|Are all Glioma Cells Cancer Stem Cells?
|Cruz Mabel1, Siden Åke1,2, Tasat Deborah Ruth3 and Yakisich J. Sebastian1*
|1Department of Clinical Neuroscience R54, Karolinska Institute, Stockholm, Sweden
|2Department of Neurology, Karolinska University Hospital, Stockholm, Sweden
|3Universidad Nacional de San Martín Buenos Aires, Argentina
||Dr. Yakisich J. Sebastian, Ph.D.,
Department of Clinical Neuroscience,
Karolinska University Hospital,
Tel: +46 8 585 89 533,
Fax: +46 8 585 83810,
|Received May 06, 2010; Accepted June 16, 2010; Published June 16, 2010
|Citation: Cruz M, Siden Å, Tasat DR, Yakisich JS (2010) Are all Glioma Cells Cancer Stem Cells? J Cancer Sci Ther 2: 100-106. doi:10.4172/1948-5956.1000032
|Copyright: © 2010 Cruz M, et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
|The cancer stem cell theory proposes that there is a small but constant subpopulation of cancer cells with stem
cell properties responsible for the self renewal capacity and unlimited proliferation of tumor as well as increased resistance
to antineoplastic drugs. Targeting these cells might constitute an effective way to cure cancer. Regarding gliomas, by
analysing proliferation kinetics of cultures containing mixed subpopulations and experimental data from literature on
glioma cell lines, we propose a model (Stemness Phenotype Model) in which all glioma cells have stem cells properties
but their phenotype varies depending on the environmental conditions. This model provides an alternative explanation
to different and sometimes controversial experimental findings and might be a useful guide for future research in the
field of gliomas and stem cell biology.
|Cancer; Glioma; Stem cells; Proliferation; Stem cell
|Limitless replicative potential is one of the hallmarks of cancer
cells and normal stem cells: while normal cells have limited replication
capacity, stem cells and cancer cells can be propagated indefinitely
(Hanahan and Weinberg, 2000; Rubin, 2002). Stem cells have also
long-term self-renewal ability and capacity to give rise to one or
more types of differentiated progeny (Nicolis, 2007). The existence of
cancer stem cells was reported in several cancer cell lines (Setoguchi
et al., 2004) suggesting that cancers are maintained by cancer stem
cells making them an important therapeutic target. Increasing data
support the existence of cancer stem cells in gliomas and several
publications reported the existence of glioma stem cells (GSCs) in
available glioma cell lines as well as in patient derived cell lines
(Fukaya et al., 2010; Kondo et al., 2004; McCord et al., 2009; Qiang et
al., 2009; Ropolo et al., 2009; Singh et al., 2004; Zheng et al., 2007; Zhou et al., 2009). In general, resistance of gliomas and other brain
tumors to therapy and relapse after standard treatments is attributed
to the presence of stem-like cells (Charles and Holland, 2010). Criteria
for defining GSCs include properties such as multipotentiality,
self-renewal, indefinite proliferation in vitro (Limitless Replicative
Potential), and tumorigenicity in vivo. Regarding the percentage
of glioma stem cells in cell lines, the data have been controversial.
The best studied cell line is the glioma C6 cell line where authors
reported values from 0.4 to 100% (Table 1). These marked differences
can be attributed to technical procedures such as among others,
isolation methods, cell culture conditions, stem cell markers (Table
1). Probably, the best experimental approach to define the percentage
of GSCs in the C6 glioma cell line was the work published by Zheng
et al. (2007) (Zheng et al., 2007), where isolated single cells were
expanded as clonal lines and representative subclones evaluated
for their ability to sustain tumor growth in mouse. In that study
67 out of 67 subclones were able to form a new tumor in mouse
indicating that most (probably 100%) of cells were cancer stem cells.
The authors also proposed a model where stem cells always divided
symmetrically (Figure 4 in Zheng et al. (2007)). Shen et al. (2008) found by means of a tumor sphere culture system and a single-cell
subsphere generation assay that the majority C6 cells (>80%) have
stem cell properties (Shen et al., 2008) while analysis of stem cells
markers and the “side population” showed that a small percentage
of C6 cells has stem cell properties (Kondo et al., 2004; Zhou et al., 2009). The current dogma in the cancer field is that cell lines contain
at least two subpopulations of cells: cancer stem cells and non-cancer
stem cells. Elucidating the real percentage of each subpopulation in
cell lines and cancers in vivo is not only of academic interest but has
also important implications for the development of new therapeutic
modalities for cancer treatment.
||Table 1: Detection of GSCs in established glioma cell lines.
|Are all glioma cells in cell lines stem cells?
|It is widely accepted that GSCs proliferate slower than non-GSCs
and experimental data support this statement. A recent study showed
that the average population doubling time (PDT) for non GSCs and
GSCs is approximately 28-30h and 55-60h respectively (Ropolo et al.,
2009). In order to bring together experimental data and the stem cell
concept we analysed the theoretical possibility of coexistence of two
different subpopulations in a same cell line. A simple mathematical
analysis predicts that existing cells lines containing at least two
subpopulations with different cell cycle length, when propagated for
a high number of passages, should be enriched for the fastest dividing
subpopulation initially present at the moment of isolation. Even small
differences in the cell cycle length will produce a cell line enriched
for the fastest subpopulation (Yakisich, 2009) . Since the C6 glioma
cell line has been propagated thousands of times (Kondo et al., 2004)
it is expected that the slowly proliferating stem cells subpopulations
might have disappeared at some point. However, the C6 glioma cell
line was apparently isolated and expanded from a single cell (Benda
et al., 1968). Then, to explain the presence of GSCs and non-GSC
in this cell line, the primordial C6 cell must have been a GSC that,
by definition of SC, was able to originate non-GSCs (this assumption
does not explain Zheng’s experiment, see below). Other cell lines and
patient derived cell lines were obtained by propagating mixed cell
populations instead of single cells. Frequently cell biologists isolate
cell lines from tumoral tissue. Therefore, in these cases, the primary cell culture might contain different cell subpopulations with different
proliferative kinetics. After continuous passages, the expected
progressive enrichment of the fastest growing subpopulation opens,
in the stem cells field, some controversial questions. It is likely that
glioma cells lines after extensive passages should contain either pure
non-GSCs (that proliferate fast) or pure GSCs (that proliferate slowly)
but not both. Since several cell lines, as mentioned above, were
isolated from expanding mixed tumors, it is expected that most of
the cell lines should contain only non-GSCs (if they outgrow GSCs).
The second possibility is that the expansion of the culture selected
for stem cells because non-GSCs did not survive the procedure or
they spontaneously stop dividing after they reach the Hayflick limit.
|This theoretical analysis raises the following question: how
can glioma cell lines maintain a constant but rare subpopulation of
GSCs? The model proposed by Zheng et al. (2007) (Figure 4 therein)
concludes that “a cell line cannot maintain a rare but constant fraction
of stem cells unless the stem cells divide symmetrically”(Zheng et al.,
2007). The model shows that when starting a culture containing 99
non-GSCs and 1 GSC (1%), after four cell divisions the culture would
contain 1564 non-GSC and 16 GSCs (1%). This might be true only
for non-GSCs and GSCs dividing at the same rate. Zheng’s model
assumes that stem and non stem cells have similar proliferation
rates. When the slower PDT of GSCs is considered, this scenario
changes completely and fails to reconciliate experimental data. Table
2 shows the expected outcome for the percentage of two different GSCs with PDTs (24h and 36h) present in a cell culture when starting
at a ratio non-GSCs: GSC= 99:1 and compares it to non-GSC (PDT=
24h). It clearly shows that the slower proliferating cell subpopulation
(PDT 36h) will gradually decrease and eventually disappear when
co-incubated with a subpopulation of shorter PDT (24h). Using the
same type of analysis showed in Table 2, a faster subpopulation (PDT
= 18h) will undergo 8 divisions within 144h and its percentage will
increase up to 4%. Undoubtedly, unless GSCs and nonGSCs have the
same PDT, one cell subpopulation will sooner or later outgrow the
other. Thus, it is unlikely that mixed non-GSCs and GSCs cultures will
keep a constant but rare fraction of stem cells as the model based in
independent cell subpopulations proposes.
||Table 2: Percentage of GSCs in a mixed culture compared to non-GSCs of similar
(Blue) or longer PDT (Red) than the GSC subpopulation.
|The symmetrical and asymmetrical cell division models
|Can the symmetrical or asymmetrical nature of cell division
explain the experimental findings?
|In symmetrically dividing cells (GSCs producing only GSCs and
non-GSCs producing only non-GSCs), isolation of single cells will
produce only pure GSCs clones and subclones or pure non-GSCs
clones and subclones (Figure 1A-B). To account for a small population
of non-GSCs, one can assume that a fraction of GSCs “differentiate”
producing non-GSCs (Figure 1B, box). In this case, it becomes a type
of asymmetrical division after all. Once a non-GSC (that divides faster
than the GSCs) is generated it will outgrow the GSC population and
at the end the culture will become a pure non-GSC cell line. More
important, this assumption does not explain Zheng’s experiment
where 100% of isolated cells had stem cell properties.
|Asymmetrical division of non-GSCs is not possible by definition
(Figure 1C). In the case that GSCs always divide asymmetrically
rendering a GSC and a non-GSC, the clonal expansion of a single
GSC will give a culture containing always 1 GSC and increasing
number of non-GSC (Figure 1D). Of course, one can argue that in
the asymmetrical cell division model, the unique GSC present in
the culture might at some point divides symmetrically. (Figure 1D box). The only “advantage” of this model is that it will explain the
presence of more than 1 GSC cell in the clonal line. Due to the
initial high number of non-GSCs and the slower PDT of GSCs, it is
expected that the clonal expansion of a single GSC cell will produce a
culture composed mostly of non-GSCs. Once again, this model does not explain why 100% of GSCs after single cell isolation have stem
cell properties. In summary, both “symmetrical division” followed
by differentiation of a certain fraction of GSCs into non-GSCs and
“asymmetrical division” followed by symmetrical division in a fraction
of GSC might explain the occurrence of a fraction of GSCs. To explain
the constant percentage of GSCs, one should assume a very delicate
balance between non-GSC generation (produced by symmetric
division or differentiation), cell death and proliferation rate of the
|In 2007, Blagosklonny, in order to explain the constant fraction
of stem cells, postulated the existence of “stemloids”, defined
basically as stem cells that proliferate fast (Blagosklonny, 2007).
Again, to maintain a constant fraction of stem cells, “stemloids”
must proliferate at the same rate as non-GSC otherwise, “stemloids”
depending whether they proliferate slower or faster than non- GSCs
will eventually either disappear from the cell culture or outgrow the
non-GSCs rendering either a pure “stemloids” cell line or a pure non-
GSCs cell line. In spite of these assumptions, none of the models
explain Zheng’s experiment because all of them predict the existence
of non-GSCs that will generate clones without SC properties.
|The “stemness phenotype” model (SPM)
|The term “stemness” is by itself controversial (Leychkis et
al., 2009). In this paper, in order to follow the criteria usually
adopted to define cancer stem cells, we use the term “stemness”
as the property of having the potential for limitless replication, selfrenewal,
multilineage differentiation, and tumorigenicity. Any model
regarding the stem cell presence in culture should explain at least 1)
the variable percentage of GSCs in the C6 glioma cell line reported
by several groups
(Table 1, 2) Zheng’s experiment, 3) the
conserved tumorigenic property of the cell line over time that might be due to the persistence of at least a rare but constant fraction of
GSCs in cell lines.
|All the models discussed above assume that there are at least
two different subpopulations (2 compartment models). The stemloid
hypothesis proposes the existence of three subpopulations (non-
GSCs, GSCs and stemloids). We propose a “1 compartment model”
that we call “Stemness Phenotype Model, SPM” where there is only
one cancer cell type. These cells are cells with different stemness
phenotype due to random biological variation (Figure 2I). In
cultures having thousands to millions of cells, individual cells varies
phenotypically (e.g. expression of stem cell markers, sensitivity to
drugs, resistance to apoptosis) due to random variability giving rise
to the apparent presence of different subpopulations (e.g. SP fraction,
CD133+ fraction, drug resistant cells). The stemness depends on the
environment where the cells grow and can range from a phenotype
resembling a non-GSC to a pure GSC. A prediction of this model is
that there are cells having “intermediate phenotypes” between both
extremes. This seems to be the case since e.g., some CD133- -a non-
GSC trait- cells have self renewal ability that is a GSC trait (Kelly et
al., 2009). Thus, all cells have stem cell potential but they require
a permissive environment to express specific traits. Moreover,
these different phenotypes can interconvert into each other when
permissive environmental changes occur (Figure 2II). In 2006, Hill
suggested that “it is very possible that, in cancer cells, degrees of
stemness exist and that these are variably expressed depending on
the environment to which the cells are exposed” (Hill, 2006 ). In
agreement with Hill’s conclusion, the SPM proposes that the culture
condition modifies the stemness phenotype and dictates the apparent
rate of GSCs/nonGSC phenotype. Our hypothesis expands Hill’s
conclusions based in theoretical analysis of proliferation kinetics of
mixed populations (see above) and new experimental data generated
later in several laboratories. At a constant culture condition (e.g.
routine serum containing media) the majority of cells adopt a non-
GSCs phenotype. Strictly speaking, not all cells in vitro grow under
the same, constant culture conditions. For instance, incubation time
changes the availability of nutrients and cells should adapt to these
changes. In cultures favouring the GSC phenotype (e.g. Serum free
+ EGF + FGF) cells having a GSCs phenotype proliferate while other
cells stop dividing and eventually die due to the harsh conditions.
Cells adapt to the changing environment and drastic changes from
one environment to another might trigger programmed cell death
and only few cells might survive. For example, shifting from serum
containing media to serum free media (or similar stem cell media)
might be harmful for most cells with non-GSC phenotype. Milder
environmental changes might shift the percentage of cells having
one phenotype to another. In support of this “adaptation process”, it
was recently reported that the in vivo environment changed the selfrenewal
capacity of C6 cells (Shen et al., 2008), and the expression of
CD133 (Griguer et al., 2008; Qiang et al., 2009; Soeda et al., 2009).
|Interestingly, it was reported that the human lung carcinoma cell
line DLKP contains 3 distinct subpopulations. On prolonged cultures
two of them can interconvert to the third one and the growth and
attachment properties of the clones themselves varied under the
different assay conditions (McBride et al., 1998). Due to the fact that
they have different proliferation kinetics, it is likely that the three
subpopulations are different phenotypes of the same cell and not true
different subpopulations. Unfortunately, the presence of stem cells
has not been studied in this cell line. However, the fact that in a cell
line a subpopulation having a certain phenotype can interconvert to
another phenotype is a strong argument favoring the interconversion
among different cell phenotypes proposed by the SPM (Figure 2II).
||Figure 2: The Stemness phenotype model. I) All cancer cells have stem cell potential and divide symmetrically. Clonal expansion of a single cancer cell generates,
due to random biological variation cells having different phenotypes (“stemness”) ranging from a pure GSC phenotype (red oval) to a pure non-GSC phenotype (green
oval) depending on the environment (e.g. A, B or C). Environment A represents an environment similar to stem cell media and B represents an environment similar
to standard culture conditions. Environment C represents an intermediate condition. II) In this example, three different phenotypes (pure SC, red ovals; pure non-SC,
green ovals and an intermediate phenotype, red-green ovals) are represented in different “niches”. All three phenotypes can potentially interconvert at variable degrees
(arrows) into each phenotype when permissive changes in the microenvironment occur. III) In vivo, the percentage of each fraction depends on tumor microenvironments
(A, B, C) that might promote specifi c phenotypes.
|How does the SPM brings together experimental data in a
single model? The SPM proposes that there are no true different
subpopulations of GSCs and non-GSCs but a single cell type that
can interconvert to each other depending on the environmental
conditions. By doing so, GSCs don’t need to be constantly generated
in cell lines and explains why GSCs don’t disappear even when a cell
line (e.g. the C6 cell line) has been sub-cultivated thousands of times.
Instead, the constant presence is explained by the ability of cancer
cells to adopt different phenotypes according to the environment
(e.g. different culture conditions and isolation methods used in
experiments showed in (Table 1). The same argument explains the
conserved tumorigenic property of the C6 glioma cell line over time.
The adaptation to changing environments is also critical to interpret
Zheng’s experiment where 67 out of 67 subclones obtained after
isolation and expansion of single cells in serum containing media were
able to induce tumors: Single cells expanded in serum free media
did not form clones (but remained quiescent and viable) or formed
clones with limited growth but, after shifting to serum containing
media, both types of cells were able to form typical tumorigenic
clones (Zheng et al., 2007). Thus, their data indicate that most C6
are “cancer stem cells” and the environment dictates the phenotype.
One can assume that serum containing media favor proliferation of
all cells that allows successful expansion of single cells generating
phenotypic diversity in the expanded clonal culture. The phenotypic
diversity we postulate is supported by the fact that a) Clones formed from CD133- single cells generate descendant having a mixture of
CD133- and CD133+cells (Zheng et al., 2007). b) Both CD133- and
CD133+ cells have also self renewal capacity (Chen et al., 2010; Kelly et
al., 2009), c) non-SP cells can generate both SP and non-SP cells (Fong
et al., 2010; Platet et al., 2007) apparently depending on the culture
conditions since there are studies where non-SP cells only produce
non-SP cells (Kondo et al., 2004). d) expression of CNS markers (GFAP,
Nestin; and NES) vary with culture conditions (Prestegarden et al.,
2010) providing strong in vitro experimental evidence that tumor
microenvironment might be an important factor in determining the
percentage of cells expressing certain stem cell marker and perhaps,
affecting the stemness of the cell as our model suggest. Nestin has
been used as a glioma stem cell marker (Table 1). Another line of
evidence supporting our hypothesis comes from the existence of the
“side population” (SP) in glioma cell lines. In several experiments,
GSCs have been isolated from the SP cell fraction that has been
found to vary between cell lines (Table 1) and it is thought that this
fraction are stem cells (Fukaya et al., 2010). To persist, the SP fraction
must have proliferation kinetic (e.g. PDT) equal or very similar to
the rest of the cells present in the culture. Otherwise, it will either
progressively disappear or outgrow the culture. The percentage of
the SP fraction increased in serum free media containing both PDGF
and bFGF but not in either PDGF or bFGF alone (Kondo et al., 2004).
More important, both SP cells as well as non-SP cells can repopulate
SP and non-SP cells (Fong et al., 2010; Platet et al., 2007). The SP fraction from the SK-MG-1 cell line has greater proliferative ability
(~ 10 times) than non-SP cells when growing in neurosphere media
containing EGF and FGF2 (Fukaya et al., 2010) clearly indicating that
the SP phenotype is culture condition dependent. In our model,
the SP population is simply a fraction of cells with higher stemness
phenotype. When incubated in stem cell media (that might not be
adequate for the rest of the cells) they survive and generate a cell line
with stem-like properties. In conclusion the SPM explains why cell
lines obtained from either single cell isolation (e.g. C6 cells) or mixed
primary cultures produce cell lines with a mixture of phenotypes.
|Finally, if we also assume that most cancers might originate from
a single cell that becomes a cancer cell (Martínez-Climent et al., 2006),
then models based in 2 or 3 compartments should explain not only
the origin of each cell subpopulations but also their persistence in
several cell lines. The stemloid hypothesis should explain the origin
of each of the three cancer cell types: non-GSCs, GSC and stemloids.
It is easy to imagine that a single cell may be responsible for the
initial cancer and the different phenotypes arise as the tumor grows
and create microenvironments favouring one or another phenotype.
It is important to notice that the original C6 glioma cell line was
isolated from a clonal strain (Benda et al., 1968) providing evidence
that a single original tumoral cell was indeed a stem cell able to
generate the different phenotypes (and subpopulations) present in
the C6 cell line. In 2009, Gupta et al proposed a “plasticity model”
enabling bidirectional interconvertibility between CSCs and non-CSCs
(Gupta et al., 2009). Although at a first glance it seems very similar
to ours, Gupta’s model proposes a unidirectional hierarchy from
stem cells to post-mitotic differentiated cells. This unidirectionality
predicts that, contrary to Zhengs’s experimental findings, a fraction
of isolated single cells will not have stem cell properties. In our
model, cancer cells can interconvert regardless of the phenotype
providing that a permissive environment allows the transition (Figure
2II). In summary, our hypothesis provides an explanation to several
controversial experimental data regarding the presence of GSCs in
cell lines and provides and alternative way to reconciliate the cancer
stem cell hypothesis.
|Stem cells in tumors
|The same analysis we used for cell lines can be extrapolated to
this in vivo situation. While pure symmetrical division of a primordial
stem cell predicts that any glioma in vivo should have 100% of GSCs,
pure asymmetrical division will generate a tumor with only 1 GSC
and increasing number of non-GSCs. Contrary to the situation in cell lines, the three possible models discussed above- GSCs that divides
symmetrically + differentiation; GSCs that divides asymmetrically and
symmetrically, (Figure 1, Figure1B and Figure1D) and the SPM (Figure 2) may occur in
vivo because tumors in situ are not subjected to routine passages. All
three models support the concept that a tumor can be originated from
a single cell and produce cells with different phenotypes. However,
the first two models by predicting the existence of non-GSCs, can’t
explain Zheng’s experiments unless we assume 1) that only GSCs are
able to propagate in vitro when establishing a cell line and 2) once the
culture is expanded and non-GSCs appear either by differentiation of
GSCs (Figure 1B) or spontaneous symmetrical division (Figure 1D), a
very delicate balance exists to prevent the outgrowth of the faster
cell subpopulation and maintain a low but constant fraction of GSCs.
Instead, the SPM proposes that in vivo all glioma cells have stem cell
properties but different stemness phenotype (but not true non-GSCs)
depending on the microenvironment (Figure 2III).
||Figure 1: Predicted composition of a cell culture after expansion of a single cell
according to the mode of cell division. (A) By symmetrical division a non-GSC
(green oval) will produce a culture composed of only non-GSCs. (B) A GSC
(red oval) will also produce only GSCs and, non-GSCs may arise only by differentiation
of GSCs (box) resembling asymmetrical division. (C) Asymmetrical
division of a non-GSC does not occur by defi nition. (D) Clonal expansion of a
single GSC will produce a cell culture containing increasing number of non-GSC
and only one GSC unless the GSC at some point divides symmetrically (box).
|Are all cancer cells in vivo stem cells? A key issue that need to
be addressed in order to find which of the above mentioned models
better explain what occurs in vivo is whether tumors always contain
stem cells (“constant presence”) or not (“variable presence”) and
what is the percentage (ranging between > 0 – 100%) of stem cells
in a given tumor that contains GSCs (e.g. “constant percentage”
or “variable percentage”). At least six different situations can be
account (A-F in Table S1). With few exceptions literature data
provide evidence that cancer stem cells (CSCs) have been isolated
from almost all gliomas specimens when they have searched for them
(Table 3). The few negative results could be due to technical reasons
(e.g. in some tumors, the cells cannot withstand the process required
to generate a single-cell suspension for culturing) rather than a true
lack of GSCs in the tumor. This strongly suggests that the presence
of CSCs might be constant and excludes situations D-F limiting the
possibilities to only three situations (A-C, Table S1). Consistent with
situations B or C, early studies detecting stem cell markers and/or
isolating spheres from primary cultures showed that a relatively
small percentage (no more than 25%) is probably stem cells (Al-Hajj
and Clarke, 2004; Yuan et al., 2004 ). The fact that CD133- cells can
give origin to both CD133- and CD133+ cells (Zheng et al., 2007)
and the recent finding that CD133- cells express a truncated variant
of the CD133 protein not recognized by some antibodies (Osmond
et al., 2010) makes this approach unreliable. The use of the SP or
Nestin as specific stem cell markers for isolating/detecting GSCs in
tumors are also unreliable (See above). If we consider the increasing
evidence that the non-SP fraction and CD133- cells, can generate the SP fraction and CD133+ cells respectively, the natural conclusion
is that not only the SP fraction or CD133+ cells but all glioma cells
might have stem cell properties. Therefore, at present the growing
experimental evidence favours a model where tumors have “constant
presence” and “constant percentage= 100%” (situation A in Table S1)
that fits the SPM. The two other models (Figure 1B and Figure 1D) do not
explain this situation because they predict the existence of non-stem
cells. The model in Figure 1B (without assuming that a fraction of
GSCs “differentiate” producing non-GSCs (box), by predicting that all
the progeny are the same, do not explain the different phenotypes
(degree of stemness) found in tumors. The ideal experiment to
address this key issue should be by isolating single cells from a
huge numbers of fresh glioma tumors, and by determination of the
percentage of the number of clones having stem cell properties. It
is expected that if the tumor is a mixture of GSCs and non-GSCs
a certain number of clonal lines will proliferate but they will never
have stem cell properties. The main concern is that non-GSC might
not survive the hard condition required to propagate single cells. The
number of isolated single cells that do not proliferate in vitro might
be informative to address this concern. In addition, isolated single
cells can be grown in different culture media that might promote
non-GSCs growth and overcome this limitation.
||Table 3: Isolation of cancer stem cells from human brain tumors. With few exceptions, cancer stem cells has been isolated from 100% of all tumors.
|Experimental data already provide evidence supporting the
idea that a single cell with stem cell properties can originate a
new tumor in vivo and rule out the need of two or more different
subpopulations for tumor growth. Thus, it is also likely that tumors
in vivo originate from a single cell (stem cell, stem-like cell or non-
GSC that mutated and acquired stem cell properties by reactivating
latent stem cell program (Nicolis, 2007) and all the descendants are
indeed cells having stem cell properties but very different phenotype
due to microenvironmental tumor heterogeneity such as perivascular
niches that migh promote stemness (Charles et al., 2010).
|Implications of the SP model
|As discussed above, the original hope that killing GSCs will
eradicate the tumor has been challenged by the fact that surviving
non-GSC might be able to induce tumor relapse because the Hayflick
limit is not a barrier for preventing symptomatic tumoral masses
(Hanahan and Weinberg, 2000; Withers and Lee, 2006). All models
predict that survival of a single cell might induce tumor relapse.
Since models B and D in Figure 1 predict that non-GSCs are being
generated constantly, these newly formed cells are at early stage of
the Hayflick limit (approximately 60-70 divisions) and able to divide
enough number of times to generate a tumor mass responsible for
symptoms. It is estimated that the volume of 1x109 cells (the result
of 30 divisions from a single cell) will produce a tumor volume of
approximately 1 ml (Withers and Lee, 2006) and 60-70 divisions can in
theory produce large tumoral masses (Hanahan and Weinberg, 2000).
Targeting specific cell subpopulation based on stem cell markers (e.g.
CD133+) is also unlikely to be successful because cells lacking a specific
stem cell marker can also produce (by inter conversion) descendants
with stem cell properties. From the clinical point of view, if true, the
SPM will change the strategy for finding new therapeutic strategies at
the preclinical level. Instead of isolating SP cells or CD133+ cells from
a cell line as source of stem cells to identify vulnerabilities, it might
be equally or even more informative, to analyse established glioma
cell lines or patient derived cell lines growing in different culture
conditions that generates different phenotypes resembling in vivo microenvironments such as hypoxic regions (Heddleston et al., 2009; Kim et al., 2009; Seidel et al., 2010) or perivascular niches (Charles et
al., 2010) that contributes to the maintenance of stemness. It might also be more successful to design therapies that kill all glioma cells in
such environments than to identify a potential specific stem cell killer
agent that will only kill a subpopulation of cells having a particular
|This work was supported by grants from the Swedish Research Council and
the Karolinska Institute.
- Al-Hajj M, Clarke MF (2004) Self-renewal and solid tumor stem cells. Oncogene
- Bao S, Wu Q, McLendon RE, Hao Y, Shi Q, et al. (2006) Glioma stem cells
promote radioresistance by preferential activation of the DNA damage
response. Nature 444: 756-760.
- Benda P, Lightbody J, Sato G, Levine L, Sweet W (1968) Differentiated rat glial
cell strain in tissue culture. Science 161: 370-371.
- Bexell D, Gunnarsson S, Siesjö P, Bengzon J, Darabi A (2009) CD133+ and
nestin+ tumor-initiating cells dominate in N29 and N32 experimental gliomas. Int J Cancer 125: 15-22.
- Blagosklonny MV (2007) Cancer stem cell and cancer stemloids: from biology
to therapy. Cancer Biol Ther 6: 1684-1690.
- Brehar FM, Bleotu C, Stefan LM, Buzgariu W, Chivu M, et al. (2009) Isolation
and partial characterization of a new human glioblastoma cell line. Chirurgia
(Bucur) 104: 453-461.
- Charles N, Ozawa T, Squatrito M, Bleau AM, Brennan CW, et al. (2010) Perivascular nitric oxide activates notch signaling and promotes stem-like
character in PDGF-induced glioma cells. Cell Stem Cell 6: 141-152.
- Charles NA, Holland EC (2010) TRRAP and the maintenance of stemness in
gliomas. Cell Stem Cell 6: 6-7.
- Chen R, Nishimura MC, Bumbaca SM, Kharbanda S, Forrest WF, et al. (2010) A hierarchy of self-renewing tumor-initiating cell types in glioblastoma. Cancer
Cell 17: 362-375.
- Chua C, Zaiden N, Chong KH, See SJ, Wong MC, et al. (2008) Characterization
of a side population of astrocytoma cells in response to temozolomide. J
Neurosurg 109: 856-866.
- Fong D, Yeh A, Naftalovich R, Choi TH, Chan MM (2010) Curcumin inhibits the
side population (SP) phenotype of the rat C6 glioma cell line: Towards targeting
of cancer stem cells with phytochemicals. Cancer Lett 293: 65-72.
- Fukaya R, Ohta S, Yamaguchi M, Fujii H, Kawakami Y, et al. (2010) Isolation of
cancer stem-like cells from a side population of a human glioblastoma cell line,
SK-MG-1. Cancer Lett 291: 150-157.
- Galli R, Binda E, Orfanelli U, Cipelletti B, Gritti A, et al. (2004) Isolation and
characterization of tumorigenic, stem-like neural precursors from human
glioblastoma. Cancer Res 64: 7011-7021.
- Griguer CE, Oliva CR, Gobin E, Marcorelles P, Benos DJ, et al. (2008) CD133
is a marker of bioenergetic stress in human glioma. PLoS ONE 3: e3655.
- Gupta PB, Chaffer CL, Weinberg RA (2009) Cancer stem cells: mirage or
reality? Nat Med 15: 1010-1012.
- Hanahan D, Weinberg RA (2000) The hallmarks of cancer. Cell 100: 57-70.
- Heddleston JM, Li Z, McLendon RE, Hjelmeland AB, Rich JN (2009) The
hypoxic microenvironment maintains glioblastoma stem cells and promotes
reprogramming towards a cancer stem cell phenotype. Cell Cycle 8: 3274-
- Hill RP (2006) Identifying cancer stem cells in solid tumors: case not proven. Cancer Res 66: 1891-1895.
- Ignatova TN, Kukekov VG, Laywell ED, Suslov ON, Vrionis FD, et al. (2002) Human cortical glial tumors contain neural stem-like cells expressing astroglial
and neuronal markers in vitro. Glia 39: 193-206.
- Kelly JJ, Stechishin O, Chojnacki A, Lun X, Sun B, et al. (2009) Proliferation of
human glioblastoma stem cells occurs independently of exogenous mitogens. Stem Cells 27: 1722-1733.
- Kim Y, Lin Q, Zelterman D, Yun Z (2009) Hypoxia-regulated delta-like 1 homologue enhances cancer cell stemness and tumorigenicity. Cancer Res
- Kondo T, Setoguchi T, Taga T (2004) Persistence of a small subpopulation of
cancer stem-like cells in the C6 glioma cell line. Proc Natl Acad Sci USA 101:
- Leychkis Y, Munzer SR, Richardson JL (2009) What is stemness? Stud Hist
Philos Biol Biomed Sci 40: 312-320.
- Martínez-Climent JA, Andreu EJ, Prosper F (2006) Somatic stem cells and the
origin of cancer. Clin Transl Oncol 8: 647-663.
- McBride S, Meleady P, Baird A, Dinsdale D, Clynes M (1998) Human lung
carcinoma cell line DLKP contains 3 distinct subpopulations with different
growth and attachment properties. Tumour Biol 19: 88-103.
- McCord AM, Jamal M, Williams ES, Camphausen K, Tofi lon PJ (2009) CD133+ glioblastoma stem-like cells are radiosensitive with a defective DNA damage
response compared with established cell lines. Clin Cancer Res 15: 5145-
- Nicolis SK (2007) Cancer stem cells and “stemness” genes in neuro-oncology. Neurobiol Dis 25: 217-229.
- Osmond TL, Broadley KW, McConnell MJ (2010) Glioblastoma cells negative
for the anti-CD133 antibody AC133 express a truncated variant of the CD133
protein. Int J Mol Med 25: 883-888.
- Platet N, Mayol JF, Berger F, Hérodin F, Wion D (2007) Fluctuation of the SP/
non-SP phenotype in the C6 glioma cell line. FEBS Lett 581: 1435-1440.
- Prestegarden L, Svendsen A, Wang J, Sleire L, Skaftnesmo KO, et al. (2010) Glioma cell populations grouped by different cell type markers drive brain tumor
growth. Cancer Res 70: 4274-4279.
- Qiang L, Yang Y, Ma YJ, Chen FH, Zhang LB, et al. (2009) Isolation and
characterization of cancer stem like cells in human glioblastoma cell lines. Cancer Lett 279: 13-21.
- Ropolo M, Daga A, Griffero F, Foresta M, Casartelli G, et al. (2009) Comparative
analysis of DNA repair in stem and nonstem glioma cell cultures. Mol Cancer
Res 7: 383-392.
- Rubin H (2002) The disparity between human cell senescence in vitro and
lifelong replication in vivo. Nat Biotechnol 20: 675-681.
- Seidel S, Garvalov BK, Wirta V, von Stechow L, Schänzer A, et al. (2010) A
hypoxic niche regulates glioblastoma stem cells through hypoxia inducible
factor 2 alpha. Brain 133: 983-995.
- Setoguchi T, Taga T, Kondo T (2004) Cancer stem cells persist in many cancer
cell lines. Cell Cycle 3: 414-415.
- Shen G, Shen F, Shi Z, Liu W, Hu W, et al. (2008) Identification of cancer stemlike
cells in the C6 glioma cell line and the limitation of current identification
methods. In Vitro Cell Dev Biol Anim 44: 280-289.
- Singh SK, Clarke ID, Terasaki M, Bonn VE, Hawkins C, et al. (2003) Identification
of a cancer stem cell in human brain tumors. Cancer Res 63: 5821-5828.
- Singh SK, Hawkins C, Clarke ID, Squire JA, Bayani J, et al. (2004) Identification
of human brain tumor initiating cells. Nature 432: 396-401.
- Soeda A, Park M, Lee D, Mintz A, Androutsellis-Theotokis A, et al. (2009) Hypoxia promotes expansion of the CD133-positive glioma stem cells through
activation of HIF-1alpha. Oncogene 28: 3949-3959.
- Withers HR, Lee SP (2006) Modeling growth kinetics and statistical distribution
of oligometastases. Semin Radiat Oncol 16: 111-119.
- Yakisich JS (2009) The oretical basis for the measurement of small differences
in the length of the cell cycle between two cell populations. The Open Biology
Journal 2: 95-100.
- Yi L, Zhou ZH, Ping YF, Chen JH, Yao XH (2007) Isolation and characterization
of stem cell-like precursor cells from primary human anaplastic oligoastrocytoma. Mod Pathol 20: 1061-1068.
- Yuan X, Curtin J, Xiong Y, Liu G, Waschsmann-Hogiu S, et al. (2004) Isolation
of cancer stem cells from adult glioblastoma multiforme. Oncogene 23: 9392-
- Zheng X, Shen G, Yang X, Liu W (2007) Most C6 cells are cancer stem cells:
evidence from clonal and population analyses. Cancer Res 67: 3691-3697.
- Zhou XD, Wang XY, Qu FJ, Zhong YH, Lu XD, et al. (2009) Detection of cancer
stem cells from the C6 glioma cell line. J Int Med Res 37: 503-510.