of side population cells
Next, KHOS cells were sorted into side population and non side population fractions and 1000
cells of each population were subcutaneously injected into immuno-compromised mice. Mice
were treated with 400mg/kg of triparanol for four weeks after which they were sacrificed and
tumor formation was assessed. Blockade of the hedgehog pathway reduced both the frequency
and size of tumors derived from side population cells when compared to tumors derived from
non-SP cells (Figure 3.3 A, B). No statistically significant differences in the weight of the mice
were detected (treated versus control), suggesting that effects were not due to a cytotoxic impact
on the animals (Figure 3.3 C). Tumors exposed to triparanol showed decreased expression of
GLI-1 when compared to untreated tumors (Figure 3.3 D). Interestingly, despite the lack of
differences seen in tumor formation frequency between the two populations of cells, only side
population cells were sensitive to chemical blockade of the hedgehog pathway, suggesting that
exploitation of this pathway may preferentially target side population cells.
77
3.3.5 Blockade of Hedgehog Signaling decreases the cellularity of side population derived tumors
Xenografted tumors were formalin fixed, paraffin embedded and stained with hematoxylin and
eosin. The grafted tumors had nearly identical cytological appearances; however, SP tumors
exposed to triparanol showed decreased cellularity when compared to DMSO controls. No
differences were seen between treated and control groups in the non-side population cell derived
tumors (Figure 3.4 A, B). Single cell suspensions generated from harvested tumors were stained
with Hoechst 33342 and analyzed for side population cells by flow cytometery. Similar to our in
vitro findings, in vivo blockade of the hedgehog pathway resulted in diminishment of the
proportion of side population cells when compared to untreated controls (Figure 3.4 C).
3.4 Discussion
We sought to identify a biological property that could be exploited to therapeutically target the
tumorigenic potential of mesenchymal side population cells. To do so, we demonstrated that
osteosarcoma side population cells have increased expression of the downstream hedgehog
signaling target genes, Gli-1 and PTCH1 when compared to non-side population cells. In vitro
blockade of hedgehog signaling with the chemical agent triparanol diminished the proportion of
side population cells found in osteosarcoma cell lines. Both side population and non-side
population cells were able to induce tumor formation in primary transplants in immuo-deficient
NOD/SCID mice; however, only side population derived tumors were sensitive to inhibition of
hedgehog signaling. Specifically, triparanol treatment resulted in a decrease in both the tumor
frequency and size of side population derived tumors. Importantly, inhibition of hedgehog
signaling had no impact on tumors derived from non-side population cells. Furthermore, SP
tumors exposed to triparanol showed decreased cellularity and flow cytometry analysis revealed
diminished numbers of side population cells within these tumors when compared to untreated
lesions. Taken together, this suggests that the hedgehog pathway may influence the development
of side population derived osteosarcoma tumors; and as such, targeting this pathway may be an
effective treatment modality for this mesenchymal neoplasm.
The mechanism by which hedgehog blockade suppresses tumor growth in osteosarcoma side
population cells is currently being investigated. We can postulate that triparanol treatment has
the potential to inhibit self-renewal, alter proliferation/differentiation, increase apoptosis, or any
78
combination of these three events within the side population. The capacity of SP cells to serially
engraft when treated with triparanol would focus on the influence of hedgehog blockade on the
self-renewal potential of these cells. Flow cytometry cell cycle analysis of DNA content coupled
with immunohistochemistry of tumor tissue sections (ki-67, tunnel staining) could reveal
differences in proliferative capacity and/or apoptotic status of side population cells treated with
triparanol. Both in vitro and in vivo treatment of cells treated with triparanol resulted in a
diminished proportion of side population cells. As mentioned, one possible explanation for this
observation is the impairment of SP cell proliferation and/or self-renewal. Alternatively,
inhibition of hedgehog signaling may impact the expression of the ABCG family of transporters,
a key determinant of the side population phenotype[29, 30]. Expression analysis of these
transporters will yield important information to explain the diminished proportion of SP cells in
the presence of triparanol. However, regardless of the mechanism, it is important to note that
tumors derived from non-side population cells were not effected by hedgehog blockade
demonstrating a biologically distinct property of side population cells.
The most common location of osteosarcomas in young adults is in areas of rapid bone
growth[31]. In these areas, osteoblasts, which are thought to derive from an early mesenchymal
precursor, undergo a well-regulated terminal differentiation process. Given the location, cellular
origins, and histological heterogeneity of osteosarcoma; it can be postulated that they arise from
the aberrant differentiation of mesenchymal or osteoblastic precursor[32]. Furthermore, in mice,
conditional loss of p53 within osteoblast precursors resulted in the formation of osteosarcomas
with 100% penetrance[33]. Also, as side population cells isolated from primary mesenchymal
neoplasms have stem-like characteristics[24], it is possible, that osteosarcoma side population
cells may also share this feature. It would be interesting to speculate whether triparanol
specifically inhibits the growth of a stem-like population within osteosarcomas. However, one
major caveat to our stud, is the use of osteosarcoma cell lines as a model system to investigate
the tumorigenic phenotype of side population cells in relation to non-side population cells within
osteosarcomas. The KHOS and HOS-MN cell lines contained a similar percentage of SP cells to
primary osteosarcoma tumors. Furthermore, both populations formed tumors upon primary
transplantation into NOD/SCID mice, thereby demonstrating some similarity of the cell lines to
their primary counterparts[24]. However, unlike primary mesenchymal neoplasms, no
differences were detected in either tumor frequency or tumor size (data not shown) between the
79
side population and non-side population derived tumors, highlighting important behavioral
differences between cell lines and primary neoplastic tumors. In glioblastomas, tumor stem cells
isolated from primary human tumors showed marked differences in gene expression patterns and
in vivo tumor biology compared to matched glioma cell lines[34]. As such, it will be important to
determine our observations can be recapitulated in primary osteosarcoma tumors.
While studies examining a broad range of epithelial carcinomas demonstrate hedgehog blockade
results in the inhibition of tumor size the exact cause of this phenomenon is currently being
debated. As oncogenic cells are not independent entities, but rather act in concert with the tissue
microenvironment, it can be hypothesized that the impact of hedgehog blockade on tumor
formation being acts on extrinsic cellular influences, rather than intrinsic cell signaling. In some
studies, the decrease in tumor size seen in epithelial carcinomas exposed to hedeghog blockade
was attributed to an inhibition of signal responsiveness in the surrounding stromal cells rather
than of the epithelial derived cancers[35, 36]. The exact means by which the inhibition of
hedgehog signaling in the murine stroma impacts the development of human carcinomas has yet
to be elucidated. Regardless of the mechanism of action, this study raises the intriguing
possibility that in xenografted studies the surrounding stroma, may be at least in part, more
sensitive to the loss of hedgehog signaling and this results in the subsequent inhibition of tumor
growth. Given this observation, it can be postulated that osteosarcoma side population cells have
a heavier reliance on the surrounding stroma and blockade of hedgehog signaling may result in
the diminished tumor volumes seen in the side population but not non-side population derived
tumors. Examination of the expression of GLI-1 in the surrounding stroma through the use of
mouse specific primers may help to support this hypothesis.
Blockade of the hedgehog pathway has proved to be successful in the treatment of
medulloblastomas; specifically, chemical blockade of the Smoothened receptor results in a
dramatic reduction of metastatic tumor burden[28, 37]. However, the use of hedgehog inhibitors
for the treatment of osteosarcomas may prove to be more elusive. As hedgehog signaling is
critical mediator of normal bone development during endochondral ossification[38], the use of
hedgehog blocking agents in young patients whose growth plates are still active may prove to be
problematic. Studies demonstrate that transient inhibition of this pathway in young mice causes
permanent defects in bone structures[39]. Given these findings, the use of hedgehog blockade in
80
young osteosarcoma patients may not be ideal; however, in it may be an option for older patients
whose the growth plates are no longer in their most active state.
81
3.5 Figures
Figure 3.1 Side population cells are present in osteosarcoma cell lines
(A) MG63, HOS, HOS-MN, and KHOS cell lines were stained with Hoechst 33342 dye. To
detect for SP, cells were analyzed by using dual wavelength analysis (blue, 424-444nm; red,
675nM) after excitation with 350nm UV light. Representative staining samples of each cell line
are shown. The SP cells are outlined in red and shown as a percentage of the total cell
population. In the presence of verapamil, the percentage of SP cells is greatly diminished. (B)
Graphical data represents the mean percentage from 3 independent trials and error bars represent
95% confidence intervals. A 95% confidence interval that does not cross the mean of a
comparison is a statistically significant difference at p<0.05.
82
HO Blue-A
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
0.61
HO Blue-A
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
0.044
HO Blue-A
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
0.017
HO Blue-A
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
9.63e-3
HO Blue-A
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
0.58
HO Blue-A
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
0.099
HO Blue-A
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
0.73
HO Blue-A
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
0.39
!"#$ %&' %&'(!) *%&'
Hoechst Red
Hoe
chst
Blu
e
0.0
0.5
1.0
1.5
2.0
A
B
!"#$
%&'
%&'(!
)
*%&'
% S
ide
Popu
latio
n
83
Figure 3.2 Hedgehog signaling in osteosarcoma side population cells
(A, B) HOS-MN and KHOS cells were stained with Hoechst 33342 dye and sorted into side
population and non-side population cells. RNA was isolated and real time RT-PCR was
performed using primers specific to human GLI-1 and PTCH. There is a 3 fold increase in
expression of GLI-1 and a 2 fold increase in expression of PTCH in side population cells when
compared to non-side population cells. Expression of GLI-1 and PTCH was determined by
taking the ratio of gene expression over the expression of perspective housekeeping genes (data
not shown). Fold increases were then determined by taking the ratio of GLI-1 and PTCH
expression of side population cells over their expression in non-side population cells. Data
shown is the mean percentage from 3 independent trials and error bars represent 95% confidence
intervals. A 95% confidence interval that does not cross the mean of a comparison is a
statistically significant difference at p<0.05. (C) KHOS cells were treated with 11.25µM of
triparanol in DMSO for 24 hours. Treatment group exhibited a decrease in the proportion of side
population when compared to DMSO carrier control. Data shown is a representative sample from
three independent trials.
84
0 1000 2000 3000 4000
FL 7 Lin: FL 7 Hst-Red
0
1000
2000
3000
4000
FL 6 Lin: FL 6 Hst-Blue
0.34
0 1000 2000 3000 4000
FL 7 Lin: FL 7 Hst-Red
0
1000
2000
3000
4000
FL 6 Lin: FL 6 Hst-Blue
0.025
DMSO Triparanol
Hoechst Red
Hoe
chst
Blu
e
0
1
2
3
4
5
Fold
Incr
ease
SP/
nonS
P G
li-1
KHOS
KHOS
HOS-MN
! "
Fold
Incr
ease
SP/
nonS
P PT
CH
#
0.0
0.5
1.0
1.5
2.0
2.5
3.0
KHOS HOS-MN
85
Table 3.1 Proportion of tumors formed in NOD/SCID mice from injection of various
numbers of side population and non-side population cells.
HOS-MN and KHOS side population and non-side population cells were subcutaneously
injected into immuno-deficient NOD/SCID mice. After 6 weeks mice were euthanized and tumor
formation was assessed. The number of cells that were injected into each mouse is denoted. The
number of tumors formed and the number of injections that were performed are indicated for
each population. The frequency of tumor formation is indicated as percentages. Chi-squared
analysis demonstrated no statistical significance (p value < 0.05) in the frequency of tumor
formation between side population and non-side population cells.
86
Cell Type
HOS-MN
Cell
Dose
(number of primary
mice with
tumors)/(total number
injected)
Frequency of
tumor formation
(percent)
Side Population 5x103 11/15 73%
1x 104 16/16 100%
Non Side Population 5x103 11/14 78%
1x 104 12/16 75%
Cell Type
KHOS
Cell
Dose
(number of primary
mice with
tumors)/(total number
injected)
Total number of
mice with
tumors (percent)
Side population 1x102 3/8 37.5%
1x103 5/9 55%
5x103 9/13 69%
1x 104 16/16 100%
Non Side Population 1x102 1/8 12.5%
1x103 6/10 60%
5x103 13/17 76%
1x 104 14/16 87%
87
Figure 3.3 Blockade of hedgehog signaling inhibits in vivo tumor formation of osteosarcoma side population cells
(A) Mice injected with 1000 KHOS side population and 1000 KHOS non side population cells
were treated with 400mg/kg of triparanol or DMSO by oral gauvage for four weeks after which
they were sacrificed and tumor formation was assessed. Fewer tumors formed in triparanol
treated mice when compared to control mice (DMSO). No difference in tumor frequency was
detected between treated and control groups of mice injected with non-side population cells.
Differences in tumor frequency was determined using a Fisher’s exact test with two tailed p
values <0.05 considered statistically significant. (B) Triparanol treated side population tumors
also weighed less when compared to control group. No difference in weight was detected in
treated or untreated tumors that formed from the non-side population cells. Data shown is the
mean percentage of tumor weight and error bars represent 95% confidence intervals. A 95%
confidence interval that does not cross the mean of a comparison is a statistically significant
difference at p<0.05. (C) No difference in mice weight of mice detected between the two groups
at the time of sacrifice.
88
Number of Cells
SP Non-SP
Tumor Formation (frequency)
1000 5/7 (71%)6/8 (75%)
KHOS
Number of Cells
SP Non-SP
Tumor Formation (frequency)
1000 5/7 (71%)2/7 (28%)
KHOS
0
200
400
600
800
1000
Tum
or w
eigh
t (m
g)
DM
SO
Trip
aran
ol
DM
SO
Trip
aran
ol
0
5
10
15
20
25
30
35
mou
se w
eigh
t (g)
DM
SO
Trip
aran
ol
SP non-SP SP
0.0
0.2
0.4
0.6
0.8
1.0
SP non-SP
Fold
dec
reas
e Tr
ipar
anol
/DM
SO G
li-1
!"
#
$
89
Figure 3.4 Side population tumors treated with triparanol have decreased cellularity
(A) Harvested tumors were formalin fixed, stained with hematoxylin and eosin, and nuclei were
counted. Representative slides of the tumors are shown. (B) SP tumors subjected to triparanol
had fewer numbers of nuclei when compared to DMSO controls. No difference in cellularity was
detected between treated and control groups from tumors derived from non-side population cells.
Graphical data represents as the mean number of nuclei counted in each field of vision at 400X
magnification. Error bars represent 95% confidence intervals. A 95% confidence interval that
does not cross the mean of a comparison is a statistically significant difference at p<0.05. (C)
Real time RT-PCR revealed that treatment with triparanol resulted in the decrease in expression
of GLI-1 in SP derived tumors compared to untreated controls. Data is represented as a ratio of
gene expression of triparanol treated tumors relative to DMSO treated tumors. (D) Harvested
tumors derived from SP cells were dissociated, stained with Hoechst 33342 dye, and analyzed
with a flow cytometer. SP tumors exposed to triparanol had a lower proportion of SP cells when
compared to DMSO controls.
90
0
50
100
150
200
250
side
pop
ulat
ion
non
-sid
e po
pula
tion
DMSO Triparanol
Num
ber o
f Cel
ls
DM
SO
Trip
aran
ol
DM
SO
Trip
aran
ol
SP non-SP
!
"
Hoechst Red
Hoe
chst
Blu
e
DMSO Triparanol#
0 1000 2000 3000 4000
FL 7 Lin: FL 7 Hst-Red
0
1000
2000
3000
4000
FL 6 Lin: FL 6 Hst-Blue
0.38
0 1000 2000 3000 4000
FL 7 Lin: FL 7 Hst-Red
0
1000
2000
3000
4000
FL 6 Lin: FL 6 Hst-Blue1.12
91
3.6 Materials and Methods
3.6.1 Cell lines
MG63, HOS, HOS-MN, and KHOS cells were all obtained from the American Type Culture
Collection (Rockville, MD). Cells were routinely maintained in DMEM (Wisent) supplemented
with 10% fetal bovine serum (Wisent). Cells were routinely maintained and cultured at 37oC
with 5% CO2 in a humidified chamber until subjected to fluorescent activated cells sorting
(FACS).
3.6.2 Flow cytometry
For side population analysis 1.0 x 106 cells/mL were treated either alone or with 2.5 µg/mL of
Hoechst 33342 dye (Sigma) for 90 min at 37oC, or in combination with 50µmol/L verapamil
(Sigma) as previously reported[21]. To detect for SP, cells were analyzed by using a dual
wavelength analysis (blue, 424-444nm; red, 675nM) after excitation with 350nm UV light
(MoFlow, Cytomation). After staining all cells were then counterstained with 1µg/mL of
propidium iodide (Molecular Probes) and PI positive (non-viable) cells were excluded from
analysis. Cell staining was quantified using LSRII flow cytometer (Becton Dikinson).
3.6.3 RNA extraction and real time-RT-PCR
For sorted cells, after staining with Hoechst 33342 dye, cells were sorted in SP and non-SP
fractions. Sorted cells were rinsed one with 1X PBS and centrifuged at 1200rpm for 10 minutes.
RNA was extracted from cell pellets using RNeasy Mirco Kit (Qiagen) according to
manufacturer’s instructions. RNA quality was assessed with a Bioanalyzer (Agilent
Technologies) For xenografted tumor tissues; tumors were flash frozen with liquid nitrogen and
disociated with a mortar and pestle. RNA was then extracted using Trizol (Invitrogen) according
to manufacturer’s instructions and quality was assessed with a Bioanalyzer (Agilent
Technologies). cDNA was generated using 1st Strand Synthesis Kit (Invitrogen) according to
manufacturer’s instructions. For real-time RT-PCR gene specific Taq-Man fluorogenic probes
for human Gli-1 and Ptch was used. We performed standard quantitative RT-PCR reactions for
GLI-1 and PTCH on the ABI prism (Applied Biosystems) sequence detection system.
Asparagine synthetase (AS) and glyceraldehyde-3-phosphate (GAPDH) were used as internal
92
control genes. The reactions were performed in triplicate in 20ul reaction volume using TaqMan
Universal PCR Master Mix (ABI) on a 96 well plate format. Expression of GLI-1 and PTCH in
individual populations of cells was first determined by taking the ratio of gene expression over
their perspective housekeeping gene. Fold increases between the two groups were determined by
taking the ratio of side population expression over non-side population expression.
3.6.4 Xenograft models
Sorted SP and non-SP cells were collected and cells were re-suspended in PBS at concentrations
ranging from 100-10,000 cells/50µl. Cells were then mixed with 50µl of Matrigel (Becton
Dickinson). This cell:matrigel suspension was then subcutaneously injected in to eight to ten
week old NOD/SCID mice. Mice were observed for up to 6 weeks after which they were
euthanized and tumor formation was assessed. Tumors that formed were removed and samples
from each tumor were harvested for FACS and histology. For histology, tumors were removed
and samples were paraffin embedded, formalin fixed, sectioned, and stained for hemotoxylin and
eosin. Staining was visualized using a Leica light microscope
3.6.5 Dissociation of xenografted tumors
Primary tumor samples were manually minced and all visible clumps removed. Enzymatic
digestion followed at 37oC for 45 minutes with constant rotation using 10mg/ml of collagenase
IV (Worthington), 2.4 U/ml of Dispase (Becton Dickinson), 0.05% trypsin (Wisent). Further
manual dissociation was performed by passing the cell slurry through an 18 gauge needle. Cells
were then centrifuged at 1400 rpm for 5 minutes and washed three times in PBS. After washing,
cells were strained through 70µm filters to remove remaining clumps. Single cell suspensions
were then stained with Hoechst 33342 dye and subjected to FACS
3.6.6 In vitro and in vivo blockade of hedgehog signaling
For in vitro experiments, cells were treated with 11.25uM of triparanol and control cells were
treated with carrier, DMSO. After 24 hours cells were detached with 0.05% trypsin in PBS
(Wisent), stained with Hoechst 33342 dye and subjected to FACS analysis (see above). Two
weeks prior to the initial injections treatment of mice began. For in vivo experiments, triparanol
was dissolved in olive oil and mice were orally gauvaged with 10mg/kg of triparanol every other
day for 4 weeks. By gauvage, one group of mice was treated with triparanol, three times a week,
93
at a dosage of 400mg/kg. The control mice were treated with the carrier, olive oil. Mice were
treated for four weeks after with they were euthanized and tumors were harvested. Harvested
tumors were either dissociated to single cell suspension were generated (see below), paraffin
embedded, formalin fixed, sectioned, and stained for hemotoxylin and eosin (staining was
visualized using a Leica light microscope ) or 3) frozen for RNA extraction (see above).
94
3.7 References
1. Zhou, B.B., et al., Tumour-initiating cells: challenges and opportunities for anticancer
drug discovery. Nat Rev Drug Discov, 2009. 8(10): p. 806-23.
2. Piccirillo, S.G., et al., Bone morphogenetic proteins inhibit the tumorigenic potential of
human brain tumour-initiating cells. Nature, 2006. 444(7120): p. 761-5.
3. Hoey, T., et al., DLL4 blockade inhibits tumor growth and reduces tumor-initiating cell
frequency. Cell Stem Cell, 2009. 5(2): p. 168-77.
4. Jiang, J. and C.C. Hui, Hedgehog signaling in development and cancer. Dev Cell, 2008.
15(6): p. 801-12.
5. Taipale, J. and P.A. Beachy, The Hedgehog and Wnt signalling pathways in cancer.
Nature, 2001. 411(6835): p. 349-54.
6. Hahn, H., et al., Mutations of the human homolog of Drosophila patched in the nevoid
basal cell carcinoma syndrome. Cell, 1996. 85(6): p. 841-51.
7. Johnson, R.L., et al., Human homolog of patched, a candidate gene for the basal cell
nevus syndrome. Science, 1996. 272(5268): p. 1668-71.
8. Dahmane, N., et al., Activation of the transcription factor Gli1 and the Sonic hedgehog
signalling pathway in skin tumours. Nature, 1997. 389(6653): p. 876-81.
9. Unden, A.B., et al., Human patched (PTCH) mRNA is overexpressed consistently in
tumor cells of both familial and sporadic basal cell carcinoma. Cancer Res, 1997.
57(12): p. 2336-40.
10. Bhardwaj, G., et al., Sonic hedgehog induces the proliferation of primitive human
hematopoietic cells via BMP regulation. Nat Immunol, 2001. 2(2): p. 172-80.
11. Trowbridge, J.J., M.P. Scott, and M. Bhatia, Hedgehog modulates cell cycle regulators in
stem cells to control hematopoietic regeneration. Proc Natl Acad Sci U S A, 2006.
103(38): p. 14134-9.
95
12. Zhao, C., et al., Hedgehog signalling is essential for maintenance of cancer stem cells in
myeloid leukaemia. Nature, 2009. 458(7239): p. 776-9.
13. Dierks, C., et al., Expansion of Bcr-Abl-positive leukemic stem cells is dependent on
Hedgehog pathway activation. Cancer Cell, 2008. 14(3): p. 238-49.
14. Bar, E.E., et al., Cyclopamine-mediated hedgehog pathway inhibition depletes stem-like
cancer cells in glioblastoma. Stem Cells, 2007. 25(10): p. 2524-33.
15. Liu, S., et al., Hedgehog signaling and Bmi-1 regulate self-renewal of normal and
malignant human mammary stem cells. Cancer Res, 2006. 66(12): p. 6063-71.
16. Peacock, C.D., et al., Hedgehog signaling maintains a tumor stem cell compartment in
multiple myeloma. Proc Natl Acad Sci U S A, 2007. 104(10): p. 4048-53.
17. Tiet, T.D., et al., Constitutive hedgehog signaling in chondrosarcoma up-regulates tumor
cell proliferation. Am J Pathol, 2006. 168(1): p. 321-30.
18. Hopyan, S., et al., Dysregulation of hedgehog signalling predisposes to synovial
chondromatosis. J Pathol, 2005. 206(2): p. 143-50.
19. Hopyan, S., et al., A mutant PTH/PTHrP type I receptor in enchondromatosis. Nat Genet,
2002. 30(3): p. 306-10.
20. Warzecha, J., et al., Inhibition of osteosarcoma cell proliferation by the Hedgehog-
inhibitor cyclopamine. J Chemother, 2007. 19(5): p. 554-61.
21. Wu, C., et al., Side population cells isolated from mesenchymal neoplasms have tumor
initiating potential. Cancer Res, 2007. 67(17): p. 8216-22.
22. Cooper, M.K., et al., Teratogen-mediated inhibition of target tissue response to Shh
signaling. Science, 1998. 280(5369): p. 1603-7.
23. Skubitz, K.M. and D.R. D'Adamo, Sarcoma. Mayo Clin Proc, 2007. 82(11): p. 1409-32.
24. Tang, N., et al., Osteosarcoma development and stem cell differentiation. Clin Orthop
Relat Res, 2008. 466(9): p. 2114-30.
96
25. Yauch, R.L., et al., A paracrine requirement for hedgehog signalling in cancer. Nature,
2008. 455(7211): p. 406-10.
26. Shaw, A., J. Gipp, and W. Bushman, The Sonic Hedgehog pathway stimulates prostate
tumor growth by paracrine signaling and recapitulates embryonic gene expression in
tumor myofibroblasts. Oncogene, 2009.
27. Rudin, C.M., et al., Treatment of medulloblastoma with hedgehog pathway inhibitor
GDC-0449. N Engl J Med, 2009. 361(12): p. 1173-8.
28. Yauch, R.L., et al., Smoothened mutation confers resistance to a Hedgehog pathway
inhibitor in medulloblastoma. Science, 2009. 326(5952): p. 572-4.
29. de Crombrugghe, B., V. Lefebvre, and K. Nakashima, Regulatory mechanisms in the
pathways of cartilage and bone formation. Curr Opin Cell Biol, 2001. 13(6): p. 721-7.
30. Kimura, H., J.M. Ng, and T. Curran, Transient inhibition of the Hedgehog pathway in
young mice causes permanent defects in bone structure. Cancer Cell, 2008. 13(3): p. 249-
60.
31. Lee, J., et al., Tumor stem cells derived from glioblastomas cultured in bFGF and EGF
more closely mirror the phenotype and genotype of primary tumors than do serum-
cultured cell lines. Cancer Cell, 2006. 9(5): p. 391-403.
97
CHATPER 4
The development of aggressive fibromatosis (desmoid tumor) is
influenced by mesenchymal progenitor cells
I preformed all experiments described in the chapter with the following exceptions:
• Puviindran Nadesan helped in managing the mouse colony • William Stanford generated the Sca-1-/- mice
98
Chapter 4
4 The development of aggressive fibromatosis (desmoid tumor) is influenced by mesenchymal progenitor cells
4.1 Abstract
The cellular origins from which most tumors arise are poorly defined. In pre-neoplastic
conditions, this knowledge can lead to the development of strategies to suppress tumor
development. Aggressive fibromatosis, also known as desmoid tumor, is a locally invasive soft
tissue tumor, whose cellular origin is undefined, but has mesenchymal cell characteristics. We
found that aggressive fibromatosis contain a subpopulation of cells that exclude Hoescht dye and
express cell surface markers found on mesenchymal progenitor cells (MPCs), demonstrating
mesenchymal progenitor features within a subset of aggressive fibromatosis cells. Using a mouse
model genetically predisposed to aggressive fibromatosis (Apcwt/1638N), we found that the number
of tumors that developed was directly proportional to the number of MPCs present as measured
by colony forming units-fibroblastic (CFU-F). Sca1-/- mice, which develop fewer MPCs, were
crossed with Apcwt/1638N mice. Apcwt/1638N/Sca1-/- mice developed substantially fewer aggressive
fibromatosis tumors. In contrast, Sca-1 deficiency had no affect on the formation of epithelial
derived intestinal polyps. Finally, MPCs isolated from Apcwt/1638N mice had the capacity to induce
aberrant cellular growth when injected into immuno-compromised mice. Taken together, this
suggests that MPCs can influence in the formation and development of aggressive fibromatosis.
99
4.2 Introduction
Mesenchymal tumors display a great deal of cellular heterogeneity with a subpopulation of cells,
side population cells (SP) showing enhanced tumor initiating potential. These observations
suggest that this tumor type may be organized into a cellular hierarchy with SP cells behaving
like cancer stem cells driving tumorigenesis in established lesions[1, 2]. However, these findings
neither prove nor insinuate that mesenchymal neoplasms are derived from a normal
mesenchymal progenitor that has sustained oncogenic mutations nor do they suggest that cells
with stem-like characteristics contribute to de novo tumor formation.
Some mesenchymal neoplasms are associated with specific translocations that result in the
expression of functional fusion proteins that often contribute to oncogenesis. For instance,
rearrangement of the EWSR1 gene on chromosome 22 to the ETS gene family member, FLI-1, on
chromosome 11 generates the EWS-FLI-1 fusion protein commonly found in Ewing’s sarcoma.
Also, chromosomal translocations resulting in the generation of the FUS-CHOP fusion protein
are observed in myxoid lipsarcomas. Interestingly, over-expression of both EWS-FLI-1 and
FUS-CHOP proteins in murine mesenchymal progenitor cells results in the induction of tumors
strongly resembling the neoplastic lesions from which they are affiliated[4, 5]. Importantly,
tumorigenesis is only driven by super physiologic expression of these fusion proteins in
mesenchymal progenitor cells, indicating cell type specificity in this malignant process. Data
also demonstrates that after long-term culture, both murine and human mesenchymal progenitor
cells can undergo spontaneous transformation producing tumors resembling fibrosarcomas[6, 7].
Also, conditional expression of a translocation specific to synovial sarcomas in muscle
progenitors, but not mature myoblasts has the capacity to induce the formation of synovial
sarcomas in mice[8], further supporting the notion that that mesenchymal precursors give rise to
mesenchymal neoplasms.
Aggressive fibromatosis, also know as desmoid tumor, are a locally invasive soft tissue tumors,
generally arising in connective tissues. While these lesions infiltrate into surrounding normal
tissues, they do not metastasize to distant sites. Histological and cytologic analysis of the tumors
reveal that they are composed of bipolar fibroblastic cells that expresses the intermediate
filament, vimentin; but lack expression of epithelial markers such as E-cadherin. The location,
100
cellular morphology and histological profile of these tumors suggest that they derive from
mesenchymal sources; however the cellular origins of aggressive fibromatosis have yet to be
elucidated[3, 9, 10].
The molecular etiology of aggressive fibromatosis is well known[11, 12]. These lesions can
occur as sporadic tumors or as part of pre-neoplastic conditions, such as familial adenomatous
polyposis (FAP) and familial infiltrative fibromatosis (FIF). Patients with familial adenomatous
polyposis develop gastrointestinal lesions that typically progress to cancer by the third decade of
life [13, 14]. In both familial adenomatous polyposis and familial infiltrative fibromatosis, the
lesions are associated with a mutation in the APC gene, while in sporadic aggressive
fibromatosis, most tumors contain mutations in CTNNB1, the gene that codes for β-catenin[15,
16]. In both cases, genetic aberrations ultimately result in the stabilization of β-catenin and the
up-regulation of β-catenin/TCF/LEF-1 dependant transcriptional activity[3, 17]. In mice,
mutations in the WNT/β-catenin pathway can also result in the formation of murine aggressive
fibromatosis. For example, the Apcwt/1638N mouse harbors a targeted mutation in the Apc gene
resulting in the expression of a truncated non-functional version of the APC protein consequently
leading to the up-regulation of β-catenin/TCF/LEF-1 dependant transcriptional activity. These
mice development a large numbers of aggressive fibromatosis tumors as well as gastrointestinal
lesions and as such, are a well-characterized animal model that closely approximates the human
disease [18, 19].
Here we use a mouse model of familial adenomatous polyposis (Apcwt/1638N) to show the
influence of normal mesenchymal progenitor cells on the formation of aggressive fibromatosis.
Importantly, we demonstrate that modification of the numbers of mesenchymal progenitor cells
directly impacts the number of tumors that were found in Apcwt/1638N mice. Furthermore, we
demonstrate that genetic the signature of aggressive fibromatosis side population cells is similar
to that of mesenchymal progenitor cells.
101
4.3 Results
4.3.1 Aggressive fibromatosis contain a subpopulation of cells with progenitor properties
Previous reports demonstrate side population cells isolated from various tissues, including
mesenchymal neoplasms, have stem like characteristics[20]. Within aggressive fibromatosis
tumors, SP cells compose 0.2-1.8% of the total population (Figure 4.1 A); suggesting the
presence of stem-like cells within this tumor type. To further verify this finding, we examined
tumors for the presence of cell surface makers known to help in the identification of
mesenchymal progenitor cells (Section 1.6.3). Cells from primary aggressive fibromatosis
tumors were stained for CD146 and Stro-1 and expression was analyzed using flow cytometry.
Staining of cells from three independent tumors revealed the presence of both markers on
aggressive fibromatosis cells (Figure 4.1B, C). Furthermore, when co-stained with Hoechst
33342 dye, we found an enrichment of CD 146 cells within the SP, suggesting that within the SP
fraction, there is an enrichment of progenitor like cells (Figure 4.1 D).
4.3.2 Positive correlation between numbers of aggressive fibromatosis and CFU-Fs in Apcwt/1638n mice
We next sought to determine if there was a relationship between the number of mesenchymal
progenitors and the development of AF tumors. To study the in vivo role of MPCs in tumor
development we utilized a well-established mouse model of AF. Apcwt/1638n mice carry a targeted
mutation at codon 1638 of the murine APC gene. Mice heterozygous for the Apc1638N mutation
(Apcwt/1638N) develop high numbers of aggressive fibromatosis with complete penetrance [18].
The inherent variability in the number of AF tumors that form in each individual Apcwt/1638n mice
allowed us to determine if a relationship existed between the numbers of tumors and the number
of MPCs. Colony forming units-fibroblastic (CFU-F), which have the potential to differentiate
down various mesenchymal lineages, is a well established measure of MPCs present in the bone
marrow, and as such, we used the number of CFU-Fs as a surrogate for MPCs. 12 Apcwt/1638n
male mice were sacrificed at 6 months of age and tumors were counted. Tumors from individual
mice ranged from 10-35. Femurs from each mouse were removed and bone marrow cells were
plated to assess CFU-F formation. We found that mice with low numbers of tumors had fewer
CFU-Fs when compared to with high numbers of tumors (Figure 4.2 A). We have previously
reported than we can modulate the number of AF tumors in APC 1638n mice by mating the mice
102
to either Timp Tg or Rhamm deficient mice[22, 23]. To confirm our findings that the CFU-F
alter with number of tumors, we measured the CFU-F numbers in these mice. Consistent with
our findings in APC 1638n mice we found is a positive correlation between numbers of CFU-Fs
and number of AF that develop (Figure 4.2 B,C) suggesting that there is a relationship between
mesenchymal progenitor cells, as measured by CFU-F, and tumor formation.
4.3.3 Mesenchymal progenitors are involved in the development of aggressive fibromatosis
Mice lacking the Sca-1 antigen develop fewer numbers of mesenchymal progenitor cells over
time, as marked by progressively fewer CFU-F[24]. Using a previously described breeding
strategy we mated Apcwt/1638n mice to Sca-1-/- mice and generated Apcwt/1638n mice that were either
wildtype, heterozygous, or completely lacking the Sca-1 gene[22]. At 6 months of age mice were
sacrificed and tumor development was assessed. We found that Apcwt/1638N/Sca-1+/+ mice had the
greatest numbers of tumors, averaging 24 tumors per male mice. In contrast, Apcwt/1638N/Sca-1-/-
littermate controls developed significantly fewer tumors averaging 12 tumors per male mouse
(Figure 4.3 A). Consistent with previously published reports male mice developed more tumors
than female mice (6,9,10). However, as with the males, Apcwt/1638n/Sca-1+/+ female mice formed
more tumors than Apcwt/1638n/Sca-1-/- littermate controls (Figure 4.3 A). Tumors from each
genotype were also stained with hematoxylin and eosin. We found that tumors derived from
Apcwt/1638N/Sca-1-/- had fewer cells compared to Apcwt/1638N/Sca-1+/+ mice (Figure 4.3 B,C)
suggesting that modulation of the number of mesenchymal progenitor cells impacts both the
number of aggressive fibromatosis that form as well as the cellularity of the tumors.
4.3.4 Mesenchymal but not epithelial derived tumors are impacted by alteration of MPCs
In addition to the formation of mesenchymal neoplasms, Apcwt/1638N mice also develop intestinal
polyps and skin cysts [19]. Interestingly, while the numbers of aggressive fibromatosis were
reduced in Apcwt/1638N/Sca-1-/- mice when compared to Apcwt/1638N/Sca-1+/+ littermate controls, no
differences were observed in the numbers of intestinal polyps or skin cysts (Figure 4.4) in either
male or female mice. Thus demonstrating that in Apcwt/1638N epithelial neoplasms were not
affected by the ablation of Sca-1.
103
4.3.5 Mesenchymal precursors from Apcwt/1638N have the capacity to initiate tumor formation
We next determined if MPCs with an oncogenic mutations within murine AFs had tumorigenic
potential. To address this question, we isolated bone marrow stromal cells from 8 week old
Apcwt/1638N and Apcwt mice. 3.0 X 106 cells from both genotypes were subcutaneously injected
into immuno-deficient NOD/SCID mice. After 12 weeks mice were sacrificed and tumor
formation was assessed. We detected the formation of aberrant cellular growth in all mice
injected with cells derived from Apcwt/1638N (Figure 4.5). In contrast, none of the mice injected
with cells derived from wild type mice generated tumors. Histological examination of the lesions
revealed the presence cells of mesenchymal origins, specifically, cells resembling those found in
the bone marrow. Interestingly, cells on the outer edge of the lesions (arrowhead) appear to be
spindle shaped, resembling fibroblasts found within aggressive fibromatosis. We observed that
mesenchymal progenitors with oncogenic mutations have the capacity to deregulate cellular
growth and the resulting tumors appear to contain cells with mesenchymal origins.
4.4 Discussion
The presence of stem-like cells in aggressive fibromatosis tumors is evidenced by the expression
of MPC surface markers and the ability of a subset of cells (SP) to exclude Hoechst dye. This
suggests that within pre-existing AF tumors, there is a population of cells with stem-like features.
However, this observation does not imply that tumors arise from normal stem cells and as such,
we sought to identify the cells responsible for oncogenesis prior to tumor formation[3, 4]. We
observed a positive correlation between MPCs numbers and AF tumor formation and we
demonstrated that the loss of MPCs caused a reduction in the numbers of AF tumors in mature
mice. We also showed that MPCs derived form mice predisposed to AF tumor formation have
the capacity to initiate aberrant cellular growth when subcutaneously injected into immuno-
compromised mice. Taken together, these findings suggest that development of aggressive
fibromatosis is influenced by a mesenchymal precursor. The observation that MPCs numbers did
not impact the formation of intestinal neoplasms, which derive from epithelial precursors,
strengthens the specificity of the association between aggressive fibromatosis and MPCs.
Not only did we demonstrate a reduction in the number of tumors from Apcwt/1638n /Sca-1-/-, but
we also showed that tumors derived from these mice contain fewer numbers of cells when
104
compared to Apcwt/1638n/Sca-1+/+ mice. Mice lacking Sca-1 develop age dependant osteoporosis
and this phenotype is caused not by the loss of differentiated osteoprogenitors, but rather by a
deficit in the numbers of MPCs. Interestingly, no differences in MPC frequency is detected in
young Sca-1-/- mice; however aged Sca-1-/- mice have much lower numbers of MPCs when
compared to their wildtype counterparts. This deficit and the subsequent development of age
dependant osteoporosis is attributed to a impairment in the capacity of MPCs cells to self-
renew[25]. It has been postulated that Sca-1 may play a role in balancing the signals between
differentiation and self-renewal in stem cells[26]. Given these observations, decreased tumor
cellularity in Apcwt/1638n /Sca-1-/- mice may be attributed to a diminished capacity of oncogenic
MPCs to self-renew.
Alternatively, aggressive fibromatosis, defined as a fibro-proliferative disorder, may be
attributed to a deregulation of MPC differentiation. MPCs are multi-potent with the capacity to
differentiate into muscle, fat, bone, cartilage, and fibroblastic cells. Isolation of single cell
derived clonal populations reveals that lineage commitment is not a random process, but rather,
organized into a cellular hierarchy, with the quinti-potent, self-renewing MPC at the apex and the
restricted fibroblast at the base[26]. Deregulation of these well-orchestrated events can alter the
lineage commitment of cells within this hierarchy[27-29]. For example, conditional deletion of
β-catenin in limb and head mesenchyme during early embryonic development results in an arrest
of osteoblastic differentiation as osteochondroprogentiors preferentially differentiate into
chondrocytes as opposed to osteoblasts[29]. Conversely, ectopic canonical Wnt signaling
enhances osteoblastic differentiation of these progenitor cells[27]. While this demonstrates that
deregulation of canonical Wnt signaling is implicated in the regulation of
osoteochondroprogenitors, the impact of this signaling pathway on cells higher up in the MPC
hierarchy has not been fully elucidated. As altered β-catenin signaling is a known molecular
determinant of AF tumor formation, it can be postulated that distortion of this pathway may tilt a
large proportion of mesenchymal precursors to forfeit towards the committed fibroblast, a
defining feature and cell population of AF tumors. The reduction in the number of MPCs in
Apcwt/1638n /Sca-1-/-, may therefore lower the numbers of cells mandated down an impaired
differentiation pathway.
Depletion of Sca-1+ cells in the mammary gland, results in the loss of regeneration potential in
mammary gland reconstitution experiments demonstrating the ability of Sca-1 to prospectively
105
identify a population of mammary epithelial cells. [30]. Interestingly, in mouse models of breast
cancer, over-expression of components of the Wnt signaling pathway results is an expansion of
Sca-1/keratin 6 progenitor cells. In contrast, in mouse models where tumorigenesis is driven by
the over-expression of H-ras, Neu, or middle T antigen, no expansion of Sca-1/keratin 6
progenitor cells is detected[31, 32]. The molecular etiology of AF is well known where the up-
regulation of β-catenin signaling is a hallmark of these neoplasms. Taken together, this suggests
that the β-catenin pathway may contribute, in part, to tumorigenesis by expanding the progenitor
cell compartment allowing for an increase in the number of cells that may be responsible for
tumor maintenance. As such, loss of Sca-1 results in fewer numbers of progenitors susceptible to
this aberrant expansion.
In summary, the loss of Sca-1 may diminish the number of aggressive fibromatosis tumors
through multiple mechanisms. Firstly, the reduction in the number of mesenchymal progenitors,
which act as a potential candidate cell of origin for these neoplasms, results in fewer cells with
the potential to form tumors. This may be particularly important in mesenchymal neoplasms
susceptible to β-catenin signaling such as AF for this loss is amplified in the tumorigenic process
by reducing the number of progenitor cells available for both cellular expansion and distorted
lineage commitment. Secondly, in Sca-1 null mice, the remaining MPCs have an impaired ability
to self-renew, resulting in the eventual loss of tumors by targeting those cells responsible for the
maintenance of the malignant tissue. This provides strong evidence that functional mesenchymal
progenitor cells are important in the development and progression of aggressive fibromatosis. .
This data strongly suggests the influence of progenitor cells in this neoplasm. The identification
of the primary cell of origin in aggressive fibromatosis is a key step towards an understanding of
the pathology of this disease. Here we provide evidence to support a model in which cells with
mesenchymal stem cell like characteristics play a role on both play a role in both the initiation
and maintenance of aggressive fibromatosis tumors. Raising the intriguing possibility that
protecting the stem cells in patients with FAP can prevent AF and understanding mesenchymal
stem cell biology can be employed to develop new treatments.
106
Figure 4.1 Human AF tumors contain progenitor cells
(A) Cells from primary human AF tumors were stained with Hoechst 33342 dye and were
analyzed by flow cytometry. The representative sample depicted shows the SP cells as outlined
and shown as a percentage of the total cell population. In the presence of verapamil, the
percentage of SP cells is greatly diminished. (B,C) Cells from representative aggressive
fibromatosis samples were stained for CD146 and Stro-1. Positively stained cells are outlined
and shown as a percentage of the total cell population. Isotype controls are shown to the left of
each experimental plot. (D) CD146 is enriched in the SP fraction. Cells from a representative
aggressive fibromatosis were stained co-stained with Hoechst 33342 dye and with an antibody
against CD146 antigen. SP cells are outlined and CD146 staining was examined in this fraction.
43% of SP cells express CD146.
107
0 1000 2000 3000 4000
FL 7 Lin: FL 7 Hst Red
0
1000
2000
3000
4000
FL 6 Lin: FL 6 Hst Blue
1.8
0 1000 2000 3000 4000
FL 7 Lin: FL 7 Hst Red
0
1000
2000
3000
4000
FL 6 Lin: FL 6 Hst Blue
0.45
Hoechst Red
Hoechst R
ed
Hoecsht Hoecsht+Verapamil
Stro-1Isotype
A
101
102
103
104
105
PE-A
0
1000
2000
3000
4000
SSC-H0
101
102
103
104
105
<PE-A>
0
1000
2000
3000
4000
SSC-H 12.2
101
102
103
104
105
<PE-A>
0
1000
2000
3000
4000
FSC-H 43.7
<DAPI-A>
CD146Isotype
0 1000 2000 3000 4000
<Hoechst Red-A>
0
1000
2000
3000
4000
<DAPI-A>
0.2
CD146Hoechst Red
Ho
ech
st
Re
d
B
C
D
108
Figure 4.2 Correlation of CFU-F with numbers of AF tumors.
A) Apcwt/1638n with fewer numbers of CFU-Fs develop fewer numbers of AF. Apcwt/1638N mice
were sacrificed at 6 months of age and tumors were counted. In tandem, colony forming units-
fibroblastic from each sacrificed mouse was assessed. Data shown represents a total of 12 mice
with calculated linear regression line and correlation coefficient displayed. B) CFU-F were
assessed from different genetically engineered strain of mice. Mice that form more tumors
(Apcwt/1638n/Timp-Tg) have higher numbers of CFU-F when compared to mice that develop
fewer numbers of AF (Apcwt/1638n/Rhamm-/-). Data represents the mean number of CFU-F per
mouse. 6 mice per given genotype were used. Error bars represent 95% confidence intervals. A
95% confidence interval that does not cross the mean of a comparison is a statistically significant
difference at p<0.05. C) Representative pictures demonstrating the differences in the number of
CFU-F between Apcwt/1638n/Rhamm-/-, Apcwt/1638n, Apcwt/1638n/Timp-Tg mice
109
Increasing Number of Aggressive Fibromatosis in Mice
CF
U-F
Apc/Apc 1638N
Rhamm-/-Apc/Apc 1638N Apc/Apc 1638N
Timp-1 Tg
CF
U-F
Apc
/Apc
163
8N
Rha
mm
+/+
Apc
/Apc
163
8N
Rha
mm
-/-
Apc
/Apc
163
8N
Ti
mp-
1 Tg
Apc
/Apc
163
8N
Ti
mp-
1 W
T
Number of AF tumors
0
50
100
150
200
0
50
100
150
200
A B
C
#C
FU
-F
APC 1638N
110
Figure 4.3 Modulation of MPCs impacts tumor development
A) Loss of Sca-1 reduces numbers of AF tumors in Apcwt/1638N mice. Apcwt/1638N/Sca-1+/+,
Apcwt/1638N/Sca-1+/-, and Apcwt/1638N/Sca-1-/- mice were sacrificed at 6 months of age and AF were
scored. In both males and females, Apcwt/1638N/Sca-1-/- mice formed fewer tumors than
Apcwt/1638N/Sca-1+/+ mice. Data represents mean number of tumors per mice (total of 15 mice per
given genotype). Error bars represent 95% confidence intervals. A 95% confidence interval that
does not cross the mean of a comparison is a statistically significant difference at p<0.05. B)
Tumors derived from Apcwt/1638N/Sca-1-/- mice have decreased cellularity when compared to
Apcwt/1638N/Sca-1+/+ littermate controls. Graphical data represents as the mean number of nuclei
counted in each field of vision at 400X magnification 3 mice for each given genotype and 4
fields for each section were counted. Error bars represent 95% confidence intervals. C)
Representative H & E slides of the tumors from each genotype show diminished numbers of
nuclei in tumors derived from Apcwt/1638N/Sca-1-/- when compared to Apcwt/1638N/Sca-1+/+ mice.
111
0
5
10
15
20
25
30
35
40
0
2
4
6
8
10
12
0
30
60
90
120
150
Num
ber o
f AF
Tum
ors
Num
ber o
f AF
Tum
ors
Num
ber o
f Cel
ls
WT +/- -/-
Sca-1 Expression
WT +/- -/-
Sca-1 Expression
WT +/- -/-
Sca-1 Expression
Male Mice Female Mice
Sca-1 WT Sca-1 +/- Sca-1 -/-
H&
E
A
B
C
112
Figure 4.4 Progenitor statuses in Apcwt/1638N/Sca-1-/- Mice
A) Apcwt/1638N/Sca-1-/- have fewer MPCs than Apcwt/1638N/Sca-1+/+ mice. There were fewer
numbers of CFU-F in Apcwt/1638N/Sca-1-/- mice when compared to Apcwt/1638N/Sca-1+/+ . Data
represents mean number of colony-forming units-fibroblastic isolated from three independent
mice from each genotype and error bars represent 95% confidence intervals. A 95% confidence
interval that does not cross the mean of a comparison is a statistically significant difference at
p<0.05. B) Side population analysis of whole bone marrow revealed a similar decrease in the
percentage of SP cells in Apcwt/1638N/Sca-1-/- compared to Apcwt/1638N/Sca-1-/- mice. Treatment of
cells with verapamil results in the inhibition of the SP population. Representative FACS plots
for each of the three genotypes are shown. C) Graphical data represents the mean number
percentage of SP cells from 3 independent mice (3 mice per given genotype). Error bars
represent 95% confidence intervals.
113
CF
U-F
Apc/Apc
1638N
Sca
-1+/+
Apc/Apc
1638N
Sca
-1-/-
0.0
0.5
1.0
1.5
2.0
WT +/- -/-
Sca-1 Expression
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
HO Blue-A
1.09
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
HO Blue-A
0.38
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
HO Blue-A
1.82
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
HO Blue-A
0.2
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
HO Blue-A
0.54
0 1000 2000 3000 4000
HO Red-A
0
1000
2000
3000
4000
HO Blue-A
0.033
0
20
40
60
80
100
120
% S
P C
ells in B
M
Hoechst R
ed
Hoechst Red
A
B
114
Figure 4.5 Loss of Sca-1 does not impact the formation of epithelial lesions
A) Apcwt/1638N/Sca-1+/+ and Apcwt/1638N/Sca-1-/- develop similar numbers of epithelial derived
lesions. Mice from each of the three genotypes were sacrificed at 6 months of age and the
number of intestinal polyps and skin cysts were counted. No significant differences were
observed in the numbers of intestinal polyps or skin cyst between Apcwt/1638N/Sca-1+/+,
Apcwt/1638N/Sca-1+/-, Apcwt/1638N/Sca-1-/- mice. Data represents the mean number of tumors per
mice (15 mice per given genotype) and error bars represent 95% confidence intervals. A 95%
confidence interval that does not cross the mean of a comparison is a statistically significant
difference at p<0.05.
115
0
1
2
3
4
5
6
7
8
0
1
2
3
4
5
6
7
8
WT +/- -/-
Sca-1 Expression
WT +/- -/-
Sca-1 Expression
Male Mice Female Mice
# o
f In
testinal P
oly
ps
# o
f In
testinal P
oly
ps
116
Figure 4.6 Stromal cells with oncogenic mutations have tumor initiating potential
A) Bone marrow stromal cells isolated from Apcwt/1638N and Apc wt/wt mice were cultured for 2
weeks, after which they were harvested and subcutaneously injected into NOD/SCID mice. After
3 months, mice were sacrificed and tumor formation assessed. Only stromal cells isolated from
Apcwt/1638N mice had the potential to induce aberrant cellular growth. B) Representative H&E
staining of lesions was performed for histological examination. Cells within the centre of the
lesions resemble bone marrow cells while cells lining the outer edge of the lesions appear to be
bipolar spindle shaped cells similar to fibroblasts (arrowheads).
117
APC 1638NWT
H&E
5X 10X
!
"
118
4.5 Materials and Methods
4.5.1 Primary tumors
Local ethical approval was obtained for all human tissue samples collected. Human and mouse
AF tumors were dissociated into single cells as previously reported [1]. Dissociated cells were
not cultured, but rather, used immediately afterwards for flow cytometry.
4.5.2 Flow cytometry
For side population analysis 1.0 x 106 cells/mL were treated either alone or with 2.5 µg/mL of
Hoechst 33342 dye (Sigma) for 90 min at 37oC, or in combination with 50µmol/L verapamil
(Sigma) as previously reported[1]. To detect for SP, cells were analyzed by using a dual
wavelength analysis (blue, 424-444nm; red, 675nM) after excitation with 350nm UV light
(MoFlow, Cytomation). For staining of mesenchymal progenitor markers, 1.0 X 106 cells
dissociated cells were re-suspended in 100µl of PBS supplemented with 2% fetal bovine serum
(Wisent) PE-conjugated CD146 (Becton Dickinson) and 0.1 µg/µl of Stro-1 (R&D Systems). For
visualization of Stro-1 staining, after incubation with primary antibody, cells were stained for 30
min at 4oC with FITC-conjugated IgM (Jackson labs). For co-staining of SP and CD146, cells
were initially incubated with Hoechst dye (see above) after which they were washed 2X with
PBS and then incubated with PE-conjugated CD146 (see above). After staining, cells were
washed 2X with PBS and then counterstained with 1µg/mL of propidium iodide (Molecular
Probes). PI positive (non-viable) cells were excluded from analysis. Cell staining was quantified
using LSRII flow cytometer (Becton Dikinson).
4.5.3 Generation of genetically engineered mice
The generation and phenotype of the APCWT/1638n and Sca-1-/- mice have been previously reported
([18], [24]. These mice were crossed to produce Apcwt/1638n/Sca-1+/+, Apcwt/1638n /Sca-1+/-, and
Apcwt/1638n /Sca-1-/- mice using a previously reported breeding strategy [22]. The generation and
phenotype of Apcwt/1638n/Rhamm-/-and Apcwt/1638n/Timp(Tg) mice has been previously reported
[22, 23].
119
4.5.4 Scoring of tumors
Generation of these crosses allowed for the comparison of littermate controls between the
various genotypes. 15 male and 15 female mice of each genotype were sacrificed at 6 months of
age and the number of AF tumors, intestinal polyps, and skin cysts were scored as previously
reported [22]. For immunohistochemistry, representative tumor samples were formalin fixed,
paraffin embedded, sectioned, and stained with hematoxalin and eosin. To determine the number
of cells present in the tumors, a blinded study was performed counting the number of nuclei
present in stained sections. Specifically, stained sections were viewed under a light microscope
(Leica) and 3 representative samples from each genotype were used to count the number of
nuclei within the field of view (400X magnification).
4.5.5 Cell Culture
Mesenchymal stromal cells were isolated as previously described[30]. Briefly, mice were
euthanized at 8 weeks of age, femurs and tibias were removed, bone marrow was aspirated and
cells were plated in MesenCult® MSC Basal Medium for Mouse Mesenchymal Stem Cells
supplemented with Mesenchymal Stem Cell Stimulatory Supplements (StemCell Technologies).
After 72 hours the medium was changed to remove non-adherent cells. For colony forming units
fibroblastic, stromal cells were cultured for 7 days after which they were stained with crystal
violet (Sigma) (.05% w/v in methanol) and colonies greater than 1mm were counted[31].
4.5.6 Xenograft models
For xenografted tumors stromal cells were grown for 14 days after which the cells were
trypsinized and harvested for injections. 3.0 X 106 cells were re-suspended in 50 µl of 1XPBS
supplemented with 2% FBS. This suspension was then mixed with 50µl of Matrigel (Becton
Dickinson) and subcutaneously injected into immuno-deficient NOD-SCID mice with a 25
gauge needle as previously described[1, 32]. Mice were observed for 12 weeks after which they
were euthanized and tumor formation was assessed. Tumors were removed and samples were
paraffin embedded, formalin fixed, sectioned, and stained for hemotoxylin and eosin. Staining
was visualized using a Leica light microscope.
120
4.5.7 Gene profiling
For human samples, after staining with Hoechst 33342 dye, cells were sorted in SP and non-SP
fractions. Sorted cells were rinsed one with 1X PBS and centrifuged at 1200rpm for 10 minutes.
RNA was extracted from cell pellets using RNeasy Mirco Kit (Qiagen) according to
manufacturer’s instructions. RNA quality was assessed with a Bioanalyzer (Agilent
Technologies) and cDNA was generated and hybridized onto Affymetrix Human Genome
U133.0 2.0 gene chips. We normalized the raw data using robust multi-array average (RMA)
algorithm (Irizarry et al. 2000). LPE (local-pooled-error test) (Jain et. al. 2003) was used to
identify differentially expressed genes between the side population and non-side population
fractions. Benjamini and Hochberg (BH) multiple testing procedure was used to evaluate false
discovery rate (FDR) (Benjamini and Hochberg, 1995). After which, analysis of gene expression
was performed using Parktec Genotyping Suite and Ingenuity Systems Software.
121
4.6 References
1. Wu, C., et al., Side population cells isolated from mesenchymal neoplasms have tumor
initiating potential. Cancer Res, 2007. 67(17): p. 8216-22.
2. Wu, C. and B.A. Alman, Side population cells in human cancers. Cancer Lett, 2008.
268(1): p. 1-9.
3. Riggi, N., et al., Development of Ewing's sarcoma from primary bone marrow-derived
mesenchymal progenitor cells. Cancer Res, 2005. 65(24): p. 11459-68.
4. Riggi, N., et al., Expression of the FUS-CHOP fusion protein in primary mesenchymal
progenitor cells gives rise to a model of myxoid liposarcoma. Cancer Res, 2006. 66(14):
p. 7016-23.
5. Rosland, G.V., et al., Long-term cultures of bone marrow-derived human mesenchymal
stem cells frequently undergo spontaneous malignant transformation. Cancer Res, 2009.
69(13): p. 5331-9.
6. Li, H., et al., Spontaneous expression of embryonic factors and p53 point mutations in
aged mesenchymal stem cells: a model of age-related tumorigenesis in mice. Cancer Res,
2007. 67(22): p. 10889-98.
7. Haldar, M., et al., A conditional mouse model of synovial sarcoma: insights into a
myogenic origin. Cancer Cell, 2007. 11(4): p. 375-88.
8. Shields, C.J., et al., Desmoid tumours. Eur J Surg Oncol, 2001. 27(8): p. 701-6.
9. Lewis, J.J., et al., The enigma of desmoid tumors. Ann Surg, 1999. 229(6): p. 866-72;
discussion 872-3.
10. Alman, B.A., et al., Increased beta-catenin protein and somatic APC mutations in
sporadic aggressive fibromatoses (desmoid tumors). Am J Pathol, 1997. 151(2): p. 329-
34.
122
11. Lips, D.J., et al., The role of APC and beta-catenin in the aetiology of aggressive
fibromatosis (desmoid tumors). Eur J Surg Oncol, 2009. 35(1): p. 3-10.
12. Kotiligam, D., et al., Desmoid tumor: a disease opportune for molecular insights. Histol
Histopathol, 2008. 23(1): p. 117-26.
13. Galiatsatos, P. and W.D. Foulkes, Familial adenomatous polyposis. Am J Gastroenterol,
2006. 101(2): p. 385-98.
14. Scott, R.J., et al., Familial infiltrative fibromatosis (desmoid tumours) (MIM135290)
caused by a recurrent 3' APC gene mutation. Hum Mol Genet, 1996. 5(12): p. 1921-4.
15. Hegde, M.R. and B.B. Roa, Detecting mutations in the APC gene in familial
adenomatous polyposis (FAP). Curr Protoc Hum Genet, 2006. Chapter 10: p. Unit 10 8.
16. Tejpar, S., et al., Predominance of beta-catenin mutations and beta-catenin dysregulation
in sporadic aggressive fibromatosis (desmoid tumor). Oncogene, 1999. 18(47): p. 6615-
20.
17. Tejpar, S., et al., Tcf-3 expression and beta-catenin mediated transcriptional activation in
aggressive fibromatosis (desmoid tumour). Br J Cancer, 2001. 85(1): p. 98-101.
18. Fodde, R., et al., A targeted chain-termination mutation in the mouse Apc gene results in
multiple intestinal tumors. Proc Natl Acad Sci U S A, 1994. 91(19): p. 8969-73.
19. Smits, R., et al., Apc1638N: a mouse model for familial adenomatous polyposis-
associated desmoid tumors and cutaneous cysts. Gastroenterology, 1998. 114(2): p. 275-
83.
20. Challen, G.A. and M.H. Little, A side order of stem cells: the SP phenotype. Stem Cells,
2006. 24(1): p. 3-12.
21. Tolg, C., et al., Genetic deletion of receptor for hyaluronan-mediated motility (Rhamm)
attenuates the formation of aggressive fibromatosis (desmoid tumor). Oncogene, 2003.
22(44): p. 6873-82.
123
22. Kong, Y., et al., Matrix metalloproteinase activity modulates tumor size, cell motility,
and cell invasiveness in murine aggressive fibromatosis. Cancer Res, 2004. 64(16): p.
5795-803.
23. Bonyadi, M., et al., Mesenchymal progenitor self-renewal deficiency leads to age-
dependent osteoporosis in Sca-1/Ly-6A null mice. Proc Natl Acad Sci U S A, 2003.
100(10): p. 5840-5.
24. Holmes, C., et al., Longitudinal analysis of mesenchymal progenitors and bone quality in
the stem cell antigen-1-null osteoporotic mouse. J Bone Miner Res, 2007. 22(9): p. 1373-
86.
25. Holmes, C. and W.L. Stanford, Concise review: stem cell antigen-1: expression, function,
and enigma. Stem Cells, 2007. 25(6): p. 1339-47.
26. Sarugaser, R., et al., Human mesenchymal stem cells self-renew and differentiate
according to a deterministic hierarchy. PLoS One, 2009. 4(8): p. e6498.
27. Day, T.F., et al., Wnt/beta-catenin signaling in mesenchymal progenitors controls
osteoblast and chondrocyte differentiation during vertebrate skeletogenesis. Dev Cell,
2005. 8(5): p. 739-50.
28. Glass, D.A., 2nd, et al., Canonical Wnt signaling in differentiated osteoblasts controls
osteoclast differentiation. Dev Cell, 2005. 8(5): p. 751-64.
29. Hill, T.P., et al., Canonical Wnt/beta-catenin signaling prevents osteoblasts from
differentiating into chondrocytes. Dev Cell, 2005. 8(5): p. 727-38.
30. Li, Y., et al., Evidence that transgenes encoding components of the Wnt signaling
pathway preferentially induce mammary cancers from progenitor cells. Proc Natl Acad
Sci U S A, 2003. 100(26): p. 15853-8.
31. Welm, B.E., et al., Sca-1(pos) cells in the mouse mammary gland represent an enriched
progenitor cell population. Dev Biol, 2002. 245(1): p. 42-56.
124
32. Shackleton, M., et al., Generation of a functional mammary gland from a single stem cell.
Nature, 2006. 439(7072): p. 84-8.
33. Peister, A., et al., Adult stem cells from bone marrow (MSCs) isolated from different
strains of inbred mice vary in surface epitopes, rates of proliferation, and differentiation
potential. Blood, 2004. 103(5): p. 1662-8.
34. Phinney, D.G., et al., Plastic adherent stromal cells from the bone marrow of commonly
used strains of inbred mice: variations in yield, growth, and differentiation. J Cell
Biochem, 1999. 72(4): p. 570-85.
35. Cheon, S.S., et al., beta-Catenin stabilization dysregulates mesenchymal cell
proliferation, motility, and invasiveness and causes aggressive fibromatosis and
hyperplastic cutaneous wounds. Proc Natl Acad Sci U S A, 2002. 99(10): p. 6973-8.
125
Chapter 5
5 Summary, conclusions, and future directions
126
5.1 Summary
The prospective isolation of a tumor initiating cells from mesenchymal neoplasms and the
characterization of these cells is described in Chapters 2 and 3. Specifically, we demonstrated
that side population cells are present in a broad range of mesenchymal neoplasms and we
observed a positive correlation between the percentage of SP cells and tumor grade. In vivo
xenograft assays demonstrated that these cells were enriched for tumor initiating potential when
compared to their non-SP counterparts. Furthermore, only SP cells had the capacity to initiate
tumor formation upon serial transplantation (Chapter 2). In osteosarcomas, a specific
mesenchymal neoplasm, only SP cells were sensitive to chemical blockade of the Hedgehog
pathway. Specifically, treatment of osteosarcoma cell lines with an inhibitor to this pathway
resulted in the diminished capacity for SP cells to induce tumors in immuno-deficient mice
(Chapter 3).
The identification of a candidate cell of origin for the mesenchymal neoplasm, aggressive
fibromatosis is described in Chapter 4. We established that aggressive fibromatosis contain SP
cells and these cells have a genetic signature similar to mesenchymal progenitor cells. Genetic
alteration of the number of mesenchymal progenitor cells in mouse models of the tumor directly
influenced the numbers of mesenchymal tumors found in mature mice, but not in epithelial
derived lesions. In addition, mesenchymal progenitor cells with oncogenic mutations had the
capacity to induce aberrant cellular growth with a phenotype resembling cells of the
mesenchymal lineage (Chapter 4).
5.2 Conclusions
The existence of over 100 different subclasses of sarcomas highlights the large degree of
diversity within this tumor type and contributes to its poorly understood disease mechanism. The
inherent heterogeneous nature of these tumors coupled with the prospective identification of SP
cells with enhanced tumor initiating potential provides strong support that cells within these
tumors are organized in a cellular hierarchy with the mesenchymal tumor stem cell at the apex
and its differentiated progeny composing the heterogeneous tissue of the tumor body. The
biologically distinct properties of these cells, which include a stem-like phenotype, further
supports this hypothesis. Identification of the signaling pathways in the SP population that can be
127
manipulated to inhibit tumor growth demonstrates one biologically discreet feature of these cells
with important clinical implications for tumor treatment.
While the prospective identification of stem like cells from mesenchymal neoplasms
provides evidence for the presence of progenitor cells in pre-existing tumors; this observation
does not identify the cell of origin for these neoplasms. Mouse models used in our studies
demonstrate the influence of mesenchymal progenitor cells in etiology of aggressive
fibromatosis. No impact was detected in epithelial derived tumors, showing cell type specificity
for this class of tumors. Furthermore, the oncogenic capacity of MPCs to induce aberrant
growths containing cells with histological features of mesenchymal cell types strengthens this
cellular association and helps to shed insight into the fundamental differences between
mesenchymal neoplasms and their epithelial counterparts.
5.3 Future Directions
5.3.1 The side population assay: considerations for its use in the isolation of cancer stem cells/tumor initiating cells
5.3.1.1 Criticisms and limitations of the side population assay
Several criticisms have been raised concerning the use of Hoechst dye as a means of isolating
stem like cells. Firstly, as Hoechst dye binds to DNA, staining renders the assay toxic to live
cells. Therefore, it can be argued that SP cells are not stem-like cells, but rather, only a
population of cells that are able to escape the lethal effects of Hoechst. Also, as dye efflux is a
dynamic process, minute variables in staining times, dye concentration, and cellular
concentrations can vastly affect the SP phenotype. Furthermore, cytometry gating strategies used
to isolate SP cells lack the consistency of gating strategies used when staining with markers[1,
2]. Taken together, these problems can lead to cross contamination of the SP and the non-SP
fractions ultimately resulting in the production of confounding data.
Studies using Hoechst dye to isolate tumor initiating cells have attempted to address such
concerns. For example, in primary mesenchymal tumors non-SP cells have the ability to form
tumors in the primary round of injections, however, these tumors fail to engraft after secondary
transplantation; presumably due to the inability of the non-SP cells to self-renew. However, the
initial capacity of tumor formation indicates that non-SP cells are indeed viable even after
128
retaining Hoechst dye[3]. In further support of this finding, tumor initiating potential of
unstained but sorted MCF-7 breast carcinoma cells were compared to positively stained/non-SP
cells. Both populations of cells had similar in vivo tumorigenic potential and in vitro colony
forming potential. Taken together, this suggests dye toxicity does not account for the increased
“stemness” of SP cells[4].
There has been conflicting data regarding the capacity of SP cells to enrich for TICs. This may
be due, in part, to variation in gating strategies used by individual labs. While inhibitors against
ABC transporters, such as verapamil, are used as a control and ensure the capture of true SP
cells, this methodology does not always yield ideal results as positively stained cells can remain
within the negative or SP gate. Interestingly, gating within different regions of the SP fraction
has been shown to give rise to cells with differing “stemness” potential. That is cells in the
lowest quadrant have increased “stemness” potential when compared to cells in the upper
quadrant[1]. Taken together, this suggests that more stringent gating strategies are necessary for
elucidating the true nature of SP cells. However, it should be noted that all isolation strategies
have their shortfalls and perhaps the combination of different isolation methods are required to
enhance the purity of cancer stem cells.
SP cells isolated from tumors are enriched for cells with tumor initiating potential, however, the
exact nature of these cells have yet to be elucidated and it is unlikely that this population is
exclusively composed of tumor initiating cells. More likely, SP cells either represent a small
population of tumor initiating cells or conversely, tumor initiating cells represent a small fraction
of SP cells. While as few as 100 SP cell from primary mesenchymal neoplasms can initiate
tumor formation, ideally, a single SP cell would have the potential to form a tumor in order to
definitively determine the relationship between the two populations. However, this does not
mean that SP cells only represent a fraction of the total tumor initiating population, as mouse
xenograft assays may not capture the tumorigenic potential of all human SP cells.
Regardless of the unknown mechanisms by which SP cells initiate tumor formation, they
represent a population of clinically relevant cells as not only are they enriched for cells with
tumor initiating potential, but also, they are resistant to chemotherapeutic drugs. As such, this
population likely plays a critical role in tumor maintenance and reoccurrence. Unlike cell surface
markers used to identify cancer stem cells, SP cells are present in tumors from different cellular
129
origins. Further knowledge of this population will therefore be important for the advancement
effective treatment modalities for wide variety of neoplasms.
5.3.1.2 An alternative hypothesis for defining side population cells
Hoechst 33342 staining intensity is not solely based on active dye removal, but rather on a broad
range of factors. For instance, before efflux begins, Hoechst 33342 binds to AT rich regions
within the minor groove of DNA; therefore, chromatin structure, conformation, and DNA
content all contribute to a given SP profile[5]. The dynamic nature of DNA structure, controlled
by both cell cycle and epigenetic events, raises the possibility that dye efflux may be a minor
contributor to the SP profile with the kinetic properties of dye to DNA binding playing a larger
role than currently credited for this phenotype[6].
Our studies focused on characterization of SP cells in primary mesenchymal neoplasms.
However, in some samples, we observed differences in the non-SP staining profile, specifically
the presence of two distinct populations, non-SPlow and non-SPhigh was detected. This phenotype
was noted in a few primary osteosarcoma samples and in the transformed osteosarcoma cell
lines, KHOS and HOS-MN. Interestingly, although the percentage of SP cells in these samples
was relatively low (below 1%), the primary tumors that exhibited this non-SP profile had more
aggressive tumors as measured by poor patient outcome (unpublished data). Given this
observation, it can be hypothesized that tumors containing dual non-SP high fractions may
represent aggressive neoplasms with characteristics similar to the transformed osteosarcoma cell
lines. It is interesting to speculate how the non-SP may be involved in this phenotype.
Presumably, the non-SP high cells have increased DNA content, and represent highly
proliferating cells that contribute acutely to the aggressiveness of a particular sample. However,
based on the cancer stem cell hypothesis the non-SP high cells should represent transiently
amplifying cells with no capacity to self-renew, and should therefore, not contribute to chronic
maintenance of the disease.
An alternative hypothesis could be formulated by examining DNA content in relation to genomic
stability rather than cell cycle. Genomic stability is a defining factor in the development and
progression of cancer[7, 8]. Non-SP cells may have increased genomic instability resulting in
chromosomal aberrations that may manifest as two distinct non-SP populations that represent
varying karyotypes with the tumor. Samples that exhibited this non-SP phenotype were derived
130
from either primary osteosarcomas or osteosarcoma cell lines. Importantly, this mesenchymal
neoplasm displays a high degree of chromosomal abnormalities with ploidy number ranging
from haploidy to near hexaploidy[9, 10]. As such the non-SPhigh fraction may represent a
population with a high ploidy number relative to the non-SPlow fraction.
Based on this hypothesis, the non SP fraction high cells may be less genetically stable and
therefore lack the cellular longevity of their SP counterparts, however, they given the poor
prognosis of the patients, points to another theory. It is also possible that within the non-SP high
cells, a small fraction of cells may acquire mutations giving them a stem cell like phenotype
thereby helping to replenish the stem cell compartment and thus contribute to poor disease
outcome[11]. Serial transplantation of the non-SPhigh population into immuno-compromised mice
and examination of the capacity of non-SPhigh cells to generate SP cells could easily address these
questions. Regardless of the mechanism, this suggests future experiments examining the non-SP
profile may reveal important the molecular mechanisms that confer the biological properties of
these cells.
5.3.2 Characterization of side population cells in mesenchymal neoplasms
5.3.2.1 Identifying molecular determinants of sarcoma side population cells
To date the molecular mechanisms that confer stem like characteristics to SP cells is not well
defined. Many experimental approaches could be used to address this question, initially to
further characterize these cells array analysis may allow for the identification of novel proteins
that are involved in conferring the SP phenotype.
miRNAs are involved in maintaining both embryonic and tissue stem cells with changes in
specific miRNAs associated with ES cell self-renewal and differentiation[12-14]. Furthermore,
not only are miRNAs involved in normal developmental processes, but they can also function as
either tumor suppressors or oncogenes, thereby regulating the tumorigenic processes[15]. Given
the broad range of their functions, a comparison of miRNA expression between SP and non-SP
cells may provide a global indication of how SP cells function and also identify novel
mechanisms by which this phenomenon occurs.
131
5.3.2.2 Identifying novel sarcoma stem cell markers
Theoretically, the ability to manipulate the intrinsic characteristics that differentiate stem-like
cells from non stem cells as a means of isolation appears to be tantalizingly simple. In the case of
cancer stem cells where markers vary not only between different tumor types, but also within
different subclasses of tumors, the universality of these assays makes them highly appealing,
effectively alleviating the dependency on identifying markers. Unfortunately, to date, data using
these strategies as a means of isolating stem cells remains relatively sparse and the majority of
the CSC literature is based on the selective expression of cell surface markers[16], however,
these two strategies need not be mutually exclusive. In the absence of known CSC markers, the
universality of the SP phenotype in tumor tissues can be used as a means for identifying such
markers. For example, as the SP enriches for cells with tumorigenic potential, these same cells
may also express novel TIC antigens. As such, SP cells can be used to generate novel antibodies
against potential CSC markers. Alternatively, combining markers that weakly enrich for TICs
with SP cells may further purify the fraction beyond that of each individual assay. Finally, unlike
sphere forming assays, which may only be a measure of anchorage independence, a hallmark of
cancer cells, the SP assay confers more specificity. Importantly, the identification of a novel
marker may not only help in the prospective isolation of a single mesenchymal cancer stem cell,
but it may also identify a marker that can be used in the isolation of normal mesenchymal
progenitor cell for which there is currently no marker or pattern of markers known to
successfully prospectively identify and isolate which has long hindered the field.
5.3.2.3 Metastatic potential of sarcoma side population cells
Epithelial to mesenchymal transformation (EMT) occurs when epithelial cells acquire a cellular
phenotype usually associated with mesenchymal cells. This process occurs during normal
embryonic development and in normal physiological processes such as wound healing[17]. EMT
entails a wide range of cellular changes, notably; epithelial cells undergo gross changes in their
cytoskeleton including the loss of adherens junctions, the loss of cortical actin, and the induction
of stress fibers[18]. The ability of malignant cells to disseminate from the primary tumor is one
hallmark of tumor cells and the subversion of the transcriptional regulation of EMT by neoplastic
cells allows for the acquisition of the migratory phenotype required for metastasis[19]. While the
132
role of EMT in metastasis is well documented, its involvement in cancer stem cells is less well
defined.
It can be postulated that in order metastatic entities to occur, they must be initially seeded by a
cancer stem cell with the potential to self-renew and differentiate thereby generating a new
colony at a distant anatomical site[20]. Herman et al also demonstrated that within a population
of human pancreatic cancer stem cells, there exists a subpopulation of cells with increased
potential to metastasis[21]. Also, in an immortalized breast cancer cell line, cells with
mesenchymal phenotypes had more stem like properties defined by an increased expression of
stem cell markers associated with breast cancer stem cells an increased capacity to form
mammary spheres, an increased ability to grow in soft agar colonies and importantly, an
increased efficiency to form tumors in immuno-compromised mice. Furthermore, forced
expression of the transcription factor snail, the master regulator of EMT results in similar
observations indicating the importance of EMT demonstrating a link between cancer stem cells
and metastatic potential[22].
In light of these findings, it would be interesting to determine if metastatic tumors had a higher
percentage of SP cells or conversely, if SP cells have an increased potential to metastasize.
Microarray analysis to screen for markers associated with a mobile or invasive phenotype could
be undertaken. In addition in vitro experiments, to measure both mobility and cellular phenotype
could be performed. However, given the loss of the SP cells during prolonged culture periods,
presumably through the induction of cellular differentiation, these experiments would be limited
to assays with short incubation times. Although, it is important to note that sarcomas are derived
from mesenchymal cells and these studies would therefore, differ from their epithelia
counterparts. As such in vitro experiments may not yield biologically significant results in this
context.
In vivo experiments to determine if SP cells have and increased capacity to metastasize would
also address this question. This would be particularity well modeled in the osteosarcomas in
which metastasize to the lungs is common and contributes to poor patient prognosis. As non-SP
cells have the potential to form tumors upon primary injections, albeit at a lower frequency than
SP cells, tail vein injections could be performed and lung colonies could be measured to
determine the metastatic potential of SP cells compared to non-SP cells. However, to definitively
133
address this question, ideal, if we could mark a single SP cells to determine the metastatic
potential.
5.3.3 Clinical significance of sarcoma side population cell
5.3.3.1 Prognostic value of sarcoma side population cells
The focus of cancer stem cell research relies on the biology of tumor initiating cells; however,
whether this biological phenomena will be clinically significant has yet to be definitively
elucidated. Studies examining the brain tumor cancer stem cell marker CD133 show promise as
expression of the antigen correlates with patient survival in gliomas [23, 24] To date, there is no
know prognostic marker to predict clinical outcome in sarcomas. In osteosarcoma, size of the
tumor upon discovery is still the best know predictive indicator[25]. This highlights the
importance of the findings that there was a positive correlation between percentage of SP cells
and tumor grade for SP cells may have the potential to be used as a prognostic determinant in
sarcoma patients. In this particular tumor type screening for the presence of these cells prior to
patient treatment coupled with long term follow up of these patients would yield invaluable data
regarding the importance of SP as a prognostic tool. For example, long term prospective studies
on the SP will determine if those tumors with a high percentage of these cells are more resistant
to conventional chemotherapy as these cells are not only enriched for cells with tumor initiating
potential but they also have the capacity to efflux such drugs. Hence, altering treatment strategies
based on the presence or absence of SP cells may ultimately result in more efficient treatment
and improved patient outcomes. Alternatively, diminishing the population of SP cells within a
given tumor may also be a potential treatment strategy. In the mesenchymal neoplasm,
aggressive fibromatosis, the presence of interferon results in an increase in the number of SP
cells demonstrating chemical agents are capable of shifting SP numbers within a given
tumor[26]. As such, agents capable of this may be used in conjunction with conventional
chemotherapy thereby decreasing or eliminating tumorigenic cells and preventing tumor
reoccurrence.
Given the broad range of mesenchymal neoplasms examined for our side population studies,
examination of individual subclasses in isolation may yield allow for the more subtle
information.
134
5.3.4 Tumor microenvironment
Tumors reside in a specialized microenvironment composed of normal cells and extracellular
matrix, both of which provide regulatory input governing the development and progression of
neoplastic lesions. While the role of the tissue microenvironment on malignant cell proliferation
and metastasis is well established, less is known about its impact on neoplastic cells organized in
a cellular hierarchy[27]. Certainly, in normal tissue, the surrounding ecosystem deeply alters the
fate of stem cells playing a critical role in orchestrating the signals required for cell fate
determination and self-renewal[28, 29]. As the normal stem niche directly impacts its cellular
residents, it can be reasoned that the tumor niche also influences the behavior of cancer stem
cells[30, 31].The observation that SP cells are enriched for tumor initiating potential in
mesenchymal neoplasms provides a useful means to examine the role of the niche on the
behavior of malignant stem-like cells
5.3.4.1 The influence of the sarcoma niche on its side population cell residents
Work in the hematopoietic field suggests a possible role for the niche in regulating CSC
maintenance, for example, specialized microenvironments of bone marrow endothelial cells
appear to be required for homing and engraftment of leukemic stem cells[32]. Furthermore, both
extracellular matrix components and signaling molecules in the HSC microenvironment can
promote cell survival in AML, providing resistance to chemotherapeutic agents[33]. In addition,
brain cancer stem cells have been shown to reside in a vascular niche that secretes factors that
promote their long-term survival; furthermore, increasing the number of endothelial cells in brain
tumor xenografts expands the proportion of self-renewing cells in the tumor[34].
We demonstrated that chemical blockade of Hedgehog signaling inhibits SP cell tumor formation
in osteosarcomas (Chapter 3). However, whether this phenomenon is due to intrinsic or extrinsic
influences of the signaling pathway cannot be elucidated from our studies. Work from other labs
examined a broad range of epithelial carcinomas and demonstrated hedgehog blockade resulted
in the inhibition of tumor size due to an inhibition of signal responsiveness in the surrounding
stromal cells rather than of the epithelial derived cancers[35, 36]. This raises the possibility that
alteration of the niche can impact the behavior of sarcoma side population cells.
135
However, it should be noted that transplantation of human cells into mouse recipients has
limitations as there is increasing speculation regarding the efficiency of the murine stroma to
support the growth of human cells. For example studies demonstrate that alteration of variables
such as the use of matrigel, strain of NOD/SCID mice, and length of incubation time can
dramatically increase the frequency of tumor initiating cells in human melanomas[37]. This
observation raises speculation regarding the efficiency of the murine stroma to support the
growth of human cells and brings into question in the use of xenograted models to study CSCs.
As such, mouse models may be a more efficient means to examine the role of the niche on the
behavior of tumor initiating cells. Furthermore, this strategy allows for the use of a wide range of
pre-existing genetically modified mice. For example, in breast cancer, sustained expression in
vivo in the mammary gland of mice of matrix metalloproteinase-3, which destroys the basement
membrane can lead to epithelial tumorigenesis. While the isolation of a murine sarcoma stem
cell has yet to be identified, Sca-1 prospectively identifies a population of cells from murine
osteosarcomas that are enriched for tumor initiating potential[38]. Injection of these cells into
genetically engineered mice lacking hedgehog expression in the stroma may begin to address the
impact of the tumor environment of sarcoma stem cells.
Importantly, the sarcoma niche has yet to be identified. As mentioned, there are 100 different
sarcoma subtypes, which in turn, are localized to various locations within the body. For example,
soft tissue sarcomas are found in surrounding connective tissue while sarcomas of the bone are
generally found in bone/ cartilage. As such, while subcutaneous injection of SP cells does allow
for tumor engraftment, this site of injection might not accurately recapitulate the true sarcoma
niche and examination within the context of the different subclasses of tumors may yield more
valuable information. Manipulation of the tumor niche may inhibit the growth of neoplastic
lesions by altering the phenotype of cancer stem cell, ultimately inducing reversion of
tumorigenesis [39].
5.3.4.2 Influence of hypoxia on side population cells
Just as the surrounding normal tissue and extracellular milieu establishes a niche with the
capacity to alter neoplastic progression, tumors themselves generate a unique microenvironment
that influences its cellular constituents. Specifically, as malignant tissue expands, the distance
between the center of a tumor mass and the surrounding normal stroma containing capillary
136
vessels increases, thereby generating an oxygen gradient with cells furthest from blood vessels
exposed to the lowest levels of oxygen. Tumors exposed to hypoxic environments, or low O2,
created from this physical distance, are more resistant to radiation, highlighting the importance
of this phenomenon in tumor biology.
Recent studies have begun to address the influence of a hypoxic environment on biologically
distinct populations of cells that constitute a heterogeneous neoplastic lesion. Given the
observation that low oxygen levels maintain the pluripotency of embryonic stem cell by blocking
differentiation, it can be reasoned that cancer stem cells would also be impacted from similar
conditions. Not surprisingly, an in vitro hypoxic environment was found to increase the number
of cells expressing cancer stem cell markers in brain tumors. The mechanisms that confer these
properties have yet to be fully elucidated, however, current studies reveal that the hypoxia
inducible family (HIF) of proteins are key regulators in this process. For example, hypoxic
inducible factors were found to regulate the tumorigenic capacity of glioma stem cells[40]. Less
is known about the impact of hypoxia in mesenchymal neoplasms, however in normal
mesenchymal stem cells, a hypoxic environment improves the success isolation of these cells,
suggesting that it plays a role in stem cell maitenance. This may be due to the hypoxic
environment created in the bone marrow where mesenchymal stem cells reside. In light of these
observations, it would be interesting to determine if a hypoxic environment would alter the
cellular hierarchy in mesenchymal neoplasms by impacting the behavior of SP cells.
The hypoxic environment may promote self-renewal and block differentiation of mesenchymal
tumor initiating cells. Alternatively, the intrinsic properties of these cells may allow for survival
of external insults generated by a hypoxic niche. SP cells by definition have the potential to
efflux hydrophilic substances, which may include the acidic substances created in the tumor
microenvironment. Not surprisingly, a hypoxic environment leads to the increased expression of
ABCG2 transporters; therefore, as tumor tissues grow beyond their blood supply, SP cells are
better equipped to sustain themselves within the toxic surroundings. This protective
characteristic likely contributes to the stem cell like phenotype of not only sarcoma SP cells, but
to all SP cells.
Initial experiments to test this hypothesis could be performed by culturing SP cells under
hypoxic conditions to determine if this alters the percentage of SP cells. Importantly, if hypoxic
137
conditions do indeed increase, or maintain SP cells this would provide a useful tool to
maintaining cells in culture as one current drawback to primary cultures is the loss of stem cells
after prolonged exposure to culture conditions. This would therefore, not only provide useful
information regarding the biology of sarcomas, would help to overcome technical difficulties
within this field.
5.3.5 Mesenchymal progenitor cells and their involvement in the development of aggressive fibromatosis
5.3.5.1 Cell of origin?
At six months of age, a decrease in the numbers of mesenchymal progenitor cells in
Apcwt/1638N/Sca-1-/- mice correlated to diminished numbers of aggressive fibromatosis tumors
(Chapter 4); however, it is unclear if this result was due to impaired de novo tumorigenesis or
altered fate determination of pre-existing tumor cells. As Sca-1 null mice have normal numbers
of mesenchymal progenitor cells at two months of age, when aggressive fibromatosis begin to
develop[41], this suggests that de novo tumor formation should not be impacted. It has been
postulated that Sca-1 may play a role in balancing the signals between differentiation and self-
renewal in stem cells[42]. Work with primary myoblasts isolated from Sca-1 null mice
demonstrates that these mice exhibit increased proliferation and reduced numbers of
undifferentiated cells[43]. Taken together, it can be hypothesized that mice lacking Sca-1 have a
signaling defect that results in increased differentiation in conjunction with an inhibition of self-
renewal of stem cells ultimately leading to the depletion in the number of stem-like cells as mice
age. Given the role Sca-1 plays in the self –renewal of MPCs, decreased tumor cellularity may be
attributed to a diminished capacity of onogenic MPCs to self renew. As such, it would be
interesting to observe the Apcwt/1638N/Sca-1-/- mice over a longer period of time to determine if
tumor numbers would be further diminished when compared to younger mice. In addition,
examination the differentiation potential of progenitors within the aggressive fibromatosis in
Apcwt/1638N/Sca-1-/-would reveal any differences in linage commitment as increased
differentiation may leave diminish the number of progenitor cells remaining to maintain the
neoplastic lesion
We demonstrate that mesenchymal progenitor cells isolated from Apcwt/1638N have the potential to
induce aberrant cellular growth when transplanted into immuno-deficient mice (Chapter 4).
138
While this suggests that MPCs may be the cellular origin of aggressive fibromatosis; these
transplantation experiments may not represent the tumorigenic process that occurs under normal
physiologic conditions. To determine if aggressive fibromatosis tumors arise from MPCs, in vivo
linage tracking of MPCs from in Apcwt/1638N mice would be required. This has been effectively
demonstrated in colon cancer. However, one again, the limitation to this experiment is the lack of
markers known to identify a mesenchymal stem cell, unlike intestinal stem cells in which BMI1,
and Lgr5 mark intestinal stem cells[44, 45]. However, conditional expression of a translocation
specific to synovial sarcomas in muscle progenitors, but not mature myoblasts has the capacity to
induce the formation of synovial sarcomas in mice[46], aids in supporting the notion that that
mesenchymal precursors give rise to mesenchymal neoplasms.
5.3.5.2 Involvement of early progenitors or differentiated progeny?
Currently, the commonly accepted tests for the identification of MSCs include the capacity to
form colony forming unit fibroblastic (CFU-F) in culture, analysis of surface marker profiles,
and multi-lineage potential, particularly osteogenesis, chondrogenesis and adipogeneiss[47]. In
the hematopoietic system the identification of cell surface markers on both stem and committed
progeny has let to the stratification of cells into well-established hierarchy. Recent in vitro data
suggests that clonogenic MSC are organized into a cellular hierarchy in which differentiation
down multiple lineages occurs in a regulated fashion [48]. In both the hematopoietic and
neuronal systems, it has been demonstrated that both stem cells and committed progenitors have
the capacity to induce tumorigenesis (See Section 1.5). To address this question in our model
system, forced differentiation of Apc1638n/wt stromal cells down various mesenchymal lineages
pathways would produce committed progenitors. Injection of these cells into immuno-
compromised mice to assess the tumor initiating abilities of these cells relative to earlier
progenitor would reveal valuable information regarding the which cells are capable of
oncogeniesis. Importantly, this technique would circumvent the need for identifying cell marker
on committed progenitors. The ambiguity regarding the existence of the MSC stems from the
lack of identifying markers, various locations of isolation, and disparity in methods used to
culture and expand cells. This has resulted in the isolation of a heterogeneous population of
“MSC” that exhibit variable phenotypes, and as such, hampers the precision of defining such a
population of cells. However, regardless of the state of the progenitor cells, we have identified a
139
mesenchymal precursor that has the potential to induce aberrant cellular growth, demonstrating
the influence of these cells in the oncogenic process.
5.4 Concluding Remarks
We identified a population of cells within mesenchymal neoplasms that are enriched for tumor
initiating potential. This observation begins to unravel the mechanisms by which mesenchymal
tumorigenesis occurs. The studies demonstrating that osteosarcoma side population cells
differentially expresses components of the hedgehog signaling prove that these cells are indeed
biological distinct from non-side population cells and provides a useful foundation from which
further investigation on the characterization of side population cells can begin. This work
provides valuable information on the pathogenesis of mesenchymal neoplasm and validates
future exploration in studying the role of mesenchymal progenitor cells in this tumor type.
140
5.5 References
1. Montanaro,F.,etal.,DemystifyingSPcellpurification:viability,yield,andphenotype
aredefinedbyisolationparameters.ExpCellRes,2004.298(1):p.144‐54.
2. Challen,G.A.andM.H.Little,Asideorderofstemcells:theSPphenotype.StemCells,
2006.24(1):p.3‐12.
3. Wu,C.,etal.,Sidepopulationcellsisolatedfrommesenchymalneoplasmshavetumor
initiatingpotential.CancerRes,2007.67(17):p.8216‐22.
4. Zhou,J.,etal.,ActivationofthePTEN/mTOR/STAT3pathwayinbreastcancerstem
likecellsisrequiredforviabilityandmaintenance.ProcNatlAcadSciUSA,2007.
104(41):p.16158‐63.
5. Lalande,M.E.andR.G.Miller,Fluorescenceflowanalysisoflymphocyteactivation
usingHoechst33342dye.JHistochemCytochem,1979.27(1):p.394‐7.
6. Ibrahim,S.F.,etal.,Kineticanalysesasacriticalparameterindefiningtheside
population(SP)phenotype.ExpCellRes,2007.313(9):p.1921‐6.
7. Grady,W.M.andS.Markowitz,Genomicinstabilityandcolorectalcancer.CurrOpin
Gastroenterol,2000.16(1):p.62‐7.
8. Heng,H.H.,etal.,Stochasticcancerprogressiondrivenbynonclonalchromosome
aberrations.JCellPhysiol,2006.208(2):p.461‐72.
9. Sandberg,A.A.andJ.A.Bridge,Updatesonthecytogeneticsandmoleculargeneticsof
boneandsofttissuetumors:osteosarcomaandrelatedtumors.CancerGenet
Cytogenet,2003.145(1):p.1‐30.
10. Marina,N.,etal.,Biologyandtherapeuticadvancesforpediatricosteosarcoma.
Oncologist,2004.9(4):p.422‐41.
141
11. Shackleton,M.,etal.,Heterogeneityincancer:cancerstemcellsversusclonal
evolution.Cell,2009.138(5):p.822‐9.
12. Hatfield,S.D.,etal.,StemcelldivisionisregulatedbythemicroRNApathway.Nature,
2005.435(7044):p.974‐8.
13. Shcherbata,H.R.,etal.,TheMicroRNApathwayplaysaregulatoryroleinstemcell
division.CellCycle,2006.5(2):p.172‐5.
14. Croce,C.M.andG.A.Calin,miRNAs,cancer,andstemcelldivision.Cell,2005.122(1):
p.6‐7.
15. Lu,J.,etal.,MicroRNAexpressionprofilesclassifyhumancancers.Nature,2005.
435(7043):p.834‐8.
16. Lobo,N.A.,etal.,Thebiologyofcancerstemcells.AnnuRevCellDevBiol,2007.23:p.
675‐99.
17. Chaffer,C.L.,E.W.Thompson,andE.D.Williams,Mesenchymaltoepithelialtransition
indevelopmentanddisease.CellsTissuesOrgans,2007.185(1‐3):p.7‐19.
18. Hay,E.D.,Anoverviewofepitheliomesenchymaltransformation.ActaAnat(Basel),
1995.154(1):p.8‐20.
19. Thiery,J.P.,Epithelialmesenchymaltransitionsintumourprogression.NatRev
Cancer,2002.2(6):p.442‐54.
20. Dalerba,P.,R.W.Cho,andM.F.Clarke,Cancerstemcells:modelsandconcepts.Annu
RevMed,2007.58:p.267‐84.
21. Hermann,P.C.,etal.,Distinctpopulationsofcancerstemcellsdeterminetumor
growthandmetastaticactivityinhumanpancreaticcancer.CellStemCell,2007.
1(3):p.313‐23.
22. Mani,S.A.,etal.,Theepithelialmesenchymaltransitiongeneratescellswithproperties
ofstemcells.Cell,2008.133(4):p.704‐15.
142
23. Zeppernick,F.,etal.,StemcellmarkerCD133affectsclinicaloutcomeinglioma
patients.ClinCancerRes,2008.14(1):p.123‐9.
24. Pallini,R.,etal.,Cancerstemcellanalysisandclinicaloutcomeinpatientswith
glioblastomamultiforme.ClinCancerRes,2008.14(24):p.8205‐12.
25. Bieling,P.,etal.,Tumorsizeandprognosisinaggressivelytreatedosteosarcoma.JClin
Oncol,1996.14(3):p.848‐58.
26. Tjandra,S.S.,etal.,IFN{beta}signalingpositivelyregulatestumorigenesisin
aggressivefibromatosis,potentiallybymodulatingmesenchymalprogenitors.Cancer
Res,2007.67(15):p.7124‐31.
27. Tlsty,T.D.andL.M.Coussens,Tumorstromaandregulationofcancerdevelopment.
AnnuRevPathol,2006.1:p.119‐50.
28. Xie,T.andL.Li,Stemcellsandtheirniche:aninseparablerelationship.Development,
2007.134(11):p.2001‐6.
29. Li,L.andT.Xie,Stemcellniche:structureandfunction.AnnuRevCellDevBiol,2005.
21:p.605‐31.
30. Bissell,M.J.andM.A.Labarge,Context,tissueplasticity,andcancer:aretumorstem
cellsalsoregulatedbythemicroenvironment?CancerCell,2005.7(1):p.17‐23.
31. Sneddon,J.B.andZ.Werb,Location,location,location:thecancerstemcellniche.Cell
StemCell,2007.1(6):p.607‐11.
32. Sipkins,D.A.,etal.,Invivoimagingofspecializedbonemarrowendothelial
microdomainsfortumourengraftment.Nature,2005.435(7044):p.969‐73.
33. DeToni,F.,etal.,AcrosstalkbetweentheWntandtheadhesiondependentsignaling
pathwaysgovernsthechemosensitivityofacutemyeloidleukemia.Oncogene,2006.
25(22):p.3113‐22.
143
34. Calabrese,C.,etal.,Aperivascularnicheforbraintumorstemcells.CancerCell,2007.
11(1):p.69‐82.
35. Yauch,R.L.,etal.,Aparacrinerequirementforhedgehogsignallingincancer.Nature,
2008.455(7211):p.406‐10.
36. Shaw,A.,J.Gipp,andW.Bushman,TheSonicHedgehogpathwaystimulatesprostate
tumorgrowthbyparacrinesignalingandrecapitulatesembryonicgeneexpressionin
tumormyofibroblasts.Oncogene,2009.
37. Quintana,E.,etal.,Efficienttumourformationbysinglehumanmelanomacells.
Nature,2008.456(7222):p.593‐8.
38. Berman,S.D.,etal.,MetastaticosteosarcomainducedbyinactivationofRbandp53in
theosteoblastlineage.ProcNatlAcadSciUSA,2008.105(33):p.11851‐6.
39. Kenny,P.A.andM.J.Bissell,Tumorreversion:correctionofmalignantbehaviorby
microenvironmentalcues.IntJCancer,2003.107(5):p.688‐95.
40. Li,Z.,etal.,Hypoxiainduciblefactorsregulatetumorigeniccapacityofgliomastem
cells.CancerCell,2009.15(6):p.501‐13.
41. Smits,R.,etal.,Apc1638N:amousemodelforfamilialadenomatouspolyposis
associateddesmoidtumorsandcutaneouscysts.Gastroenterology,1998.114(2):p.
275‐83.
42. Holmes,C.andW.L.Stanford,Concisereview:stemcellantigen1:expression,
function,andenigma.StemCells,2007.25(6):p.1339‐47.
43. Mitchell,P.O.,etal.,Sca1negativelyregulatesproliferationanddifferentiationof
musclecells.DevBiol,2005.283(1):p.240‐52.
44. Barker,N.,etal.,Identificationofstemcellsinsmallintestineandcolonbymarker
geneLgr5.Nature,2007.449(7165):p.1003‐7.
144
45. Sangiorgi,E.andM.R.Capecchi,Bmi1isexpressedinvivoinintestinalstemcells.Nat
Genet,2008.40(7):p.915‐20.
46. Haldar,M.,etal.,Aconditionalmousemodelofsynovialsarcoma:insightsintoa
myogenicorigin.CancerCell,2007.11(4):p.375‐88.
47. Dominici,M.,etal.,Minimalcriteriafordefiningmultipotentmesenchymalstromal
cells.TheInternationalSocietyforCellularTherapypositionstatement.Cytotherapy,
2006.8(4):p.315‐7.
48. Sarugaser,R.,etal.,Humanmesenchymalstemcellsselfrenewanddifferentiate
accordingtoadeterministichierarchy.PLoSOne,2009.4(8):p.e6498.