Efficient Inhibition of Ovarian Cancer by Truncation Mutantof FILIP1L Gene Delivered by Novel Biodegradable

Cationic Heparin-Polyethyleneimine Nanogels

Chuan Xie,1,* Ma-ling Gou,2,* Tao Yi,1 Hongxin Deng,2 Zheng-yu Li,1 Ping Liu,1 Xiao-rong Qi,1

Xiang He,1 Yuquan Wei,2 and Xia Zhao1,2

Abstract

Filamin A interacting protein 1-like (FILIP1L), which was reported to be consistently absent in ovarian cancercell lines, has been identified to hold therapeutic potential for inhibiting tumor growth, and its COOH-terminaltruncation mutant (FILIP1LDC103) was found to be more potent than the wild-type. The use of polymericnanoparticles to deliver functional gene intraperitoneally holds much promise as an effective therapy for ovariancancer. In this study, a recombinant plasmid expressing FILIP1LDC103 (FILIP1LDC103-p) was constructed, andbiodegradable cationic heparin-polyethyleneimine (HPEI) nanogels were prepared to deliver FILIP1LDC103-pinto human ovarian cancer SKOV3 cells. The expression of FILIP1LDC103 in vitro and in vivo was determinedusing RT-PCR and western blot analysis. Moreover, a SKOV3 intraperitoneal ovarian carcinomatosis model wasestablished to investigate the antitumor activity of HPEI + FILIP1LDC103-p complexes in nude mice. Tumorweights were evaluated during the treatment course. Cell proliferation and apoptosis were evaluated by Ki-67immunochemical staining and TUNEL assay, respectively, and the antiangiogenic effect of FILIP1LDC103-p wasassessed by CD31 immunochemical staining and alginate-encapsulated tumor cell assay. FILIP1LDC103-p couldbe efficiently transfected into SKOV3 cells by HPEI nanogels. Intraperitoneal administration of HPEI +FILIP1LDC103-p complexes led to effective growth inhibition of ovarian cancer, in which tumor weight de-creased by almost 72% in the treatment group compared with that in the empty-vector control group. Mean-while, decreased cell proliferation, increased tumor cell apoptosis, and reduction in angiogenesis were observedin the HPEI + FILIP1LDC103-p group compared with those in the control groups. These results indicated thatHPEI nanogels delivering FILIP1LDC103-p might be of value in the treatment against human ovarian cancer.

Introduction

Filamin a Interacting protein 1-like (FILIP1L) is con-sidered to have an important potential for regulating

tumor cell proliferation, apoptosis, and angiogenesis. It is alsoknown to be down-regulated in ovarian cancer-1 (DOC1),which was originally found to be present in all the humannormal ovarian epithelial cell lines, but was consistently ab-sent in ovarian cancer cell lines (Mok et al., 1994). A previousstudy has reported that overexpression of FILIP1L in humanumbilical vein endothelial cells (HUVECs) could result in theinhibition of cell migration and proliferation, as well as theinduction of cell apoptosis. Its different truncation mutants

had various activities in increasing apoptosis and inhibitingproliferation, in which a COOH-terminal truncation mutant(FILIP1L mutant 1-790, hereafter called FILIP1LDC103)showed a more potent activity than wild-type and othertruncation mutants. In addition, targeted expression of FI-LIP1LDC103 led to the inhibition of tumor growth in nudemice bearing M21 human melanoma (Kwon et al., 2008).However, until now, studies on the antitumor activity of FI-LIP1LDC103 gene in other types of human cancers, especiallyhuman ovarian cancer, are extremely lacking.

Ovarian cancer is the most common and lethal malignancyof the female reproductive tract and presently ranks fifth incausing female cancer-related mortality ( Jermal et al., 2010).

1Gynecological Oncology of Biotherapy Laboratory, Department of Gynecology and Obstetrics, West China Second Hospital, SichuanUniversity, Chengdu 610041, People’s Republic of China.

2State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, Sichuan University, Chengdu 610041, People’s Republic ofChina.

*Chuan Xie and Ma-ling Gou contributed equally to this work.

HUMAN GENE THERAPY 22:1413–1422 (November 2011)ª Mary Ann Liebert, Inc.DOI: 10.1089/hum.2011.047

1413

Although much progress has been made in ovarian cancertherapy, the 5-year survival rate of women with this diseasehas not improved substantially (Raki et al., 2006). Thus,finding novel therapeutic approaches is essential.

Gene therapy, whose potential in cancer treatment has beenwidely recognized (Anderson et al., 2004; Edelstein et al., 2004;Veiseh et al., 2009), holds great promise for the treatment of awide variety of diseases. However, except for the target gene,the gene delivery system is the crucial factor affecting theactivity of gene therapy. Nonviral vectors hold many advan-tages over viral carriers, including relative safety, low im-munogenicity and toxicity, and ease of large plasmid DNAproduction (Young and Mautner, 2001; Zhdanov et al., 2002;Gao et al., 2008; Montier et al., 2008; Morille et al., 2008; Suzukiet al., 2010). Polyethyleneimine (PEI), which was used as agene vector as far back as 1951, becomes progressively one ofthe most effective nonviral gene carrier agents (Boussif et al.,1995; Lungwitz et al., 2005; Neu et al., 2005). Currently, thecommercial PEI25K (25 kg/mol) has been widely consideredas a ‘‘gold standard’’ to assess the transfection efficiency ofnew polymer-based gene carriers. However, PEI, whosetransfection efficiency increases along with an increase of cy-totoxicity, is not biodegradable. Moreover, PEI has the short-coming of inducing obvious aggregation of erythrocytes andhemolysis (Kunath et al., 2003; Neu et al., 2005). In our labo-ratory, PEI2K was chemically conjugated by heparin to form anovel biodegradable cationic nanogel (Fig. 1a), named HPEI.Furthermore, in our previous study, the novel biodegradablecationic HPEI nanogels were used to effectively deliver aplasmid expressing vesicular stomatitis virus matrix protein(pVSVMP) into C-26 cells in vitro and in vivo, and the growthof abdominal and pulmonary metastases in C-26 colon carci-noma in BALB/c mice was efficiently suppressed bypVSVMP/HPEI complexes (Gou et al., 2010). In this work,biodegradable cationic HPEI nanogels were prepared andthen used as a gene carrier to deliver a recombinant plasmidencoding FILIP1LDC103 to treat an SKOV3 intraperitonealovarian carcinomatosis in nude mice, with the aim to evaluatethe potential effect of FILIP1LDC103 against ovarian cancer.

Materials and Methods

Cell culture

HUVECs were isolated from human umbilical vein vas-cular wall as described previously (Baudin et al., 2007). In

brief, human umbilical vein vascular wall was digested withcollagenase IV at 37�C for 10 min, and the homogenate wascentrifuged at 750 g for 10 min. Cells were suspended andseeded on fibronectin-coated plates and cultured in Earle’ssalts medium supplemented with 10% fetal calf serum (FCS).The human epithelial serous cystadenocarcinoma cell lineSKOV3 was obtained from the American Type CultureCollection (ATCC, Rockville, MD) and grown in RPMI 1640medium (GIBCO, Carlsbad, CA) supplemented with 10%FCS, 2 mM L-glutamine, 100 U/ml penicillin, and 100 lg/mlstreptomycin. Cells were maintained in a humidified incu-bator with 5% CO2 atmosphere at 37�C.

RT-PCR and plasmid construction

The primers of FILIP1LDC103 were designed based on itscDNA sequence with upstream primer 5¢-CGC GGA TCC

AAG ATG GTG GTG GAT GAA CAG CA-3¢ and down-stream primer 5¢-CCG CTC GAG TCA CGC ATG CTT GGCACT GAT TT-3¢. The incorporated 5¢BamHI and 3¢XhoI re-striction sites are shown in bold, whereas the protective baseis shown in italics. RT-PCR was performed as follows: in-cubation at 50�C for 30 min, and denaturation at 94�C for2 min, followed by a standard PCR regimen of 94�C for30 sec, 55�C for 30 sec, and 72�C for 3 min (30 cycles). Fivemicroliters of each RT-PCR product was analyzed by agarosegel electrophoresis. The housekeeper gene b-actin served asinternal control.

pVAX1 plasmid (Invitrogen, San Diego, CA) expressingFILIP1LDC103, named FILIP1LDC103-p, was constructed inour laboratory and is illustrated schematically in Fig. 2A. Inbrief, cultured HUVECs were harvested and total RNA wasextracted using TRIzol reagent (Invitrogen) according to themanufacturer’s protocol. The RNA sample was then sub-jected to RT-PCR for amplification of the encoding region ofmFILIP1LDC103, using a One Step RNA PCR Kit (AMV;Takara Bio Inc., Kyoto, Japan) with the above primers andconditions of reaction. The amplified mFILIP1LDC103 DNA(2,370 bp) and eukaryotic expression vector pVAX1 weredigested with BamHI and XhoI, respectively, and then ligatedovernight at 16�C using T4 DNA ligase, followed by trans-formation into competent Escherichia coli DH5a. The resultantrecombinant plasmid was double-digested to ensure itscorrectness, and the identity of FILIP1LDC103 cloned into theexpression plasmid pVAX1 was finally confirmed by auto-mated DNA sequencing. Plasmid without FILIP1LDC103

FIG. 1. Preparation scheme and character-ization of HPEI nanogels. (a) Preparationscheme of HPEI nanogels. Reaction betweenheparin and PEI occurs when catalyzed byEDC/NHS. (b) Image from TEM of HPEInanogels. Color images available online atwww.liebertonline.com/hum

1414 XIE ET AL.

was used as an empty-vector control (hereafter called E-p).Colonies of Escherichia coli containing FILIP1LDC103 or E-pwere cultured in Luria-Bertani broth containing 100 lg/mlkanamycin. Large-scale plasmid DNA was purified usingan EndoFree Plasmid Giga kit (Qiagen, Valencia, CA).The DNA was dissolved in sterile endotoxin-free water, ad-justed eventually to 1.0 mg/ml, and then stored at - 20�C forfuture use.

Preparation of HPEI and transfection of plasmid

Biodegradable HPEI nanogels were synthesized as previ-ously described (Gou et al., 2010). In brief, 0.05 g of heparinwas first dissolved in 2-(N-morpholino)ethanesulfonic acid(MES) buffering agent (100 ml, 50 mM); 0.02 g of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC) and 0.03 g of N-hydroxysuccinimide (NHS) were subsequently added intothe above solution agents to activate the carboxylic acidgroups of heparin. After 20 ml of PEI2K solution (7.5 mg/ml)was dropped into the above solution, the reaction was car-ried out at room temperature overnight. Thereafter, the re-sulting HPEI nanogels were dialyzed in double-distilledwater for 3 days. Later, HPEI nanogels were filtered using asyringe filter, then adjusted to a final concentration of1.0 mg/ml, and stored at 4�C.

SKOV3 ovarian cancer cells (2.0 · 105) were grown onsix-well plates in RPMI 1640 medium and cultured for 24 hr.Twenty micrograms of HPEI and 2 lg of FILIP1LDC103-por E-p were diluted in 1 ml of RPMI 1640 medium with-out antibiotic and serum, respectively, and then combinedat a ratio of 10:1. The combinations were transfected to cellswith 50–60% confluence. Meanwhile, the medium alone wasused as a control agent. After 4 to 6 hr, the medium wasreplaced by 2 ml of RPMI 1640 medium. After 48 hr ofincubation, the cells and supernatants were collected forfurther assay.

Western blot analysis

Cells or tumor tissues were lysed in RIPA lysis buffer con-taining 50 mM Tris-HCl (pH 7.4), 0.25% sodium deoxycholate,150 mM NaCl, 1% NP-40, 1 mM NaF, 1 mM Na3V4, and 1 mMcocktail (Sigma, St. Louis, MO). Protein concentrations weredetermined using the Bradford assay (Bio-Rad, Hercules, CA).Equal amounts of protein were resolved on a 10% SDS-PAGE,and then electroblotted to polyvinylidene difluoride (PVDF)membranes (Millipore, Bedford, MA), with GAPDH as theloading control. Then the PVDF membranes were blocked with5% nonfat milk for 2 hr and incubated with rabbit polyclonalantibodies against FILIP1LDC103 (1: 800 dilution; Santa CruzBiotechnology, Santa Cruz, CA). Then the membranes wereincubated with goat anti-rabbit secondary antibody at 1:5,000dilution (Abcam, Cambridge, MA) in PBS with Tween 20 for1 hr. The blot bands were developed using the enhanced che-miluminescence detection system (Pierce Biotech Inc., Rock-ford, IL) according to the manufacturer’s instruction.

Tumor xenograft model and animal treatment

All animal research procedures were approved by the In-stitutional Animal Care and Use Committee of Sichuan Uni-versity (Chengdu, People’s Republic of China). Female athymicnude mice were used to establish an intraperitoneal carcino-matosis model according to the previous study (Lin et al., 2007).In brief, SKOV3 cells (5 · 106) in 100ll of RPMI 1640 mediumwere injected subcutaneously into the dorsal sides of four mice.When the diameter was about 1 cm, tumors were collected andthen minced into tiny particles. Tumor particles were mixedwith the RPMI 1640 solution to reach a final volume of 10 ml,and 20 mice were inoculated intraperitoneally with 0.5 ml of theabove mixture. Mice were assigned randomly to one of thefollowing groups (five per group): (a) untreated, 100ll of 5%glucose solution; (b) HPEI, 50 ll of HPEI in 50ll of 5% glucosesolution; (c) HPEI + E-p, 5lg of pVAX1/50ll of HPEI in 50ll of

FIG. 2. Construction and identification of recombinant plasmid FILIP1LDC103-p. (A) Schematic illustration of constructionof recombinant plasmid FILIP1LDC103-p. Amplified FILIP1LDC103 was cloned into the expression vector pVAX1 followingdouble-digestion. As a control, pVAX1 without FILIP1LDC103 was used as an empty vector (E-p). (B) Detection of theexpression of FILIP1LDC103 in the SKOV3 cell line and HUVECs. Expression of FILIP1LDC103was detected in HUVECs; bycontrast, there was no expression in SKOV3 cells. Lane 1, SKOV3 cell line; lane 2, HUVECs. (C) Identification of plasmidFILIP1LDC103-p by BamHI/XhoI digestion. a, residual plasmid FILIP1LDC103-p after 2 hr of double-digestion; b, pVAX1from double-digestion; c, target gene FILIP1LDC103 from double-digestion. (D) Identification of FILIP1LDC103 cloned intoexpression vector pVAX1 by PCR.

MUTANT FILIP1L GENE AS THERAPY FOR OVARIAN CANCER 1415

5% glucose solution; and (d) HPEI + FILIP1LDC103-p, 5lg ofFILIP1LDC103-p/50ll of HPEI in 50ll of 5% glucose solution.Intraperitoneal administration was initiated 5 days after inoc-ulation. The HPEI/DNA complexes were prepared and incu-bated at room temperature for 30 min before administration.Mice received therapy every 3 days and were sacrificed after 12treatments. Intraperitoneal tumors were resected and weighedimmediately to assess the antitumoral efficacy.

Alginate-encapsulated tumor cell assay

To explore inhibition of angiogenesis, the alginate-encap-sulation assay was done. In brief, SKOV3 cells were re-suspended in a 1.5% solution of alginate (Sigma) and addeddropwise into a solution of 250 mM CaCl2; an alginate beadwas formed containing 1 · 105 cells. Four beads were thenimplanted subcutaneously in the back of nude mice. Eightmice were then grouped and treated as described earlier.Treatment was initiated on the same day of implantingbeads. After 2 weeks, mice were injected intravenously with0.2 ml of a 50 mg/kg fluorescein isothiocyanate (FITC)–dextran (Sigma) solution. Alginate beads were photographedafter being exposed surgically, and then rapidly removed20 min after FITC-dextran injection. The uptake of FITC-dextran was measured as described (Hoffmann et al., 1997).

Immunohistochemistry

Deparaffinized tumor sections were immersed in 10 mMcitrate buffer (pH 6.0) and then heated in an autoclave for5 min in saturated steam for antigen retrieval. Endogenousperoxidase activity was quenched in 3% H2O2 for 10 min, andnonspecific binding sites of reagents were subsequentlyblocked with homeotypic nonimmunoglobulin of secondaryantibody at 37�C for 20 min. Then tumor sections were incu-bated with the primary antibody (R&D Systems, Minneapolis,MN) at 4�C overnight. Tumor sections were incubated withthe biotinylated secondary antibody at 37�C for 40 min, fol-lowed by sequential incubation with streptavidin–biotin–horseradish peroxidase complex for 40 min at 37�C. Colori-metric detection was performed with diaminobenzidine.

TUNEL assay for apoptotic cells

TUNEL staining was performed to analyze apoptotic cellsin tumor tissues using apoptotic cell kits according to themanufacturer’s protocol (Promega, Madison, WI). The apo-ptosis index was calculated by analyzing the average per-centage of green fluorescence-positive cells in 10 randomfields from different sections at · 400 magnification.

Evaluation of toxicity

The potential treatment-related toxicity and side effects suchas weight changes of mice, diarrhea, toxic death, or behaviorwere observed and evaluated during the whole treatmentcourse. Sections of organs such as heart, liver, spleen, lung, andkidney were stained with hematoxylin and eosin (HE) andobserved by different pathologists in a blinded manner.

Statistical analysis

All results are expressed as the means – SD, and one-wayANVOA for multiple-group comparisons was used to ana-

lyze differences among four groups. A p value below 0.05was considered statistically significant.

Results

Cloning FILIP1LDC103 cDNAand plasmid construction

To obtain FILIP1LDC103 cDNA fragments, total RNA wasextracted from harvested HUVECs, and then RT-PCR wasperformed. One percent agarose gel electrophoresis revealedone band about 2,370 bp in size as expected. We obtained afragment of approximately 2,370 bp in length, as shown inFig. 2B, consistent with the expected size.

The purified PCR product was first double-digested withBamHI/XhoI, and then FILIP1LDC103 was cloned into theBamHI/XhoI site of the pVAX1 vector treated with theabove enzymes. The identity of recombinant plasmidFILIP1LDC103-p was confirmed by PCR (Fig. 2D), double-digestion (Fig. 2C), and direct automated DNA sequencing.

Preparation and characterization of HPEI

Catalyzed by EDC and NHS, biodegradable cationic HPEInanogels were formed when PEI2K was chemically conju-gated by heparin, as shown in Fig. 1a. According to the imagefrom a transmission electron microscope (TEM; Fig. 1b), cat-ionic HPEI nanogels loading FILIP1LDC103-p were mono-dispersed and spherical with a diameter of about 28 nm.

Expression of FILIP1LDC103 in vitro

First, we used RT-PCR to test the expression of the FI-LIP1LDC103 gene in human ovarian cancer SKOV3 cells.Semiquantitative RT-PCR indicated the complete absence ofFILIP1LDC103 transcript expression in SKOV3 ovarian cancercells, whereas it was expressed in HUVECs, as shown in Fig. 2B.

Next, to investigate whether FILIP1LDC103 was expressedin human ovarian cancer SKOV3 cells transfected byHPEI + FILIP1LDC103-p complexes, SKOV3 cells were see-ded in six-well plates and then transfected with HPEI +FILIP1LDC103-p complexes, HPEI + E-p complexes, HPEInanogels, and 5% glucose (untreated group). After 48 hr ofincubation, the in vitro expression of the FILIP1LDC103 genein transfected SKOV3 cells was confirmed using RT-PCR andwestern blot analysis. As shown in Fig. 3A, the expression ofFILIP1LDC103 in SKOV3 cells transfected with HPEI +FILIP1LDC103-p could be detected, whereas there was noexpression of FILIP1LDDC103 in SKOV3 cells from the un-treated, HPEI, or HPEI + E-p groups.

Expression of FILIP1LDC103 in vivo

To examine the expression of FILIP1LDC103 in vivo, anintraperitoneal ovarian carcinoma model was established innude mice, which were then treated with HPEI +FILIP1LDC103-p, HPEI + E-p, HPEI, or 5% glucose. Tumorswere collected for RT-PCR and western blot on the third dayafter the final treatment. The level of FILIP1LDC103 mRNAand expression of FILIP1LDC103 protein were detected intumor tissues from the HPEI + FILIP1LDC103-p group,whereas there was no expression in the HPEI, E-p, and 5%glucose groups (Fig. 3B). In addition, immunochemicalstaining of FILIP1LDC103 in tumor tissues also suggested

1416 XIE ET AL.

expression of FILIP1LDC103 in the HPEI + FILIP1LDC103-pcomplex–treated group as shown in Fig. 3C.

HPEI + FILIP1LDC103-p complexes inhibitedintraperitoneal ovarian cancer xenograft growthin nude mice

To evaluate the effect of HPEI + FILIP1LDC103-p on sup-pressing ovarian cancer growth in vivo, we established anintraperitoneal xenograft model of human ovarian cancer.Nude mice were grouped and treated as aforementioned,then sacrificed at the termination of the animal experiment.Intraperitoneal tumor nodules were collected totally andcarefully; the average weights were 1.23 – 0.18 g, 1.25 – 0.13 g,1.04 – 0.14 g, and 0.29 – 0.11 g in the untreated, HPEI, HPEI +E-p, and HPEI + FILIP1LDC103-p groups, respectively ( p <0.01; Fig. 4B and C). HPEI + FILIP1LDC103-p reduced tumorweight by almost 72% compared with the HPEI + E-p group.There was no significant difference in tumor weight amongthe three control groups ( p > 0.05).

Intraperitoneal tumor nodules in a wide range of sizeswere scattered predominantly in the pelvis, as well as belowthe liver, in the three control groups (Fig. 4A). Ascites werefound in three mice in the untreated group, in four mice inthe HPEI-treated group, and in two mice in the HPEI + E-p–treated group, including hemorrhagic ascites in one, two, andone from the untreated, HPEI, and HPEI + E-p groups, respec-tively (Table 1). None of the mice in the HPEI + FILIP1LDC103-p group developed ascites, and intraperitoneal tumor nodulesof mice in this group were all scattered in the pelvis.

Suppression of angiogenesis by HPEI +FILIP1LDC103-p in vivo

To explore potential mechanisms underlying the antitu-mor effect of HPEI + FILIP1LDC103-p in vivo, and as anti-angiogenesis is generally considered to be an importantmechanism for cancer therapy, frozen sections were stained

by antibody of CD31 to investigate the antiangiogenic effectof HPEI + FILIP1LDC103-p. Angiogenesis in tumor tissueswas evaluated by microvessel density (MVD) in differentsections stained with an antibody reactive to CD31, whichhad high specific affinity for vascular endothelial cells. CD31immunochemical staining revealed a significant reduction ofMVD in tumor tissues of the HPEI + FILIP1LDC103-p groupin comparison with tissues of the untreated, HPEI, andHPEI + E-p groups (5.10 – 2.73 vs. 26.70 – 6.13 vs. 27.00 – 6.75vs. 24.70 – 5.50, respectively; p < 0.01; Fig. 5A); there was alsono significant difference among the numbers of microvesselsof the different control groups ( p > 0.05).

In addition, the capability of antiangiogenesis was alsodetected using the alginate-encapsulated tumor cell assay;newborn blood vessels in alginate beads from nude micetreated with HPEI + FILIP1LDC103-p were apparently fewerthan those in the other control groups. Moreover, the levelsof FITC-dextran uptake were significantly low in mice trea-ted with HPEI + FILIP1LDC103-p compared with levels inthe three control groups (0.83 – 0.21 lg/bead vs. 5.43 –0.79 lg/bead vs. 5.19 – 0.76 lg/bead vs. 4.98 – 0.96 lg/bead,respectively; p < 0.01; Fig. 5B).

The above results suggested that angiogenesis of humanovarian cancer xenograft was effectively inhibited in nudemice treated with HPEI + FILIP1LDC103-p. Therefore, wemight conclude that the antitumor activity of HPEI +FILIP1LDC103-p is due, at least partially, to the inhibition oftumor neovascularization in ovarian cancer.

Inhibition of cell proliferationby HPEI + FILIP1LDC103-p in vivo

To explore further the role of HPEI + FILIP1LDC103-ptherapy in ovarian cancer in vivo, we performed Ki-67 im-munohistochemistry to investigate whether the antitumor ef-fect of HPEI + FILIP1LDC103-p correlated with the decreasedcell proliferation. The proliferation index was determined by aratio of the Ki-67-positive cells to the total number of cancer

FIG. 3. Expression of FI-LIP1LDC103-p in vitro andin vivo. (A) Identification ofthe in vitro expression of FI-LIP1LDC103 in transfectedSKOV3 cells by RT-PCR andwestern blot analysis. FI-LIP1LDC103 expression wasdetected in SKOV3 cells trea-ted with recombinant plas-mid FILIP1LDC103-p; bycontrast, no expression of FI-LIP1LDC103 was observed inSKOV3 cells that were un-treated or treated with HPEIand HPEI + E-p. The house-keeper gene b-actin served asinternal control. The produc-tion of FILIP1LDC103 proteinwas further confirmed bywestern blot analysis; the37-kDa GAPDH served asinternal control. (B) In vivoexpression of FILIP1LDC103 was determined by RT-PCR and western blot analysis. (C) Immunostaining of FILIP1LDC103 intumor tissue. Original magnification, · 400. Color images available online at www.liebertonline.com/hum

MUTANT FILIP1L GENE AS THERAPY FOR OVARIAN CANCER 1417

cells. The most significant reduction of Ki-67 expression wasobserved in the tumors of mice receiving HPEI +FILIP1LDC103-p treatment when compared with the threecontrols groups (18.50 – 6.67 vs. 46.40 – 14.92 vs. 45.10 – 15.89vs. 37.40 – 13.77, respectively; p < 0.05; Fig. 6A).

Induction of cell apoptosis by HPEI + FILIP1LDC103-pin vivo

As tumor growth may be considered an imbalance be-tween cell proliferation and apoptosis, we also applied the

TUNEL assay to explore possible mechanisms underly-ing the antitumor effect of HPEI + FILIP1LDC103-p in vivo.The TUNEL assay showed many apoptotic cells (TUNEL-positive nuclei) in the tumor sections of the HPEI +FILIP1LDC103-p complexes group. However, there were fewsuch cells in the three control groups. The apoptotic indexrevealed that tumors in the HPEI + FILIP1LDC103-p grouphad significantly more TUNEL-positive nuclei than tumorsin the 5% glucose, HPEI, or HPEI + E-p groups (55.80 – 9.40vs. 6.60 – 3.10 vs. 6.90 – 3.18 vs. 10.70 – 3.92, respectively; p <0.01; Fig. 6B).

FIG. 4. FILIP1LDC103-p inhibited intraperitoneal ovarian cancer xenograft growth in nude mice. (A) Nude mice wereeuthanized at the termination of the animal experiment. Intraperitoneal tumor nodules in a wide range of sizes were scatteredpredominantly in the pelvis, as well as below the liver. (B) All the intraperitoneal tumor lesions were collected totally andcarefully, and then weighed immediately to assess the antitumoral efficacy. HPEI + FILIP1LDC103-p reduced tumor weightby almost 72% compared with the HPEI + E-p group. a, untreated group (5% glucose solution); b, HPEI group; c, HPEI + E-pgroup; d, HPEI + FILIP1LDC103-p group. (C) Tumor from HPEI + FILIP1LDC103-p–treated mice showed significant differ-ence compared with tumors from the other control groups (**p < 0.01). The results are presented as means – SD. Color imagesavailable online at www.liebertonline.com/hum

Table 1. Characterization of Intraperitoneal Xenografts of Human Ovarian Cancer in Nude Mice

Groups (n = 5)Number of

nodulesDiameter of

nodules (mm)Number of hemorrhagic

ascites/ascitesaAverage weight(g, means – SD)

Untreated 8–14 2–10 1/3 1.23 – 0.18HPEI 10–15 3–13 2/4 1.25 – 0.13HPEI + E-p 7–13 1–11 1/2 1.04 – 0.14HPEI + FILIP1LDC103-p 3–6 1–4 0/0 0.29 – 0.11

aVolume of ascites or hemorrhagic ascites found in nude mice was all less than 3 ml.

1418 XIE ET AL.

Observation of toxicity

Because animal weight was considered a parameter forevaluation of physical status, we measured the weight ofmice every 3 days (Fig. 7A) and observed no significantdifferences among the four groups. Moreover, HE histolog-ical staining of spleen, liver, lung, kidney, and heart did notdemonstrate any significant pathologic differences (Fig. 7B).

Discussion

Previous study has shown that growth of xenograftedhuman malignant melanoma can be suppressed by intro-ducing the different mutant-type genes of FILIP1L, and theCOOH-terminal truncation mutant 1-790 (FILIP1LDC103)demonstrated a most potent activity (Kwon et al., 2008).However, studies on the antitumoral efficacy of the FI-LIP1LDC103 gene in human ovarian cancer have been lack-ing until now. In the present study, we constructed arecombinant plasmid expressing FILIP1LDC103 (FI-

LIP1LDC103-p), which was then used to effectively inhibitintraperitoneal xenograft growth of human ovarian cancer innude mice, with biodegradable cationic HPEI nanogels as anovel delivery system. Meanwhile, treatment with HPEI +FILIP1LDC103-p complexes in nude mice was shown to bewithout detectable toxicity.

The FILIP1L gene, whose translation product has 893amino acids, is located on the region of chromosome 3 inhumans. It was identified as a potentially important regu-lator of angiogenesis activity. Overexpression of FILIP1Lcould lead to inhibition of proliferation and induction ofapoptosis, whereas FILIP1LDC103 is more potent than wild-type FILIP1L in mediating these activities (Kwon et al., 2008).Results obtained using the microarray technique revealedthat FILIP1L played a possible role in the pattern of sup-pressing tumor (Ross et al., 2000; Su et al., 2001). A sustainedincrement of FILIP1L gene expression over the time coursewas observed in HUVECs in response to 5-fluorouraciltreatment, suggesting that the gene might be related to cell

FIG. 5. Inhibition of angiogenesis in vivo. (A) Sections of tumor tissues were examined using immunohistochemical analysiswith anti-CD31 antibody. The MVD of tumor tissues from the HPEI + FILIP1LDC103-p group showed a significant decreasecompared with the MVD of the other groups (**p < 0.01). (B) Vascularization of alginate beads. Nude mice were treated on thesame day as implantation of the alginate beads containing 1 · 105 SKOV3 cells; beads were then surgically removed. FITC-dextran uptake of alginate beads from the HPEI + FILIP1LDC103-p group indicated a significant decrease compared withuptake from the other groups (**p < 0.01). The results are presented as means – SD (n = 5). Color images available online atwww.liebertonline.com/hum

MUTANT FILIP1L GENE AS THERAPY FOR OVARIAN CANCER 1419

death (Tandle et al., 2005). Furthermore, FILIP1L has a novelfunction in cell senescence, which has an important role ininhibiting the occurrence of an immortal phenotype, a crucialfeature in tumorigenesis (Schwarze et al., 2002). FILIP1L wasreported to be present in human normal ovarian epithelial cells,but consistently absent in ovarian cancer cell lines (Mok et al.,1994). Consequently, we may wonder whether transferring theFILIP1L gene into ovarian cancer cells by a suitable deliverysystem could be an approach for treating ovarian cancer.

Over the last decade, many functional genes have beenreported to effectively inhibit ovarian cancer in vitro andin vivo (Malek et al., 2008; Nguyen et al., 2010), but the ab-sence of efficient and safe gene delivery technologies is amajor obstacle to the clinical use of these gene treatments.Currently, many delivery systems have been applied in genetransfer; these are generally classified as nonviral and viralvectors. Previous studies have demonstrated that nonviralvectors have many advantages over viral vectors (Young andMautner, 2001; Zhdanov et al., 2002; Gao et al., 2008; Montieret al., 2008; Morille et al., 2008; Suzuki et al., 2010). In this

work, we developed a novel nonviral gene delivery systembased on PEI. It has the advantages of PEI and resolves theproblem of unbiodegradability and the transfection efficien-cy–dependent cytotoxicity of PEI. In our study, heparin wasused to conjugate PEI into biodegradable cationic nanogelsfor the following reasons. First, it is relatively stable in vitroand easy to degrade through enzymolysis and hydrolysisin vivo. Second, as a natural polysaccharide, it is nontoxic andbiocompatible. Third, it could improve the biocompatibilityof the nanogels. In our previous study, the transfection effi-ciency of HPEI was found to be comparable to, but less toxicthan, that of PEI25k. Analyses of the erythrocyte aggregationand hemolysis showed that HPEI had a better blood com-patibility than PEI25K. Moreover, HPEI nanogels were stablein vitro and could be quickly degraded into the low-molecular-weight chemical compound PEI excreted in the urine (Gouet al., 2010). Based on these advantages, HPEI nanogels wereused to deliver a recombinant plasmid to evaluate the anti-tumor activity of FILIP1LDC103, as well as the efficacy andsafety of this delivery system.

FIG. 6. Effects of expression of FILIP1LDC103-p on cell apoptosis and proliferation in vivo. (A) More Ki-67-positive cellswere identified in the untreated, HPEI, and HPEI + E-p groups than in the HPEI + FILIP1LDC103-p group. The percentage ofKi-67-positive cells showed a significant reduction in the HPEI + FILIP1LDC103-p group compared with that in the othercontrol groups (**p < 0.01). (B) The TUNEL-positive nuclei in the HPEI + FILIP1LDC103-p group were significantly increasedwhen compared with those in the other three groups (**p < 0.01). The results are presented as means – SD (n = 5). Color imagesavailable online at www.liebertonline.com/hum

1420 XIE ET AL.

In the current study, we have demonstrated that a re-combinant plasmid encoding FILIP1LDC103 can signifi-cantly suppress intraperitoneal xenograft growth of humanovarian cancer in nude mice. This provides proof, in princi-ple, that FILIP1LDC103 possesses antitumor activity againsthuman ovarian cancer in vivo. To elucidate the antitumormechanism in vivo of FILIP1LDC103, inhibition of angio-genesis including the alginate assay and CD31 immunohis-tochemistry was performed. The results indicated that theMVD of tumor tissue sections was significantly reduced byFILIP1LDC103-p and that angiogenesis was significantly in-hibited in the FILIP1LDC103-p group compared with that inthe other controls. Data from proliferation and apoptosisanalysis using Ki-67 staining and the TUNEL techniqueshowed that expression of FILIP1LDC103 resulted in signif-icant inhibition of cell proliferation and induction of cellapoptosis compared with the control therapies. These resultsdemonstrated that the antitumor activity of FILIP1LDC103-pin xenografted ovarian cancer might be ascribed to success-fully sustained expression of FILIP1LDC103 in vivo. How-ever, the exact mechanism of its antitumor activity requiresfurther investigation. Previous study had shown that whenthe gene expression of FILIP1L was silenced, endothelialmonocyte-activating polypeptide II (EMAP II) failed to showa modulatory effect on the expression of KLF4, ADM, SOCS3,and TNFAIP3 (Tandle et al., 2005), suggesting that FILIP1Lmay play a role in mediating the effect of EMAP II. Similarly,the up-regulation of gene expression of TC-1 and KLF4 inresponse to endostatin was observed with silencing of thegene encoding FILIP1L at the mRNA level using small in-terfering RNA; it is suggested that FILIP1L might be up-stream of TC-1 and KLF4 in a pathway mediating theantiangiogenic response to endostatin treatment (Mazzanti

et al., 2004). These data suggest that FILIP1L might play aprimary role in mediating some common response pathwaysobserved in endothelial cells after exposure to angiogenicinhibitors.

In this study, using HPEI to deliver FILIP1LDC103-p totreat nude mice bearing ovarian cancer, we found that in-traperitoneal xenografted ovarian cancer could be effectivelysuppressed; also, evaluation of toxicity indicated that treat-ment with the HPEI + FILIP1LDC103-p complex seemed tobe safe and without detectable toxicity at the dose we used.Nevertheless, the toxicity of this delivery system in otherresearch requires further investigation, and a study on theantitumor activity of FILIP1LDC103 in other cancers may berequired. Future investigations into these points might beinteresting.

In conclusion, our results suggested that FILIP1LDC103-pcould be effectively transfected into human ovarian cancerSKOV3 cells by HPEI nanogels, and biodegradable cationicHPEI nanogels might be a safe and effective delivery system.FILIP1LDC103, as a potential angiogenesis inhibitor, couldefficiently inhibit the growth of xenografted ovarian cancerthrough the inhibition of angiogenesis and tumor cell pro-liferation and the induction of tumor-cell apoptosis, andHPEI nanogel-delivered FILIP1LDC103-p may have a po-tentially promising application against ovarian cancer.

Acknowledgments

This work was supported by the National Key Basic Re-search Program (973 Program) of China (2011CB910703).

Author Disclosure Statement

The authors have declared no conflicts of interest.

FIG. 7. Observation of tox-icity in mice treated withHPEI + FILIP1LDC103. (A)Toxicity-dependent bodyweight was not observed inmice treated with HPEI +FILIP1LDC103-p. There wereno significant differences inbody weight among the fourgroups ( p > 0.05). The resultsare presented as means – SD.(B) HE histological stainingof liver, heart, spleen, lung,and kidney from mice treatedwith HPEI + FILIP1LDC103-pdid not demonstrate any sig-nificant pathologic differ-ences (n = 5). Color imagesavailable online at www.liebertonline.com/hum

MUTANT FILIP1L GENE AS THERAPY FOR OVARIAN CANCER 1421

References

Anderson, D.G., Peng, W., Akinc, A., et al. (2004). A polymerlibrary approach to suicide gene therapy for cancer. Proc. Natl.Acad. Sci. U.S.A. 101, 16028–16033.

Baudin, B., Bruneel, A., Bosselut, N., and Vaubourdolle, M.(2007). A protocol for isolation and culture of human umbilicalvein endothelial cells. Nat. Protoc. 2, 481–485.

Boussif, O., Lezoualc’h, F., Zanta, M.A., et al. (1995). A versatilevector for gene and oligonucleotide transfer into cells in cul-ture and in vivo: polyethylenimine. Proc. Natl. Acad. Sci.U.S.A. 92, 7297–7301.

Edelstein, M.L., Abedi, M.R., Wixon, J., and Edelstein, R.M.(2004). Gene therapy clinical trials worldwide 1989–2004—anoverview. J. Gene Med. 6, 597–602.

Gao, Y., Xu, Z., Chen, S., et al. (2008). Arginine-chitosan/DNAself-assemble nanoparticles for gene delivery: in vitro charac-teristics and transfection efficiency. Int. J. Pharm. 359, 241–246.

Gou, M., Men, K., Zhang, J., et al. (2010). Efficient inhibition of C-26 colon carcinoma by VSVMP gene delivered by biode-gradable cationic nanogel derived from polyethyleneimine.ACS Nano 4, 5573–5584.

Hoffmann, J., Schirner, M., Menrad, A., and Schneider, M.R.(1997). A highly sensitive model for quantification of in vivotumor angiogenesis induced by alginate-encapsulated tumorcells. Cancer Res. 57, 3847–3851.

Jermal, A., Siegel, R., Xu, J., and Ward, E. (2010). Cancer statis-tics, 2010. CA Cancer J. Clin. 60, 277–300.

Kunath, K., von Harpe, A., Fischer, D., et al. (2003). Low-molecular-weight polyethyleneimine as a non-viral vector for DNA deliv-ery: comparison of physicochemical properties, transfectionefficiency and in vivo distribution with high-molecular-weightpolyethyleneimine. J. Control. Release 89, 113–125.

Kwon, M., Hanna, E., Lorang, D., et al. (2008). Functional charac-terization of filamin A interacting protein 1-like, a novel candi-date for antivascular cancer therapy. Cancer Res. 68, 7332–7341.

Lin, X.J., Chen, X.C., Wang, L., et al. (2007). Dynamic progressionof an intraperitoneal xenograft model of human ovarian can-cer and its potential for preclinical trials. J. Exp. Clin. CancerRes. 26, 467–474.

Lungwitz, U., Breunig, M., Blunk, T., and Gopferich, A. (2005).Polyethylenimine-based non-viral gene delivery systems. Eur.J. Pharm. Biopharm. 60, 247–266.

Malek, A., Bakhidze, E., Noske, A., et al. (2008). HMGA2 gene isa promising target for ovarian cancer silencing therapy. Int. J.Cancer 123, 348–356.

Mazzanti, C.M., Tandle, A., Lorang, D., et al. (2004). Early ge-netic mechanisms underlying the inhibitory effects of en-dostatin and fumagillin on human endothelial cells. GenomeRes. 14, 1585–1593.

Mok, S.C., Wong, K.K., Chan, R.K., et al. (1994). Molecularcloning of differentially expressed genes in human epithelialovarian cancer. Gynecol. Oncol. 52, 247–252.

Montier, T., Benvegnu, T., Jaffres, P.A., et al. (2008). Progressin cationic lipid-mediated gene transfection: a series of bio-inspired lipids as an example. Curr. Gene Ther. 8, 296–312.

Morille, M., Passirani, C., Vonarbourg, A., et al. (2008). Progressin developing cationic vectors for non-viral systemic genetherapy against cancer. Biomaterials 29, 3477–3496.

Neu, M., Fischer, D., and Kissel, T. (2005). Recent advances inrational gene transfer vector design based on poly(ethyleneimine) and its derivatives. J. Gene Med. 7, 992–1009.

Nguyen, T.M.B., Subramanian, I.V., Xiao, X., et al. (2010). Adeno-associated virus-mediated delivery of kringle 5 of humanplasminogen inhibits orthotopic growth of ovarian cancer.Gene Ther. 17, 606–615.

Raki, M., Rein, D.T., Kanerva, A., and Hemminki, A. (2006).Gene transfer approaches for gynecological diseases. Mol.Ther. 14, 154–163.

Ross, D.T., Scherf, U., Eisen, M.B., et al. (2000). Systematic vari-ation in gene expression patterns in human cancer cell lines.Nat. Genet. 24, 227–235.

Schwarze, S.R., DePrimo, S.E., Grabert, L.M., et al. (2002). Novelpathways associated with bypassing cellular senescence inhuman prostate epithelial cells. J. Biol. Chem. 277, 14877–14883.

Su, A.I., Welsh, J.B., Sapinoso, L.M., et al. (2001). Molecularclassification of human carcinomas by use of gene expressionsignatures. Cancer Res. 61, 7388–7393.

Suzuki, R., Namai, E., Oda, Y., et al. (2010). Cancer gene therapyby IL-12 gene delivery using liposomal bubbles and tumoralultrasound exposure. J. Control. Release 142, 245–250.

Tandle, A.T., Mazzanti, C., Alexander, H.R., et al. (2005). En-dothelial monocyte activating polypeptide-II induced geneexpression changes in endothelial cells. Cytokine 30, 347–358.

Veiseh, O., Kievit, F.M., Gunn, J.W., et al. (2009). A ligand-mediated nanovector for targeted gene delivery and trans-fection in cancer cells. Biomaterials 30, 649–657.

Young, L.S., and Mautner, V. (2001). The promise and potentialhazards of adenovirus gene therapy. Gut 48, 733–736.

Zhdanov, R.I., Podobed, O.V., and Vlassov, V.V. (2002). Cationiclipid-DNA complexes—lipoplexes—for gene transfer andtherapy. Bioelectrochemistry 58, 53–64.

Address correspondence to:Dr. Xia Zhao

Gynecological Oncology of Biotherapy LaboratoryDepartment of Gynecology and Obstetrics

West China Second HospitalSichuan University, No. 20, Section 3

South People’s RoadChengdu 610041, SichuanPeople’s Republic of China

E-mail: xia-zhao@126.com

Received for publication March 21, 2011;accepted after revision April 22, 2011.

Published online: April 22, 2011.

1422 XIE ET AL.