Targeted endostatin-cytosine deaminase fusion gene therapy plus 5-fluorocytosine
suppresses ovarian tumor growth
Yuh-Pyng Sher1,3, Chun-Mien Chang1, Chiun-Gung Juo5, Chun-Te Chen7, Jennifer L.Hsu7,8,
Chen-Yuan Lin2,3, Zhenbo Han7 , Shine-Gwo Shiah6, and Mien-Chie Hung1,4,7,8
1
Center for Molecular Medicine, 2Division of Hematology and Oncology, China Medical University
Hospital, Taichung 404, Taiwan
3
Graduate Institute of Clinical Medical Science, 4Graduate Institute for Cancer Biology, China
Medical University, Taichung 404, Taiwan
5
Molecular Medicine Research Center, Chang Gung University, Tao-Yuan 333, Taiwan
6
National Institute of Cancer Research, National Health Research Institutes, Miaoli 350, Taiwan
7
Department of Molecular and Cellular Oncology, The University of Texas MD Anderson Cancer
Center, Houston, Texas 77030, USA
8
Asia University, Taichung 413, Taiwan
Running Title: SV-hEndoyCD suppresses ovarian tumor growth
Keywords: Ovarian cancer, anti-angiogenesis, VISA, gene therapy
Abbreviations: hEndoyCD, human endostatin-yeast cytosine deaminase fusion protein; 5-FC,
5-fluorocytosine; 5-FU, 5-fluorouracil; Luc, luciferase; SV, Survivin-VISA; VISA,
VP16-GAL4-WPRE integrated systemic amplifier
*Correspondence should be addressed to Mien-Chie Hung, Department of Molecular and
Cellular Oncology, University of Texas MD Anderson Cancer Center, 1515 Holcombe
Boulevard, Houston, Texas 77030. Phone: 713-792-3668; Fax: 713-794-0209. E-mail:
mhung@mdanderson.org
1
Abstract
There are currently no effective therapies for cancer patients with advanced ovarian
cancer, so developing an efficient and safe strategy is urgent. To ensure cancer-specific
targeting, efficient delivery, and efficacy, we developed an ovarian cancer-specific construct
(Survivin-VISA-hEndoyCD) composed of the cancer specific promoter survivin in a
transgene amplification vector (VISA) to express a secreted human endostatin-yeast cytosine
deaminase fusion protein (hEndoyCD) for advanced ovarian cancer treatment. hEndoyCD
contains an endostatin domain that has tumor targeting ability for anti-angiogenesis and a
cytosine deaminase domain that converts the prodrug 5-fluorocytosine (5-FC) into the
chemotherapeutic drug, 5-fluorouracil.
Survivin-VISA-hEndoyCD was found to be highly
specific, selectively express secreted hEndoyCD from ovarian cancer cells, and induce cancer
cell killing in vitro and in vivo in the presence of 5-FC without affecting normal cells. In
addition, Survivin-VISA-hEndoyCD plus 5-FC showed strong synergistic effects in
combination with cisplatin in ovarian cancer cell lines. Intraperitoneal treatment with
Survivin-VISA-hEndoyCD coupled with liposome attenuated tumor growth and prolonged
survival in mice bearing advanced ovarian tumors. Importantly, there was virtually no severe
toxicity when hEndoyCD is expressed by Survivin-VISA plus 5-FC compared with CMV
plus 5-FC. Thus, the current study demonstrates an effective cancer-targeted gene therapy
that is worthy of development in clinical trials for treating advanced ovarian cancer.
2
Introduction
Ovarian cancers is the fifth leading of cancer deaths in women and associated with the
highest rate of mortality in patients with advanced-stage, high-grade serous ovarian cancer
with a 5-year survival at about 30% due to fact that most ovarian cancer patients are
diagnosed with an advanced disease as there are no clinical symptoms in the early stages (1).
Standard care for ovarian cancer is surgery and adjuvant chemotherapy, but drug-resistant
cancer recurs in about 25% patients within 6 months (2). Because currently no effective
therapy is available for patients with advanced ovarian cancer, the development of new and
effective methods for treatment of advanced ovarian cancer is necessary.
Targeted gene therapy is an attractive strategy due to the benefit of tumor-specific
expression of therapeutic genes (3). Survivin is upregulated in most human tumors including
those of the lungs and ovaries (4-6) and involved in cancer progression and treatment
resistance (7). Recently, a study from the Cancer Genome Atlas (TCGA) reported that p53 is
mutated in 95% of high-grade serous ovarian adenocarcinomas (8) . Since wild type p53 is
known to negatively regulate the survivin promoter, but not mutant p53 (9), we hypothesized
that the survivin promoter may serve as a cancer-specific promoter for ovarian cancer gene
therapy. Using a similar approach from a previously developed lung cancer gene therapy (10),
we constructed a Survivin-VISA-based expression vector for ovarian cancer. The VISA
(VP16-GAL4-WPRE integrated systemic amplifier) transgene expression platform is
comprised of a two-step amplification system (11, 12) and a woodchuck hepatitis virus
posttranscriptional regulatory element (WPRE). The promoter drives the expression of
VP-16-GAL4, a synthetic transcriptional activator, which then binds to its responsive element
(G5E4T) to drive expression of the transgene. The addition of the WPRE element was shown
to increase transgene expression (13, 14). As demonstrated by several studies, the VISA
system enhances transgene expression to a level comparable to that of CMV in cancer cells
but remains inactive in normal tissues (10, 15-17).
3
Because ovarian cancer growth and peritoneal dissemination are angiogenesis-dependent
(18, 19), angiogenic inhibitors may be a promising therapeutic approach to suppress tumor
growth by blocking formation of new blood vessels. Endostatin, a 20 kDa fragment of
collagen XVIII, is an endogenous antiangiogenic protein which functions by inhibiting
endothelial cell proliferation and migration, inducing endothelial cell apoptosis (20, 21),
inhibiting MMP-2 activity, blocking the binding of VEGF, and stabilizing cell-cell and
cell-matrix adhesions to prevent the loosening structures required during angiogenesis (22,
23). Although clinical trials using endostatin in cancer patients have only had sporadic
positive results (24), endostatin plays an important role in antiangiogenesis.
To further improve endostatin for higher therapeutic efficacy, we selectively linked a
prodrug conversion system to endostatin to establish highest anti-tumor activity in the tumor
microenvironment and low toxicity in normal tissues in the presence of a prodrug. It has
previously been shown that endostatin selectively targets neovascular endothelial cells,
suppressing tumor growth (25). In this report, we generated a secreted form of the human
endostatin-yeast cytosine deaminase fusion protein (hEndoyCD) that targets neovascular
endothelial cells (by hEndo) and converts the pro-drug, 5-fluorocytosine (5-FC), into the
chemotherapeutic drug, 5-fluorouracil (5-FU), at the tumor site (by yCD). To boost the tumor
specific targeting effect, we selectively expressed hEndoyCD fusion protein by integrating it
into the Survivin-VISA vector to generate Survivin-VISA-hEndoyCD (SV-hEndoyCD). Our
study demonstrates that hEndoyCD gene therapy plus 5-FC has significant anti-cancer and
tumor-specific targeting effects without systemic toxicity in normal. Thus, the current study
provides a promising strategy worthy of further development in clinical trials for treating
advanced ovarian cancer by a cancer-targeted gene therapy.
4
Results
Survivin-VISA retains its specificity towards ovarian cancer cells and prolongs
transgene expression
To investigate whether the survivin promoter is suitable as a universal cancer promoter
for cancer targeted gene therapy in ovarian cancer treatment, three promoters, survivin,
survivin combined with VISA system (Survivin-VISA), and CMV were used to drive
luciferase expression to detect their specificity (Fig. 1a). Although the survivin promoter was
active in ovarian cancer cells (P = 0.16), its activity was much weaker than that of the CMV
promoter (Fig. 1b). While the survivin promoter activity was greatly enhanced in the VISA
vector (Survivin-VISA or SV) compared with the survivin promoter alone and even showed
stronger reporter activity that is comparable to or higher than the CMV promoter in most of
the ovarian cancer cell lines (Fig. 1b), the reporter activity of VISA vector remained weak in
several normal cell lines tested (p = 0.03), indicating that the SV vector is selectively
enhanced for transgene expression in ovarian cancer cells but not in normal cells.
Interestingly, p53 is mutated in about 95% of high-grade serous ovarian adenocarcinomas (8).
Since p53 has been described to negatively regulate the survivin promoter and that mutant
p53 cannot inhibit the survivin promoter (9), we proposed that survivin could serve as a
cancer-targeted promoter to drive therapeutic gene in ovarian cancer.
To further measure the kinetics of transgene expression driven by the SV vector in
ovarian cancer cells, the luciferase activity in PA1 cells after transient transfection was
measured daily for five days. The duration of luciferase gene expression was significantly
prolonged in PA1 cells transfected with SV (Fig. 1c). We calculated the area under the
luciferase activity curve, which indicates the total expression index (TEI). By setting TEI for
CMV promoter as 1, we found that the TEI for SV was 7.5-fold higher, suggesting that the
5
SV vector has stronger and more stable reporter activity than CMV in PA1 cells.
Cancer-specific targeting of Survivin-VISA in vivo
Because metastatic ovarian cancers are nearly always confined to the peritoneal cavity,
the 2009 National Comprehensive Cancer Network (NCCN) Practice Guidelines in Oncology
for stage II-IV ovarian cancer recommended intraperitoneal (i.p.) chemotherapy for patients
who were optimally debulked (< 1 cm) or intravenous (i.v.) taxane/carboplatin for a total of
6–8 cycles (26, 27). To investigate whether the Survivin-VISA vector still retains activity and
specificity through i.p. delivery in vivo, the CMV-Luc, and Survivin-VISA-Luc (SV-Luc)
constructs were coupled with liposome for delivery to tumor-free (without tumor injection) or
tumor-bearing (with tumor injection) through i.p. injection (28). Bioluminescent imaging
revealed strong luciferase signals in the abdomen of mice treated with CMV-Luc construct in
both tumor-free and tumor-bearing mice. In contrast, the signals were detected primarily in
tumor-bearing mice treated with SV-Luc construct but low or undetectable in tumor-free mice
(Fig. 2a). To precisely identify the source of the signals, mice were sacrificed immediately
after in vivo imaging, and their organs were dissected and proteins extracted to measure ex
vivo luciferase activity. In the CMV-Luc group, luciferase activity was high in the tumor, liver,
spleen, and colon, whereas in the SV-Luc group, the activity was high in tumors, but very low
in all other organs, which reflect our observations from in vivo imaging (Fig. 2b). These
results demonstrate that the transgene expression in the SV vector is more cancer specific
than the CMV vector and suggest that this vector may be suitable for gene therapy.
hEndoyCD expressed from Survivin-VISA retains endostatin activity and is active in
5-FC to 5-FU conversion
To examine the possibility of using human endostatin-yeast cytosine deaminase
(hEndoyCD)
fusion
protein
for
cancer-targeted
6
gene
therapy,
we
constructed
CMV-hEndoyCD and Survivin-VISA-hEndoyCD (SV-hEndoyCD) plasmids. The vector
alone without VISA and hEndoyCD gene was used as the control (Ctrl). The expression of
hEndoyCD driven by CMV or SV after transient transfection was detected in PA1 cell lysates
by immunoblotting using an anti-endostatin antibody (Fig. 3a). The secreted fusion proteins
from the media were also detected by endostatin ELISA kit (data not shown). In addition, to
detect whether secreted hEndoyCD retains its antiangiogenic activity, the media of PA1 cells
from transient transfection of the indicated plasmids were used for endothelial tube assays
with human umbilical vein endothelial cells (HUVEC). As shown in Figure 3b, tube
formation was significantly suppressed in the presence of media from cells transfected with
CMV-hEndoyCD and SV-hEndoyCD but not the control vector.
Next, we examined whether hEndoyCD retains its enzymatic activity in converting 5-FC
prodrug to cytotoxic 5-FU. Cells were transiently transfected with indicated plasmids, and
5-FC was added one day later. We used LC/MS/MS to detect 5-FU in culture media of PA1
cells (Fig. 3c). The results demonstrated that both CMV- and SV-hEndoyCD successfully
converted 5-FC to 5-FU compared with no 5-FU detection in cells transfected with the
control vector. In addition, the concentration of 5-FU from culture media in SV-hEndoyCD
group was higher than in the CMV-hEndoyCD group, suggesting it would have a stronger
“bystander effect”, affecting not just the cells transduced with therapeutic gene but also the
untransfected neighboring cells. We further determined the cytotoxic activities of the
indicated plasmids plus 5-FC treatment in PA1 cells by crystal violet staining to measure
whole cell viability (Fig. 3d). In the control group without DNA transfection, 5-FC treatment
alone in PA1 cells did not cause any cell death. We normalized the cell viability of transfected
cells to the control group and found that both CMV-hEndoyCD and SV-hEndoyCD
demonstrated higher killing effect in ovarian cancer cell lines compared with the normal cell
lines (P < 0.001). However, CMV-hEndoyCD had stronger killing effect than SV-hEndoyCD
in normal cell lines (Fig. 3d). According to the results shown in Fig. 3c, the higher 5-FC
7
conversion to 5-FU in the media of the SV-hEndoyCD group should have exerted stronger
cytotoxicity than the CMV-hEndoyCD group in PA1 cells for better bystander effect.
However, the results showed similar percentage of survival in Fig. 3d. To further investigate
whether it was caused by saturated 5-FC concentration, we tested 5-FC in a dose-dependent
manner in PA1 cells transfected with the indicated plasmids using different concentrations of
5-FC and measured the percent of survival. Figure 3e shows that the percent of survival in
SV-hEndoyCD group was similar at 5-FC concentrations ranging from 0.0078 mg/ml
(64-fold less) to 0.5 mg/ml. In contrast, CMV-hEndoyCD treatment at low concentrations of
5-FC (0.0078 and 0.002 mg/ml) had less therapeutic effect compared with SV-hEndoyCD (P
< 0.01) (Fig. 3e). Taken together, these results indicate that the hEndoyCD driven by the
Survivin-VISA vector contains both antiangiogenic activity and 5-FC prodrug-converting
enzyme activity to selectively kill ovarian cancer cells in vitro
hEndoyCD gene therapy exhibits synergistic cytotoxicity in combination with cisplatin
While the combination of 5-FU and cisplatin is widely used in the clinic for human
carcinomas, especially for head and neck cancer (29), such chemotherapeutic regimens have
seldom been reported to be effective against advanced ovarian cancer (30). To investigate
whether there is a synergistic effect when SV-hEndoyCD is combined with chemotherapy,
we first examined the effect of 5-FU and cisplatin in PA1 ovarian cancer cells. The
combination index (CI), which determines the degree of synergism for treatment using two
drugs, was calculated by a commercial analysis program, CalcuSyn. A CI of <1, 1, or >1 is
indicative of synergistic, additive, or antagonistic effects, respectively. As shown in Figure 4a,
we first measured the CI values for the combination treatment of 5-FU and cisplatin in
ovarian cancer cells. This combination demonstrated higher cytotoxic effect than single drug
treatment (upper panel, Fig. 4a). All of the CI values at different effective dose (ED) were
among 0.1 to 0.2, indicating a strong synergistic interaction with combination treatment. In
8
addition, the median-effective dose (Dm) that causes 50% cytotoxicity was at 0.24 M of
5-FU, whereas the two-drug combination reached the same effect at 20-fold less of 5-FU
(lower panel, Fig. 4a). We then tested the combination of SV-hEndoyCD/5-FC and cisplatin.
Similar to the combination of 5-FU and cisplatin, higher cytotoxic effect was observed in the
combination than in the single drug treatment, which had CI values that ranged from 0.4 to
0.5. Again, the Dm of 5-FC dose in SV-hEndoyCD/5-FC treatment was 3.47 M, which was
reduced to 0.95 M in the combination treatment, indicating a synergistic effect for
SV-hEndoyCD/5-FC and cisplatin combination (Fig. 4b).
Survivin-VISA-hEndoyCD inhibits tumor growth and prolongs survival in mouse
xenograft models
To evaluate the antitumor effects of SV-hEndoyCD in human ovarian cancer in vivo, we
established a metastatic ovarian tumor animal model by inoculating human ovarian cancer
cells, SKOV3ip1-luc (16) in nude mice by i.p. injection. After tumors developed, we then
treated mice with i.p. delivery of liposomal DNA complex and 5-FC and monitored the
luciferase signals using IVIS imaging system. The signal increased much more slowly in
mice treated with CMV-hEndoyCD and SV-hEndoyCD than the control vector (p < 0.05),
indicating that both CMV-hEndoyCD and SV-hEndoyCD inhibited tumor growth (Fig. 5a).
SV-hEndoyCD also significantly prolonged the survival time (p = 0.01) whereas
CMV-hEndoyCD prolonged the survival time but not significantly compared with the control
(p = 0.12) (Fig. 5b). To determine the tumor cytotoxic effect of hEndoyCD in vivo, we
intratumorally delivered the DNA-liposome complex and 5-FC by to treat mice bearing
subcutaneous (s.c.) PA1 tumors. Interestingly, the tumor growth was significantly suppressed
in the SV-hEndoyCD group compared with the control and CMV-hEndoyCD groups (Fig. 5c).
To further investigate the therapeutic effect of SV-hEndoyCD in a metastatic ovarian tumor
animal model, mice were inoculated with PA1 cells and then treated with DNA-liposome
9
complex by i.p. injection and i.p. delivery of 5-FC. Similar to the results shown in Figure 5c,
i.p. tumor from SV-hEndoyCD/5-FC-treated group was significant smaller than in
CMV-hEndoyCD/5-FC-treated group (Fig. 5d). In addition, we observed higher hEndoyCD
expression in mice tumor tissues from the SV-hEndoyCD group than that from
CMV-hEndoyCD group in which hEndoyCD expression was only detectable in one of the
two mice tumor tissue samples two days after the DNA treatment. We also detected a higher
amount
of
5-FU
in
the
tumor
tissues
from
SV-hEndoyCD/5-FC-
than
CMV-hEndoyCD/5-FC-treated mice (Fig. 5e; 44 ng/g vs. 7.8 ng/g of tissue, respectively).
These results indicate that SV-hEndoyCD/5FC treatment has therapeutic benefit for advanced
ovarian cancer treatment.
SV-hEndoyCD plus 5-FC treatment has no severe acute toxicity
Next, we wanted to compare the safety profile between SV-hEndoyCD and
CMV-hEndoyCD treatment. A single high dose of 50 g plasmid DNA was injected through
tail vein of Balb/c mice with i.p. delivery of 5-FC one day later. The serum levels of aspartate
aminotransferase (AST) and alanine aminotransferase (ALT) were measured by liver function
assays. In the CMV-hEndoyCD treatment group, the serum levels of AST and ALT were
significantly higher than either SV-hEndoyCD or control group but decreased by Day 2 (Fig.
6a). In addition, one mouse with the highest concentration of AST and ALT in the
CMV-hEndoyCD group died on day 1 while all animals in the control and SV-hEndoyCD
groups survived (Fig. 6b). These results indicate that the SV-hEndoyCD treatment is safer
than the CMV-hEndoyCD.
Discussion
In this study, we showed that hEndoyCD has dual functions in anti-angiogenesis and
conversion of prodrug 5-FC to toxic drug 5-FU in ovarian cancer cells. Most current
10
chemotherapies use non-discriminatory approaches that kill both cancer cells and
non-cancerous surrounding cells. To avoid the non-specific treatment and achieve safer
approach for treating advanced ovarian cancer, we used a previously established the
cancer-targeted therapeutic vector SV-VISA to express hEndoyCD, a fusion protein can
selectively target to endothelial cells in vitro and to the tumor site in vivo. These effects
would reduce the nonspecific distribution of expressed hEndoyCD for less toxicity in normal
tissue as targeted expression of hEndoyCD ensures 5-FC is converted to 5-FU only at the
tumor site. Moreover, because hEndoyCD can be secreted from the cells, it also targets
surrounding endothelial cells in tumor vasculature and induce both endothelial and cancer
cell apoptosis. It has previously been shown that the number of endothelial cells in tumor
tissues is greatly reduced in the endostain-cytosine deaminase-treated group than in the
control group by using CD31-positive staining (25).
Several studies have focused on the improvement of the CD/5-FC system such as by
replacing the bacterial-derived (31, 32) with yeast-derived CD domain to enhance the activity
of 5-FC/5-FU conversion. In addition, the combination of yeast CD and uracil
phosphoribosyltransferase (UPRT) genes efficiently catalyzes the direct conversion of 5-FC
into the toxic metabolites 5-FU and 5-fluorouridine-5’-monophosphate to bypass the natural
resistance of certain human tumor cells to 5-FU (33). Such modifications have been used in
our previous study (25, 34). In targeted gene therapy system, however, there was no
significant improvement using the modified CD/UPRT, which might be attributed to the
decreased transfection efficiency of a larger plasmid (Supplementary Fig. 1). Therefore, in
this study, only the yeast CD was used in the hEndoyCD expression construct.
Because the development of peritoneal metastases in malignant ascites is
angiogenesis-dependent and a main cause of morbidity in ovarian cancer (18, 19, 35),
inhibition of angiogenesis appears to be a promising strategy to complement existing
treatment methods. Recently, a mutant endostatin with P125A substitution was reported with
11
better binding to the endothelial surface and increased antiangiogenic activity compared with
the native protein (36). The mutant P125A-endostatin domain of hEndoyCD was also used in
our study, but such modification did not improve antiangiogenic activity significantly in vitro
(data not shown).
Targeted gene therapy driven from CMV and SV were assessed in two different cancer
cells established xenograft models with ip or sc tumor inoculation and treatment (Fig. 5). The
SV-hEndoyCD/5-FC
treatment
had
better
therapeutic
effect
than
Ctrl
and
CMV-hEndoyCD/5-FC in all animal models. However, in SKOV3-ip1 xenograft model,
CMV-hEndoyCD/5-FC treatment showed tumor killing effect during therapeutic period (Fig.
5a), whereas no therapeutic effect in both PA1 xenograft models, ip and sc (Figs. 5c and d). It
could be caused by variables such as differences in cancer cell characteristics. In addition, the
transfection efficiency of plasmids in the SKOV3-ip1 cells was six-fold higher than in the
PA1 cells, suggesting better benefits of gene therapy to express transgene for SKOV3-ip1
cells than PA1 cell treatment, even under the CMV promoter. The tumor growth inhibition
effects in the xenograft models were more apparent than what we observed in cell lines. Since
hEndoyCD driven from SV vector is continually expressed to convert pro-drug 5-FC to
cytotoxic 5-FU in the tumor tissues, the in vivo model allows for longer therapeutic benefits
from the treatment which could not be observed in a short treatment widow in the in vitro cell
system.
Although 5-FU is not a standard regimen for ovarian cancer, our study demonstrate that
5-FU converted from hEndoyCD/5-FC treatment provides promising therapeutic effects in
ovarian cancer cells in vitro and in vivo and has synergistic treatment effect with cisplatin in
vitro. Taken together, our study demonstrates that cancer-targeted gene therapy such as
SV-hEndoyCD may be considered as an alternative treatment option for advanced ovarian
cancer.
12
Materials and Methods
Cell Lines. Human ovarian cancer cell lines (2774C10 and HeyA8), immortalized normal
mammary epithelial (184A1 and MCF-10A), normal lung fibroblasts (WI-38), and
immortalized normal lung epithelial (HBE4-E6/E7) were obtained from the American Type
Culture Collection (Manassas, VA). Immortalized normal lung epithelial cells (HBEC-3KT)
were kindly provided by Dr. John D. Minna (37). NOE99 and NOE115 were primary cultures
of normal ovarian epithelium cells (16). PLC14F and BMF were primary culture of human
lung fibroblast cells in our laboratory. All primary cultures were obtained under protocols
approved by the institutional review board. Other human ovarian cancer cell lines, ES2,
MDAH2774, PA1, A59, and A59-4, were kindly provided by the National Health Research
Institutes in Taiwan. SKOV3.ip1 was established as previously described (16, 38). Before use,
all cell lines were tested and found to be free of Mycoplasma infection.
Evaluation of Promoter Activity and Cytotoxicity. Survivin-VISA plasmid used in this study
was constructed as previously described (10). To test the promoter activity, we used a
Dual-Luciferase reporter assay to normalize transfection efficiency according to the
manufacturer’s instructions (Promega, Madison, WI). To investigate the promoter specificity
in vivo, the tissues in the abdomen of mice (liver, spleen, colon, intestine and tumor) were
dissected and extracted in Passive Lysis Buffer for luciferase assay. To construct the
therapeutic plasmid, fusion protein coding sequence was first cloned into the pET28a(+)
expression vector by directly amplifying the cytosine deaminase gene (from Saccharomyces
cerevisiae) fused to the C-terminus of human endostatin. We then added immunoglobulin
light chain signal peptides to the N-terminus of this fusion gene to express a protein that can
be secreted. The human endostatin-yeast cytosine deaminase fusion gene (hEndoyCD) was
further constructed into a vector containing kanamycin resistant gene to generate
13
pUK21-CMV-hEndoyCD and pUK21-Survivin-VISA-hEndoyCD. Secreted fusion proteins
in the media of cells transfected with these plasmids can be detected by the Human
Endostatin Immunoassay (R&D systems, MN). To measure the cytotoxicity of hEndoyCD in
the presence of 5-FC, hEndoyCD-transfected cells (on Day 1) were treated with 0.5 mg/ml
5-FC one day later (on Day 2). The percentage of cells that survived after the treatment of
hEndoyCD in the presence of 5-FC was performed by crystal violet staining on day 5 in by
fixing cells with 1% glutaraldehyde for 30 min, staining with 0.5% crystal violet for 30 min,
washing with water, and allowing to air-dry. Subsequently, Sorenson's solution (30 mM
trisodium citrate, 00.6% HCl, and 47.5% ethanol) was added to elute the dye, and the optical
density was read on a microplate autoreader (Bio-Tek Instruments, Winooski, VT) at 570 nm.
Ovarian Cancer Animal Model. Nude mice were purchased from National Laboratory Animal
Center (Taiwan) at 6-8 weeks of age. For the subcutaneous ovarian cancer model, PA1 cells
(1 x 107) were injected subcutaneously at a single site on the right flank of the animals. The
presence of subcutaneous tumors in animals was calculated by using the formula V (mm3) = a
x b2/2, where a is the largest dimension and b is the perpendicular diameter. For metastatic
ovarian cancer models, SKOV3-ip1 (1x106) or PA1 (5x106) human ovarian cancer cells were
inoculated in nude mice by intraperitoneal (i.p.) injection. Therapeutic plasmids preparation
for animal treatment was as previously described (10). The treatment was repeated twice
weekly for 3 weeks. The day after the DNA-liposome complex injection, 5-FC (500 mg/kg)
was administered to mice by i.p. injection. All animal experiments were carried out under
protocols approved by the Institute Animal Care and Use Committee of China Medical
University and Hospital.
IVIS and Quantification. In vivo imaging was carried out as previously described (10). To
detect promoter specificity, promoter constructs were complexed with liposome (HLDC) for
14
delivery to mice by i.p. injection. The luciferase activity was imaged by IVIS Imaging system
(Xenogen, Alameda, CA) two days after DNA/liposome injection and analyzed using Living
Imaging software.
Endothelial Tube Assay. The detailed procedures have been previously described (25). In
brief, a suspension of 5x103 HUVEC cells was seeded into a Matrigel-coated 96-well plate
and cultured with concentrated conditional media collected from PA1 cells transfected with
the indicated plasmids. All assays were performed in triplicate. Five fields were viewed to
calculate the number of tubes formed.
Detection of 5-FU by LC/MS/MS. 5-FU was detected by using hydrophilic interaction liquid
chromatography (HILIC) coupled with mass spectrometry (39). The UPLC-tandem mass
spectrometry (LC/MS/MS) system consisted an Acquity (Waters, Milford, MA, USA)
coupled to an AB Sciex QTRAP 5500 mass spectrometer (AB Sciex, Foster City, CA, USA)
equipped with an electrospray ion source for ion production. Data acquisition and integration
were controlled by Analyst® Software. The chromatographic separation was performed on a
150 mm x 2.1 mm, 1.8 m HILIC column (waters) maintained at 45oC. The flow rate was 0.4
ml/min. The mobile phase A was water with 10 mM ammonium formate and mobile phase B
was acetonitrile. The linear gradient of phase B was decreased from 92% to 85% in 1 min,
and to 80% in 1 min then to 75% in 1 min, and then back to 92% in 0.5 min and kept for 3.5
min for equilibrium. The MRM experiments were conducted by monitoring the precursor ion
to product ion transitions in the negative ion mode for 5FU from m/z 129.0 (Q1) to m/z 42.0
(Q3).
To determine the concentration of converted 5-FU in PA1 cells transfected with
indicated plasmids with 5-FC added one day post transfection, cell lysates were extracted
using 200 l of methanol containing 5-bromouracil (BrU; 100 ng/ml) as internal standard.
15
After homogenization, samples were ultracentrifuged and supernatant collected for detection.
For 5-FU detection in culture media, 25 l of cultured media was mixed with 200 l of
acetonitrile containing BrU. After centrifugation, the supernatant was collected for detection.
5-FU standard solutions were prepared in methanol ranging from 0.01 to 100 ng/ml.
Synergistic analyses. The synergism between two drugs was quantified by combination index
(CI) using the CalcuSyn program (Biosoft, Cambridge, United Kingdom). A CI of <1, 1, or
>1 is indicative of synergistic, additive, or antagonistic effects, respectively (40). To calculate
the CI, constant ratio combination design is performed by choosing the two drugs at their
equipotent ratio (i.e., at the ratio of their IC50) and then fixing the ratio for drugs
combination. After the ratio is set, the mixture of the two drugs is serially diluted by 4-fold in
a constant ratio to obtain a good dosage range. Various doses of 5-FU and cisplatin were
combined at a constant ratio of 1:1.08 to generate dose-response curves of fraction affected
(cytotoxicity) to evaluate the effects of the drug combinations by MTT assay. In
SV-hEndoyCD/5-FC and cisplatin combination assays, SV-hEndoyCD plasmids were
transiently transfected in ovarian cancer cells, and one day later, various doses of 5-FC and
cisplatin were added to cells in a fixed ratio (9.3:1).
Acute toxicity analysis. The assay was performed as previously described (10). Briefly, to
detect the effect of systemic administration of DNA-liposome complex in high doses, female
Balb/c mice received 50 g of plasmid coupled with DNA by i.v. injection. Blood samples
from mice were analyzed for AST, and ALT levels by assay kits (Roche, Mannheim,
Germany).
Statistical Analysis. Student’s t test was used to compare the difference between two groups
and all statistical tests were two sided. Survival curves were obtained by the Kaplan-Meier
16
method. The difference of survival time between two groups was analyzed with the log-rank
test. The significance level was set at P < 0.05.
Conflicts of interest
All authors have no conflicts of interest.
Acknowledgements
This
work
was
supported
by
grants
from
National
Science
Council
(NSC100-3112-B-039-003), Private University (NSC99-2314-B-039-029-MY3 to Y.-P.S. and
NSC99-2632-B to M.-C.H.), Department of Health Cancer Research Center of Excellence
(DOH99-TD-C-111-005), and The Sister Institution Fund of China Medical University and
Hospital and The University of Texas MD Anderson Cancer Center.
17
References
1. Jemal A, Siegel R, Xu J, Ward E. Cancer statistics, 2010. CA Cancer J Clin. 2010; 60:
277-300.
2. Miller DS, Blessing JA, Krasner CN, Mannel RS, Hanjani P, Pearl ML, et al. Phase II
evaluation of pemetrexed in the treatment of recurrent or persistent platinum-resistant ovarian
or primary peritoneal carcinoma: a study of the Gynecologic Oncology Group. J Clin Oncol.
2009; 27: 2686-91.
3. Gunther M, Wagner E, Ogris M. Specific targets in tumor tissue for the delivery of
therapeutic genes. Curr Med Chem Anticancer Agents. 2005; 5: 157-71.
4. Altieri DC. Validating survivin as a cancer therapeutic target. Nat Rev Cancer. 2003; 3:
46-54.
5. Li F, Ling X. Survivin study: an update of "what is the next wave"? J Cell Physiol. 2006;
208: 476-86.
6. Sui L, Dong Y, Ohno M, Watanabe Y, Sugimoto K, Tokuda M. Survivin expression and
its correlation with cell proliferation and prognosis in epithelial ovarian tumors. Int J Oncol.
2002; 21: 315-20.
7. Zaffaroni N, Pennati M, Daidone MG. Survivin as a target for new anticancer
interventions. J Cell Mol Med. 2005; 9: 360-72.
8. Bell D, Berchuck A, Birrer M, Chien J, Cramer DW, Dao F, et al. Integrated genomic
analyses of ovarian carcinoma. Nature. 2011; 474: 609-15.
9. Mirza A, McGuirk M, Hockenberry TN, Wu Q, Ashar H, Black S, et al. Human survivin
is negatively regulated by wild-type p53 and participates in p53-dependent apoptotic pathway.
Oncogene. 2002; 21: 2613-22.
10. Sher YP, Tzeng TF, Kan SF, Hsu J, Xie X, Han Z, et al. Cancer targeted gene therapy of
BikDD inhibits orthotopic lung cancer growth and improves long-term survival. Oncogene.
2009; 28: 3286-95.
11. Iyer M, Wu L, Carey M, Wang Y, Smallwood A, Gambhir SS. Two-step transcriptional
amplification as a method for imaging reporter gene expression using weak promoters. Proc
Natl Acad Sci U S A. 2001; 98: 14595-600.
12. Zhang L, Johnson M, Le KH, Sato M, Ilagan R, Iyer M, et al. Interrogating androgen
receptor function in recurrent prostate cancer. Cancer Res. 2003; 63: 4552-60.
13. Glover CP, Bienemann AS, Heywood DJ, Cosgrave AS, Uney JB. Adenoviral-mediated,
high-level, cell-specific transgene expression: a SYN1-WPRE cassette mediates increased
transgene expression with no loss of neuron specificity. Mol Ther. 2002; 5: 509-16.
14. Zufferey R, Donello JE, Trono D, Hope TJ. Woodchuck hepatitis virus
posttranscriptional regulatory element enhances expression of transgenes delivered by
retroviral vectors. J Virol. 1999; 73: 2886-92.
18
15. Lang JY, Hsu JL, Meric-Bernstam F, Chang CJ, Wang Q, Bao Y, et al. BikDD eliminates
breast cancer initiating cells and synergizes with lapatinib for breast cancer treatment. Cancer
Cell. 2011; 20: 341-56.
16. Xie X, Hsu JL, Choi MG, Xia W, Yamaguchi H, Chen CT, et al. A novel hTERT
promoter-driven E1A therapeutic for ovarian cancer. Mol Cancer Ther. 2009; 8: 2375-82.
17. Xie X, Xia W, Li Z, Kuo HP, Liu Y, Ding Q, et al. Targeted expression of BikDD
eradicates pancreatic tumors in noninvasive imaging models. Cancer Cell. 2007; 12: 52-65.
18. Olson TA, Mohanraj D, Carson LF, Ramakrishnan S. Vascular permeability factor gene
expression in normal and neoplastic human ovaries. Cancer Res. 1994; 54: 276-80.
19. Yoneda J, Kuniyasu H, Crispens MA, Price JE, Bucana CD, Fidler IJ. Expression of
angiogenesis-related genes and progression of human ovarian carcinomas in nude mice. J
Natl Cancer Inst. 1998; 90: 447-54.
20. Dixelius J, Larsson H, Sasaki T, Holmqvist K, Lu L, Engstrom A, et al.
Endostatin-induced tyrosine kinase signaling through the Shb adaptor protein regulates
endothelial cell apoptosis. Blood. 2000; 95: 3403-11.
21. Sim BK, MacDonald NJ, Gubish ER. Angiostatin and endostatin: endogenous inhibitors
of tumor growth. Cancer Metastasis Rev. 2000; 19: 181-90.
22. Dixelius J, Cross M, Matsumoto T, Sasaki T, Timpl R, Claesson-Welsh L. Endostatin
regulates endothelial cell adhesion and cytoskeletal organization. Cancer Res. 2002; 62:
1944-7.
23. Kim YM, Jang JW, Lee OH, Yeon J, Choi EY, Kim KW, et al. Endostatin inhibits
endothelial and tumor cellular invasion by blocking the activation and catalytic activity of
matrix metalloproteinase. Cancer Res. 2000; 60: 5410-3.
24. Abdollahi A, Hahnfeldt P, Maercker C, Grone HJ, Debus J, Ansorge W, et al.
Endostatin's antiangiogenic signaling network. Mol Cell. 2004; 13: 649-63.
25. Ou-Yang F, Lan KL, Chen CT, Liu JC, Weng CL, Chou CK, et al. Endostatin-cytosine
deaminase fusion protein suppresses tumor growth by targeting neovascular endothelial cells.
Cancer Res. 2006; 66: 378-84.
26. Hydzik C. Treatment of ovarian cancer with intraperitoneal chemotherapy. Oncology
(Williston Park). 2009; 23: 15-20.
27. Markman M. An update on the use of intraperitoneal chemotherapy in the management
of ovarian cancer. Cancer J. 2009; 15: 105-9.
28. Templeton NS, Lasic DD, Frederik PM, Strey HH, Roberts DD, Pavlakis GN. Improved
DNA: liposome complexes for increased systemic delivery and gene expression. Nat
Biotechnol. 1997; 15: 647-52.
29. Kish JA, Ensley JF, Jacobs JR, Binns P, al-Sarraf M. Evaluation of high-dose cisplatin
and 5-FU infusion as initial therapy in advanced head and neck cancer. Am J Clin Oncol.
1988; 11: 553-7.
19
30. Tanaka T, Masuda H, Naito M, Tamai H. Pretreatment with 5-fluorouracil enhances
cytotoxicity and retention of DNA-bound platinum in a cisplatin resistant human ovarian
cancer cell line. Anticancer Res. 2001; 21: 2463-9.
31. Kievit E, Bershad E, Ng E, Sethna P, Dev I, Lawrence TS, et al. Superiority of yeast
over bacterial cytosine deaminase for enzyme/prodrug gene therapy in colon cancer
xenografts. Cancer Res. 1999; 59: 1417-21.
32. Zhang M, Li S, Nyati MK, DeRemer S, Parsels J, Rehemtulla A, et al. Regional delivery
and selective expression of a high-activity yeast cytosine deaminase in an intrahepatic colon
cancer model. Cancer Res. 2003; 63: 658-63.
33. Erbs P, Regulier E, Kintz J, Leroy P, Poitevin Y, Exinger F, et al. In vivo cancer gene
therapy by adenovirus-mediated transfer of a bifunctional yeast cytosine deaminase/uracil
phosphoribosyltransferase fusion gene. Cancer Res. 2000; 60: 3813-22.
34. Chen CT, Yamaguchi H, Lee HJ, Du Y, Lee HH, Xia W, et al. Dual targeting of tumor
angiogenesis
and
chemotherapy
by
endostatin-cytosine
deaminase-uracil
phosphoribosyltransferase. Mol Cancer Ther. 2011; 10: 1327-36.
35. Parsons SL, Lang MW, Steele RJ. Malignant ascites: a 2-year review from a teaching
hospital. Eur J Surg Oncol. 1996; 22: 237-9.
36. Subramanian IV, Bui Nguyen TM, Truskinovsky AM, Tolar J, Blazar BR, Ramakrishnan
S. Adeno-associated virus-mediated delivery of a mutant endostatin in combination with
carboplatin treatment inhibits orthotopic growth of ovarian cancer and improves long-term
survival. Cancer Res. 2006; 66: 4319-28.
37. Ramirez RD, Sheridan S, Girard L, Sato M, Kim Y, Pollack J, et al. Immortalization of
human bronchial epithelial cells in the absence of viral oncoproteins. Cancer Res. 2004; 64:
9027-34.
38. Yu D, Wolf JK, Scanlon M, Price JE, Hung MC. Enhanced c-erbB-2/neu expression in
human ovarian cancer cells correlates with more severe malignancy that can be suppressed by
E1A. Cancer Res. 1993; 53: 891-8.
39. Pisano R, Breda M, Grassi S, James CA. Hydrophilic interaction liquid
chromatography-APCI-mass spectrometry determination of 5-fluorouracil in plasma and
tissues. J Pharm Biomed Anal. 2005; 38: 738-45.
40. Chou TC, Talalay P. Quantitative analysis of dose-effect relationships: the combined
effects of multiple drugs or enzyme inhibitors. Adv Enzyme Regul. 1984; 22: 27-55.
20
Figure Legends
Fig. 1. Survivin promoter is active in in ovarian cancer cells. (a) Schematic diagram of
engineered promoter-driven luciferase constructs. (b) The transcriptional activities of the
survivin, Survivin-VISA and CMV promoters were measured in human ovarian cancer and
normal cell lines by cotransfection with the indicated plasmid DNA and pRL-TK for dual
luciferase assay. The relative luciferase activity shown here represents the dual luciferase
activity ratio (firefly/renilla luciferase) by setting the CMV activity as 1. The p53 status is
listed under the indicated ovarian cancer cell lines (W, wild type; M, mutant; N/A, unknown).
(c) The kinetics of luciferase activity driven by CMV or Survivin-VISA (SV) were measured
in PA1 cells. Indicated plasmids were transiently transfected in PA1 cells, and then luciferase
activity measured daily for five days by dual luciferase assay. TEI, total expression index was
measured by setting the area under the curve in CMV-luc group as 1.
Fig. 2. Survivin-VISA drives selective expression of transgene. (a) Indicated plasmid plus
liposome was administered to tumor-free and tumor-bearing (intraperitoneal) mice by i.p.
injection of 100 g DNA per mouse. Luciferase activity was detected by the noninvasive
imaging system after 2 days. The quantified signal (photons/sec) from whole body is shown
under the images. (b) Tissue and tumor specimens from mice were dissected and total protein
extracted for luciferase activity assay. RLU, relative light unit. Error bars indicate SEM.
Fig. 3. Endostatin-cytosine deaminase is active and converts 5-FC to 5-FU. (a)
hEndoyCD driven by CMV or Survivin-VISA vector is indicated by CMV-hEndoyCD and
SV-hEndoyCD, respectively. hEndoyCD driven by CMV and SV was probed with
anti-endostatin antibody in PA1 ovarian cancer cell lysate after DNA transient transfection by
Western blot. Tubulin was used as protein loading control. (b) Anti-angiogenesis effect of
21
hEndoyCD. The concentrated culture media from indicated DNA that was transfected into
PA1 cells was added to human umbilical vein endothelial cells (HUVEC) in endothelial tube
assay. Tubular formations were quantified by counting the branch points in four randomly
selected fields. Top, enlarged tubular morphogenesis. ** p value < 0.01. (c) One day after
indicated plasmids were transiently transfected into PA1 cells, 5-FC was added to cells, and
the concentration of 5-FU was detected from culture media or cell pellet at different time
points by LC/MS/MS analysis. (d) In vitro cell killing effect of hEndoyCD driven by CMV or
SV. Indicated therapeutic plasmids were transient-transfected in PA1 cells by Lipofectamine
2000. One day later, 5-FC (0.5 mg/ml) was added to cells and incubated for 3 more days.
Relative cell viability was measured by crystal violet staining with the cell viability of the
vector group set as 100%. The p values of therapeutic effect between the ovarian cancer cell
lines and normal cell lines are shown. (e) In vitro cell killing effect of hEndoyCD was
performed as described in (d) but with a serial dilution of 5-FC. ** p value < 0.01.
Fig. 4. SV-hEndoyCD and cisplatin combination demonstrates synergistic therapeutic
effect. (a) The combination Index (CI) was calculated to identify potential synergistic
cytotoxic effect of combination therapy. 5-FU and cisplatin were added to cells in a fixed
ratio (1:1.08). (b) SV-hEndoyCD plasmids were transiently transfected in ovarian cancer cells,
and one day later, 5-FC and cisplatin were added to cells in a fixed ratio (9.3:1). Relative cell
viability was measured by MTT assay. The CI determines the degree of the interaction of
drugs: <1, synergistic; 1, additive; >1, antagonistic. ED50: 50% effective dose; ED75: 75%
effective dose; ED90: 90% effective dose. Dm: the median-effective dose. The Dm value in
two drugs combination presents the median-effect dose of former one.
Fig. 5. hEndoyCD inhibits tumor growth and prolongs survival in ovarian cancer
animal models. (a) Nude mice that received i.p. injection of 1x106 SKOV3ip1-luc human
22
ovarian cancer cells were with 25 g of DNA-liposome complexes and one day later with
5-FC treatment by i.p. injection for a total 6 times within 3 weeks. Arrows indicate the day of
drug administration. The photo signals were quantified by IVIS to determine the tumor size.
Error bars indicate SEM. N indicates the mouse number for each group. * p value < 0.05. (b)
Kaplan-Meier
survival
analysis.
The
p
values
for
CMV-hEndoyCD-
and
SV-hEndoyCD-treated groups are 0.12 and 0.01, respectively. (c) Nude mice that were
subcutaneously (s.c.) injected with 1x107 PA1 human ovarian cancer cells were treated with
25 g of DNA-liposome complex by intratumoral injection, and one day later, 5-FC was
administered by i.p. injection for a total 6 times treatment within 3 weeks when sc tumors
were formed. Arrows indicate the therapy time points. (d) Nude mice with i.p. injection of
5x106 PA1 human ovarian cancer cells were treated in the same dosage and schedule as
described in (a). Tumors were removed and tumor weight measured after 3 months. The
results are shown from two independent experiments. N=10 per group. (e) PA1 tumors
isolated from mice were analyzed for hEndoyCD protein expression level detected two days
after the indicated DNA and 5-FC treatment. Tumor tissues from two mice in each group are
labeled as No. 1 and No. 2. The dissected tumor tissues with detected hEndoyCD protein
expression were used to measure the 5-FU amount. ** p value < 0.01. Error bars indicate
SEM.
Fig. 6. Acute toxicity assay of therapeutic plasmids treatment in immunocompetent
mice. Female Balb/c mice were given single dose of 50 g plasmid DNA in a liposomal
complex via the tail vein injection (N=5 mice/group) and one day later with 5-FC (500 mg/kg)
by i.p. injection. (a) Serum levels of AST and ALT in mice were measured after the plasmid
DNA/5-FC injection. Error bars indicate SEM. A dash line indicates the basal level of AST
and ALT in normal mice. * p value < 0.05. (b) Kaplan-Meier survival analysis.
23