Oncotarget8250
www.impactjournals.com/oncotarget
www.impactjournals.com/oncotarget/ Oncotarget, 2017, Vol. 8, (No. 5), pp: 8250-8263
Stattic and metformin inhibit brain tumor initiating cells by
reducing STAT3-phosphorylation
Verena Leidgens
1
, Judith Proske
1,*
, Lisa Rauer
1,*
, Sylvia Moeckel
1
, Kathrin Renner
2
,
Ulrich Bogdahn
1
, Markus J. Riemenschneider
3
, Martin Proescholdt
4
, Arabel
Vollmann-Zwerenz
1
, Peter Hau
1
, Corinna Seliger
1
1
Department of Neurology and Wilhelm Sander-NeuroOncology Unit, University Hospital Regensburg, Regensburg, Germany
2
Department of Internal Medicine III, University Hospital Regensburg, Regensburg, Germany
3
Department of Neuropathology, Regensburg University Hospital, Regensburg, Germany
4
Department of Neurosurgery, University Hospital Regensburg, Regensburg, Germany
*
These authors have contributed equally to this work
Correspondence to: Corinna Seliger, email: corinna.seliger@klinik.uni-regensburg.de
Keywords: glioma, BTIC, STAT3, Stattic, metformin
Received: August 20, 2016 Accepted: November 21, 2016 Published: December 24, 2016
ABSTRACT
Glioblastoma (GBM) is the most common and malignant type of primary brain
tumor and associated with a devastating prognosis. Signal transducer and activator
of transcription number 3 (STAT3) is an important pathogenic factor in GBM and can
be specically inhibited with Stattic. Metformin inhibits GBM cell proliferation and
migration. Evidence from other tumor models suggests that metformin inhibits STAT3,
but there is no specic data on brain tumor initiating cells (BTICs).
We explored proliferation and migration of 7 BTICs and their differentiated
counterparts (TCs) after treatment with Stattic, metformin or the combination thereof.
Invasion was measured in situ on organotypic brain slice cultures. Protein expression
of phosphorylated and total STAT3, as well as AMPK and mTOR signaling were
explored using Western blot. To determine functional relevance of STAT3 inhibition
by Stattic and metformin, we performed a stable knock-in of STAT3 in selected BTICs.
Inhibition of STAT3 with Stattic reduced proliferation in all BTICs, but only
in 4 out of 7 TCs. Migration and invasion were equally inhibited in BTICs and TCs.
Treatment with metformin reduced STAT3-phosphorylation in all investigated BTICs
and TCs. Combined treatment with Stattic and metformin led to signicant additive
effects on BTIC proliferation, but not migration or invasion. No additive effects on
TCs could be detected. Stable STAT3 knock-in partly attenuated the effects of Stattic
and metformin on BTICs.
In conclusion, metformin was found to inhibit STAT3-phosphorylation in BTICs
and TCs. Combined specic and unspecic inhibition of STAT3 might represent a
promising new strategy in the treatment of glioblastoma.
INTRODUCTION
High-grade gliomas, especially glioblastomas
(GBM), are highly complex and heterogeneous primary
brain tumors, accounting for about 30% of all tumors of
the central nervous system [1]. Glioblastomas are nearly
uniformly fatal with median overall survival ranging
between 14.6 and 26.3 months in patients treated within
clinical studies [2, 3]. Brain tumor initiating cells (BTICs)
represent cancer stem-like progenitor cells, which are not
only implicated in tumor initiation, but also in recurrence
and progression [4–6]. BTICs are characterized by self-
renewal, clonogenicity, pluripotency, and closely resemble
the histopathological phenotype of parental tumors after
implantation of these cells into athymic mice [6].
Persistent activation of STAT3 (signal transducer
and activator of transcription number 3) has been detected
in many cancers [7], including gliomas, and is correlated
Research Paper
Oncotarget8251
www.impactjournals.com/oncotarget
with poor survival [8]. This was conrmed by correlation
of strong expression of STAT3 phosphorylated at Y705
in GBM specimens with a more aggressive phenotype
and shorter overall survival [9]. A series of elegant
studies has demonstrated an important role of STAT3
in gliomas in vivo and in vitro. Evidence has emerged,
that STAT3 is required by BTICs to maintain their stem-
like characteristics [10]. RNA interference of STAT3
sufciently led to growth arrest, inhibited neurosphere-
formation and could induce apoptosis in BTICs [11].
Hence, STAT3 inhibitors have become a major interest
in neuro-oncology. The STAT inhibitor Stattic [12] was
shown to selectively inhibit STAT3 [13]. However,
it is still unclear whether Stattic binds directly to the
phosphorylation site at Y705 or if it acts by altering the
conformation of the SH2 domain, because it binds to
Cys687 on the opposite side of the phosphopeptide
binding face [13].
Metformin (1,1-dimethylbiguanide hydrochloride)
is a biguanide drug mainly prescribed in the treatment of
type 2 diabetes [14]. Metformin has also antineoplastic
effects and may reduce the risk of certain cancer types in
diabetic patients [15, 16]. Several in vitro studies revealed
anti-proliferative effects of metformin on cancer stem cells
[17], glioma initiating cells [18, 19], and human GBM
lines [20]. Known mechanisms of action of metformin
are the inhibition of complex I of the respiratory chain
[21], resulting in activation of AMPK (adenosine
monophosphate-activated protein kinase) and the
inhibition of mTOR (mammalian Target Of Rapamycin)
[22]. Interestingly, metformin has been shown to reduce
STAT3-phosphorylation in a study investigating triple-
negative breast cancer [23]. Similar effects could also be
shown in two established GBM lines [24].
The primary aim of our study was to better
characterize the effects of STAT3 inhibition on primary
BTICs and their differentiated counterparts. In addition
to specic STAT3 inhibition with Stattic, we investigated
whether metformin inhibits STAT3-phosphorylation in
BTICs and whether additive effects can be achieved by
combining Stattic with the approved and clinically well-
tolerated anti-diabetic drug metformin.
RESULTS
Characteristics of brain tumor initiating cells
All primary BTICs used here were derived from
patients, who had undergone resection of WHO grade IV
gliomas at the Neurosurgery Department of the University
Hospital Regensburg [19, 25]. Primary cell lines were
established and used in low (typically below passage 8)
passage numbers to assure maximum resemblance to original
tumor cells. O
6
-methylguanine-DNA methyltransferase
(MGMT) methylation status varied in-between the lines,
while all lines were isocitrate dehydrogenase 1-(IDH1) wild
type in culture. BTIC-13 lost its IDH1 R132H mutation
under culture conditions (Table 1). Cells were kept as BTICs
under serum free conditions, and after withdrawal of growth
factors and addition of 10% serum to the cell culture medium
as differentiated tumor cells (TCs).
Stattic treatment reduces proliferation and
migration of BTICs and TCs
First, we investigated the effects of different doses of
Stattic on proliferation (Figure 1) and migration (Figure 2)
of BTICs and TCs. Proliferation was assessed after 48 and
96 h to ensure sufcient proliferation while simultaneously
avoiding cell death due to high conuence. Migration was
analyzed at 16, 24, 40 and 48 h, to provide also early time
points (16, 24 h), when migration is not confounded by
proliferation. The reduction of proliferation and migration
caused by Stattic was dose-dependent and cell line-
dependent. High doses of Stattic (10 – 15 μM) inhibited
proliferation in all BTIC lines and in 4 out of 7 respective
TC lines signicantly and in two other lines noticeably
(Figure 1 and Supplementary Figure 1). Protein expression
of pSTAT3 at Y705 was dose-dependently reduced after
Stattic treatment (24 h), while total STAT3 was not affected,
as assessed by Western blot (Figure 1F). When investigating
migration at early time points (24 h), high doses of Stattic
restricted migration in 6 out of 7 BTIC lines and in all
TC lines (Figure 2 and Supplementary Figure 2). When
comparing sensitivity of BTICs and TCs, BTICs were
signicantly more sensitive to Stattic than TCs regarding
proliferation (Figure 1E), but also differed according to
their basal proliferative capacity, which is observable when
comparing the DMSO controls. Migration did not differ
between BTICs and TCs (Figure 2E).
STAT3-overexpression enhances proliferation
and migration
We assessed effects of STAT3-overexpression
on proliferation, migration, and Stattic sensitivity in
exemplary cell lines (BTIC-8 and -13). We chose BTIC-
8 and -13 due to their good response to Stattic as wild
type cells, which unmasks a reduction of inhibitory
effects more easily than in cells that already respond
less in the wild type state. Markedly increased levels of
STAT3 were conrmed via qRT-PCR (Figure 3A, 3B)
and Western blot (Figure 3C) upon transfection with the
STAT3 construct. Stattic treatment led to weaker effects
on STAT3-phosphorylation in STAT3-overexpressing cells
(Supplementary Figure 3). Upon STAT3-overexpression,
BTIC-8 revealed signicantly increased proliferation after
96 h (Figure 3D). Sensitivity to Stattic appeared slightly
weaker in both lines for some Stattic concentrations, but
the results did not reach statistical signicance. Migration
was enhanced due to increased STAT3 expression
(Figure 3F, 3G), but sensitivity to Stattic did not differ
Oncotarget8252
www.impactjournals.com/oncotarget
Figure 1: Proliferation is reduced upon Stattic treatment. Exemplary proliferation of BTIC-8 A. and -13 B. and respective TC-8
C. and -13 D. upon Stattic treatment at indicated concentrations (2.5, 5, 10, 15 μM). E. Proliferation of BTICs (summarized for BTIC-7,
-8, -10, -11, 12, -13, -18) was affected signicantly more than that of the respective TCs (summarized for TC-7, -8, -10, -11, 12, -13, -18).
F. Stattic treatment (24 h) reduced phosphorylation of STAT3 as exemplarily shown in Western blot analysis of BTIC-11. Corresponding
GAPDH controls are indicated by use or not use of the asterisk.
Table 1: Patient characteristics
Histology
WHO
Grade
MGMT meth. IDH1 Age Gender OS (months)
BTIC-7 Primary GBM IV – wt 52 f 16.3
BTIC-8 Primary GBM IV – wt 52 f 4.0
BTIC-10 Primary GBM IV + wt 46 m 18.8
BTIC-11 Primary GBM IV + wt 55 m 17.5
BTIC-12 Gliosarcoma IV + wt 69 m 16.5
BTIC-13 Secondary GBM IV + R132H 42 m 8.5
BTIC-18 Primary GBM IV – wt 49 m 20.5
Seven different BTICs and their differentiated counterparts were used for in vitro assays. All BTICs were derived from
WHO grade IV gliomas. While MGMT methylation status differed between lines, IDH mutation was negative for all cells
in culture (R132H mutation in BTIC-13 was lost under culture conditions).
Oncotarget8253
www.impactjournals.com/oncotarget
substantially between wild type and knock-in (Figure 3F,
3G).
BTIC motility is affected by Stattic in an in situ
3D model
Spheroids of lentivirally transduced and uorescence
tagged BTIC-8, BTIC-10, BTIC-12 and BTIC-13 were
implanted on OBSCs (organotypic brain slice cultures).
Migration areas were analyzed over 14 days in vitro (=div)
and normalized to 0-h spheroid expansion areas. Treatment
of OBSCs with Stattic (15 μM) led to signicantly reduced
invasion in all investigated BTICs and TCs (Figure 4A-
4D). In BTIC-8 and TC-8 Stattic treatment led to slightly
reduced invasion between day 0 and day 7, but the cells
were hardly detected anymore at day 14, indicating
cytotoxicity. Exemplary pictures of invading BTIC-13 are
shown in Figure 4E. BTICs tended to migrate farther than
TCs, but both groups had similar migratory potential after
treatment with 15 μM Stattic.
Metformin inhibits STAT3-phosphorylation
Based on the results above and published effects
of metformin on BTICs [18, 19], we next investigated
STAT3-phosphorylation at Y705 and S727 after
Figure 2: Migration is restricted upon Stattic treatment. Exemplary migration of BTIC-8 A. and -13 B. and respective TC-8 C.
and -13 D. upon Stattic treatment at indicated concentrations (2.5, 5, 10, 15 μM). E. Migration of BTICs (summarized for BTIC-8, -11, -13,
-18) and respective TCs (summarized for TC-8, -11, -13, -18) was equally affected.
Oncotarget8254
www.impactjournals.com/oncotarget
Figure 3: Effects of STAT3-overexpression in BTICs. A, B. STAT3-overexpression was conrmed by qRT-PCR and Western
blot C. Corresponding GAPDH controls are indicated by use or not use of the asterisk. D, E. Proliferation of wild type and STAT3-
overexpressing BTIC-8 and BTIC-13 upon Stattic treatment. F, G. Migration of wild type and STAT3-overexpressing BTIC-8 and BTIC-13
upon Stattic treatment.
treatment with increasing doses of metformin (Figure 5).
Following a 48 h treatment, metformin inhibited STAT3-
phosphorylation in all investigated BTICs and TCs (Figure
5 and Supplementary Figure 4). Signaling pathways
known to be inuenced by metformin (activation of
AMPK and inhibition of mTOR) are shown as positive
control in BTIC-11. STAT3 knock-in opposed (low
dose) or weakened (high-dose) the effects of metformin
treatment, as shown exemplarily in BTIC-8 (S4D).
Combined treatment with Stattic and metformin
increases functional effects
Based on the inuence of metformin on STAT3-
phosphorylation we investigated the effects of combined
treatment with Stattic and metformin on proliferation
(Figure 6) and migration (Figure 7) of BTICs and TCs.
Low dose combination treatment (1 mM metformin
and 2.5 μM Stattic) resulted in strong and additive
proliferation inhibition in all BTIC lines (Figure 6 and
Supplementary Figure 5). In contrast, no additive effects
were observed for TC lines (Figure 6 and Supplementary
Figure 5). Compared to single treatment, the combination
did not result in enhanced restriction of migration as seen
in BTICs as well as TCs (Figure 7 and Supplementary
Figure 6). In addition, the combination of metformin and
Stattic did not lead to additive effects regarding invasion
as investigated in BTIC-8 using the aforementioned OBSC
model (Supplementary Figure 7).
DISCUSSION
For the rst time, we were able to demonstrate that
metformin in combination with specic STAT3 inhibition
by Stattic additively reduced proliferation in a multitude
of primary human BTICs. Furthermore, we observed
that metformin inhibited phosphorylation of STAT3 in
all BTICs and TCs. On the contrary, the combination of
Stattic and metformin affected migration of BTICs and
TCs to a lower extent in all primary lines. Stable knock-
in of STAT3 led to a slight attenuation of the observed
effects, while proliferation and partly migration was
enhanced.
Many studies have been performed to identify and
understand aberrant regulation of STAT3 and its impact
on proliferation, migration, and invasion. Among those,
Bromberg et al. [26] reported that dominant-negative
STAT3 abrogated oncogenic transformation, whereas
constitutively activated STAT3C [27] mutants induced it.
Furthermore, STAT3 was shown to be essential in human
glioma cells to maintain their tumor initiating capacity and
the ability to invade the normal brain [10]. These ndings
indicated STAT3 as a promising target for anticancer
therapy.
Stattic was used in concentrations ranging from
2.5 to 15 μM to reveal minimal dosages required for
proliferation and migration restriction in vitro. Although
in vitro studies reported cell cytotoxicity even with low
Stattic concentrations, in vivo animal studies revealed no