Livin contributes to tumor hypoxia-induced resistance to cytotoxic
therapies in glioblastoma multiforme
1
2
3
4
Chia-Hung Hsieh1,2,3*, Yu-Jung Lin1, Chung-Pu Wu4, Hsu-Tung Lee5, Woei-Cherng
5
Shyu6,7*, Chi-Chung Wang8*
6
7
1
8
Taiwan
9
2
Graduate Institute of Basic Medical Science, China Medical University, Taichung,
Department of Medical Research, China Medical University Hospital, Taichung,
10
Taiwan
11
3
Department of Biomedical Informatics, Asia University, Taichung, Taiwan
12
4
Department of Physiology and Pharmacology, Chang Gung University, Tao-Yuan,
13
Taiwan
14
5
Department of Neurosurgery, Taichung Veterans General Hospital, Taichung, Taiwan
15
6
Department of Neurology, Center for Neuropsychiatry, China Medical University
16
and Hospital, Taichung, Taiwan
17
7
Graduate Institute of Immunology, China Medical University, Taichung, Taiwan
18
8
Graduate Institute of Basic Medicine, Fu Jen Catholic University, New Taipei,
19
Taiwan
20
21
*Corresponding authors:
1
1
2
Chia-Hung Hsieh and Woei-Cherng Shyu, China Medical University and Hospital,
3
No. 91, Hsueh-Shih Road, Taichung 404, Taiwan. Phone: 886-4-22052121; Fax: 886-
4
4-22333641; E-mail: chhsiehcmu@mail.cmu.edu.tw (C-H Hsieh);
5
shyu9428@gmail.com (W-C Shyu). Chi-Chung Wang, Fu Jen Catholic University,
6
No.510, Zhongzheng Road, New Taipei City 24205, Taiwan. Phone: 886-2-
7
29052039; E-mail: 075006@mail.fju.edu.tw
8
9
Conflict of interest: The authors have declared that no conflict of interest exists.
10
Running title: Livin in tumor hypoxia-induced therapy resistance
11
Key words: tumor hypoxia, inhibitor of apoptosis proteins, Livin, hypoxia inducible
12
factor-1, cell-permeable peptide
13
Word count: 4643
14
Figures: 6
15
16
Translational Relevance
17
18
19
This study shows tumor hypoxia can induce Livin upregulation, subsequent antiapoptosis as well as resistance to cytotoxic therapies through a HIF-1α-dependent
2
1
pathway. The induction of Livin in response to tumor hypoxia is therefore associated
2
with bad prognosis. This may be particularly relevant in the selection of cancer
3
patients prone to receive targeting Livin inhibitors as adjunct treatments. The cell-
4
permeable peptide targeted to Livin derived from this study also provides a tractable
5
path for clinical translation in improvement of Livin-mediated therapy resistance.
6
7
Abstract
8
9
Purpose: Tumor hypoxia is one of the crucial microenvironments to promote therapy
10
resistance (TR) in glioblastoma multiforme (GBM). Livin, a member of the family of
11
inhibitor of apoptosis proteins, contributes anti-apoptosis. However, the role of tumor
12
hypoxia in Livin regulation and its impact on TR are unclear.
13
Experimental Design: Livin expression and apoptosis for tumor hypoxic cells
14
derived from human glioblastoma xenografts or in vitro hypoxic stress-treated
15
glioblastoma cells were determined by western blotting, immunofluorescence imaging
16
and annexin V staining assay. The mechanism of hypoxia-induced Livin induction
17
was investigated by chromatin immunoprecipitation assay and reporter assay. Genetic
18
and pharmacological manipulation of Livin were utilized to investigate the role of
19
Livin on tumor hypoxia-induced TR in vitro or in vivo.
3
1
Results: The upregulation of Livin expression and downregulation of caspase activity
2
were observed under cycling and chronic hypoxia in glioblastoma cells and
3
xenografts, concomitant with increased TR to ionizing radiation and temozolomide.
4
However, knockdown of Livin inhibited these effects. Moreover, hypoxia activated
5
Livin transcription through the binding of hypoxia-inducible factor 1α to the Livin
6
promoter. The targeted inhibition of Livin by the cell-permeable peptide (TAT-Lp15)
7
in intracerebral glioblastoma-bearing mice demonstrated a synergistic suppression of
8
tumor growth and increased survival rate in standard-of-care treatment with radiation
9
plus temozolomide.
10
Conclusions: These findings indicate a novel pathway that links upregulation of
11
Livin to tumor hypoxia-induced TR in GBM and suggest that targeting Livin using
12
cell-permeable peptide may be effective therapeutic strategy for tumor
13
microenvironment-induced TR.
14
15
Introduction
16
17
Hypoxia is well evidenced within most of solid tumors (1). The acute, intermittent,
18
or cycling hypoxia is associated with inadequate blood flow, whereas chronic hypoxia
19
is the consequence of increased oxygen diffusion distance resulting from tumor
4
1
expansion (2). These hypoxic areas can either promote cell death or provoke an
2
adaptive response leading to the selection for death resistance (1, 3, 4). Once tumor
3
cells become adaptive to hypoxia, there are more resistant to apoptosis and less
4
responsive to cancer therapy. Tumor hypoxia is thought to be a crucial role in
5
pathological characteristics of glioblastoma multiforme (GBM), including
6
invasiveness, necrosis and microvascular hyperplasia. It also contributes the
7
resistance to chemotherapy, immunotherapy and radiotherapy due to the possibility of
8
dysregulation of apoptotic pathway or other mechanisms (5). However, the exact
9
mechanisms triggered by hypoxia that lead to anti-apoptosis are not well known.
10
11
Tumor resistance to apoptosis is one of hallmarks of cancer (6). Evasion of
12
apoptosis may contribute to carcinogenesis, tumor progression and also to therapy
13
resistance (TR). It has been reported that several gene families are involved in the
14
negative regulation of apoptosis, including the inhibitor of apoptosis proteins (IAPs)
15
(7, 8). IAPs are a family of antiapoptotic proteins with 70 amino acid baculoviral
16
repeats (BIR) including NAIP, c-IAP1 (MIHB, HIAP-2), c-IAP2 (HIAP-1, MIHC,
17
API2), XIAP (hILP, MIHA, ILP-1), Survivin, BRUCE (apollon), ILP-2 and Livin
18
(BIRC7, ML-IAP, KIAP) (9, 10). These proteins can bind and suppress upstream or
19
downstream caspases and inhibit apoptosis. Among these proteins, Livin is more
5
1
recently identified and selectively binds on the endogenous IAP antagonist SMAC
2
and caspase-3, caspase-7, and caspase-9 (11-13). In addition, a variety of
3
malignancies overexpressed Livin, which is correlated with tumor progression, TR
4
and poor patient’s outcome in some tumors (14). However, the impact of tumor
5
microenvironment on Livin expression and its regulatory mechanisms remain unclear.
6
7
In this study, we hypothesized that Livin may represent one of hypoxia-
8
responsive genes and play a role in tumor hypoxia-induced anti-apoptotic effect at
9
tumor microenvironment. Our results show that both cycling hypoxia and chronic
10
hypoxia can induce Livin upregulation, subsequent anti-apoptosis as well as
11
resistance to cytotoxic therapies in glioblastoma cells and xenografts through a
12
hypoxia-inducible factor 1 (HIF-1) α-dependent pathway. Genetic inactivation and
13
pharmacological inhibition of Livin in vitro or in vivo suppress tumor hypoxia-
14
induced TR and generate a synergistic suppression of anti-tumor growth and tumor
15
cell death. These data highlight tumor hypoxia-induced Livin expression may
16
represent a pathway for resistance of some tumors to radio- and chemotherapeutics.
17
Besides, the cell-permeable peptide targeted to Livin derived from this study provides
18
a translational path for clinical improvement of standard-of-care therapies in GBM
19
patients.
6
1
2
Materials and Methods
3
4
Molecular and Cellular Assays
5
6
Several molecular and cellular assays were used in this study. These included the
7
following: Quantitative reverse transcription-PCR (qPCR) and Western blot analyses
8
to assess the expression of the hypoxia-inducible factor (HIF)-1α, HIF-2α and Livin
9
proteins; HIF-1 overexpression and knockdown, promoter analysis, luciferase assay
10
and chromatin immunoprecipitation (ChIP) to investigate the regulation of Livin
11
promoter; genetic inactivation or pharmacological inhibition of Livin, caspase-3
12
activity and apoptosis assays, cells irradiation, clonogenic survival and cytotoxicity
13
assays to determine the role of Livin in hypoxia-mediated resistance to ionizing
14
radiation (IR) and temozolomide (TMZ). Detailed procedures can found in the
15
Supplementary Materials and Methods.
16
17
Identification of hypoxia subtypes in glioblastoma xenografts
18
19
The hypoxia subtypes in glioblastoma xenografts were distinguished according
7
1
to the physiological and molecular characteristics of cycling hypoxia. Cycling
2
hypoxia tended to occur in regions with blood perfusion and HIF-1 activation whereas
3
chronic hypoxia located in areas without blood perfusion but with HIF-1 activation.
4
Perfusion marker, Hoechst 33342, staining and HIF-1 activation labeling together
5
with immunofluorescence imaging and fluorescence-activated cell sorting were used
6
to isolate hypoxic tumor subpopulations from human glioblastoma xenografts. Cells
7
with positive Hoechst 33342 staining and GFP expression were potential cycling
8
hypoxic cells. However, cells that were positive for GFP expression but negative for
9
Hoechst 33342 were mostly chronic hypoxic cells. The detail method are described in
10
Supplementary Materials and Methods.
11
12
Animal model
13
14
Eight-week-old male nude mice (Balb/c nu/nu) were purchased from the Animal
15
Facility of the National Science Counsel (NSC) and were used to establish the
16
orthotopic glioblastoma xenograft model according to the published methods (15).
17
Briefly, 2 × 105 GBM8401/SFFV-LucGFP, GBM8401 and U251 cells were harvested
18
by trypsinization and injected into the left basal ganglia of anesthetized mice. The
19
tumors developed at 12 days after tumor implantation for evaluating the efficiency of
8
1
therapy studies and at 18 days after tumor implantation for in vivo TUNEL and
2
clonogenic survival assays. All animal experiments were conducted according to
3
Institutional Guidelines of China Medical University after acquiring permission from
4
the local Ethical Committee for Animal Experimentation.
5
6
In vivo treatment
7
8
Intracerebral gioblastoma-bearing mice were randomly assigned to four different
9
therapeutic groups: control (vehicle treatment), 2 fractions of stereotactic radiation (4
10
Gy each) plus TMZ (5 mg/kg) by oral gavage 1 hour prior to radiation, intravenous
11
injection of TAT-K22 or TAT-Lp15 (20 mg/kg) prior to the combination of TMZ and
12
radiation treatment. All of treatments were administered daily on 2 consecutive days.
13
All of treatments were carried out at day 12 after tumor cell injection for
14
bioluminescent imaging and animal survival assays and at day 18 after tumor cell
15
injection for TUNEL staining and in vivo/in vitro clonogenic survival assays. Tumor
16
progression was monitored by bioluminescence imaging and mice were monitored
17
daily for survival. Animals were killed at the onset of neurologic signs or any type of
18
distress.
19
9
1
Bioluminescent imaging (BLI)
2
3
Mice were imaged with the IVIS Imaging System 200 Series (Caliper) to record
4
bioluminescent signal emitted from the engrafted tumors. Mice were anesthetized
5
with isoflurane and received intraperitoneal injection of D-Luciferin (Caliper) at a
6
dose of 270 µg/g body weight. Imaging acquisition was performed at 15 minutes after
7
intraperitoneal injection of luciferin. For BLI analysis, regions of interest
8
encompassing the intracranial area of signal were defined using Living Image
9
software, and the total number of photons per second per steradian per square
10
centimeter were recorded. To facilitate comparison of growth rates, each mouse’s
11
luminescence readings were normalized against its own luminescence reading at day
12
12, thereby allowing each mouse to serve as its own control.
13
14
Animal irradiation and in vivo/in vitro clonogenic survival assay
15
16
Mice were locally irradiated (4.25 Gy/minute) at a dose of 4 Gy under anesthesia
17
using a highenergy X-ray linear accelerator (Varian). Tumors were then excised,
18
minced and dissociated for 30 minutes at 37˚C in Hanks' balanced salt solution
19
containing 166 U/mL collagenase XI, 0.25 mg/mL protease and 225 U/mL DNase.
10
1
Cells were recovered by straining through an 80 μm mesh and centrifuging at 500 g
2
and were resuspended in culture medium. Cells were counted by hemocytometer
3
using trypan blue and assessed by in vitro clonogenic survival assay.
4
5
Statistical analysis
6
7
One-way analysis of variance with post hoc Scheffe analyses and Kaplan-Meier
8
Survival Analysis with Tarone-Ware statistics were carried out using the SPSS
9
package (version 18.0). The differences between control and experimental groups
10
were determined by the two-sided, unpaired Student t test. P < 0.05 was considered
11
significant.
12
13
Results
14
15
Tumor hypoxia induces Livin expression
16
17
We first analyzed the time course of Livin mRNA in GBM8401 cells with or
18
without hypoxic stress (<1% O2). By 3 hours, Livin mRNA had increased
19
significantly compared with the normoxic control (Fig. 1A). Moreover, Livin mRNA
11
1
and protein levels were all increased in the glioblastoma cells, U251, U87 and
2
GBM8401 cells, at 24 hours after hypoxic treatment (<1% O2) (Figure 1B and 1C).
3
Besides, glioblastoma cells experienced an increase in the Livin expression under
4
both non-interrupted and cycling hypoxic stress (Supplementary Fig. S1 and S2). To
5
better verify endogenous tumor hypoxia-mediated Livin expression in the solid tumor,
6
Hoechst 33342 staining and HIF-1 activation labeling together with
7
immunofluorescence imaging and fluorescence-activated cell sorting were utilized to
8
isolate hypoxic tumor subpopulations from human glioblastoma xenografts as
9
described in our previous studies (16-18). The GBM8401/hif-1-r tumors showed
10
highly heterogeneous Hoechst 33342 staining and HIF-1 activation on
11
immunofluorescence images (Fig. 1D and Supplementary Fig. S3). The Livin staining
12
indicated that Livin expression occurs in the cycling hypoxic (Hoechst 3342+ and
13
GFP+) and chronic hypoxic (Hoechst 3342- and GFP+) areas. Next, the total RNA of
14
the hypoxic cell subpopulations derived from disaggregated orthotopic GBM8401/hif-
15
1-r and U87/hif-1-r xenografts (18) were further analyzed for Livin and vascular
16
endothelial growth factor (VEGF), an HIF-1 target gene, expression by qPCR. The
17
expression of Livin and VEGF increased significantly in cycling and chronic hypoxic
18
cells compared with that in normoxic cells (Fig. 1E). Besides, the results from
19
immunofluorescence staining of HIF-1α and Livin human glioblastoma specimens
12
1
also demonstrate that Livin expression co-localized with high expression of HIF-1α
2
(Fig. 1F and Supplementary Fig. S4), suggesting that endogenous hypoxia leads to the
3
induction of Livin in glioblastoma.
4
5
On the other hand, incubation of normoxic U251, U87 and GBM8401 cells for
6
24 hours with desferrioxamine (100 μM) or cobalt chloride (50 μM), which mimic
7
hypoxia by inducing transcription from HIF-1-dependent genes, increased Livin
8
mRNA to a similar level as that found during hypoxia (Supplementary Fig. S5).
9
Moreover, the transcriptional activation of Livin in the hypoxia-treated cells increased
10
significantly after treatment, with maximal activation occurring 48 hours after
11
hypoxic stress (<1% O2) in GBM8401 cells which stably transfected with 2510-bp
12
Livin promoter-driven luciferase reporter gene (Livin-Luc) (Supplementary Fig. S6).
13
Taken together, these results indicate tumor hypoxia, either chronic or cycling
14
hypoxia, can induce Livin expression and hypoxia-induced Livin expression is
15
associated with HIF-1 activation.
16
17
Hypoxia-induced Livin expression is dependent on HIF-1α
18
13
1
In an attempt to gain insight into the mechanisms of hypoxia-induced Livin
2
induction, we examined Livin mRNA and protein levels in GBM8401 and U87 cells
3
lacking HIF-1α or HIF-2α expression. After lentiviral transduction with shRNAs
4
against HIF-1α and HIF-2α, GBM8401 and U87 cells lost hypoxia-induced HIF-1α or
5
HIF-2α expression compared with that in cells transduced with control scrambled
6
shRNA (Fig. 2A). In nonspecific shRNA control GBM8401 or U87 cells, hypoxic
7
stress stimulated a significant increase in Livin mRNA and protein levels (Fig. 2B and
8
2C). However, knockdown of HIF-1α, but not HIF-2α, significantly abrogated
9
hypoxia-induced Livin expression. Besides, the hypoxia-induced transcriptional
10
activation of Livin was inhibited by knockdown of HIF-1α in U251, U87 and
11
GBM8401 cells and treatment with YC-1 (a HIF-1α inhibitor) in HEK-293 and
12
GBM8401 cells (Fig. 2D and Supplementary Fig. S7). On the contrary, coexpression
13
of HIF-1α-oxygen-dependent degradation domain (ODD) deletion mutant and Livin-
14
Luc significantly enhanced the reporter activity but not control plasmids. These
15
results indicate that HIF-1α is a crucial transcription factor for hypoxia-mediated
16
Livin induction.
17
18
HIF-1α binds directly to the Livin promoter to regulate its expression
19
14
1
Next, we sought to determine whether HIF-1α binds to the Livin promoter for
2
hypoxia-induced its expression. The bioinformatics analysis identified two HIF-1α-
3
binding sites in the Livin promoter sequence from −2000 to +100 (Fig. 3A),
4
suggesting that HIF-1α might regulate Livin expression by directly binding to its
5
promoter. To pinpoint the exact binding motifs, we introduce point mutations into the
6
hypoxia response element (HRE)-1 and HRE-2 of Livin-Luc. The ablation of HRE-1,
7
but not HRE-2, on the Livin promoter abrogated hypoxia-mediated Livin induction
8
(Fig. 3B). ChIP assays also confirmed the binding of HIF-1α to the Livin promoter
9
(Fig. 3C and 3D). Collectively, these results suggest that HIF-1α regulates Livin
10
transcription by directly binding to the Livin promoter in a hypoxia-dependent
11
fashion.
12
13
Hypoxia-mediated Livin expression results in antiapoptosis
14
15
Given the recognized role of Livin as an apoptosis suppressor, we determined
16
whether glioblasmtoma cells with hypoxia-mediated Livin up-regulation became
17
more resistant to apoptotic insults. Staurosporine (STA) was chosen as the apoptotic
18
trigger. Tetracycline-regulatable lentiviral vectors encoding shRNAs were used to
19
stably and specifically knockdown Livin induction in GBM8401 and U87 cells under
15
1
hypoxia (Fig. 4A and 4B). The caspase activities and apoptotic cells were all
2
decreased significantly in cycling or non-interrupted hypoxia-treated cells compared
3
to normoxic cells (Fig. 4C and 4D), suggesting hypoxic cells are resistant to STA-
4
induced apoptosis. After the pretreatment of doxycycline (Dox), however, either
5
cycling or non-interrupted hypoxia-induced antiapoptotic effects were inhibited by
6
Livin knockdown. These results indicate that Livin plays an essential role in tumor
7
hypoxia-induced antiapoptosis in glioblasmtoma cells.
8
9
10
Hypoxia-mediated Livin expression promotes radioresistance and
chemoresistance
11
12
Radiotherapy plus concomitant and adjuvant temozolomide (TMZ) is standard-of-
13
care treatment for patients with GBM (19). We tested whether hypoxia-induced Livin
14
expression is associated with the development of radiation and TMZ resistance via
15
clonogenic survival assay and MTT assay. Both non-interrupted and cycling hypoxia
16
pretreatment significantly increased cell resistance to ionizing radiation and TMZ
17
compared with normoxic controls in GBM8401 and U251 cells (Fig. 5A and 5B;
18
Supplementary Fig. S8). But, loss of Livin significantly inhibits hypoxia-induced
19
radioresistance and chemoresistance and increases the radiosensitivity and
16
1
cytotoxicity of glioblastoma cells to TMZ, suggesting Livin is a critical effector
2
involved in hypoxia-induced radioresistance and chemoresistance in glioblastoma
3
cells.
4
5
Next, we generated therapeutic peptide to inhibit Livin activity and test whether
6
this peptide can block hypoxia-induced radioresistance and chemoresistance in
7
glioblastoma cells. To bind Livin, we constructed a cell-permeable peptide comprising
8
the twenty-three amino acid-terminal residues of Livin-binding peptide, Lp15, derived
9
from a randomized peptide expression library (20). The truncated Lp15 and control
10
peptide, K22, were rendered cell-permeant by fusing each to the cell-membrane
11
transduction domain of the human immunodeficiency virus-type 1 (HIV-1) Tat protein
12
to obtain a 35~37 amino acid peptide (TAT-Lp15) and control peptide (TAT-K22),
13
respectively (Fig. 5C). The fluorophore dansyl chloride was conjugated to TAT-Lp15
14
and TAT-K22. Confocal microscopy showed that cells treated with either TAT-Lp15-
15
dansyl or TAT-K22-dansyl exhibited fluorescence in their cytoplasm (Fig. 5D),
16
indicating intracellular peptide uptake. TAT-Lp15-dansyl accumulation was detectable
17
in cells within 15 minutes of incubation, peaked during 60 minutes, and remained
18
detectable for 8 hours after washing the peptide from medium (Fig. 5E). To assess
19
whether the TAT-Lp15 peptide can sensitize glioblastoma cells toward apoptosis, a
17
1
series of concentrations of TAT-Lp15 peptide were treated to Livin-postive
2
(GBM8401) cells with or without Livin knockdown and Livin negative (H1299) cells
3
followed by STA treatment. TAT-Lp15 peptide enhanced STA-induced caspase
4
activities and apoptosis in a dose-dependent manner in GBM8401 cells but not in
5
H1299 cells or Livin knockdown GBM8401 cells (Supplementary Fig. S9, S10, S11
6
and S12). Besides, TAT-Lp15 peptide, but not of a control peptide, significantly
7
increased the radiosensitivity and cytotoxicity of U87, U251 and GBM8401 cells to
8
TMZ and decreased hypoxia-induced radioresistance and chemoresistance in
9
glioblastoma cells (Fig. 5F and 5G). The combination treatment of GBM8401 cells
10
with TAT-Lp15 and 4-Gy irradiation could also yield a synergistic effect on tumor cell
11
killing (Supplementary Fig. S13).
12
13
Livin blockage enhances the efficiency of radiation plus TMZ treatment in
14
glioblastoma xenografts
15
16
Finally, we investigated whether the targeted inhibition of Livin by therapeutic
17
peptide represents a viable approach for inhibition of tumor microenvironment-
18
mediated therapy resistance in GBM. We first determined whether TAT-Lp15 could be
19
delivered into the brain in the intact animal. Fluorescence imaging demonstrated that
18
1
brains from nude mice injected intravenously with a 500 μM dose of either TAT-
2
Lp15-dansyl or TAT-K22-dansyl, exhibited strong fluorescence in the cortex (Fig.
3
6A), indicating TAT-Lp15-dansyl or TAT-K22-dansyl enters the brain upon peripheral
4
administration. To mimic the clinical setting, we sought to test the efficacy of TAT-
5
Lp15 when added to standard-of-care treatment with radiation plus TMZ. BLI was
6
utilized to assess intracranial tumor growth in the orthotopic GBM8401/SFFV-
7
LucGFP xenograft model (16, 17, 21). Consistent with the in vitro studies, mice that
8
had TAT-Lp15 treatment prior to initiation of the combination of TMZ and radiation
9
treatment had significantly better tumor growth delay and survival rate than those
10
pretreated with TAT-K22 or vehicle (Fig. 6B, 6C and 6D). The combination treatment
11
of GBM8401 xenografts with TAT-Lp15 and standard-of-care treatment with radiation
12
plus TMZ could yield a synergistic effect in therapeutic benefits. Moreover, TUNEL
13
assays were done on tumors collected from mice submitted to the above experimental
14
conditions on day 3 after treatments. TAT-Lp15 treatment during the combination of
15
TMZ and radiation treatment increased the extent of TMZ and radiation-induced
16
apoptosis of tumor cells (Fig. 6E and 6F). By pimonidazole/annexin V doubling
17
staining flow cytometry assay, TAT-Lp15 treatment also enhanced the TMZ and
18
radiation-induced apoptosis in tumor hypoxic cells (Fig. 6G). Besides, pretreatment
19
with TAT-Lp15, but not TAT-K22, produced a significant reduction in clonogenicity
19
1
after TMZ and radiation treatment in both GBM8401 and U251 xenografts (Fig. 6H
2
and 6I). Taken together, these results suggest that targeted inhibition of Livin by
3
therapeutic peptide enhances the efficiency of standard-of-care therapies in vivo in
4
orthotopic glioblastoma xenografts.
5
6
Discussion
7
8
9
Hypoxic tumor microenvironment has long been considered as major challenge in
cancer therapy and is one of the primary causes for cancer recurrence in GBM (22).
10
Understanding the underlying mechanism governing the nature of therapeutic
11
resistance is therefore critical and may uncover therapeutic interventions. It is a well-
12
known fact that chronic hypoxia contributes to tumor microenvironment-mediated
13
therapy resistance. Although tumor cycling hypoxia is now a well-recognized
14
phenomenon in animal and human solid tumors, however, relatively few studies to
15
date have specifically investigated its role and mechanism on therapy resistance. In
16
this study, we identified a novel hypoxia responsive gene, Livin, promotes anti-
17
apoptosis and therapeutic resistance in glioblastoma. Despite the potential role of
18
other hypoxia-dependent anti-apoptotic regulation in therapeutic resistance, we
19
discovered a novel aspect by which HIF-1α transactivation of Livin expression is
20
1
required for tumor chronic and cycling hypoxia-induced anti-apoptosis and
2
therapeutic resistance. On the basis of our findings, we propose a model in which
3
genetic mutation or hypoxic tumor microenvironment increases HIF-1α accumulation
4
through promoting protein synthesis or inhibiting protein degradation. The
5
cytoplasmic HIF-1α then translocates to the nucleus, recognizes a cognate sequence
6
on the Livin promoter, induces Livin expression, and promotes inhibition of caspase
7
activities and anti-apoptosis.
8
9
IAP family of proteins are known to act as caspase inhibitors, blocking caspase
10
activation and further inhibiting apoptosis (9). Previous studies have shown that a
11
number of different IAPs were highly expressed in malignant gliomas including XIAP
12
(23), apollon (24) and survivin (25), with an adverse correlation with patient
13
outcomes. Although less study are carried out in brain tumors, recent studies reported
14
that Livin and survivin are up-regulated in glioma stem cells (GSC) derived from
15
human glioblastoma and astrocytoma tissues or glioblastoma cell lines but are the
16
most increased in just differentiated GSC (26, 27). These findings imply that Livin
17
may contribute to anti-apoptosis and therapeutic resistance in glioblastoma. Here, we
18
first demonstrate that the tumor microenvironment-mediated expression pattern of
19
Livin.
To directly distinguish and isolate hypoxic tumor subpopulations from solid
21
1
tumors, we used a reliable protocol of cycling- and chronic-hypoxic cell identification
2
derived from our previous studies that allows subsequent immunofluorescence
3
imaging and flow cytometric analysis of the biosignature of these cells (16). Our
4
results indicate that both of probable cycling and chronic hypoxic areas are important
5
in tumor microenvironment-mediated Livin induction. However, Livin expression
6
tends to occur in probable cycling hypoxic areas with high HIF-1 activation and blood
7
perfusion within the tumor microenvironment in glioblastoma xenografts. Besides, in
8
vitro studies also demonstrate that cycling hypoxia induced more Livin expression
9
compared with non-interrupted hypoxia. These differential Livin expressions suggest
10
that HIF-1α is a major regulator of tumor microenvironment-mediated Livin induction
11
because Livin is one of HIF-1α target genes. Moreover, the results of Livin expression
12
co-localized with high expression of HIF-1α in human glioblastoma specimens also
13
support this notion.
14
15
Despite frequent reports of Livin overexpression in human cancers and its
16
intrinsic nature in inhibition of caspase activities (28), the impact of Livin contributed
17
to tumor microenvironment-mediated anti-apoptosis and therapeutic resistance
18
remains unknown. In contrast to Livin, other IAPs, IAP-2 (29) and surviving (30),
19
have been identified as hypoxia responsive genes. Although the detail mechanism is
22
1
still unclear, IAP-2 can be induced by severe hypoxia in several mammalian cells via
2
HIF-1-independent mechanisms. Besides, hypoxia can induce survivin expression
3
through the binding of HIF-1α to the survivin promoter. Therefore, IAP-2 and
4
survivin are thought to be a role for hypoxia-mediated death resistance. Herein, we
5
provide the first evidence to show that hypoxia stimulates HIF-1α, but not HIF-2α, to
6
bind to the human Livin promoter and regulate its transcription. Moreover, our results
7
provide clear evidence that hypoxic tumor microenvironment, either chronic or
8
cycling hypoxia, enhances the expression and function of Livin and further promotes
9
anti-apoptosis and resistance to ionizing radiation and TMZ in glioblastoma. Genetic
10
inactivation or pharmacological inhibition of Livin sensitize cells to cytotoxic
11
therapies-induced apoptosis in hypoxic condition or endogenous tumor
12
microenvironment. These finding are important in the selection of optimal genetic or
13
pharmacological approaches to inhibit tumor microenvironment-mediated therapeutic
14
resistance via the blockade of this mechanism.
15
16
Numerous studies suggest that Livin is an attractive target for cancer treatment.
17
The reason is based on several independent pieces of evidence. First, the anti-
18
apoptotic ability of Livin can contribute to the apoptosis-resistant phenotype of cancer
19
cells (13). Second, Livin has been considered critical for tumor progression included
23
1
its function in cell proliferation, migration, invasion and epithelial-mesenchymal
2
transition (EMT) (31). Third, Livin is preferentially expressed in tumors but not in
3
normal tissues (14). In this study, we provide a novel rationale for Livin as a potential
4
therapeutic target in cancer, which is hypoxic tumor microenvironment can induce up-
5
regulation of Livin and further contributes to inhibition of tumor apoptosis and
6
therapeutic resistance. The strategies to specific Livin intervention for novel cancer
7
therapies have been proposed and developed in the past decade such as small
8
interfering RNA (32) and peptide strategies (20). Unfortunately, the use of small
9
interfering RNA is still therapeutically impractical. The ideal of targeted inhibition of
10
Livin by specific targeting peptides is still in cell model and not readily translatable to
11
preclinical or clinical trial.
12
13
In the present study, we generate a cell-permeable peptide, TAT-Lp15, for targeting
14
Livin and inhibiting its function by fusing Tat domain from HIV-1 with the truncated
15
peptide derived from linear peptides that specifically bind to Livin as shown in
16
previous study (20). Our findings demonstrate for the first time the feasibility of using
17
cell-penetrating peptide for targeting Livin to sensitize cancer cells for pro-apoptotic
18
stimuli both in vitro and in vivo. We have shown that the addition of TAT-Lp15
19
sensitizes glioblastoma cell lines to the pro-apoptotic effects of ionizing radiation and
24
1
TMZ in vitro. Moreover, the results derived from in vivo studies demonstrate that the
2
combination of TAT-Lp15 treatment and cytotoxic therapies has a significant
3
synergistic effect in anti-glioblastoma growth and tumor cell killing above and beyond
4
that seen with the standard-of-care therapy of radiation plus TMZ. In the absence of a
5
proapoptotic stimulus, however, TAT-Lp15 treatment doesn’t increase the spontaneous
6
apoptosis rate of normal brain tissues (data not shown), suggesting that this
7
therapeutic peptide may have promising therapeutic potential to sensitize tumors for
8
the cytotoxic therapies without affecting normal tissues. Besides, our results also
9
reflect that the addition of an Livin targeting peptide appeared to be sufficient to
10
initiate apoptosis and enhance therapeutic efficiency when combined with cytotoxic
11
therapy even glioblastomas contain multiple abnormalities of the apoptotic pathway
12
upstream of the Livin or other IAPs dysregulation from genetic mutation or tumor
13
microenvironment.
14
15
For the molecularly targeted therapy in intracranial tumors, the blood-brain barrier
16
(BBB) is an impediment to drug bioavailability. Here, we find that TAT-Lp15 is able
17
to cross even the intact BBB in normal brain. Therefore, this peptide offers a feasible
18
path for translation to the clinic. Previous preclinical studies validating Livin as
19
therapeutic targets in gliomas have utilized Livin siRNA (33). However, it is not
25
1
feasible to use the siRNA employed due to off-target effects, stimulation of the innate
2
immune response and inefficient delivery to target organ (34). Besides, a series of
3
Smac mimetic peptides or compounds derived from the N-terminal tetra-peptide of a
4
mitochondrial protein called Smac/DIABLO (second mitochondria-derived activator
5
of caspase/direct IAP binding protein with low pI) were designed as IAP antagonists
6
(35). Although these inhibitors can binds to type-II BIR domains of IAPs and interfere
7
with the interaction between IAPs and active caspases, the recent study demonstrates
8
that the mechanism of these compounds sensitized cells to apoptosis is primarily
9
indirect due to autoubiquitination and proteasomal degradation of cIAPs rather than
10
depression of caspase activity (36). It is still unclear whether these compounds are
11
able to inhibit Livin function. Although the detail mechanism of Livin-binding
12
peptides in inhibition of Livin function is still unknown, no obvious sequence
13
homologies to Smac or caspases and normal Livin expression in intracellular
14
expression of the Livin-targeting peptides suggest these peptides have distinct
15
mechanisms to resentize cells to apoptosis (20). Therefore, the cell-permeable peptide
16
targeted to Livin derived from this study provides a unique therapeutic drug in
17
treatment of cancer with Livin expression.
18
19
Acknowledgments
26
1
2
3
We thank the National RNAi Core Facility, Academia Sinica, Taiwan, for
technical support.
4
5
Grant Support
6
7
Grant support was provided by the National Science Council of the Republic of
8
China (Grant No. 102-2314-B-039-029-MY3) and grants CMU101-S-17 and
9
CMU103-BC-6 from China Medical University.
10
11
References
12
13
1.
Harris AL. Hypoxia--a key regulatory factor in tumour growth. Nature reviews
14
Cancer. 2002;2:38-47.
15
2.
16
angiogenesis and radiotherapy response. Nature reviews Cancer. 2008;8:425-37.
17
3.
18
and molecular aspects. Journal of the National Cancer Institute. 2001;93:266-76.
19
4.
Dewhirst MW, Cao Y, Moeller B. Cycling hypoxia and free radicals regulate
Hockel M, Vaupel P. Tumor hypoxia: definitions and current clinical, biologic,
Ruan K, Song G, Ouyang G. Role of hypoxia in the hallmarks of human cancer.
27
1
Journal of cellular biochemistry. 2009;107:1053-62.
2
5.
3
malignant glioma microenvironment: regulation and implications for therapy. Current
4
molecular pharmacology. 2009;2:263-84.
5
6.
Hanahan D, Weinberg RA. The hallmarks of cancer. Cell. 2000;100:57-70.
6
7.
Igney FH, Krammer PH. Death and anti-death: tumour resistance to apoptosis.
7
Nature reviews Cancer. 2002;2:277-88.
8
8.
9
international du cancer. 2009;124:511-5.
Oliver L, Olivier C, Marhuenda FB, Campone M, Vallette FM. Hypoxia and the
Fulda S. Tumor resistance to apoptosis. International journal of cancer Journal
10
9.
Deveraux QL, Reed JC. IAP family proteins--suppressors of apoptosis. Genes &
11
development. 1999;13:239-52.
12
10. Wei Y, Fan T, Yu M. Inhibitor of apoptosis proteins and apoptosis. Acta
13
biochimica et biophysica Sinica. 2008;40:278-88.
14
11. Vucic D, Stennicke HR, Pisabarro MT, Salvesen GS, Dixit VM. ML-IAP, a novel
15
inhibitor of apoptosis that is preferentially expressed in human melanomas. Current
16
biology : CB. 2000;10:1359-66.
17
12. Lin JH, Deng G, Huang Q, Morser J. KIAP, a novel member of the inhibitor of
18
apoptosis protein family. Biochemical and biophysical research communications.
19
2000;279:820-31.
28
1
13. Kasof GM, Gomes BC. Livin, a novel inhibitor of apoptosis protein family
2
member. The Journal of biological chemistry. 2001;276:3238-46.
3
14. Liu B, Han M, Wen JK, Wang L. Livin/ML-IAP as a new target for cancer
4
treatment. Cancer letters. 2007;250:168-76.
5
15. Sarkaria JN, Carlson BL, Schroeder MA, Grogan P, Brown PD, Giannini C, et al.
6
Use of an orthotopic xenograft model for assessing the effect of epidermal growth
7
factor receptor amplification on glioblastoma radiation response. Clin Cancer Res.
8
2006;12:2264-71.
9
16. Hsieh CH, Shyu WC, Chiang CY, Kuo JW, Shen WC, Liu RS. NADPH oxidase
10
subunit 4-mediated reactive oxygen species contribute to cycling hypoxia-promoted
11
tumor progression in glioblastoma multiforme. PloS one. 2011;6:e23945.
12
17. Hsieh CH, Wu CP, Lee HT, Liang JA, Yu CY, Lin YJ. NADPH oxidase subunit 4
13
mediates cycling hypoxia-promoted radiation resistance in glioblastoma multiforme.
14
Free radical biology & medicine. 2012;53:649-58.
15
18. Chou CW, Wang CC, Wu CP, Lin YJ, Lee YC, Cheng YW, et al. Tumor cycling
16
hypoxia induces chemoresistance in glioblastoma multiforme by upregulating the
17
expression and function of ABCB1. Neuro-oncology. 2012;14:1227-38.
18
19. Stupp R, Mason WP, van den Bent MJ, Weller M, Fisher B, Taphoorn MJ, et al.
19
Radiotherapy plus concomitant and adjuvant temozolomide for glioblastoma. The
29
1
New England journal of medicine. 2005;352:987-96.
2
20. Crnkovic-Mertens I, Bulkescher J, Mensger C, Hoppe-Seyler F, Hoppe-Seyler K.
3
Isolation of peptides blocking the function of anti-apoptotic Livin protein. Cellular
4
and molecular life sciences : CMLS. 2010;67:1895-905.
5
21. Hsieh CH, Chang HT, Shen WC, Shyu WC, Liu RS. Imaging the impact of Nox4
6
in cycling hypoxia-mediated U87 glioblastoma invasion and infiltration. Molecular
7
imaging and biology : MIB : the official publication of the Academy of Molecular
8
Imaging. 2012;14:489-99.
9
22. Amberger-Murphy V. Hypoxia helps glioma to fight therapy. Current cancer drug
10
targets. 2009;9:381-90.
11
23. Wagenknecht B, Glaser T, Naumann U, Kugler S, Isenmann S, Bahr M, et al.
12
Expression and biological activity of X-linked inhibitor of apoptosis (XIAP) in human
13
malignant glioma. Cell death and differentiation. 1999;6:370-6.
14
24. Chen Z, Naito M, Hori S, Mashima T, Yamori T, Tsuruo T. A human IAP-family
15
gene, apollon, expressed in human brain cancer cells. Biochemical and biophysical
16
research communications. 1999;264:847-54.
17
25. Chakravarti A, Noll E, Black PM, Finkelstein DF, Finkelstein DM, Dyson NJ, et
18
al. Quantitatively determined survivin expression levels are of prognostic value in
19
human gliomas. Journal of clinical oncology : official journal of the American Society
30
1
of Clinical Oncology. 2002;20:1063-8.
2
26. Jin F, Zhao L, Zhao HY, Guo SG, Feng J, Jiang XB, et al. Comparison between
3
cells and cancer stem-like cells isolated from glioblastoma and astrocytoma on
4
expression of anti-apoptotic and multidrug resistance-associated protein genes.
5
Neuroscience. 2008;154:541-50.
6
27. Jin F, Zhao L, Guo YJ, Zhao WJ, Zhang H, Wang HT, et al. Influence of
7
Etoposide on anti-apoptotic and multidrug resistance-associated protein genes in
8
CD133 positive U251 glioblastoma stem-like cells. Brain research. 2010;1336:103-
9
11.
10
28. Yan B. Research progress on Livin protein: an inhibitor of apoptosis. Molecular
11
and cellular biochemistry. 2011;357:39-45.
12
29. Dong Z, Venkatachalam MA, Wang J, Patel Y, Saikumar P, Semenza GL, et al.
13
Up-regulation of apoptosis inhibitory protein IAP-2 by hypoxia. Hif-1-independent
14
mechanisms. The Journal of biological chemistry. 2001;276:18702-9.
15
30. Chen YQ, Zhao CL, Li W. Effect of hypoxia-inducible factor-1alpha on
16
transcription of survivin in non-small cell lung cancer. Journal of experimental &
17
clinical cancer research : CR. 2009;28:29.
18
31. Li F1, Yin X, Luo X, Li HY, Su X, Wang XY, et al. Livin promotes progression
19
of breast cancer through induction of epithelial-mesenchymal transition and activation
31
1
of AKT signaling. Cellular Signalling. 2013;25:1413-22.
2
32. Crnkovic-Mertens I, Hoppe-Seyler F, Butz K. Induction of apoptosis in tumor
3
cells by siRNA-mediated silencing of the livin/ML-IAP/KIAP gene. Oncogene.
4
2003;22:8330-6.
5
33. Yuan B, Ran B, Wang S, Liu Z, Zheng Z, Chen H. siRNA directed against Livin
6
inhibits tumor growth and induces apoptosis in human glioma cells. Journal of neuro-
7
oncology. 2012;107:81-7.
8
34. Chen SH, Zhaori G. Potential clinical applications of siRNA technique: benefits
9
and limitations. European journal of clinical investigation. 2011;41:221-32.
10
35. Dubrez L, Berthelet J, Glorian V. IAP proteins as targets for drug development in
11
oncology. OncoTargets and therapy. 2013;9:1285-304.
12
36. Darding M, Feltham R, Tenev T, Bianchi K, Benetatos C, Silke J, et al.
13
Molecular determinants of Smac mimetic induced degradation of cIAP1 and cIAP2.
14
Cell death and differentiation. 2011;18:1376-86.
15
16
17
Figure legends
18
19
Figure 1. Tumor hypoxia induces Livin expression. (A) Time course of Livin mRNA
32
1
expression after exposure of GBM8401 cells to <1% O2. Livin mRNA (B) and protein
2
(C) levels in U251, U87 and GBM8401 cells at 24 hours after normoxic (N.) and
3
hypoxic (H.) treatment (<1% O2). (D) Representative images of microscopic
4
GBM8401/hif-1-r. Top left, fluorescence overlay image of perfusion marker, Hoechst
5
33342 (blue), and HIF-1 activation labeling, GFP reporter (green), indicating areas
6
with normoxia (Hoechst 33342+ and GFP-), chronic hypoxia (Hoechst 33342- and
7
GFP+) and cycling hypoxia (Hoechst 33342+ and GFP+). Top right, fluorescence
8
overlay image of Hoechst 33342 (blue), GFP reporter (green), and Livin (red). White
9
box 1 and 2 indicate cycling hypoxic area and chronic hypoxic area, respectively.
10
Bottom left and right, magnification of white box 1 and 2. While color indicates the
11
co-localization of Hoechst 33342, GFP and Livin. Yellow color represents the co-
12
localization of GFP and Livin. (Magnification: Upper, 100X; Lower, 200X) (E) Livin
13
and VEGF mRNA levels in normoxic cells (Hoechst 3342+ and GFP-), chronic
14
hypoxic cells (Hoechst 3342- and GFP+) and cycling hypoxic cells (Hoechst 3342+
15
and GFP+) isolated from disaggregated GBM8401/hif-1-r and U87/hif-1-r xenografts.
16
Error bars denote the standard deviation within triplicate experiments. *P < 0.01
17
compared to normoxia. # P <0.01 compared to chronic hypoxia (F)
18
Immunofluorescence imaging of Livin and HIF-1α expression in primary GBMs.
19
Upper panels show the staining of DAPI (blue), HIF-1α (green), and Livin (red).
33
1
Lower panels demonstrate the overlay image of DAPI (blue), HIF-1α (green), and
2
Livin (red) and indicate the magnification of while boxes. Yellow color indicates the
3
co-localization of HIF-1α and Livin. (Magnification: Upper, 100X; Lower, 100X,
4
200X and 400X)
5
6
Figure 2. Hypoxia-induced Livin expression is dependent on HIF-1α. (A) Verification
7
of HIF-1α or HIF-2α knockdown by HIF-1α or HIF-2α shRNAs. Livin mRNA (B)
8
and protein (C) levels in GBM8401 and U87 with or without HIF-1α or HIF-2α
9
knockdown at 24 hours after hypoxic treatment (<1% O2). (D) Luciferase reporter
10
assay shows that HIF-1α shRNA inhibited the hypoxia-induced transcriptional
11
activation of Livin in U251, U87 and GBM8401 glioblastoma cells. Error bars denote
12
the standard deviation within triplicate experiments. *P < 0.001 compared to
13
normoxia. # P <0.001 compared to scramble (Scr.) shRNA.
14
15
Figure 3. HIF-1α binds directly to the Livin promoter to regulate its expression. (A)
16
Graphic representation of the putative Livin promoter. Two putative hypoxia response
17
elements (HREs) were identified. (B) Luciferase reporter plasmids carrying the wild
18
type or mutant Livin promoter regions were co-transfected with the Renilla luciferase
19
reporter plasmid into GBM8401 cells; and the cells were treated with or without
34
1
hypoxia (<1% O2) for 24 hours. (C and D) Chromatin immunoprecipitation followed
2
by real-time PCR (ChIP-qPCR) assay of HIF-1α binding in Livin promoter in
3
response to hypoxia (<1% O2) for 24 hours. Results are expressed as percentage of
4
input. Error bars denote the standard deviation among triplicate experiments. *P <
5
0.001 compared to control or normoxia. # P <0.001 compared to vehicle or wild type.
6
7
Figure 4. Hypoxia-mediated Livin expression results in antiapoptosis. Verification of
8
Livin knockdown by Tet-regulatable lentiviral knockdown system in mRNA (A) and
9
protein (B) levels. The lentiviral infected cells were treated for 48 hours with Dox
10
(0.04 μg/mL) to induce Livin knockdown. Error bars denote the standard deviation
11
among triplicate experiments. *P < 0.001 compared to no Dox treatment. Caspase-3
12
activities (C) and percent of apoptotic cells (D) in GBM8401 and U87 cells cultured
13
in normoxia, non-interrupted (Ni.) hypoxia (<1% O2) and cycling (Cy.) hypoxia (<1%
14
O2) with or without Livin knockdown in response to staurosporine (STA) treatment.
15
Cells were pretreated for 48 hours with Dox (0.04 μg/ml) to induce Livin knockdown
16
and exposed to hypoxic stresses before STA treatment. Error bars denote the standard
17
deviation among triplicate experiments. *P < 0.01 compared to control. # P <0.05
18
compared to normoxia with STA treatment.
19
35
1
Figure 5. Hypoxia-mediated Livin expression promotes radioresistance and
2
chemoresistance. (A) Radiation cell survival curves for GBM8401 cells. Cells were
3
pretreated for 48 hours with Dox to induce Livin knockdown and exposed to hypoxic
4
stress (<1% O2), either non-interrupted (Ni.) or cycling (Cy.) hypoxia, before
5
irradiation. (B) Cytotoxicity assay of Livin knockdown revealed increased
6
cytotoxicity of cells to TMZ and suppressed hypoxia-mediated TMZ resistance in
7
GBM8401 and U251 glioblastoma cells. *P < 0.01 compared to normoxia. # P <0.01
8
compared to untreated Dox groups. (C) Sequences of TAT-Lp15 amino acids and TAT-
9
K22 control peptide. (D) Visualization of intracellular accumulation of TAT-Lp15-
10
dansyl or TAT-K22-dansyl (10 μM) at 60 minutes after incubation to GBM8401 cells.
11
(E) Time course of TAT-Lp15-dansyl (10 μM) fluorescence after incubation to
12
GBM8401 cells. Fluorescence intensities were determined by flow cytometry.
13
Surviving fraction (F) and cell viability (G) for U251, U87 and GBM8401 cells with
14
or without TAT-K22 or TAT-Lp15 (30μM) and receiving hypoxic stress (H) (<1% O2)
15
followed by 4-Gy irradiation (R) or 100-μM TMZ (T). Error bars denote the standard
16
deviation among triplicate experiments. *P < 0.001 compared to control. # P <0.001
17
compared to radiation or TMZ alone.
18
radiation or TMZ groups.
※
P <0.001 compared to hypoxia combined
19
36
1
Figure 6. Livin blockage enhances the efficiency of radiation plus TMZ treatment in
2
glioblastoma xenografts. (A) Detection of TAT-Lp15-dansyl and TAT-K22-dansyl in
3
the cortex of nude mouse brain 1 hour after intravenous injection. Bars=100 μm (B)
4
Bioluminescent images from control and treated animals including the TAT-K22 or
5
TAT-Lp15 treatment, combination of TMZ (T) and radiation (R) treatment,
6
pretreatment of TAT-K22 plus the combination of TMZ and radiation treatment and
7
pretreatment of TAT-Lp15 plus the combination of TMZ and radiation treatment on
8
day 20 after tumor implantation. (C) The mean normalized BLI values associated with
9
longitudinal monitoring of intracranial tumor growth for each treatment group. Error
10
bars denote the standard deviation among 9 mice per group. (D) The corresponding
11
survival curves of GBM8401 xenograft-bearing mice for each treatment group. (E)
12
Representative sections showing apoptotic cells measured by the TUNEL assay of
13
GBM8401 xenografts at 2 days after treatments. Each mean is based on 15
14
measurements (5 measurements for each of three tumors).
15
(F) The percentages of apoptotic cells on cross-sections. *P < 0.01 compared to
16
control. #P <0.001 compared to control TAT-K22 peptide. (G) The percentages of
17
apoptotic cells in tumor hypoxic cells within GBM8401 xenografts determined by
18
pimonidazole/annexin V doubling staining flow cytometry assay. *P < 0.001
19
compared to control. #P <0.01 compared to control TAT-K22 peptide. Surviving
37
1
fraction (H) and colony images (I) in excised GBM8401 and U251 xenografts after
2
treatments. Error bars denote the standard deviation among triplicate experiments. *P
3
< 0.001 compared to combination of TMZ and radiation treatment.
38