Cancer Stem Cells

1
Cancer Stem Cells CAMB 512 February 21, 2008
2
What makes a stem cell a stem cell?

• Self-renewal - when a stem cell divides, each
  daughter cell can either retain its stem cell
  identity, or can become a progenitor. Stem cells
  appear to retain capacity for limitless division.
• Multipotency - progenitors (also called transit
  amplifying cells) have limited proliferative
  capacity and are committed to terminally
  differentiate into multiple lineages
  (multipotent). Often have high proliferative
  rates in the short term, whereas most stem cells
  divide rarely.

Stem cell
Progenitor cell
B
C
D
A
Different cell types
3
Embryonic stem cells derived from blastocysts -
E3.5 mouse embryos
Pluripotent (able to contribute to all embryonic
tissues in vivo) Can be primed to differentiate
into an increasing number of cell types in
vitro Human ES cell linestechnical, ethical,
oversight issues abound
4
What are the criteria by which stem cells can be
identified?
Paradigm - hematopoietic stem cells (HSCs) -
found in bone marrow in adults (yolk sac blood
islands, AGM, and fetal liver in developing
embryos). Self-renew and are multipotent - but,
very few in number, hard to identify. How can we
study in the laboratory?
Bone marrow biopsy
5
Adoptive transferthe idea behind Bone Marrow
Transplant in patients
One can also separate marrow cells by virtue of
the cell-surface proteins they express by
fluorescence activated cell sorting (FACS) By
testing these separated cells, identified some
that produce long-term (ie, permanent)
reconstitution of all blood cell lineages in
recipient mice - at least some of these are stem
cells. Short-term reconstitution characteristic
of progenitor cells. LTR cells express specific
markers, pump out rhodamine or Hoescht dyes
through ABC transporters (side-population
cells). Very few cells (1/10,000) in bone
marrow have LTR stem cell properties
6
Current view of hematopoietic differentiation
from HSC - gradual process of restricting fate.
CSFs and lineage-specific transcription factors,
drive expression of critical genes
7
Criteria for identifying adult stem cells - need
for caution
Self-renewal and pluripotency - require relevant
assays!! Cell purification - ideally, single
cell resolution (markers). Single LacZ
Lin-CD29hiCD24 cell repopulates a mammary fat
pad.
Shackleton et. al, (2006) Nature 439 84-88
8
Recognized fairly recently that a subset of
cancer cells seem capable of self-renewing and
producing tumors either in vitro or in vivo.
1994 - first real identification of a true
cancer stem cell (AML), which can repeatedly
confer disease in recipient mice. How do stem
cells relate to cancer?
Huntly and Gilliland (2005) NRC 5 311-321
9
Cells initiating acute myeloid leukemia after
transplantation into SCID mice
John Dick Leukemia Initiating cell CD34CD38-
1/250,000 AML cells NOT CD34CD38
CD34-CD38- Lapidot et al. (1994) Nature
367645-648.
10
Human AML originates from primitive HSCs? (but
frequency varies)
Bonet and Dick (1997) Nat. Med. 3730-737
11
(No Transcript)
12
Model for normal and AML hematopoietic system
13
Non-hematopoietic cancer stem cells
Cells from 9 breast cancer patients (pleural
effusion) either purified directly (UP) or
passaged once in mice (P) Based on previous work
looking at cell surface markers on breast ca
cells, authors purified a subpopulation of tumor
cells that express high levels of CD44
expression, low levels of CD24, etc. ONLY these
CD44, CD24-/low cells (15 of total)
reproducibly formed tumors when injected into the
mammary pads of immunodeficient mice. Other
cells did NOT form new tumors, even when 10 X
more cells were injected (exception - T7).
Consistent with idea of cancer stem cells
Al-Hajj et. al (2003) PNAS 100 3983-3988
14
Histology from the injection site
Would you have any concerns about this approach?
15
Dick (2003) PNAS 1003547-3549
16
(No Transcript)
17
Where would a cancer stem cell come from?
Transformation of normal SC? Alternative -
dedifferentiation of a committed progenitor
through which it gains ability to self-renew.
Recent evidence strongly supports this idea In
either case, multiple additional mutations are
almost certainly needed to create full-blown
cancer stem cell Progeny of CSC arent normal,
but have limited proliferative potential (similar
to normal progenitors).
Pardal, Clarke, Morrison NRC 2003
18
August 2006 - report in Cell that fibroblasts can
be converted into embryonic stem-like cells
when engineered to express only four specific
genes!! (not really true ES cellsbut pretty
close!) So, aspects of stem-ness can be
conferred on differentiated cells with minimal
genetic changes! Implications for cancer stem
cells are huge...
Takahashi and Yamanaka 2006 Cell 126 1-14
19
Fibroblasts engineered to express only four
specific genes - Oct-4, c-Myc, Sox2 and KLF4, can
be converted into cells indistinguishable from
bona fide ES cells. These iPS (induced
pluripotent cells) can generate live-born
mice! Demonstrates that true stem-ness can be
conferred on non-stem cells
From Wernig et. al (2007) Nature 448 318-325
See also Okita et. al, (2007) Nature 448
313-17 Maherali et. al (2007) Cell Stem Cell 1
55-70
20
Also 2006, Scott Armstrongs group (HMS) showed
that if they introduced a specific leukemic
translocation (MLL-AF9) into committed progenitor
cells (GMPs), and then injected these transformed
cells into mice - they get cancer (AML)! Called
cells that cause this leukemia L-GMPs. But -
the disease is transferable to new mice - hence,
gained self-renewal 1/6 cells by limiting
dilution. Culturing cells in differentiation
medium reduces the number of L-GMPs - so they can
also differentiate into other cells...
Krivtsov et. al (2006) Nature 442818-822
21
Self-renewal correlates to changes in gene
expression - 363 genes altered in the
self-renewing L-GMPs compared to initial GMPs.
91 of those 363 genes have been linked to
self-renewal in other stem cells. Implication -
one transforming event (in this case, the MLL-AF9
transgene) may produce conditions that allow for
selection of a self-renewal program... Details
will differ for different tissues, but - it may
not require a large number of genetic changes to
make a tumor initiating cell Time will tell -
22
Cancer Stem Cells Identified in Human Cancers

• AML (acute myeloid leukemia)
• medulloblastomas
• glioblastomas
• colon cancer
• ALL (acute lymphoblastic leukemia)

23
Identification of human CD133 brain tumor
initiating cells
Singh et al. (2004) Nature 432396-401
24
100 CD133 cells sufficient to induce
medulloblastoma in mice 105 CD133- insufficient
25
Human colon cancer cells initiating tumors in
SCID mice
CC-ICs are CD133
OBrien et al. (2007) Nature 445106-110
26
Limiting Dilution Analysis
27
Rare CD133 cells in colon cancer
Ricci-Vitiani et al. (2007) 44511-115
28
(No Transcript)
29
The big question - whats different about cancer
SCs (or tumor-initiating cells - TICs) from
normal SCs? Traditional therapies may be best at
debulking majority of cancer cells in tumors,
but may be less effective at eradicating cancer
SCsalso, SCs are particularly resistant to many
chemotherapeutics (ABC transporters, etc.) -
side population cells particularly good at
pumping out drugs, making them resistantobvious
problem! Multi-drug resistance (MDR)
30
Are we targeting the right cancer cells.?!? If
we could target cancer SCs, what about your skin,
blood, gut lining, etc.? How selective could we
be?
Huntly and Gilliland (2005) NRC 5311
31
Rich (2007) Cancer Res. 678980-8984
32
Wang (2007) Cell Stem Cell 1497-501
33
Are Cancer Stem Cells Necessarily Rare?
34
LSCs are frequent in mice with MLL-AF9 leukemia
Somervaille and Cleary (2006) Cancer Cell
10257-268
35
Tumor growth need not be driven by rare cancer
stem cells
Kelly et al. (2007) Science 317337
36
Bottom line
Cancers may contain cells with some capacity to
self-renew and produce multiple transformed cell
types (TICs or CSCs). Not necessarily
transformed stem cells These may represent
important targets for future therapies Genetic
instability inherent to cancer cells generates
mutations that can be selected during treatment,
resulting in more resistant form of the
disease Successful treatment of cancer -
transforming it from a fatal to a chronic disease
- will require a combination of different
treatments tailored to each cancer as it
develops. Some will target common features, some
individual ones The challenge - how do we
identify the critical Achilles heel for a given
tumor (given the large number of mutations, some
of which are important, some perhaps not), and
how do we treat it without killing normal cells?