Cell Stem Cell
Metastatic Cancer Stem Cells:
An Opportunity for Improving Cancer Treatment?
Maximilian Diehn1,2,* and Ravindra Majeti1,3,*
1Cancer Center and Institute for Stem Cell Biology and Regenerative Medicine
2Department of Radiation Oncology
3Division of Hematology, Department of Internal Medicine
Stanford University School of Medicine, Palo Alto, CA 94305, USA
*Correspondence: email@example.com (M.D.), firstname.lastname@example.org (R.M.)
incompletely understood. In this issue of Cell Stem Cell, Pang et al. (2010) isolate a subpopulation of human
colorectal CSCs that uniquely possesses metastatic potential.
A growing body of evidence has demon-
strated that many human tumors are
organized as cellular hierarchies initiated
and maintained by a subpopulation of
self-renewing cancer stem cells (CSCs)
(Dalerba et al., 2007; Jordan et al.,
2006). These CSCs have been isolated
based on their surface immunopheno-
types and the ability to serially transplant
human cancer in immunodeficient mice,
where they give rise to the heterogeneous
cell types comprising the original tumor.
Many human cancers, particularly car-
cinomas, are able to metastasize to
secondary sites, and these metastases
contribute significantly to patient mor-
bidity and mortality. Cancer stem cells
have been identified in both primary and
metastatic tumors; however, the relation-
ship between CSCs from these two sour-
ces has not been established. Are the
same CSCs responsible for both the
primary tumor and metastases? Or, in
contrast, do distinct CSC subsets exist,
one stationary and one migratory, as has
been proposed (Brabletz et al., 2005)?
Can CSCs capable of metastasis be
identified in primary tumors and can the
presence of these cells predict the occur-
rence of metastases?
Data supporting the existence of sepa-
rable populations of stationary and migra-
tory CSCs come from the investigation
of a human pancreatic cancer cell line,
in which both the CD133+CXCR4?and
CD133+CXCR4+fractions were capable
of sustaining tumor growth, but depletion
of the CD133+CXCR4+fraction abrogated
the formation of metastases (Hermann
is provocative, formal demonstration of
this model requires experiments with
primary human tumors. In the present
study, Pang and colleagues demonstrate
the existence of a subpopulation of CSCs
in primary human colorectal cancer that
uniquely possesses metastatic potential
(Pang et al., 2010).
The authors examined surface antigen
expression on primary colorectal cancers
and liver metastases and observed that
CD26 was preferentially found on the
metastases. Moreover, they determined
that all 16 metastases examined con-
tained CD26-positive cells, while such
cells were present in only 8 out of 27
primary tumors. Strikingly, 0 out of 19
tumor developed liver metastases, while
5 of the 8 patients with CD26+ cells ulti-
mately developed liver metastases (the
remaining 3 subjects were followed for
a much shorter time period). These clini-
cal data strongly link the presence of
CD26+ cells in the primary tumor to the
CD26+ cells metastatic CSCs?
To investigate this important question,
the authors established an orthotopic
xenotransplantion assay in which colo-
rectal cancer subpopulations were im-
planted into the mouse cecal wall. Both
CD26+and CD26?populations were able
to initiate growth of human tumors in the
tions established liver metastases in this
assay. Consistent with this observation,
only CD26+cells were detected in the
portal vein, and unlike CD26+cells, even
direct portal vein injection of CD26?cells
did not result in liver metastases. These
ulation of metastatic human colorectal
CSCs that can be isolated based on its
expression of CD26.
These findings have significant clinical
implications for the diagnosis and treat-
ment of human colorectal cancer. The
identification of subgroups of patients at
highest risk of metastasis is an important
area of clinical investigation because
these patients stand to gain the most
from aggressive systemic therapy. The
findings by Pang et al. suggest that anal-
could identify early-stage patients who
will ultimately develop metastases. Thus,
in large, independent patient cohorts.
Accomplishing this important verifica-
tion step will require the establishment of
routine methods to assess CD26 expres-
sion in patient specimens. Pang et al.
employ flow cytometry for this purpose,
and while this technique is used clinically
in certain situations (e.g., hematologic
malignancies), it is currently not routinely
performed on solid tumor specimens,
which are more difficult to process than
blood samples. The most straightforward
tochemistry (IHC)because thisisroutinely
performed in clinical laboratories. How-
ever, IHC may lack the necessary sensi-
tivity and dynamic range to identify
small subsets of CD26+CSCs in primary
tumors. In addition, use of quantitative
real-time PCR and other assays on bulk
tissues is complicated by the fact that
CD26 is expressed by both malignant
and normal cells, making it difficult to be
certain which cells are contributing to
the measured expression. Most likely,
Cell Stem Cell 6, June 4, 2010 ª2010 Elsevier Inc.
novel technologies and approaches will
be required to reliably assess CSC
markers, such as CD26, in clinical speci-
mens. Among the leading candidates are
devices that can purify circulating tumor
cells from peripheral blood and allow sub-
sequent analysis of marker expression
(Nagrath et al., 2007; Talasaz et al., 2009).
Since CSC responsible for metastasis will
most likely be present in the circulation,
this represents an intriguing, minimally
invasive approach that could overcome
the challenges inherent in isolating cells
from solid tissues.
Beyond assessment of CD26 expres-
sion in primary tumors, it will be important
to relate the presence of CD26+CSCs to
our existing molecular understanding of
colorectal cancer (Cunningham et al.,
2010). Specifically, it will be important
to elucidate if CD26+CSC-containing
tumors are enriched among those with
the chromosomal instability or microsa-
tellite instability phenotypes or if there is
overlap between these tumors and those
with derangements in oncogenes such
Assuming the expression of CD26 by
CSCs can be reliably measured in clinical
specimens, an important question will be
whether the new findings can be used
to select patients who will benefit from
motherapy aimed at eliminating micro-
scopic metastases. Currently, the use of
such therapies remains controversial in
patients with stage II disease (i.e., deeply
invading tumors without lymph node
metastases) (Cunningham et al., 2010).
Since ?75%–80% of stage II patients
survive long term even in the absence of
adjuvant treatment, it has been difficult
to show a clear survival benefit for the
addition of chemotherapy (Gill et al.,
2004). Extrapolating the findings of Pang
et al., it is possible that the subset of
stage II patients whose tumors contain
CD26+CSCs are most likely to benefit
from systemic treatment. To test this
could be designed which stratify patients
based on the presence of CD26+CSCs
and only administer chemotherapy to the
patients most likely to develop metas-
tases. A similar trial design is currently
being used to test if loss of heterozygosity
at 18q, which correlates with higher risk
of metastasis, and microsatellite insta-
bility, which correlates with lower risk
of metastasis, can be used to select
stage II patients for adjuvant treatment
The most exciting clinical implication of
apeutically target CD26 for the treatment
cancer. Pang et al. show that downregu-
lation of CD26 decreased the migratory
and invasive capacities of CD26+CSCs
in vitro. This result suggests that blocking
monoclonal antibodies or inhibitory small
molecules targeting CD26 could eliminate
or inactivate metastatic CSCs. Even if
such therapies were not directly cytotoxic
to CD26+CSCs, preventing their ability
to metastasize would represent a novel
therapeutic approach that could be envi-
sioned as an important component of
multimodality cancer therapy.
Brabletz, T., Jung, A., Spaderna, S., Hlubek, F.,
and Kirchner, T. (2005). Nat. Rev. Cancer 5,
Cunningham, D., Atkin, W., Lenz, H.J., Lynch, H.T.,
Minsky, B., Nordlinger, B., and Starling, N. (2010).
Lancet 375, 1030–1047.
Dalerba, P., Cho, R.W., and Clarke, M.F. (2007).
Annu. Rev. Med. 58, 267–284.
Gill, S., Loprinzi, C.L., Sargent, D.J., Thome ´, S.D.,
Alberts, S.R., Haller, D.G., Benedetti, J., Francini,
G., Shepherd, L.E., Francois Seitz, J., et al.
(2004). J. Clin. Oncol. 22, 1797–1806.
Hermann, P.C., Huber, S.L., Herrler, T., Aicher, A.,
Heeschen, C. (2007). Cell Stem Cell 1, 313–323.
M., Bruns, C.J.,and
Jordan, C.T., Guzman, M.L., and Noble, M. (2006).
N. Engl. J. Med. 355, 1253–1261.
Nagrath, S., Sequist, L.V., Maheswaran, S., Bell,
D.W., Irimia, D., Ulkus, L., Smith, M.R., Kwak,
E.L., Digumarthy, S., Muzikansky, A., et al.
(2007). Nature 450, 1235–1239.
Pang, R., Law, W.L., Chu, A.C.Y., Poon, J.T., Lam,
C.S.C., Chow, A.K.M., Ng, L., Cheung, L.W.H.,
Lan, X.R., Lan, H.Y., et al. (2010). Cell Stem Cell
6, this issue, 603–615.
Talasaz, A.H., Powell, A.A., Huber, D.E., Berbee,
J.G., Roh, K.H., Yu, W., Xiao, W., Davis, M.M.,
Pease, R.F., Mindrinos, M.N., et al. (2009). Proc.
Natl. Acad. Sci. USA 106, 3970–3975.
Cell Stem Cell
Cell Stem Cell 6, June 4, 2010 ª2010 Elsevier Inc.