Neurotherapeutics

DOI 10.1007/s13311-017-0524-0

REVIEW

Brain Cancer Stem Cells in Adults and Children: Cell Biology

and Therapeutic Implications
Tamara J. Abou-Antoun 1 & James S. Hale 2 & Justin D. Lathia 2,3,4 &
Stephen M. Dombrowski 5

# The American Society for Experimental NeuroTherapeutics, Inc. 2017

Abstract Brain tumors represent some of the most malig- specifically at multiple molecular pathways may be more
nant cancers in both children and adults. Current treatment effective in treating adult brain tumors which seem to have
options target the majority of tumor cells but do not ade- a greater propensity towards microenvironmental interac-
quately target self-renewing cancer stem cells (CSCs). CSCs tions. Ultimately, CSC targeting approaches in combination
have been reported to resist the most aggressive radiation with current clinical therapies have the potential to be more
and chemotherapies, and give rise to recurrent, treatment- effective owing to their ability to compromise CSCs main-
resistant secondary malignancies. With advancing technolo- tenance and the mechanisms which underlie their highly
gies, we now have a better understanding of the genetic, aggressive and deadly nature.
epigenetic and molecular signatures and microenvironmental
influences which are useful in distinguishing between dis- Keywords Cancer stem cells . Childhood brain tumors .
tinctly different tumor subtypes. As a result, efforts are now Glioblastoma . Epigenetics . Microenvironment . Therapeutic
underway to identify and target CSCs within various tumor implications
subtypes based on this foundation. This review discusses
progress in CSC biology as it relates to targeted therapies
which may be uniquely different between pediatric and adult Introduction
brain tumors. Studies to date suggest that pediatric brain
tumors may benefit more from genetic and epigenetic Brain tumors are a complex collection of diseases with an
targeted therapies, while combination treatments aimed anticipated 23,770 cases per year and an associated 16,050
deaths per year [1].These malignancies account for the leading
cause of cancer death in children. Among the most malignant
and aggressive forms of these tumors are glioblastoma
multiforme (GBM; World Health Organization grade IV as-
* Stephen M. Dombrowski
dombros@ccf.org
trocytoma), characterized by increased mitotic index, necrosis
and vascular proliferation [2]. GBM survival rates have
1
School of Pharmacy, Department of Pharmaceutical Sciences,
remained largely unchanged since the 1960s, with a median
Lebanese American University, Byblos, Lebanon survival time of 12-18 months [3]. Current treatment regimens
2
Department of Cellular and Molecular Medicine, Lerner Research
are palliative in nature and involve surgical resection, ionizing
Institute, Cleveland Clinic, Cleveland, OH, USA radiation, and chemotherapy, highlighting the need for more
3
Department of Molecular Medicine, Cleveland Clinic Lerner College
effective therapies that exploit the unique biology of solid
of Medicine at Case, Western Reserve University, Cleveland, OH, tumors and their microenvironment [4].
USA Cellular heterogeneity has long been appreciated as a hall-
4
Case Comprehensive Cancer Center, Cleveland, OH, USA mark in these tumors, similar to what has been observed in the
5
Department of Neurological Surgery, Section of Pediatric
normal brain. Stem cell populations reside in many tissues
Neurosurgical Oncology, Neurological Institute, Cleveland Clinic, and are responsible for tissue development and homeostasis,
Cleveland, OH, USA giving rise to diverse cell types organized in defined cellular
 Abou-Antoun et al.

hierarchies. Traditionally, stem cells have been defined by an CSCs for functional studies, multiple cell-surface marker
ability to self-renew and differentiate along multiple lineages strategies have been used, including CD133 [15], CD49f
[5]. Within the brain, neural stem/progenitor cells (NSPCs) [16], CD36 [17], A2B5 [18], CD44 [19], L1CAM [20], and
give rise to neurons and glia [6] via the generation of interme- epidermal growth factor receptor (EGFR) [21], found mostly
diate progenitor cells that have a more restricted differentia- in adult GBM. The expression of these cell-surface markers
tion potential and serve as a transit-amplifying population be- vary within patient-derived tumors and xenograft models, and
tween NSPCs and their terminal progeny. Two main stem some of these markers have been demonstrated to also be a
reservoirs or neurogenic zones have been identified in the therapeutic target as reduction in expression has resulted in
adult brain: 1) the subventricular zone adjacent to the lateral decreased self-renewal. Several transcription factors have also
ventricle, and 2) the subgranular zone of the dentate gyrus in been identified to play pivotal functional roles in the CSC
the hippocampus [7]. Within these distinct anatomical loca- subpopulations, including BMI1 [22], Olig2 [23], and SOX2
tions, there is interaction of NSPCs with other cell types, in- [24]. In addition to altered protein expression, unique epige-
cluding endothelial cells, which regulate NSPC behavior [8]. netic patterns in the form of altered DNA methylation signa-
These zones are crucial to the maintenance of NSPCs and tures, which underlie the altered protein expression, have been
highlight the potential importance of microenvironmental reg- identified in adult GBM [25].
ulation in the stem-cell state that is also likely important in the The first CSCs to be identified in a childhood cancer were
context of brain tumors. acute myeloid leukemia stem cells [26], which were found to
express the hematopoietic stem marker CD34, but not the
Cancer Stem Cells lymphocyte differentiation marker CD38 [27]. Since this ob-
servation, multiple pediatric brain tumors have been reported
Within a tumor lays a subset of self-renewing, multi-potent to harbor CSCs, including medulloblastomas [28] and high-
cancer stem cells (CSCs) that phenotypically and functionally grade gliomas (HGGs) [29]. The identification of pediatric
resemble normal stem cells and drive tumor growth and recur- brain CSCs follows the same rationale as in adults; most re-
rence. The CSC hypothesis has been influenced by the desire ports have isolated CSCs from within bulk tumors using the
to provide a model for the development and maintenance of previously reported stem markers and verified their capacity
cellular heterogeneity and inspired by the long-standing ob- to self-renew, differentiate, and recapitulate the tumor of ori-
servations that cancer has many similarities with develop- gin. Along with expression of adult brain tumor CSC markers
ment, which has compared a tumor to an aberrantly developed (including CD133, SOX2, musashi-1, BMI1), pediatric brain
organ. By leveraging in vitro functional aspects used to define tumor CSCs also express elevated maternal embryonic leucine
and enrich NSPCs [9], and the ability to form clonal, free- zipper kinase and phosphoserine phosphatase expression [15].
floating spheres in culture, CSCs were characterized directly In addition, mouse models have been developed that can dis-
from patient-derived tumors in multiple cancer types, includ- tinguish pediatric brain tumor CSCs based on the expression
ing breast [10], colon [11], brain [12], and ovarian [13]. The of CD15 [30], Nestin [65], or Sox2 [31].
CSC hypothesis provides an additional paradigm for the de- Another important property of CSC is resistance to many
velopment of cellular heterogeneity and identifies a popula- therapeutic approaches, including radiation and chemothera-
tion of cells that continue to persist, despite aggressive thera- py. These therapeutic approaches have increased efficacy to-
pies. This model does not take into account the multiple layers wards non-stem tumor cells but do not effectively target
of oncogenic mutations necessary to initiate tumor or clonal CSCs; CSCs are often enriched in treated tumors. Current
relationships that may persist during tumor growth. therapies can also impact the tumor microenvironment and
Furthermore, the CSC hypothesis provides a model for poten- generate stresses that can induce the stem cell state, including
tial lineage relationships between tumor cells but cannot de- alterations in pH, oxygen content, or nutrient supply (Fig. 2).
finitively explain the cell(s) of origin that initiate a tumor [14]. While CSCs have been identified in pediatric and adult brain
CSC studies have relied on several functional characteris- tumors, it is important to highlight that these tumors are con-
tics to assess differences with non-stem tumor cell progeny, siderably different and therefore the CSC populations within
including sustained self-renewal, persistent proliferation, dif- them may differ from each other and may represent distinct
ferentiation potential, and an increased ability to initiate tu- targets that may be utilized therapeutically for better clinical
mors (Fig. 1). Compared with CSCs, the non-stem tumor cells outcomes (Table 1).
are generally more sensitive to conventional therapy and are
unable to recapitulate the heterogeneity of the original tumor. CSC Therapeutic Resistance
Associated characteristics such as low frequency within a tu-
mor, ability to differentiate along multiple lineages, and stem CSCs are frequently refractory to therapeutic intervention and,
cell marker expression have been observed, but, importantly, as such, are able to repropagate the tumor mass following
these are not functional properties [4]. To enrich brain tumor various treatments [32, 33]. One explanation may be that
Brain Cancer Stem Cells in Adults and Children

Fig. 1 Cancer stem cells

following treatment, radiation- and chemoresistant popula- therapeutic resistance and tumor recurrence. Ionizing radia-
tions of CSCs have been selected and enriched leading to tion is delivered in wave form (i.e., x-ray or gamma ray) and

Fig. 2 Plasticity and therapeutic

implications. CSC = cancer stem
cell
 Abou-Antoun et al.

Table 1 Brain cancer stem cell characterization in pediatric and adult patients

Cancer type Stem cell distinction Gene aberrations Epigenetic aberrations Molecular drivers

Pediatric brain cancer stem cells CD133 ACVR1 mutation H3K27M MYCN
CD49f ERBB1 amplification BMI1 Wnt/β-catenin
CD140a ATRX FOXG1 Sonic hedgehog
Nestin H3F3A mutation SOX2 NOTCH
CD15 DAXX mutation Musashi-1 PI3K/Akt/mTOR
SOX2 TP53 ATRX
NF-1 BMP1
BRAF EZH2
KRAS
PDGFRA
Adult brain cancer stem cells CD133 Amplifications or gain-of-function SOX2, NOTCH/integrin signaling
CD49f mutations in: FOXM1, FOXG1, PI3K/Akt and MAPK signaling
EGFR EGFR NANOG, STAT3, TGF-β
L1CAM PDGFRA/B GLI1, ASCLI, Wnt/β-catenin
CD44 HDM2 ZFX, ZFHX4, Sonic hedgehog
CD36 PIK3CA, and PIK3R1 HOXA10, VEGFR
A2B5 Mutations or deletions of the EZH2/BMI1 L1CAM-integrinα6
tumor suppressors: FACT
PTEN HIF2α
TP53
CDKN2A
NF1
ATRX and RB1
Mutations with favorable
outcomes:
IDH1

Pediatric and adult brain cancer stem cells with the stem markers, gene, and epigenetic aberrations, as well as molecular drivers, are listed
MYCN = N-myc proto-oncogene; PI3K = phosphoinositide 3-kinase; mTOR = mechanistic target of rapamycin; TGF = transforming growth factor;
VEGFR = vascular endothelial growth factor receptor; L1CAM = L1 cell adhesion molecule; FACT = facilitates chromosome transcription; HIF =
hypoxia-inducible factor

leads to the loss of electrons (ionization) in nucleic acids, of checkpoint kinases [33]. Through these various mecha-
proteins, and water. This results primarily in DNA damage nisms, GBM CSCs are able to survive multiple insults.
and formation of toxic free radicals from water, leading to The failure of effective clinical therapies has led to the
further damage. In response to radiation, GBM CSCs have exploration of a number of novel small molecule inhibitors
been shown to possess enhanced DNA damage responses me- of various pathways relevant in GBM CSCs; however, limited
diated primarily through the actions of poly adenosine long-term benefit in the treatment of adult GBM has been
diphosphate-ribose polymerase and ataxia telangiectasia mu- reported. Currently, agents targeting diverse pathways, includ-
tated [34]. ing phosphoinositide 3-kinase (PI3K) [37], wingless (WNT)
In addition to radiation, chemotherapeutics have been uti- [38], and NOTCH [39], in adult GBM CSCs are under clinical
lized based on their ability to modify or modulate DNA repair. evaluation.
Temozolomide (TMZ), a common oral chemotherapy drug, Aside from directly targeting tumor cells, altering the tumor
alkylates/methylates guanidine DNA residues thereby leading microenvironment has been hypothesized as a therapeutic strat-
to cell death. The majority of tumor cells, and inclusively egy. CSCs are not randomly distributed within a tumor but
CSCs, have been shown to overexpress O-6-methylguanine- present in distinct anatomical niches, which contain nutrients,
DNA methyltransferase (MGMT), which removes the meth- oxygen, and physical and soluble interactions that maintain
ylations introduced by TMZ thereby repairing damaged DNA CSC self-renewal. Multiple niches have been described in
[35]. Importantly, MGMT has been shown to mediate resis- adult GBM, including the perivascular or proliferative niche,
tance to other alkylating agents such as nitrosoureas [36]. In and the hypoxic or perinecrotic niche [40]. In addition, the re-
addition to MGMT, deficiency in mismatch repair has been lationship between CSCs and their niches is dynamic as CSCs
proposed as an additional mode of resistance in GBM cells; may actively regulate niche formation and maintenance (Fig. 3).
however, no studies have specifically evaluated this response Proximity to vascular endothelial cells has been shown to
in GBM CSCs. Augmented cell-cycle checkpoint response regulate directly CSC growth, with ablation of the vasculature
has also been observed in GBM CSCs through the activities leading to tumor regression. This has been shown to depend
Brain Cancer Stem Cells in Adults and Children

Fig. 3 Autocrine and paracrine

loops

both on direct endothelial interaction (NOTCH and integrin drug-resistance [53]. In addition, the histone chaperone com-
signaling) [13, 16, 41], as well as paracrine signaling via solu- plex facilitates chromosome transcription (FACT) was recent-
ble factors such as basic fibroblast growth factor, [42], nitric ly reported to correlate with expression of CSC markers in an
oxide [43] and sonic hedgehog (Shh) [44]. Extracellular matrix adult GBM model. FACT expression was found to correlate
is another vital component of the perivascular niche. Direct with gene transcription of stem markers SOX2, OCT4,
interaction of laminins with CSC receptor integrin alpha 6 OLIG2, and NANOG, and transcriptional knock-down of
has been shown to be vital to proliferation and migration [16]. FACT or its inhibition with a small molecule (CBL0137) re-
The hypoxic niche is not well defined structurally, charac- duced the expression of these genes [54]. The overexpression
terized by low oxygen tension and increased acidity. This of forkhead box protein M1 (FOXM1), a potent metastatic
niche primarily regulates CSC behavior through the induction inducer and important regulator of NSPCs, was also found
of transcription factors hypoxia-inducible factors 1α and 2α to be important for GBM CSCs. Interestingly, irradiation of
(HIF-1α and HIF-2α, respectively) [45]. These transcription GBM CSCs led to further upregulation of FOXM1, which
factors have been shown to regulate CSC proliferation and rendered them radioresistant in a signal transducer and activa-
tumorigenicity [46], as will be described in detail below. tor of transcription 3 (STAT3)-dependent manner [55]. STAT3
has also been described to be a key GBM CSC signaling node
[56]. Finally, several other key oncogenic and stem-cell path-
Molecular Signatures of Brain CSCs ways have been implicated in adult GBM CSC maintenance,
including c-MYC [57] and AEG-1, which facilitates β-
Great efforts have been made in an attempt to understand the catenin translocation to the nucleus and activates downstream
molecular signature of brain CSCs in both adult and pediatric targets of the Wnt pathway [58].
populations, and much of this work has focused on adult In the pediatric population, most studies have focused on
GBM and childhood medulloblastoma models [47]. medulloblastoma, which, along with other embryonal brain
In adult GBM, CSCs have been reported to express various tumors, is believed to originate from NSPCs of the ventricular
tumorigenic proteins that drive self-renewal, including zone and cerebellar external germinal layer [59]. The notion
(PI3K/Akt) and mitogen-activated protein kinase [48], that medulloblastoma contains Bstem-like^ features came
transforming growth factor-β (TGF-β) [49], the Wnt/β- about from the findings that pathways such as Wnt, Shh,
catenin pathway, and Shh signaling [50]; in vivo tumorigenic- and Notch, which govern NSPC specification, proliferation,
ity such as L1CAM [51] and integrin alpha-6 [16]; angiogenic and survival, are also aberrantly activated in such tumors,
potential through upregulation of vascular endothelial growth suggesting a molecular link between NSPCs and medulloblas-
factor (VEGF)/VEGF receptor [52]; and treatment resistance toma [60–63]. It has been reported that CD133+ cells were
through Notch [39] and TGF-β signaling pathways that have reduced almost 5-fold after inhibiting Notch signaling in me-
been shown to promote DNA repair. Furthermore, overex- dulloblastoma cells and apoptotic rates following Notch
pression of the ABC-type transporters that efflux the drugs blockade were almost 10-fold higher in primitive nestin-
out of the GBM CSCs, has also been implicated in their positive cells compared with nestin– cells, thereby suggesting
 Abou-Antoun et al.

that these medulloblastoma stem cells exhibit a particular vul- that targeting HIF-1α with digoxin resulted in increased sur-
nerability to notch signaling inhibition [47]. In addition, vival in a GBM xenograft model. At the molecular level,
MYCN has been reported to be involved in the survival and digoxin decreased HIF-1α protein expression, as well as the
propagation of the aggressive medulloblastoma stem-like cells mRNA levels of VEGF and the CD34-positive vasculature
with CD133 expression, and thus targeting MYCN may be within these tumors [71].
warranted [64]. Hypoxic influences in pediatric brain CSCs have not been
Activation of the PI3K/Akt/mechanistic target of as extensively studied as in adults. One report showed that
rapamycin (mTOR) pathway has also been reported in pedi- hypoxia inhibited p53 activation and subsequent astroglial
atric medulloblastoma nestin-expressing perivascular stem differentiation of HGG precursors. The authors report that
cells. It is believed that these CSCs are radiation resistant while HGG precursors generated endogenous bone morpho-
and are directly responsible for tumor recurrence via a p53- genetic protein (BMP) signaling leading to mitotic arrest un-
dependent cell-cycle arrest and re-entry in to the cell cycle der high oxygen tension, hypoxia actively repressed this sig-
72 h postradiotherapy [65]. Moreover, MYC amplification naling [72]. These results show a novel, mutually antagonistic
and p53 disruption in cerebellar stem cells have been shown interaction between hypoxia response and neural differentia-
to be associated with uncontrolled cell proliferation and ag- tion signals in HGG proliferation, and suggest differences
gressive tumor recurrence in an orthotopic model [66]. This between normal and HGG precursors, which may be exploited
finding implicates the possibility of transforming normal cer- for pediatric brain cancer therapy. Furthermore, the expansion
ebellar stem cells into tumorigenic cells after MYC amplifica- of medulloblastoma CSCs within the hypoxic niche has been
tion and p53 disruption, suggesting that normal stem cells may observed, further implicating the role of hypoxia in inducing
become the Btumor-initiating^ cells if primed with specific stem-like transformation of cells within pediatric brain tu-
transforming mutations. mors. These cells could be targeted via an oncolytic
engineered herpes simplex virus strategy [73].

Factors Influencing Brain CSCs Immune Evasion

Microenvironmental Influences on Brain CSCs The ability of CSCs to evade the immune system may also be
an important characteristic of specific cancer subtypes. The
CSCs do not act alone but rather are part of an active micro- exact mechanisms that make brain CSCs predominantly
environment that drives tumor propagation (Fig. 3). Key prop- non-antigenic remains unclear. Various reports have shown
erties of brain tumor CSC niches include elevated hypoxia and cancers to induce immune suppression mechanisms and deac-
the interaction with infiltrating immune cell populations. tivate key immune players such as inducing T-cell apoptosis
or inhibiting their proliferation, activation of regulatory T
Hypoxic Influences in Brain CSCs (Treg) cells and deactivation of natural killer and dendritic
cells [74, 75], in an attempt to evade the immune system.
Hypoxia is associated with necrotic regions and increases the The mechanisms by which CSCs evade immune surveil-
maintenance of GBM CSCs via a variety of mechanisms in- lance include: 1) secretion of soluble factors such as argi-
cluding HIF-1α [67] and TGF-β [68]. Hypoxic regions also nase [76] and periostin [77], both of which recruit potent
contained elevated expression of CD133, alkaline phospha- anti-inflammatory tumor-associated macrophages/microglia
tase (another stem cell marker), and correlated with shorter to suppress innate and adaptive immune responses; and 2)
overall, as well as progression-free, survival in adult patients activation of cytotoxic T-cell apoptosis by secreting
with GBM [68]. HIF-1α is a potent inducer of angiogenic galectin-3 and enhancing Treg activity via TGF-β and
factors leading to the aberrant vasculature and GBM progres- STAT3 activation. In addition, CSCs can suppress the func-
sion. HIF-1α was recently demonstrated to be activated inde- tion of immune cells by simply coming into direct contact
pendent of hypoxia via a profilin-1/von Hippel-Lindau inter- with them via the cell surface expression of the pro-
action, the targeting of which exhibited reduced tumor angio- grammed death-ligand 1[78].
genesis, normal vasculature, and improved survival in a ge- It was recently reported that the CSC population within
netically engineered GBM mouse model [69]. Other hypoxia adult GBMs co-segregated with the immune-suppressive my-
regulators have been demonstrated to be important in GBM eloid-derived suppressor cells, and that the CSCs were able to
CSCs, including von Hippel-Lindau, which interacts with in- selectively drive myeloid-derived suppressor cells-mediated
hibitor of DNA binding 2 protein increasing HIF-2α levels immune suppression via macrophage migration inhibitory
[70]. Based on their importance in hypoxia and GBM CSC factor [76]. Other reports have shown interleukin (IL)-6 to
maintenance, identifying HIF-associated signaling nodes to be highly overexpressed in the CSC population, rendering
target may reduce self-renewal. It was recently demonstrated them immunosuppressive, and also enhanced the invasive
Brain Cancer Stem Cells in Adults and Children

potential of these cells, thus playing a prominent dual role in A myriad of studies have shown that CSCs in adult brain
tumor immune evasion and invasion [79]. The therapeutic tumors have genetic alterations affiliated with tumorigenesis,
efficacy of the IL-12-expressing version of oncolytic including amplifications or gain-of-function mutations in
engineered herpes simplex virus G47 (G47-mIL12) has been EGFR, PDGFRA, HDM2, PIK3CA, and PIK3R1; mutations
found to not only specifically kill CSC, but also inhibit Tregs or deletions of the tumor suppressors PTEN, TP53, CDKN2A,
and VEGF-induced neovascularization. leading to tumor re- NF1, ATRX, and RB1 [5, 88, 89]. Such mutations have ren-
gression [80]. Finally, the interaction between tumor- dered these tumor cells more malignant, immune evasive,
associated macrophages/microglia [81], TGF-β [82], stress- therapeutically resistant, and recurrence prone. The only mu-
inducible protein 1 [83], and matrix metalloproteinases has tation found to correlate with a favorable prognosis is the
been shown to intensify tumor invasion and infiltration by isocitrate dehydrogenase 1 mutation, where long-term surviv-
promoting extracellular matrix degradation. al was evidenced in adult patients with GBM harboring such
There are limited studies investigating pediatric brain tu- mutations, indicating that these tumors represent a unique
mors and immune evasion. One report showed that class of less malignant GBM [90].
indoleamine 2,3-dioxygenase 1 (IDO1) was overexpressed Genetic heterogeneity within single-cell clones has recent-
in pediatric medulloblastoma, and that cross-talk between ly been investigated via patient-derived GBM cells. Some of
mTOR and IDO1 induced immune escape in medulloblasto- these naïve patient-derived GBM clones expressed resistance
ma cells [84]. Inhibition of mTOR potently induced IDO1 to TMZ, indicating that conventional drug-resistance is inher-
expression and activity, corroborating its ability to recruit ent in these GBM clones. PTEN, EGFR, and the constitutively
Treg cells in the tumor microenvironment, which is the mech- active EGFR deletion mutant, EGFRvIII were differentially
anism by which mTOR-targeted therapy fails. More recently, expressed in three tumors, highlighting the variability of ex-
one report showed that central nervous system primitive pression in distinct known molecular GBM drivers at the clon-
neuroectodermal tumors are capable of evading immune rec- al level. This study also showed that multiple experiments
ognition by downregulating the expression of their cell surface conducted on clones of different passages, grown in identical
MHC-I and CD1d, and by overexpressing granzyme inhibi- culture conditions consistently exhibited diverse and indepen-
tors SERPINB9, SERPINB1, and SERPINB4 [85]. Another dent variations in cellular proliferation and differentiation po-
study found that genetically downregulating Treg TGF-β sig- tential [91]. These findings further support the notion that
naling nearly abolished Treg cells and inhibited medulloblas- targeting the bulk tumor mass in hopes of eradicating the
toma progression via CD8+ cytotoxic T-lymphocyte attack tumor and achieving long-term, cancer-free survival is largely
[86]. These findings suggest that medulloblastoma cells evade limited by this clonal heterogeneity.
immune recognition possibly by upregulating TGF-β signal- Once again there has been little investigation into pe-
ing of Treg cells leading to the subsequent suppression of the diatric brain tumors for identifying specific genetic alter-
immune responses and specifically T-cell-mediated immunity. ations that give rise to brain CSCs. Several groups are
Whether these immune-evasive mechanisms are the driving now studying the underlying genetic dysregulation in pe-
forces behind pediatric brain CSCs is yet to be established, diatric brain tumors, including SHH, Wnt, and Notch
and may be a potential platform to develop specific therapeu- signaling mutations in medulloblastomas [92], ERBB1
tic targeting for pediatric brain tumors. gene amplification and ACVR1, PDGFRA, and ATRX,
PPMID, and TP53 mutations in diffuse intrinsic pontine
gliomas (DIPGs) [93], and H3F3A and DAXX mutations
Genetic and Epigenetic Influences in Brain CSCs in childhood gliomas [94]. It is not yet known whether
these or other genetic aberrations may be affiliated with a
Childhood brain tumors differ vastly from adult tumors in subset of CSCs within these tumors.
their genetic, epigenetic, and protein profiles (Fig. 4, In a recent study that employed RCAS/TVA system to
Table 1). For instance, epigenetic regulation in pediatric brain induce platelet-derived growth factor (PDGF)-B overexpres-
tumors is more apparent than in adult tumors, whereas envi- sion, p53 loss, and histone 3.3 lysine to methionine mutation
ronmental and microenvironmental influences exert a greater (H3.3K27M) researchers were able to genetically engineer a
impact on adult brain tumors. If a child develops a brain tumor model of pediatric DIPG upon exposure to ectopic PDGF-B
before the age of 2 years, it is likely a result of genetic and/or ligand and p53-deficiency along with H3.3K27M overexpres-
epigenetic alterations that have induced tumorigenic transfor- sion [95]. This model could serve as a valuable tool to inves-
mation in certain cells within the developing brain, and not tigate experimentally the cell of origin and stem-cell perpetra-
due to long-term carcinogen exposure that may be mutagenic tor in pediatric HGGs. Further studies utilizing this model
in oncogenes or tumor suppressor genes [87]. Cancer arises would shed light on some of the genetic alterations that may
from mutations in tumor promoters and tumor suppressors, be the driving force of pediatric brain CSC propagation and
and this mutational background also applies to CSCs. maintenance.
 Abou-Antoun et al.

Fig. 4 Age-related tumor

frequency. PDGFRA = platelet-
derived growth factor receptor A;
ALT = alanine transaminase;
EGFR = endothelial growth factor
receptor; IDH1 = isocitrate dehy-
drogenase 1; TERT = telomerase
reverse transcriptase; NF1 =
Neurofibromatosis type 1

Epigenetic Influences involved in cell proliferation and cell–cell interaction (EMP3),

angiogenesis (PCDH-gamma-A11), cell-cycle regulation
Epigenetic regulation, which refers to regulation of gene ex- (CDK2A-p16INK4a and CDK2B-p15INK4b), inhibition of ap-
pression independent of genetic mutations, is now thought to optosis (DAPK1, TIMP3, CDH1), and drug resistance (O6-
be involved in pediatric and adult brain tumor CSCs. Cellular MGMT) [109] underlines the importance of epigenetic profiling
hierarchy in both normal [96] and neoplastic tissue is regulat- of various tumors, and further highlights the potential of targeting
ed by epigenetic mechanisms [25], including DNA methyla- epigenetic regulators to reach more effective therapies.
tion, chromatin remodeling through histone methylation, and Recent efforts have compared the epigenetic landscape of
regulatory noncoding RNAs [97]. This involves opening (me- GBM CSCs and differentiated cells and identified a set of 4
diated by methylation of H3K4) and closing (mediated by transcription factors (POU3F2, SOX2, SALL2, and OLIG2)
H3K27 methylation) of the chromatin, which is associated that are capable of inducing the differentiated GBM cells to
with gene activation and silencing, respectively. undergo stem-like transformation into tumor-propagating
Key chromatin modifiers such as mixed-lineage leukemia cells in vivo [110]. This further highlight the dynamic plastic-
1 have been demonstrated to be important in adult GBM ity that can be transcriptionally regulated within these malig-
CSCs via hypoxia-mediated HIF2α-induction [98] and acti- nant tumors, enabling their ability to transition between dif-
vation of the homeobox gene HOXA10 [99]. The polycomb ferentiated and undifferentiated states based on various exter-
genes, EZH2 and BMI1, are believed to drive this transcrip- nal or internal challenges.
tional repression by histone methylation and reports have The epigenetic regulators that influence childhood malig-
demonstrated that EZH2 silencing of the BMP pathway in- nancies have mainly been reported in DIPG. Recent studies
hibits GBM CSC differentiation, and that self-renewal capac- found mutations in H3F3A or HIST1H3B, which encode his-
ity and tumorigenicity is also lost with inhibition of EZH2 or tone variant H3.3 or H3.1, respectively, resulting in the re-
forced expression of the BMP pathway in GSCs [100]. placement of lysine residue at position 27 with methionine
Other epigenetic regulators, including SOX2 [101], FOXM1 (K27M) or the glycine residue at position 34 with arginine
[102], FOXG1 [103], NANOG [104], STAT3 [105], GLI1 [44], or valine (G34R/V) [111]. One group found that H3F3A
ASCLI [106], ZFX [107], and ZFHX4 [108], have been reported K27M mutant GBMs show significant decreases in overall
to play crucial roles in the maintenance and self-renewal potential H3K27me3 without significant changes in EZH2 expression
of adult GBM CSCs. Epigenetic silencing of glioblastoma genes [112]. Furthermore, using human embryonic stem cells to
Brain Cancer Stem Cells in Adults and Children

model pediatric gliomas that harbor the H3.3K27M histone Therapeutic Challenges in Brain CSCs
mutation, it was observed that H3.3K27M expression
synergized with loss of the tumor suppressor p53 and activa- Malignant brain tumors remain a challenge to treat for a vari-
tion of PDGFRA in NSPCs, which led to neoplastic transfor- ety of factors, including the interdependence of microenviron-
mation [113]. Another report found other mutations that target mental, genetic, and epigenetic factors that drive the CSC
the receptor tyrosine kinase–RAS–PI3K signaling, and cell- state, as discussed above. While a number of new small-
cycle regulation in 68% and 59% of pediatric DIPGs and the molecule inhibitors of receptor tyrosine kinases, anti-
non-brainstem HGGs, respectively. In addition, the recurrent angiogenic factors, antiproliferative, and proapoptotic agents
somatic mutation, ACVR1, which encodes for the bone such as PDGFRα/β, VEGFR, EGFR, PI3K, and mTOR [117]
BMP1, along with the frequent somatic mutations in histone are being evaluated in adult GBM, these have not resulted in
H3 genes, TP53 and ATRX, have been reported in both DIPGs significant improvements in the progression-free or overall
and non-brainstem HGGs [114]. Finally, by comparison < 3% survival rates. The same can be said for studies that have
of pediatric HGGs harbor the telomerase reverse transcriptase explored the use of combination therapies that use chemother-
promoter mutations, while 86% occur in adult GBM [115]. apeutic agents and/or radiotherapy along with inhibitors of
This diversity between childhood and adult brain tumors as receptor tyrosine kinases, histone deacetylases, mTOR,
well as the intratumoral diversity emphasize the importance of DNA topoisomerases, integrins, or immune modulators.
devising targeted and personalized therapies. These failures may be due, in part, to the inability to effective-
Recent reports have found that polycomb transcription fac- ly target CSCs.
tor BMI1 to be highly correlated with Shh ligand concentra- CSC-targeting strategies have shown some promise as a
tions in medulloblastoma CSCs, indicating that Shh signaling recent study has reported a 2.9-fold increase in progression-
may play a pivotal role in BMI1 expression. Furthermore, it free survival with a vaccine strategy using autologous CSCs
was determined that downstream effectors of BMI1 may be with mRNA-transfected dendritic cells in patients with GBM
contributing to the activation of Shh, thus highlighting the im- [118]. Studies such as these suggest that CSC targeting strate-
portance of this reciprocal communication on the maintenance gies may be more effective than conventional therapies and
of medulloblastoma stem cell subpopulation [22]. In medullo- thus warrants larger-scale investigation. However, future stud-
blastoma groups 3 and 4 (both non-Shh/Wnt), BMI1 and ies are likely to benefit from additional considerations that drive
FOXG1, genes known to be associated with self-renewal and therapeutic failure, including redundant signaling of overlap-
proliferation [116], are overexpressed in CSCs. BMI1 has been ping pathways involved in CSC growth/survival mechanisms.
very well characterized as a major epigenetic regulator of brain Moreover, fluid transport and retention mechanisms both at the
tumor CSC therapy resistance and self-renewal capacity in both brain–vascular (i.e., blood–brain barrier) and cellular mem-
adult and pediatric populations, and as such warrants further brane level (i.e., drug-efflux protein pumps) may also contrib-
investigation as a potential therapeutic target in these cancers. ute, in part, to drug-resistance in brain tumors. Brain tumor

Fig. 5 Resistance to cytotoxic

DNA damaging agents
 Abou-Antoun et al.

CSCs specifically are thought to be drug resistant owing to Another consideration is specific self-renewal pathway
upregulation of proteins involved in active drug efflux [119], targeting, which is currently being explored in a variety of
thus sparing CSCs from cytotoxicity and apoptosis. Other fac- tumors. Such pathways that may be considered include Wnt/
tors such as hypoxic areas of tumor cells [46], direct cell–cell β-catenin [121], Notch [122], SHH [123], EGFR, and STAT3.
communication, local secretion of the cytokines IL-6 or stromal It will be important to assess how targeting these pathways
cell-derived factor 1, DNA damage repair [33], and impacts other neural cell types, including NSPCs, as well as
microRNAs [120] are also reasons of drug resistance in potential resistance mechanisms that may emerge.
GBM. Another major consideration is the inherent plasticity As discussed above, the stem-cell state can be induced via
of CSCs, transitioning between stem and differentiated cell stress present in the microenvironment, including hypoxia,
states, as well as the rise of new CSCs from the differentiated lower pH, or metabolic stress. The advances that have been
population, places additional challenges in developing effective made to identify molecular mechanisms that drive the stem-
therapies (Fig. 2). Multimodality approaches that target growth cell state could be leverage for therapeutic development.
factors, tumorigenic pathways, epigenetic, and microenviron- Successful strategies may prevent the cellular stress response
mental factors that are responsible for CSC plasticity should be and not only target the stem cell state, but also sensitize cells to
considered. These considerations are relevant for both adult these stresses.
and pediatric tumors and leveraging the epigenetic state may
be especially effective in the pediatric setting as these tumors
rely on epigenetic regulation and cannot be as aggressively Conclusions/Summary
treated with surgery, radiation, and chemotherapy as in adult
brain tumors. Current available treatments have been shown to slow pro-
gression, but most often fail to eradicate brain tumors. It is
likely that these treatments effectively kill many tumor cells
but do not effectively target the highly malignant CSCs that
Implications for Future Developments in Brain CSC adapt rapidly to give rise to recurrent, treatment-resistant ma-
Therapeutics lignancies. Our evolving understanding of the genetic, epige-
netic, and molecular signatures and microenvironmental influ-
As a challenge to treat tumors effectively involves therapeutic ences that may be unique to CSCs will enable us to develop
resistance via the integration of microenvironmental, genetic, more effective multimodal therapies for a variety of distinct
and epigenetic factors that converge on the stem-cell state, the tumor subtypes based on these characterizations. These
development of CSC targeting strategies remains a priority for targeted therapies may likely be different between pediatric
future efforts. Considerations for therapeutic development and adult patients with brain tumors based on the genetic,
should include neutralizing the stem-cell phenotype, self- epigenetic, molecular signatures, and microenvironmental in-
renewal pathways, and transitions into the stem cell state. fluences that drive these cancers.
These therapies may take several forms, including small-
molecule inhibitors, natural products and/or diet modification, Required Author Forms Disclosure forms provided by the authors are
available with the online version of this article.
or viral delivery, but each strategy should take into consider-
ation the blood–brain barrier and achieving effective tumor
penetrance.
Embedded within the stem cell phenotype is the ability to References
self-renew and resist therapies via resistance to redox stress,
1. Ostrom QT, Gittleman H, de Blank PM et al. American Brain
efficient DNA repair capacity, metabolic reprogramming,
Tumor Association. Adolescent and young adult primary brain
ability to withstand hyponutritious and hyperinflammatory and central nervous system tumors diagnosed in the United
conditions, and the ability to expel anticancer drugs by upreg- States in 2008–2012. Neuro Oncol 2016; Suppl 1:i1-i50.
ulation of ABC drug efflux transporters (Fig. 5). Confounding 2. Vigneswaran K, Neill S, Hadjipanayis CG. Beyond the World
targeting efforts is the relatively quiescent nature to some CSC Health Organization grading of infiltrating gliomas: advances in
the molecular genetics of glioma classification. Ann Transl Med
populations, which provides an additional challenge to target 2015;3:95.
if the strategy is predicated on proliferation to generate an 3. Jelsma R, Bucy PC. The treatment of glioblastoma multiforme of
effect. An emerging consideration is the interaction between the brain. J Neurosurg 1967;27(5):388-400.
CSCs and the immune system. Immuno-oncology efforts for 4. Weller M, Cloughesy T, Perry JR, Wick W. Standards of care for
treating brain tumors would benefit from the integration of treatment of recurrent glioblastoma—are we there yet? Neuro
Oncol 2013;15:4-27.
CSCs into their models as there may be an opportunity to 5. Lathia JD, Mack SC, Mulkearns-Hubert EE, Valentim CLL, Rich
generate a more sustained immune response by concomitant JN. Cancer stem cells in glioblastoma. Genes Develop 2015;29:
targeting of CSCs. 1203-1217.
Brain Cancer Stem Cells in Adults and Children

6. Cavallucci V, Fidaleo M, Pani G. Neural stem cells and nutrients: 27. Bonnet D, Dick JE. Human acute myeloid leukemia is organized
poised between quiescence and exhaustion. Trends Endocrinol as a hierarchy that originates from a primitive hematopoietic cell.
Metab 2016; 27:756-769. Nat Med 1997; 3:730-737.
7. Rusznák Z, Henskens W, Schofield E, Kim WS, Fu Y. Adult 28. Singh SK, Clarke ID, Terasaki M et al. Identification of a cancer
neurogenesis and gliogenesis: possible mechanisms for stem cell in human brain tumors. Cancer Res 2003; 63: 5821-
neurorestoration. Exp Neurobiol 2016;25:103-112. 5828.
8. Hale JS, Sinyuk M, Rich JN, Lathia JD. Decoding the cancer stem 29. Singh SK, Clarke ID, Hide T, Dirks PB. Cancer stem cells in
cell hypothesis in glioblastoma. CNS Oncol 2013;2:10.2217/cns. nervous system tumors. Oncogene 2004; 23: 7267-7273.
13.23. 30. Read T-A, Fogarty MP, Markant SL, et al. Identification of CD15
9. Reynolds BA, Weiss S. Clonal and population analyses demon- as a marker for tumor-propagating cells in a mouse model of
strate that an EGF-responsive mammalian embryonic CNS pre- medulloblastoma. Cancer Cell 2009;15:135-147.
cursor is a stem cell. Dev Biol 1996;175(1):1-13. 31. Vanner RJ, Remke M, Gallo M, et al. Quiescent Sox2+ cells drive
10. Al-Hajj M, Wicha MS, Benito-Hernandez A, Morrison SJ, Clarke hierarchical growth and relapse in sonic hedgehog subgroup me-
MF. Prospective identification of tumorigenic breast cancer cells. dulloblastoma. Cancer Cell 2014; 26: 33-47.
Proc Natl Acad Sci U S A 2003;100:3983-3988. 32. Chen J, Li Y, Yu T-S, et al. A restricted cell population propagates
11. Ricci-Vitiani L, Lombardi DG, Pilozzi E et al. Identification and glioblastoma growth following chemotherapy. Nature 2012;488:
expansion of human colon-cancer-initiating cells. Nature 522-526.
2007;445:111-115. 33. Bao S, Wu Q, McLendon RE, et al. Glioma stem cells promote
12. Ignatova TN, Kukekov VG, Laywell ED, Suslov ON, Vrionis FD, radioresistance by preferential activation of the DNA damage re-
Steindler DA. Human cortical glial tumors contain neural stem- sponse. Nature 2006;444:756-760.
like cells expressing astroglial and neuronal markers in vitro. Glia 34. Carruthers R, Ahmed SU, Strathdee K. et al. Abrogation of
2002;39(3):193-206. radioresistance in glioblastoma stem-like cells by inhibition of
13. Bapat SA, Mali AM, Koppikar CB, Kurrey NK. Stem and ATM kinase. Mol Oncol 2015;9:192-203.
progenitor-like cells contribute to the aggressive behavior of hu- 35. Beier D, Röhrl S, Pillai DR. et al. Temozolomide preferentially
man epithelial ovarian cancer. Cancer Res 2005;65:3025-3029. depletes cancer stem cells in glioblastoma. Cancer Res 2008;68:
14. Lathia JD. Cancer stem cells: moving past the controversy. CNS 5706-5715.
Oncol 2013;2(6):465-467. 36. Passagne I, Evrard A, Depeille P, Cuq P, Cupissol D, Vian L. O(6)-
15. Hemmati HD, Nakano I, Lazareff JA, et al. Cancerous stem cells methylguanine DNA-methyltransferase (MGMT) overexpression
can arise from pediatric brain tumors. Proc Natl Acad Sci U S A in melanoma cells induces resistance to nitrosoureas and temozo-
2003; 100:15178-15183. lomide but sensitizes to mitomycin C. Toxicol Appl Pharmacol
16. Lathia JD, Gallagher J, Heddleston JM, et al. Integrin alpha 6 2006;211:97-105.
regulates glioblastoma stem cells. Cell Stem Cell 2010;6:421-432. 37. Firat E, Niedermann G. FoxO proteins or loss of functional p53
17. Hale JS, Otvos B, Sinyuk M, et al. Cancer stem cell-specific scav- maintain stemness of glioblastoma stem cells and survival after
enger receptor CD36 drives glioblastoma progression. Stem Cells ionizing radiation plus PI3K/mTOR inhibition. Oncotarget 2016;
2014;32:1746-1758. 7: 54883-54896.
18. Ogden AT, Waziri AE, Lochhead RA et al. Identification of 38. Duan R, Han L, Wang Q, et al. HOXA13 is a potential GBM
A2B5+CD133- tumor-initiating cells in adult human gliomas. diagnostic marker and promotes glioma invasion by activating
Neurosurgery 2008;62:505-514. the Wnt and TGF-β pathways. Oncotarget 2015;6:27778-27793.
19. Liu G, Yuan X, Zeng Z, et al. Analysis of gene expression and 39. Yahyanejad S, King H, Iglesias VS et al. NOTCH blockade com-
chemoresistance of CD133+ cancer stem cells in glioblastoma. bined with radiation therapy and temozolomide prolongs survival
Mol Cancer 2006;5:67. of orthotopic glioblastoma. Oncotarget 2016;7:41251-41264.
20. Bao S, Wu Q, Li Z, et al. Targeting Cancer Stem Cells through 40. Codrici E, Enciu AM, Popescu ID, Mihai S, Tanase C. Glioma
L1CAM Suppresses glioma growth. Cancer Res 2008;68:6043- stem cells and their microenvironments: providers of challenging
6048. therapeutic targets. Stem Cells Int 2016; 5728438.
21. Mazzoleni S, Politi LS, Pala M, et al. Epidermal growth factor 41. Zhu TS, Costello MA, Talsma CE, et al. Endothelial cells create a
receptor expression identifies functionally and molecularly dis- stem cell niche in glioblastoma by providing Notch ligands that
tinct tumor-initiating cells in human glioblastoma multiforme nurture self-renewal of cancer stem-like cells. Cancer Res
and is required for gliomagenesis. Cancer Res 2010;70:7500- 2011;71:6061-6072.
7513. 42. Fessler E, Borovski T, Medema JP. Endothelial cells induce cancer
22. Wang X, Venugopal C, Manoranjan B, et al. Sonic hedgehog stem cell features in differentiated glioblastoma cells via bFGF.
regulates Bmi1 in human medulloblastoma brain tumor-initiating Mol Cancer 2015;14:157.
cells. Oncogene 2012;31:187-199. 43. Charles N, Ozawa T, Squatrito M, et al. Perivascular nitric oxide
23. Ligon KL, Huillard E, Mehta S, et al. Olig2-regulated lineage- activates Notch signaling and promotes stem-like character in
restricted pathway controls replication competence in neural stem PDGF-induced glioma cells. Cell Stem Cell 2010;6.
cells and malignant glioma. Neuron 2007;53:503-517. 44. Clement V, Sanchez P, de Tribolet N, Radovanovic I, Ruiz i Altaba
24. Berezovsky AD, Poisson LM, Cherba D, et al. Sox2 promotes A. HEDGEHOG-GLI1 signaling regulates human glioma growth,
malignancy in glioblastoma by regulating plasticity and astrocytic cancer stem cell self-renewal, and tumorigenicity. Curr Biol
differentiation. Neoplasia 2014;16:193-206. 2007;17:165-172.
25. Carén H, Pollard SM, Beck S. The good, the bad and the ugly: 45. Yang L, Lin C, Wang L, Guo H, Wang X. Hypoxia and hypoxia-
epigenetic mechanisms in glioblastoma. Mol Aspects Med inducible factors in glioblastoma multiforme progression and ther-
2013;34:849-862. apeutic implications. Exp Cell Res 2012;318:2417-2426.
26. Brian J. P. Huntly and D. Gary Gilliland. Leukaemia stem cells 46. Li Z, Bao S, Wu Q, et al. Hypoxia-inducible factors regulate tu-
and the evolution of cancer-stem-cell research. Nat Rev Cancer morigenic capacity of glioma stem cells. Cancer Cell 2009;15:
2005; 5:311-321. 501-513.
 Abou-Antoun et al.

47. Xing Fan, William Matsui, Leila Khaki, et al. Notch pathway glioblastoma stem-like cells in hypoxic niches. Cell Rep 2016;
inhibition depletes stem-like cells and blocks engraftment in em- 15:2500-2509.
bryonal brain tumors. Cancer Res 2006;66:7445-7452. 68. Iwadate Y, Matsutani T, Hirono S, Shinozaki N, Saeki N.
48. Lee J, Kotliarova S, Kotliarov Y et al. Tumor stem cells derived Transforming growth factor-β and stem cell markers are highly
from glioblastomas cultured in bFGF and EGF more closely mir- expressed around necrotic areas in glioblastoma. J Neurooncol
ror the phenotype and genotype of primary tumors than do serum- 2016;129:101-107.
cultured cell lines. Cancer Cell 2006;9:391-403. 69. Fan Y, Potdar AA, Gong Y, et al. Profilin-1 phosphorylation di-
49. Ghosh D, Ulasov IV, Chen L. et al. TGFβ-responsive HMOX1 rects angiocrine expression and glioblastoma progression through
expression is associated with stemness and invasion in glioblasto- HIF-1α accumulation. Nat Cell Biol 2014;16:445-456.
ma multiforme. Stem Cells 2016;34:2276-2289. 70. Lee SB, Frattin V, Bansal M, et al. An ID2-dependent mechanism
50. Jin X, Jeon HM, Jin X et al. The ID1-CULLIN3 axis regulates for VHL inactivation in cancer. Nature 2016;529:172-177.
intracellular SHH and WNT signaling in glioblastoma stem cells. 71. Nigim F, Cavanaugh J, Patel AP, et al. targeting hypoxia-inducible
Cell Rep 2016;16:1629-1641. factor 1α in a new orthotopic model of glioblastoma recapitulating
51. Cheng L, Wu Q, Guryanova OA, et al. Elevated invasive potential the hypoxic tumor microenvironment. J Neuropathol Exp Neurol
of glioblastoma stem cells. Biochem Biophys Res Commun 2015;74:710-722.
2011;406:643-648. 72. Pistollato F, Chen HL, Rood BR, et al. Hypoxia and HIF1alpha
52. Bao S, Wu Q, Sathornsumetee S. et al. Stem cell-like glioma cells repress the differentiative effects of BMPs in high-grade glioma.
promote tumor angiogenesis through vascular endothelial growth Stem Cells 2009;27:7-17.
factor. Cancer Res 2006;66:7843-7848. 73. Friedman GK, Moore BP, Nan L, et al. Pediatric medulloblastoma
53. Xi G, Hayes E, Lewis R, et al. CD133 and DNA-PK regulate xenografts including molecular subgroup 3 and CD133+ and
MDR1 via the PI3K- or Akt-NF-κB pathway in multidrug- CD15+ cells are sensitive to killing by oncolytic herpes simplex
resistant glioblastoma cells in vitro. Oncogene 2016;35:241-250. viruses. Neuro Oncol 2016;18:227-235.
54. Dermawan JK, Hitomi M, Silver DJ, et al. Pharmacological 74. Huettner C, Czub S, Kerkau S, Roggendorf W, Tonn JC.
targeting of the histone chaperone complex FACT preferentially Interleukin 10 is expressed in human gliomas in vivo and in-
eliminates glioblastoma stem cells and prolongs survival in pre- creases glioma cell proliferation and motility in vitro. Anticancer
clinical models. Cancer Res 2016;76:2432-2442. Res 1997;17:3217-3224.
55. Maachani UB, Shankavaram U, Kramp T, Tofilon PJ, 75. Jackson C, Ruzevick J, Phallen J, Belcaid Z, Lim M. Challenges in
Camphausen K, Tandle AT. FOXM1 and STAT3 interaction con- immunotherapy presented by the glioblastoma multiforme micro-
fers radioresistance in glioblastoma cells. Oncotarget 2016;7: environment. Clin Dev Immunol 2011;732413.
77365-77377. 76. Otvos B, Silver DJ, Mulkearns-Hubert EE, et al. Cancer stem cell-
56. Liu M, Inoue K, Leng T, Guo S, Xiong Z. TRPM7 channels secreted macrophage migration inhibitory factor stimulates mye-
regulate glioma stem cell through STAT3 and Notch signaling loid derived suppressor cell function and facilitates glioblastoma
pathways. Cell Signal 2014;26:2773-2781. immune evasion. Stem Cells 2016; 34: 2026-2039.
57. Fang X, Zhou W, Wu Q, et al. Deubiquitinase USP13 maintains 77. Zhou W, Ke SQ, Huang Z, et al. Periostin secreted by glioblasto-
glioblastoma stem cells by antagonizing FBXL14-mediated Myc ma stem cells recruits M2 tumor-associated macrophages and pro-
ubiquitination. J Exp Med 2017;214:245-267. motes malignant growth. Nat Cell Biol 2015;17:170-182.
58. Hu B, Emdad L, Kegelman TP, et al. Astrocyte elevated gene-1 78. Silver DJ, Sinyuk M, Vogelbaum MA, Ahluwalia MS, Lathia JD.
regulates beta-catenin signaling to maintain glioma stem-like The intersection of cancer, cancer stem cells, and the immune
stemness and self-renewal. Mol Cancer Res 2017;15:225-233. system: therapeutic opportunities. Neuro Oncol 2016; 18: 153-
59. Oliver TG, Wechsler-Reya RJ. Getting at the root and stem of 159.
brain tumors. Neuron 2004; 42: 885-888. 79. Qiu B, Zhang D, Wang Y, et al. Interleukin-6 is overexpressed and
60. Raffel C, Jenkins RB, Frederick L, et al. Sporadic medulloblasto- augments invasiveness of human glioma stem cells in vitro. Clin
mas contain PTCH mutations. Cancer Res 1997; 57: 842-845. Exp Metastasis 2013;30:1009-1018.
61. Reifenberger J, Wolter M, Weber RG, et al. Missense mutations in 80. Cheema TA, Fecci PE, Ning J, Rabkin SD. Immunovirotherapy
SMOH in sporadic basal cell carcinomas of the skin and primitive for the treatment of glioblastoma. Oncoimmunology 2014;3:
neuroectodermal tumors of the central nervous system. Cancer e27218.
Res 1998; 58: 1798-1803. 81. Markovic DS, Vinnakota K, Chirasani S, et al. Gliomas induce
62. Zurawel RH, Chiappa SA, Allen C, Raffel C. Sporadic medullo- and exploit microglial MT1-MMP expression for tumor expan-
blastomas contain oncogenic β-catenin mutations. Cancer Res sion. Proc Natl Acad Sci U S A 2009;106:12530-12535.
1998; 58: 896-899. 82. Ye XZ, Xu SL, Xin YH, et al. Tumor-associated microglia/
63. Hallahan AR, Pritchard JI, Hansen S, et al. The SmoA1 mouse macrophages enhance the invasion of glioma stem-like cells via
model reveals that notch signaling is critical for the growth and TGF-beta1 signaling pathway. J Immunol 2012;189:444-453.
survival of sonic hedgehog-induced medulloblastomas. Cancer 83. Fonseca AC, Romao L, Amaral RF, et al. Microglial stress induc-
Res 2004; 64: 7794-7800. ible protein 1 promotes proliferation and migration in human glio-
64. Ahmad Z, Jasnos L, Gil V, et al. Molecular and in vivo character- blastoma cells. Neuroscience 2012;200:130-141.
ization of cancer-propagating cells derived from MYCN- 84. Folgiero V, Miele E, Carai A, et al. IDO1 involvement in mTOR
dependent medulloblastoma. PLOS ONE 2015;10(3):e0119834. pathway: A molecular mechanism of resistance to mTOR
65. Dolores Hambardzumyan, Oren J. Becher, et al. PI3K pathway targeting in medulloblastoma. Oncotarget 2016; 7: 52900-52911.
regulates survival of cancer stem cells residing in the perivascular 85. Vermeulen JF, van Hecke W, Spliet WG, et al. Pediatric primitive
niche following radiation in medulloblastoma in vivo. Genes Dev neuroectodermal tumors of the central nervous system differen-
2008;22:436-448. tially express granzyme inhibitors. PLOS ONE 2016;11:
66. Pei Y, Moore CE, Wang J, et al. An animal model of Myc-driven e0151465.
medduloblastoma. Cancer Cell 2012; 21: 155-167. 86. Gate D, Danielpour M, Rodriguez J Jr, et al. T-cell TGF-β signal-
67. Mineo M, Ricklefs F, Rooj AK, et al. The Long Non-coding RNA ing abrogation restricts medulloblastoma progression. Proc Natl
HIF1A-AS2 facilitates the maintenance of mesenchymal Acad Sci U S A 2014;111:E3458-E3466.
Brain Cancer Stem Cells in Adults and Children

87. Kunkle B, Yoo C, Roy D. Discovering gene-environment interac- 107. Fang X, Huang Z, Zhou W, et al. The zinc finger transcription
tions in glioblastoma through a comprehensive data integration factor ZFX is required for maintaining the tumorigenic potential
bioinformatics method. Neurotoxicology 2013;35:1-14. of glioblastoma stem cells. Stem Cells 2014;32:2033-2047.
88. Parsons DW, Jones S, Zhang X, et al. An integrated genomic 108. Chudnovsky Y, Kim D, Zheng S, et al. ZFHX4 interacts with the
analysis of human glioblastoma multiforme. Science 2008;321: NuRD core member CHD4 and regulates the glioblastoma tumor-
1807-1812. initiating cell state. Cell Rep 2014;6:313-324.
89. Cancer Genome Atlas Research Network. Comprehensive geno- 109. Liu A, Hou C, Chen H, Zong X, Zong P. Genetics and epigenetics
mic characterization defines human glioblastoma genes and core of glioblastoma: applications and overall incidence of IDH1 mu-
pathways. Nature 2008;455:1061-1068. tation. Front Oncol 2016;6:16.
90. Weller M, Pfister SM, Wick W, Hegi ME, Reifenberger G, Stupp 110. Suvà ML, Rheinbay E, Gillespie SM, et al. Reconstructing and
R. Molecular neuro-oncology in clinical practice: a new horizon. reprogramming the tumor-propagating potential of glioblastoma
Lancet Oncol 2013;14: e370-e379. stem-like cells. Cell 2014;157:580-594.
91. Meyera M, Reimandb J, Lan X, et al. Single cell-derived clonal 111. Wu G, Broniscer A, McEachron TA, et al.; St. Jude Children’s
analysis of human glioblastoma links functional and genomic het- Research Hospital-Washington University Pediatric Cancer
erogeneity. PNAS 2015;112: 851-856. Genome Project. Somatic histone H3 alterations in pediatric dif-
92. Remke M, Ramaswamy V, Taylor MD. Medulloblastoma molec- fuse intrinsic pontine gliomas and non-brainstem glioblastomas.
ular dissection: the way toward targeted therapy. Curr Opin Oncol Nat Genet 2012;44:251-253.
2013;25:674-681. 112. Venneti S, Garimella MT, Sullivan LM, et al. Evaluation of histone
93. Jones C, Baker SJ. Unique genetic and epigenetic mechanisms 3 lysine 27 trimethylation (H3K27me3) and enhancer of Zest 2
driving paediatric diffuse high-grade glioma. Nat Rev Cancer (EZH2) in pediatric glial and glioneuronal tumors shows de-
2014;14:10. creased H3K27me3 in H3F3A K27M mutant glioblastomas.
94. Appin CL, Brat DJ. Molecular genetics of gliomas. Cancer J Brain Pathol 2013;23:558-564.
2014;20:66-72. 113. Funato K, Major T, Lewis PW, Allis CD, Tabar V. Use of human
embryonic stem cells to model pediatric gliomas with H3.3K27M
95. Misuraca KL, Hu G, Barton KL, Chung A, Becher OJ. A Novel
histone mutation. Science 2014;346:1529-1533.
mouse model of diffuse intrinsic pontine glioma initiated in Pax3-
114. Wu G, Diaz AK, Paugh BS, et al. St. Jude Children's Research
expressing cells. Neoplasia 2016;18:60-70.
Hospital-Washington University Pediatric Cancer Genome
96. Smith ZD, Meissner A. DNA methylation: roles in mammalian
Project. The genomic landscape of diffuse intrinsic pontine glioma
development. Nat Rev Genet 2013;14:204-220.
and pediatric non-brainstem high-grade glioma. Nat Genet
97. Heddleston JM, Hitomi M, Venere M, et al. Glioma stem cell
2014;46:444-450.
maintenance: the role of the microenvironment. Curr Pharm Des
115. Killela PJ, Reitman ZJ, Jiao Y, et al. TERT promoter mutations
2011;17:2386-2401.
occur frequently in gliomas and a subset of tumors derived from
98. Heddleston JM, Wu Q, Rivera M, et al. Hypoxia-induced mixed- cells with low rates of self-renewal. Proc Natl Acad Sci U S A
lineage leukemia 1 regulates glioma stem cell tumorigenic poten- 2013;110:6021-6026.
tial. Cell Death Differ 2012;19:428-439. 116. Manoranjan B, Wang X, Hallett RM, et al. FoxG1 interacts with
99. Gallo M, Ho J, Coutinho FJ, et al. A tumorigenic MLL-homeobox Bmi1 to regulate self-renewal and tumorigenicity of medulloblas-
network in human glioblastoma stem cells. Cancer Res 2013;73: toma stem cells. Stem Cells 2013;31:1266-1277.
417-427. 117. Takeuchi K, Ito F. Receptor tyrosine kinases and targeted cancer
100. Lee J, Son MJ, Woolard K, et al. Epigenetic-mediated dysfunction therapeutics. Biol Pharm Bull 2011;34:1774-1780.
of the bone morphogenetic protein pathway inhibits differentiation 118. Vik-Mo EO, Nyakas M, Mikkelsen BV, et al. Therapeutic vacci-
of glioblastoma-initiating cells. Cancer Cell 2008;13:69-80. nation against autologous cancer stem cells with mRNA-
101. Gangemi RM, Griffero F, Marubbi D, et al. SOX2 silencing in transfected dendritic cells in patients with glioblastoma. Cancer
glioblastoma tumor-initiating cells causes stop of proliferation and Immunol Immunother 2013;62:1499-1509.
loss of tumorigenicity. Stem Cells 2009;27:40-48. 119. Hirschmann-Jax C, Foster AE, Wulf GG, et al. A distinct Bside
102. Joshi K, Banasavadi-Siddegowda Y, Mo X, et al. MELK- population^ of cells with high drug efflux capacity in human tu-
dependent FOXM1 phosphorylation is essential for proliferation mor cells. Proc Natl Acad Sci U S A 2004;101:14228-14233.
of glioma stem cells. Stem Cells 2013;31:1051-1063. 120. Costa PM, Cardoso AL, Custódia C, et al. MiRNA-21 silencing
103. Verginelli F, Perin A, Dali R, et al. Transcription factors FOXG1 mediated by tumor-targeted nanoparticles combined with suniti-
and Groucho/TLE promote glioblastoma growth. Nat Commun nib: a new multimodal gene therapy approach for glioblastoma. J
2013;4:2956. Control Release 2015;207:31-39.
104. Zbinden M, Duquet A, Lorente-Trigos A, Ngwabyt SN, Borges I, 121. Venugopal C, Hallett R, Vora P, et al. Pyrvinium targets CD133 in
Ruiz i Altaba A. NANOG regulates glioma stem cells and is es- human glioblastoma brain tumor-initiating cells. Clin Cancer Res
sential in vivo acting in a cross-functional network with GLI1 and 2015;21:5324-5337.
p53. EMBO J 2010;29:2659-2674. 122. Fouladi M, Stewart CF, Olson J, et al. Phase I trial of MK-0752 in
105. Sherry MM, Reeves A, Wu JK, Cochran BH. STAT3 is required children with refractory CNS malignancies: a Pediatric Brain
for proliferation and maintenance of multipotency in glioblastoma Tumor Consortium study. J Clin Oncol 2011;29:3529-3534.
stem cells. Stem Cells 2009;10:2383-2392. 123. Robinson GW, Orr BA, Wu G, et al. Vismodegib exerts targeted
106. Rheinbay E, Suva ML, Gillespie SM, et al. An aberrant transcrip- efficacy against recurrent sonic hedgehog-subgroup medulloblas-
tion factor network essential for Wnt signaling and stem cell main- toma: results from phase II pediatric brain tumor consortium stud-
tenance in glioblastoma. Cell Rep 2013;3:1567-1579. ies PBTC-025B and PBTC032. J Clin Oncol 2015;33:2646-2654.