stemcells and cancer - columbia university · breast cancer (mmtv-wnt-1) cho et al., stem cells,...
TRANSCRIPT
Stem Cells and Cancer
Piero Dalerba, MD HICCC -‐ Herbert Irving Comprehensive Cancer Center
Department of Pathology and Cell Biology Department of Medicine (Division of Diges+ve and Liver Diseases)
CELLULAR & MOLECULAR BIOLOGY OF CANCER -‐ PATH G4500-‐001 October 19, 2015
Structure of the Class
1. the concept of “stem cells”
2. the concept of “cancer stem cells”
3. biological implicaXons of the “cancer stem cell” model
4. clinical implicaXons of the “cancer stem cell” model
Historical origins of the word “stem cell” [#1]
Ernst Haeckel is famous for having introduced the concept: “Ontogenesis recapitulates phylogenesis”
Ernst Haeckel, Anthropoghenie, 1874 Ernst Haeckel (1834-‐1919)
Images from: the Wellcome Library, London
The early stages of embryonic
development are very similar across species
Historical origins of the word “stem cell” [#2] Historically, Ernst Haeckel first used the word “stem cell” to idenXfy an ancestor/progenitor cell that stands at the “stem” of a genealogic tree, used to depict either evoluXonary (i.e. phylogeneXc) or developmental (i.e. ontogeneXc) processes.
PhylogeneXc tree of life (i.e. evoluXon of life on Earth)
Ernst Haeckel, 1866
ectoderm mesoderm
endoderm
ICM (inner cell mass)
trophoblast
zygote
OntogeneXc tree of life (i.e. embryonic development)
Current use of the word “stem cell”
The meaning remains essenXally intact today, where it is mainly used in developmental biology, to idenXfy a cell capable to generate and sustain over Xme a specific set of diversified cell populaXons whose aggregate interacXon leads to the formaXon of either:
a) an enXre living organism: -‐ the zygote (toXpotent capacity) -‐ embryonic stem cells (pluripotent capacity)
b) a specific subset of its organs and Xssues: -‐ adult stem cells (oligopotent or mulXpotent capacity)
Many Xssues have a “hierarchical” organizaXon and undergo a conXnuous process of cell turnover.
oligolineage precursors
mature cells
long-term self-renewing
stem cells
short-term multipotent progenitors
specialized cell types oeen short-‐lived
undergo turnover during lifeXme (with variable kineXcs)
long-‐lived
Stem Cell ProperXes
oligolineage precursors
mature cells
long-term self-renewing
stem cells
short-term multipotent progenitors
Stem cells are defined by three key properXes:
2) The ability to form new stem cells with idenXcal, intact differenXaXon potenXal, thus maintaining the stem cell pool (self-‐renewal);
self-‐renewal
differen+a+on
1) The ability to give rise to a funcXonally heterogeneous progeny of cells, which progressively specialize according to a hierarchical process (differen+a+on);
3) The ability to modulate the two previous properXes according to environmental sXmuli and geneXc constraints (homeosta+c control).
Examples of normal Xssues that are sustained by adult stem cell populaXons [#1].
Hematopoietic system
Dalerba et al., Stem cells, cell differenXaXon and cancer. In: Abeloff’s Clinical Oncology -‐ 5th ediXon -‐ Chapter 7 (2013)
Intestinal epithelium
Radtke & Clevers, Science 307-‐1904-‐1909 (2005)
Examples of normal Xssues that are sustained by adult stem cell populaXons [#2].
CK20
MUC2
Tumors are histological “caricatures” of corresponding normal tissues
colon cancer SU#56 colon cancer UM#8 colon cancer UM#17
colon cancer UM#4 colon cancer SU#60 colon cancer MIC#69 normal colon
normal colon
David von Hansemann (1858-‐1920)
Tumors are histological “caricatures” of their corresponding normal Xssues.
Images from: Historic Book Collec+on, Humboldt Universität, Berlin
von Hansemann, Die mikroskopische Diagnose der bösar+gen Geschwülste [The microscopic diagnosis of malignant tumors] (1897)
Figure 7. Rectal Cancer […] Great similarity with the normal mucosa. Goblet cells.
normal colon colon cancer
Tumor Xssues act as “histological caricatures” of their normal counterparts.
The three-dimensional architecture and cell composition of neoplastic tissues usually
mirrors that of the corresponding normal tissues of origin, and includes a variety of cell types.
Rajendran and Dalerba, TheoreXcal and experimental foundaXons of the “cancer stem cell” model. In: Cancer Stem Cells (ed. V.K. Rajasekhar) -‐ Chapter 1 (2014)
Key points [#1] • cancer tissues are frequently heterogeneous in cell
composition; this diversity often mirrors the repertoire of specialized cell types found in normal counterparts;
• thus, in many instances, cancer tissues are not simply amorphous collections of transformed cells, but complex three-dimensional structures that retain architectural features of their parent tissues;
amorphous tumor mass
architecturally complex tumor mass
vs.
If tumor Xssues are “caricatures” of normal ones (and retain many of their architectural features), then do they also contain a pathological (e.g. mutated) stem cell populaXon that sustains their long-‐term growth?
The “Cancer Stem Cell” hypothesis
In other words, are malignant Xssues the product of pathological stem cells?
If this hypothesis is correct, what are its biological implicaXons and
how can we test it?
Biological implicaXons of the “Cancer Stem Cell” hypothesis
1) Origin of cell heterogeneity in cancer Xssues Is the cellular diversity observed within tumor Xssues generated only by geneXc mechanisms (e.g. random mutaXons) or also by epigeneXc mechanisms (e.g. mulX-‐lineage differenXaXon)?
geneXc vs. epigeneXc diversity
2) FuncXonal hierarchy in cancer cell populaXons
Are the different types of cancer cells found within malignant Xssues also different in their capacity to sustain
long-‐term Xssue expansion and proliferaXon?
tumorigenic vs. non-‐tumorigenic populaXons
Sources of cell heterogeneity in cancer Xssues [#1]
differences between various types of cancer cells are caused by differences in the repertoire of geneXc mutaXons (i.e. co-‐existence of different geneXc sub-‐clones);
“stochasXc” model (random mutaXon)
differences between various types of cancer cells are caused by differenXal acXvaXon of specialized gene-‐expression programs that “enforce” specific cell idenXXes (i.e. differenXaXon);
“cancer stem cell” model (mulX-‐lineage differenXaXon)
clone #3
clone #2 clone #1
lineage cell type #2
lineage cell type #1
lineage cell type #3
the heterogeneity is geneXc the heterogeneity is epigeneXc
Monoclonal origin of human cancer
Stanley M. Gartler (1923 -‐ present)
From: Rubin’s Pathology: clinicopathologic foundaXons of medicine -‐ 7th ediXon (2015) Strayer and Rubin – Chapter 5 “Neoplasia”.
Nature Rev. Cancer, 10:441-‐448 (2010)
Indirect evidence of mulX-‐lineage differenXaXon in human hematological malignancies
Philip Fialkow (1934-‐1996)
oligolineage precursors
long-term self-renewing
stem cells
short-term multipotent progenitors
mature cells
Image from: University of Washington
CML (chronic myelogenous leukemia)
monoclonal expansions of either A or B G6PD alleles
Sources of cell heterogeneity in cancer Xssues [#2]
IMPORTANT: the two models are not mutually exclusive, the two mechanisms co-‐exist and act together to generate cell diversity… but with different funcXonal consequences!
“stochasXc” model (random mutaXon)
“cancer stem cell” model (mulX-‐lineage differenXaXon)
vs. clone #3
clone #2 clone #1
lineage cell type #2
lineage cell type #1
lineage cell type #3
ImplicaXons of different sources of cell diversity
The different types of cancer cells (i.e. the different geneXc clones within the tumor mass) are all endowed with long-‐term growth and proliferaXon capacity. Each of them can expand as a disXnct monoclonal populaXon.
“stochasXc” model (random mutaXon)
Only a subset of cancer cells, the “cancer stem cells” (CSC), is endowed with long-‐term growth capacity, while other tumor cell types have l imited proliferaXve capacity. Tissues originated from CSC contain mulXple cell types as a result of mulX-‐lineage differenXaXon.
“cancer stem cell” model (mulX-‐lineage differenXaXon)
clone #3
clone #2 clone #1
monoclonal expansion #1
monoclonal expansion #2
monoclonal expansion #3
long-‐term growth and full reconsXtuXon of parent Xssue diversity
limited proliferaXon
cell sorting by flow cytometry
Experimental approach to test the “cancer stem cell” model
primary tumor specimen
dis-‐aggregaXon into single-‐cell suspension
evaluation of differential tumorigenic properties by
injection in NOD-SCID mice +
evaluation of the ability to re-create the phenotypic
heterogeneity of parent tumors +-
laser detector
staining with monoclonal anXbodies
visualizaXon of different cell types by flow cytometry
Experimental validaXon of the “cancer stem cell” model
Colon Cancer
EpCAM
SelecXon of surface markers for differenXal isolaXon by flow cytometry of mature/immature cell types
in the human normal colon epithelium
CD44 merge
Piero Dalerba and Shigeo Hisamori
Sorting and injection strategy for differential tumorigenicity experiments based on EpCAM/CD44 in human Colon Cancer.
2nd sort
P9
P5
2nd sort
subcutaneous injection in immunodeficient mice
(NOD/SCID, NOD/SCID/IL2Rγ-/-)
D
22.6 %
63.8 % A
92.7 %
B
96.7 %
1st sort
1st sort
C
UM-COLON#4 Linneg population
Cell dose and tumor formation
in NOD/SCID mice
dose take
EpCAMhigh/CD44+ 1000 2/3
EpCAMlow/CD44neg 10.000 0/3
CD44
Dalerba et al., P.N.A.S., 104: 10158-10163 (2007)
Reconstitution of parental EpCAM/CD44 expression profile in tumors grown from sorted EpCAMhigh/CD44+ cells.
parent xenografts
UM#4
MIC69
tumors grown from [EpCAMhigh/CD44+]
26.42 %
0.74 % *
19.84 %
3.4 %
ES
Ahi
gh/C
D44
+ ce
lls
14 % 19 %
C
F
BA
ED
Dalerba et al., PNAS, 104:10158-‐10163 (2007)
Morphology and differentiation of human CRC tumors grown from EpCAMhigh/CD44+ cells.
CK20
Muc
Normal epithelium
UM#4 [xeno]
UM#4 [ESA/CD44]
MIC69 [ESA/CD44]
MIC69 [xeno]
A
F
B
G I
DC
H L
E
Dalerba et al., PNAS, 104:10158-‐10163 (2007)
LimitaXons of classic transplantaXon experiments based on flow cytometry [#1]
Due to obvious ethical reasons, transplantaXon experiments using human cells are performed in xenogeneic hosts (i.e. immuno-‐deficient mice).
Are differences in the tumorigenic capacity of different tumor cell types due to an incompaXbility of specific cell types with the mouse microenvironment?
Breast Cancer (MMTV-Wnt-1) Cho et al., Stem Cells, 26:364-371 (2008)
Skin Cancer (squamous cell carcinoma) Malanchi et al., Nature, 452:650-653, (2008)
Mouse epithelial “Cancer Stem Cell” models show that
differences in tumorigenic capacity can exist independently of species-specific differences in
host microenvironment.
CD34+, KRT10neg: tumorigenic CD34neg, KRT10+ : non-tumorigenic
Malanchi et al., Nature, 452:650-‐653, (2008)
DMBA/TPA-induced Skin Carcinoma
Mouse models of epithelial “Cancer Stem Cells”
LimitaXons of classic transplantaXon experiments based on flow cytometry [#2]
Most experiments are performed using high numbers of purified cells (102-‐103 cells/dose).
How can we be absolutely certain that, indeed, the various cell types found in secondary tumors arise from the mulX-‐lineage differenXaXon of a single cell, and not from the parallel expansion of mulXple geneXc clones?
Lineage tracing experiments on “Cancer Stem Cells”
Analysis of a well-differentiated human colon cancer tissue
reveals the presence of multiple epithelial cell types
Analysis of monoclonal EGFP+ colon cancer xenograft tissues reveals the presence of multiple
epithelial cell types, closely recapitulating the full cellular diversity of the parent tissue
Infection of cancer cells with a lentivirus
encoding EGFP
Subcutaneous injection into a NOD/SCID/IL2Rϒ-/- mouse
and generation of a monoclonal EGFP+ colon
cancer tissue
Monoclonality of cancer tissues is verified by
confirmation of a unique lentivirus integration site
Isolation of a single (n= 1) EGFP+ human colon CSC
(EpCAMhigh/CD44+)
Dissociation into a single-cell
suspension
Rajendran and Dalerba, TheoreXcal and experimental foundaXons of the “cancer stem cell” model. In: Cancer Stem Cells (ed. V.K. Rajasekhar) -‐ Chapter 1 (2014)
UM-COLON#4 clone 8 Lin- phenotypic populations
Tested for tumorigenicity
Cell dose and tumor formation in NOD/SCID/
IL2Rγ-/- mice
500 cells 100 cells
EGFP+/EpCAM+/CD44+ 5/5 (100%) 5/5 (100%)
EGFP+/EpCAM+/CD44- 0/5 (0%) 0/5 (0%)
Cell dose and tumor formation in NOD/SCID/IL2Rγ-/- mice (EpCAM+/CD44+)
Xenograft line 10,000 5,000 1,000 500 100 50 30 10 5 1 frequency (95%CE)
UM-COLON#4 - - 3/3 12/12 12/14 3/4 13/30 2/34 0/4 1/132* 1/62 (1/44 – 1/89)
UM-COLON#8 4/4 4/4 4/4 3/4 4/10 2/10 - 1/6 - - 1/223 (1/117 – 1/427)
Clo
ne 8
CD
44-
Clo
ne 8
CD
44+
pare
nt C
D44
- pa
rent
CD
44+
A
B C
F G H
KRT20 Ki67
D E
Generation and characterization of a monoclonal colon cancer xenograft, from a single (n =1) “cancer stem cell”
Dalerba et al., Nature Biotechnology, 29:1120-1127, 2011
Key points [#2]
• the cellular diversity of cancer tissues frequently recapitulates the physiological lineage composition of their normal counterparts;
• the cellular diversity of cancer tissues is frequently epigenetic in origin, and arises as the result of a multi-lineage differentiation process, reminiscent of a stem cell hierarchical system;
• often, within cancer tissues, only a specific subset of cancer cells is endowed with the capacity to initiate the growth of a secondary tumor when transplanted; tumors originated from these cells, however, recapitulate the repertoire of heterogeneous cells types found in the parent tissues; these cells are defined “cancer stem cells”;
Integrated view of cell heterogeneity in cancer Xssues
“stochasXc” model (random mutaXon)
“cancer stem cell” model (mulX-‐lineage differenXaXon)
geneXc clone #3
geneXc clone #2
geneXc clone #1
lineage cell type #2
lineage cell type #1
lineage cell type #3
Integrated model clone #1
(undergoing mulX-‐lineage epigeneXc differenXaXon) clone #2
(undergoing mulX-‐lineage epigeneXc differenXaXon)
clone #3 (undergoing mulX-‐lineage epigeneXc differenXaXon)
Charles Darwin (1858-‐1920)
Images from: the Wellcome Library, London
Conceptual implicaXons of the “cancer stem cell” theory for “evoluXonary” or “darwinian” models of tumor progression
the “units of selecXon” in cancer Xssues are individual “cancer stem cells”
(i.e. the cells with self-‐renewal capacity) as opposed to any individual cell
clone #1 (sustained by “cancer stem cells” with geneXc
background #1 )
clone #2 (sustained by “cancer stem cells” with geneXc
background #2)
clone #3 (sustained by “cancer stem cells” with geneXc
background #3 )
Origin and biological idenXty of “cancer stem cells”
1) “cancer stem cells” vs. normal stem cells Are “cancer stem cells” mutated variants of normal stem cells
(that have lost normal homeostaXc controls) or do they correspond to more mature, differenXated cell types (that have
aberrantly acquired the capacity to self-‐renew)?
2) one vs. mulXple “cancer stem cell” idenXXes Are “cancer stem cell” populaXons homogeneous or
heterogeneous in their epigeneXc idenXXes (i.e. do they encompass one or mulXple cell types)?
TheoreXcal consideraXons in support of a stem cell origin of cancer Xssues.
• • • •
• • • • • • • •
• • • •
Cancer originates as the result of the progressive accumulaXon of mulXple geneXc mutaXons over several years (Fearon and Vogelstein, 1990). Thus, stem cells represent the ideal targets for malignant transformaXon, as they are long-‐lived and have the opportunity to accumulate sequenXal mutaXons over Xme.
• • • •
• • • • • • • •
• • • • • • • •
• • • • • • • •
• • • •
stem cells (long-‐lived)
mature cells (oeen short-‐lived)
1st mutation 2nd mutation 1+2 1
1+2
1+2
1 1
1 1
1
1 1
1
1+2
1+2
1+2 1+2
1+2
2
“Cancer stem cells” is an operaXonal definiXon.
1. Only a subset of cancer cells within each tumor is endowed with tumorigenic capacity and can be serially transplanted into immunodeficient mice.
Three observaXons define the existence of a “Cancer Stem Cell” populaXon:
3. Tumors grown from tumorigenic cells contain mixed populaXons of both tumorigenic and non-‐tumorigenic cancer cells, thus recreaXng the full phenotypic heterogeneity of the parent tumor.
2. Tumorigenic cancer cells are characterized by a disXncXve profile of surface markers and can be differenXally and reproducibly isolated from non-‐tumorigenic ones by flow cytometry.
1) it is a “working tool” to dissect the funcXonal hierarchy of cancer populaXons; 2) it implies the fulfillment of two stem-‐cell properXes (i.e. self-‐renewal and differenXaXon), but it does not imply the origin of “cancer stem cells” from normal stem cells.
Thus, the concept of “Cancer Stem Cells”…
a) are they homogeneous or heterogeneous populaXons?
b) which cell types do they contain?
c) are they composed by immature or mature cell types?
LimitaXons of transplantaXon experiments based on flow cytometry [#3]
Flow cytometry provides only limited knowledge about molecular idenXty, funcXonal status and heterogeneous composiXon of “cancer stem cells”.
+-
3) Single-cells from selected phenotypic populations are sorted into individual 96-well PCR plates
1 cell/well (96-well PCR plate)
6) Individual cell real-time PCR curves are analyzed and converted into gene-expression levels. Individual cells are associated into distinct subsets using
statistical clustering algorithms.
5) 9,216 (96x96) single-cell PCR reactions are run in parallel using a microfluidic chip
1) Primary tissues are collected from surgical specimens and dis-aggregated into single-cell suspensions.
2) Single-cell suspensions are stained with fluorochrome-conjugated monoclonal antibodies and analyzed by flow cytometry
4) RNA is reverse transcribed and loaded into a microfluidic chip
Dalerba et al., Nature Biotechnology, 29:1120-‐1127 (2011)
Analysis of Xssue cell composiXon using single-‐cell genomics
Each row identifies an
individual single-cell
Each column identifies an
individual gene
genes
cells
Fluo
resc
ence
sign
al
Ct value
1) For each of the 9,216 (96x96) SINCE-‐PCR reac;ons, an amplifica;on curve is generated and an individual threshold cycle (Ct) value is calculated.
2) Ct values are normalized and color-‐coded. Normalized Ct values (Ctnorm) are obtained by subtrac;ng from the raw Ct value (Ct) the mean Ct value for the same gene on the whole sample (Ctmean) and then by dividing by three ;mes the standard devia;on of the same gene’s Ct values distribu;on (3SDCt). Results are color-‐coded using increasingly darker shades of red for high expression values (Ct < Ctmean), increasingly darker shades of green for low expression values (Ct > Ctmean) and grey for lack of expression (Ct > 40).
3) Gene-‐expression results are ploUed using hierarchical clustering algorithms available in the MATLAB so[ware (MathWorks Inc.). Hierarchical clustering is performed on both cells and genes, to visualize simultaneously cells with similar expression paUerns and genes with similar expression profiles.
High expression
Low expression
Normalization of SINCE-PCR threshold cycles (Ct)
No expression
Ct >40 Ct mean Ct -3SD
Analysis and graphic display of single-cell PCR data
Hierarchical clustering associates genes with
similar expression patterns
Hierarchical clustering
associates cells with similar expression
profiles
For each gene, Ct values are normalized
and color-coded
red: Ct < Ct mean green: Ct > Ct mean
grey: Ct > 40
Ct value for gene “X”
n. o
f sin
gle-
cells
Ct norm = Ct - Ct mean
3 SD Ct
Ct +3SD
3SD Ct
TaqMan PCR cycle
M96 chip
Dalerba et al., Nature Biotechnology, 29:1120-‐1127 (2011)
CD44
EpCAM
CD66a CEACAM1
Analysis by flow cytometry of the human colon epithelium: dissecXon and purificaXon of disXnct cell subtypes.
Dalerba et al., Nature Biotechnology, 29:1120-1127, 2011
KRT20 CEACAM1 MUC2 Ki67
C DA B E F
Single-cell PCR analysis of human normal colon epithelium
color key - normalized C
t values
high expression
low expression
CA1+ SLC26A3+
MUC2+ TFF3high
OLFM4+ CA2high
LGR5+ ASCL2+
GUCA2B+
Goblet cells
Enterocytes
CBC (columnar basal cells)
Novel populations
Dalerba et al., Nature Biotechnology, 29:1120-1127, 2011
Sorting gate
MUC2 KRT20
UM-COLON#4 Clone 8
OLF
M4+
C
A2h
igh
LG
R5+
A
SCL2
+
MUC2+ TFF3high
Single-‐cell PCR analysis of a monoclonal colon cancer xenograe originated from a single (n =1) “cancer stem cell”
Dalerba et al., Nature Biotechnology, 29:1120-1127, 2011
Key points [#3]
• “cancer stem cell” populaXons, as currently defined using flow cytometry and operaXonal criteria, are oeen heterogeneous and contain mulXple cell types, including immature progenitor populaXons;
• it is unclear whether all or only some of the cell types within the currently defined “cancer stem cell” populaXons are endowed with “cancer stem cell” properXes;
• it also remains unclear whether the epigeneXc idenXty of “cancer stem cells” corresponds to that of normal stem cells or of more differenXated, specialized cell types.
“EvoluXon” of “Cancer Stem Cells” over disease progression?
Dalerba et al., Stem cells, cell differenXaXon and cancer. In: Abeloff’s Clinical Oncology -‐ 5th ediXon -‐ Chapter 7 (2013)
Most likely, the molecular idenXty of “cancer stem cells”
evolves during disease progression.
While, in the early stages of neoplasXc transformaXon, the first round of mutaXons is likely to occur in the long-‐lived stem cell compartment, it is also possible that, during disease progression, a second round of
mutaXons might disable controls in self-‐renewal
pathways and “unleash” self-‐renewal in rapidly dividing mulX-‐potent progenitors.
Does the “Cancer Stem Cell” hypothesis have implicaXons for the modeling of
metastasis?
TradiXonal model of Metastasis.
Dalerba et al., Ann. Rev. Med., 58:267-‐84 (2007)
primary tumor
apoptosis
Clone#1
Clone#2
Clone#3
The “classic” model predicts that primary tumors and autologous metastases will display substanXal differences (e.g. different repertoires of geneXc mutaXons).
primary tumor
Cancer Stem Cell
Differentiated cancer cells
“Cancer Stem Cell” model of Metastasis.
metastasis
The “cancer stem cell” model predicts that primary tumors and autologous metastases will show substanXal degrees of similarity.
• gene-‐expression profiling (Weigelt et al., Nat. Rev. Cancer, 5:591-‐602, 2005) • histopathology (Brabletz et al., Nat. Rev. Cancer, 5:744-‐749, 2005)
apoptosis (e.g. anoikis)
Dalerba et al., Annual Review of Medicine, 58:267-‐284 (2007)
Indeed, this has been a common (though somewhat unexpected) experimental finding:
Clinical implicaXons of the “Cancer stem cell” hypothesis
1) “Cancer stem cells” and tumor aggressiveness Is the cell composiXon of tumor Xssues associated to a more or less aggressive disease? Are tumors with abundant “cancer
stem cells” more aggressive?
2) “Cancer stem cells” and response to treatments Are the different types of cancer cells found within malignant Xssues also different in their sensiXvity to various types of
anX-‐tumor therapies?
Gene-expression arrays:
Breast Cancer Stem Cells
Vs Normal Breast Epithelium
Strategy designed to maximize the content of prognostically relevant information, including:
1) tumor-specific genes
2) cancer stem cell-specific genes
ExploiXng “cancer stem cells” to develop a prognosXc tool for human Breast Cancer paXents
186 genes Invasiveness Gene Signature
(IGS)
Liu et al., NEJM, 357:217-226 (2007)
The IGS is associated with poor prognosis in human
Breast Cancer
Breast Cancer patients whose tumor is characterized by a transcriptional profile similar to the IGS are characterized by reduced overall and disease-free survival.
population:
early breast cancer patients
from the
Netherlands Cancer Institute (NKI database)
Liu et al., NEJM, 357:217-226 (2007)
KRT20
Dalerba et al., Nature Biotechnology, 29:1120-1127, 2011
normal colon colon adenoma colon carcinoma
Sparsity test FRD < 0.0001
Mining of gene-expression databases using Boolean logic.
Expression of differenXaXon genes associates with paXent survival.
Tumor progressively exhausts its growth potenXal
Treatment is broadly cytotoxic, but does not specifically target self-‐
renewing cancer cells
Treatment targets all self-‐renewing cancer cells
Tumor is grossly reduced in size, but eventually relapses driven by self-‐
renewing cancer cells.
a)
b)
long-‐term outcome
type of treatment
ImplicaXons of the CSC model for the design and evaluaXon of anX-‐tumor treatments
Rajendran and Dalerba, TheoreXcal and experimental foundaXons of the “cancer stem cell” model. In: Cancer Stem Cells (ed. V.K. Rajasekhar) -‐ Chapter 1 (2014)
short-‐term outcome
“Stem Cell” working models to explain tumor resistance to classic anX-‐neoplasXc agents (i.e. chemotherapy, radiotherapy)
a) quiescent - G0 state (leukemia)
b) high-level expression of:
- drug pumps (leukemia) - enzymes for DNA repair (brain cancer) - scavengers of ROS (breast cancer) - anti-apoptotic pathways
a) multi-potent progenitors: - high proliferation rates b) mature - differentiated cancer cells: - exposed to high intra-cellular concentrations of cytotoxic drugs - exposed to high levels of ROS (“reactive oxygen species”) - vulnerable to extensive DNA damage - sensitive to apoptosis
Classical anti-tumor agents (e.g. agents toxic to proliferating cells in the S or M phase of the cell cycle)
Examples:
- Vinca alkaloids - Anti-metabolites - Topo-isomerase inhibitors - Ionizing radiation
Cancer Stem Cell (CSC)
a) multi-potent progenitors (not self-renewing)
b) mature - differentiated cancer cells
Tumor is disrupted in architecture, but eventually
relapses driven by self-renewing cancer cells
Treatment targets a specific lineage of cells, not encompassing
all self-renewing cancer cells
ImplicaXons of the CSC model for the design and evaluaXon of anX-‐tumor treatments
“Targeted” therapies might be aimed at cell populations devoid of self-renewal capacity, leading to inconsequential clinical results.
Key points [summary #1]
1) tumor tissues are frequently heterogeneous in cell composition; this diversity: a) is not only genetic, but also epigenetic in origin; b) it often mirrors the physiological diversity of specialized cell types found their normal counterparts; c) can arise as the result of a multi-lineage differentiation process, reminiscent of a stem cell hierarchical system;
2) in tumor tissues, the capacity to form tumors upon transplantation is frequently restricted to a subset of cancer cells, operationally defined as “cancer stem cells”; the precise molecular identity of “cancer stem cells” is still under study, and probably evolves over time, during disease progression;
Key points [summary #2]
3) analysis by gene-‐expression arrays of a bulk tumor’s transcripXonal profile and the idenXficaXon of a high degree of similarity with a gene-‐expression signature characterisXc of “cancer stem cells” is usually associated with a more aggressive disease;
4) anX-‐tumor treatments unable to target and eradicate “cancer stem cell” populaXons are likely unable to achieve long-‐term eradicaXon of tumor Xssues.
“Not everyone accepts the hypothesis of cancerous stem cells. SkepXcs say proponents are so in love with the idea that they dismiss or ignore evidence against it. […] the hypothesis was more akin to religion than to science.”
Gina Kolata “Scien+sts Weigh Stem Cells’ Role as Cancer Cause” The New York Times (December 21, 2007)
Final remarks
Suggested readings
• Ramalho-‐Santos M. and Willenbring M. On the origin of the term “stem cell” Cell Stem Cell, 1, 35–38 (2007)
• Weinberg R.A. MulX-‐step Tumorigenesis Chapter 11 -‐ In: The Biology of Cancer (2nd ediXon), 439-‐509 (2014)
• Rajendran P.S. and Dalerba P. TheoreXcal and experimental foundaXons of the “cancer stem cell” model. Chapter 1 -‐ In: Cancer Stem Cells, 3-‐16 (2014)
• Dalerba P., Clarke M.F., Weissman I.L., Diehn M. Stem Cells, Cell DifferenXaXon and Cancer Chapter 7 -‐ In: Abeloff’s Clinical Oncology (5th ediXon), 98-‐107 (2013)
Conflict of interest disclosures
• Some of the “single-‐cell genomics” techniques described in this presentaXon have been patented by my former academic insXtuXon (Stanford University) and licensed to a start-‐up pharmaceuXcal company (Quan+cel Pharmaceu+cals Inc.). As a result of the licensing agreement negoXated by Stanford University, I was awarded an aliquot of restricted stock in the company. Aside from this, I have no other financial interests in the company (e.g. no paid consultant agreements, no industry-‐sponsored research grants).