European Journal of Pharmaceutical Sciences 52 (2014) 48–61

Contents lists available at ScienceDirect

European Journal of Pharmaceutical Sciences journal homepage: www.elsevier.com/locate/ejps

Efficient siRNA delivery and tumor accumulation mediated by ionically cross-linked folic acid–poly(ethylene glycol)–chitosan oligosaccharide lactate nanoparticles: For the potential targeted ovarian cancer gene therapy Tony Shing Chau Li a, Toshio Yawata b,c, Koichi Honke a,c,⇑ a

Department of Biochemistry, Kochi Medical School, Kochi University, Oko-Cho, Nankoku, Kochi 783-8505, Japan Department of Neurosurgery, Kochi Medical School, Kochi University, Oko-Cho, Nankoku, Kochi 783-8505, Japan c Center for Innovative and Translational Medicine, Kochi Medical School, Kochi University, Oko-Cho, Nankoku, Kochi 783-8505, Japan b

a r t i c l e

i n f o

Article history: Received 30 April 2013 Received in revised form 4 October 2013 Accepted 21 October 2013 Available online 29 October 2013 Keywords: Folic acid Chitosan Nanoparticles Targeting siRNA Ovarian cancer

a b s t r a c t For effective ovarian cancer gene therapy, systemic administrated tumor-targeting siRNA/folic acid– poly(ethylene glycol)–chitosan oligosaccharide lactate (FA–PEG–COL) nanoparticles is vital for delivery to cancer site(s). siRNA/FA–PEG–COL nanoparticles were prepared by ionic gelation for effective FA receptor-expressing ovarian cancer cells transfection and in vivo accumulation. The chemical structure of FA– PEG–COL conjugate was characterized by MALDI-TOF-MS, FT-IR and 1H NMR. The average size of siRNA/ FA–PEG–COL nanoparticles was approximately 200 nm, and the surface charge was +8.4 mV compared to +30.5 mV with siRNA/COL nanoparticles. FA–PEG–COL nanoparticles demonstrated superior compatibility with erythrocytes in terms of degree of aggregation and haemolytic activity and also effects on cell viability was lower when compared with COL nanoparticles. FA grafting significantly facilitated the uptake of nanoparticles via receptor mediated endocytosis as demonstrated by flow cytometry. The in vitro transfection and gene knockdown efficiency of HIF-1a were superior to COL nanoparticles (76– 62%, respectively) and was comparable to Lipofectamine 2000 (79%) as demonstrated by RT-qPCR and Western blot. Gene knockdown at the molecular level translated into effective inhibition of proliferation in vitro. Accumulation efficiency of FA–PEG–COL nanoparticles was investigated in BALB/c mice bearing OVK18 #2 tumor xenograft using in vivo imaging. The active targeting FA–PEG–COL nanoparticles showed significantly greater accumulation than the passive targeting COL nanoparticles. Based on the results obtained, siRNA/FA–PEG–COL nanoparticles show much potential for effective ovarian cancer treatment via gene therapy. Ó 2013 Elsevier B.V. All rights reserved.

Abbreviations: ANOVA, analysis of variance; BCA, bicinchoninic acid; COL, chitosan oligosaccharide lactate; cps, centipoise; CV, coefficient of variation; DCC, dicyclohexylcarbodiimide; DCl, deuterium chloride; DEPC, diethyl pyrocarbonate; DLS, dynamic light scattering; DMEM, Dulbecco’s modified eagle’s medium; ECL, enhanced chemiluminescence; EDC, N-(3-dimethylaminopropyl)-N-ethylcarbodiimide; EPR, enhanced permeability and retention effect; EtBr, ethidium bromide; FA, folic acid; FA–PEG–COL, folic acid–poly(ethylene glycol)–chitosan oligosaccharide lactate; FBS, fetal bovine serum; FT-IR, Fourier transformed infrared spectroscopy; HIF-1a, hypoxic inducible factor-1 alpha; HRP, horseradish peroxidise; MALDI-TOF-MS, matrix-assisted laser desorption/ionization-time of flight mass spectrometry; MWCO, molecular weight cut off; NH2–PEG–COOH, a-aminopropylx-carboxypentyloxy-polyxyethylenehydrochloride; NHS, N-hydroxysuccinimide; NIRF, near-infrared fluorescence; NPs, nanoparticles; NTC, no template control; PDI, polydispersity index; PVDF, polyvinylidene fluoride; RBCs, red blood cells; RGD peptide, arginine–glycine–aspartic acid peptide; RT-qPCR, real-time quantitative polymerase chain reaction; TBE, Tris/borate/EDTA; TBS-T, Tris-buffered saline with Tween; TPP, tripolyphosphate. ⇑ Corresponding author at: Department of Biochemistry, Kochi Medical School, Kochi University, Oko-Cho, Nankoku, Kochi 783-8505, Japan. Tel.: +81 88 880 2313; fax: +81 88 880 2314. E-mail address: [email protected] (K. Honke). 0928-0987/$ - see front matter Ó 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ejps.2013.10.011

1. Introduction Ovarian cancer has the highest mortality rate and is among the most common of female malignancies in Western countries (Tagawa et al., 2012). Ovarian cancer is often undetectable in its early stages and therefore diagnosis usually occurs when surgical treatment is no longer an effective option (Hua et al., 2009). Moreover, ovarian cancer cells are known to develop resistance to standard chemotherapeutic treatments (Tagawa et al., 2012). Gene silencing via RNA interferences (RNAi) mediated by short interfering RNA (siRNA) has enormous therapeutic potential for the treatment of cancer. Many genes associated with regulation of proliferation and angiogenesis are mutated in cancer resulting in uncontrolled proliferation and these genes are a potential target for the gene silencing therapy (Dykxhoorn and Lieberman, 2006; Cho et al., 2008; Davis et al., 2010). Hypoxic inducible factor-1a (HIF-1a) is often overexpressed in cancers including ovarian cancer

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

and it is associated with tumor aggressiveness, angiogenesis, cell migration, proliferation, survival, glucose metabolism, metastasis and drug resistance through oncogene gain-of-function and tumor suppressor gene loss-of-function (Semenza, 2001; Yeo et al., 2004). Aside from induction via hypoxic conditions associated with tumor microenvironment, genetic alterations in cancer cells can induce HIF-1a expression via a non-hypoxic stimulation, leading to cancer progression (Hirota and Semenza, 2006). We here predicted that HIF-1a suppression via siRNA technique would provide effective tumoricidal outcome in human ovarian cancer cells. In fact, siRNA knockdown of the HIF-1a gene has been reported to inhibit angiogenesis and tumor cellular energy production resulting in growth suppression (Ziello et al., 2007; Chiavarina et al., 2010). However, the use of siRNA gene therapy is hindered because achieving sufficient concentration of siRNA at the tumor site(s) is difficult. siRNA has a high degradation rate in serum due to its physical characteristics, a rapid elimination by the renal pathway and a low permeability across cellular membranes (Miyagishi et al., 2004; Sioud and Sorensen, 2003; Sorensen and Sioud, 2003). One way to enhance the delivery of siRNA to the site of action is a development of a suitable delivery platform with characteristics that enables biocompatibility, a high loading capacity, protection of siRNA during transport and a high targeting ability (Creusat et al., 2012). Recently the delivery system for gene therapy has moved from viral vectors to synthetic and natural cationic polymers because viral vectors have potential to evoke immunogenic responses and can be hazardous during preparation (Liang et al., 2009). A representative cationic polysaccharide is a natural substance, chitosan (Liang et al., 2009; Chan et al., 2007; Salmaso et al., 2004). Chitosan forms ionic interactions with siRNA via a high net positive charge of its amino group. Moreover, nano-metric particles are easily formed by crosslinking chitosan with a counter ion such as tripolyphosphates (TPP), which particle provides a protection against degradation of loaded siRNA (Park et al., 2010). There are problems in the use of chitosan for gene delivery. First, excessive positive charge that remains on the surface of the nanoparticles after formulation brings about interaction with red blood cells (RBCs), opsonization and activation of immune system resulting in elimination of them. PEGylation has been reported to reduce their surface charge and thus prolong their circulating half-life (Hua et al., 2009; Dykxhoorn and Lieberman, 2006; Cho et al., 2008; Park et al., 2010). Second, achieving a sufficient concentration of siRNA at the tumor site in a timely manner is difficult, as systemically administrated chitosan nanoparticles are only passively delivered there via the enhanced permeability and retention (EPR) effect (Chan et al., 2007). Third, uptake of chitosan nanoparticles by nonspecific endocytosis in tumor cells results in a low siRNA transfection efficiency. To address these problems, an active targeting system that can also aid the uptake of nanoparticles is required. Although targeting ligands, such as RGD peptide (Park et al., 2010), epitope of Herceptin (Park et al., 2010) and fibroblast growth factor (Kalli et al., 2008), have been developed for the enhancement of targeting ability of delivery systems, high cost and handling issues are often associated with such compounds (Parker et al., 2005). In the present study we have emplyed folic acid (FA) as a targeting ligand because FA is harmless on normal cells, little immunogenic, inexpensive and stable under both in vitro and in vivo conditions (Creda-Cristerna et al., 2011; Yang et al., 2010) and furthermore the expression level of the FA receptor is high in ovarian cancer cells (Kalli et al., 2008; Parker et al., 2005; Werner et al., 2011), which promotes the receptor-mediated endocytosis of nanoparticles. The present study shows the potential utility of an siRNA delivery system with FA–PEG–COL nanoparticles encapsulating an HIF1a siRNA for a targeted ovarian cancer gene therapy.

49

2. Materials and methods 2.1. Materials 2.1.1. Chemicals Chitosan oligosaccharide lactate (COL) (3–5 kDa), sodium tripolyphosphate (TPP), folic acid P 97%, diethyl pyrocarbonate P 98% (DEPC) and fluorescein isothiocyanate (FITC) were purchased from Sigma Aldrich (St. Louis, MO). a-Aminopropyl-xcarboxypentyloxy-polyxyethylenehydrochloride (NH2–PEG– COOH) was purchased from NOF Corporation (Tokyo, Japan). Dicyclohexylcarbodiimide (DCC), N-hydroxysuccinimide (NHS), 35% deuterium chloride solution in deuterium oxide 99.5% and DNA Step Ladder (10–100 bp) were purchased from Wako Pure chemical industries Ltd (Osaka, Japan). N-(3-dimethylaminopropyl)-N-ethylcarbodiimide (EDC) was purchased from Acros organics. Alexa Fluor 647 carboxylic acid succinimidyl ester was purchased from Molecular probes (Eugene, OR), Silencer select validated siRNAs targeting against human HIF-1a (x2); Sense-strand; CCUCAGUGUGGGUAUAAGATT and CCAUAUAGAGAUACUCAAATT, Anti-sense strand; UCUUAUACCCACACUDAGGTT and UUUGAGUAUCUCUAUAUGGTT and Silencer Select Negative Control #1 siRNA were purchased from Ambion (Japan K.K., Tokyo, Japan), Lipofectamine 2000 reagent was purchased from Invitrogen (Carlsbad, CA, USA).

2.1.2. Cells and animals Human ovarian endometriod carcinoma OVK18#2 cells were purchased from RIKEN cell bank (Ibaraki, Japan), and were cultured in RPMI 1640 medium (Wako chemical industries Ltd., Osaka, Japan) or FA-free RPMI 1640 (Gibco by Life technologies, New York, USA) supplemented with 10% fetal bovine serum (FBS) (Gibco, Invitrogen Corporation, NY, USA). Mouse macrophage RAW 264.7 was cultured in Dulbecco’s Modified Eagle’s Medium (DMEM) (Wako chemical industries Ltd., Osaka, Japan) supplemented with 10% FBS and was a gift from Dr. Kotani, Kochi Glycobiology Research Centre, Kochi Medical School, Kochi University, Japan. BALBc nude/nude mice were purchased from Japan SLC Inc. (Hamamatsu, Japan). Animals were kept under pathogen-free conditions and exposed to 12 h light/dark cycle. All animal experiments were performed in accordance with guidelines approved by the ethics committee of Kochi Medical School, Kochi University. Four weeks prior to experimentations, animals were fed an alfalfa-free chow.

2.2. Methods 2.2.1. Synthesis of folic acid–poly ethylene glycol (FA–PEG) conjugate The conjugation of FA to NH2–PEG–COOH was adapted from a previously reported method with minor modifications (Liang et al., 2009; Chan et al., 2007; Salmaso et al., 2004; Park et al., 2010). Briefly, FA dissolved in dehydrated dimethyl sulfoxide (DMSO) was reacted with DCC and NHS under nitrogen atmosphere in the dark at ambient temperature for 18 h to form an activated FA (FA/NHS/DCC molar ratio, 1:5:5), the solution was filtered through a 0.2 lm Teflon syringe filter to remove insoluble dicyclohexylurea side product. NH2–PEG–COOH was added to the activated FA solution at a molar ratio of 10:1 (FA:PEG). The reaction was allowed to proceed under nitrogen atmosphere in the dark at ambient temperature for 8 h. The reactant was then precipitated with excess acetone and the precipitate was dialyzed against MilliQ water (Spectra/Por 6; MWCO, 1000) for 5 days. The precipitate was centrifuged and freeze-dried (DC 801, Yamato Scientific Co., Ltd., Tokyo, Japan).

50

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

2.2.2. Synthesis of FA–PEG–COL conjugate To conjugate FA–PEG with COL, 500 mg of COL was dissolved in 50 ml of Milli-Q water. Sixty mg of FA–PEG, 3 mg of EDC and 1 mg of NHS were then added to COL solution at ambient temperature. The reaction was allowed to proceed overnight in the dark with slow stirring. To remove un-reacted components, the reactant solution was dialyzed against Milli-Q water (Spectra/Por 3; MWCO 3500) for 5 days. The reactant was centrifuged and the supernatant was freeze-dried. 2.2.3. Fluorescein isothiocyanate (FITC) and Alexa fluor 647 labelling of COL and FA–PEG–COL conjugate The labelling of COL and FA–PEG–COL with FITC for cell uptake profile of nanoparticles was adapted from previously reported methods with minor modification (Chae et al., 2005; Huang et al., 2002; Onishi and Machida, 1999). The labelling of amine reactive Alexa fluor 647 for in vivo imaging of nanoparticles was completed with the same procedures as FITC labelling but minor adjustments were made to the amount of fluorescent label used. The mechanism of labelling was based on the reaction between the isothiocyanate group of FITC or the carboxylic acid/succinimidyl ester group of Alexa Fluor 647 and the primary amino group on COL. Fifty mg of COL or FA–PEG–COL was dissolved in 5.0 ml of DMSO/H2O co-solvent system (90/10 by vol). FITC (5.0 wt% of COL) dissolved in acetone or Alexa Fluor 647 (1.0 wt% of COL) dissolved in DMSO was added into the COL or FA–PEG–COL solution. The reaction was allowed to proceed overnight in the dark with shaking at ambient temperature. The solution mixture was precipitated with excess acetone and was centrifuged (6000g, 40 min). The supernatant was discarded and the precipitant was re-dissolved in 20 ml of Milli-Q water. Remaining un-reacted fluorescence labels were removed by dialysis against Milli-Q water for 5 days (Spectra/Por 6; MWCO 3500) before freeze-drying. The labelling efficiency of FITC was determined as previously described (Huang et al., 2002; Onishi and Machida, 1999). Briefly, fluorescence intensity reading of the FITC labelled-conjugate dissolved in a DMSO/H2O (90/10 v/v) was measured on a fluorometer (kexc 490 nm, kemi 520 nm) (Fluoroskan Ascent FL, Thermo Labsystems, Helsinki, Finland). The degree of labelling (percent) was calculated as the percent weight of FITC to weight of the FITC labelled-conjugate. The labelling efficiency of Alexa fluor 647 was determined as FITC using spectrophotometric measurement at 648 nm (UV–visible Spectrophotometer Ultrospec 2100 pro, Amersham Biosciences, Buckinghamshire, UK) instead of fluorescence intensity measurements.

2.2.6. Proton nuclear magnetic resonance spectroscopy (1H NMR) The chemical structure of FA–PEG–COL conjugate was obtained using a 400 MHz NMR spectrometer (JNM-LA 400 (Lambda 400) JEOL, Tokyo, Japan). FA–PEG–COL conjugate was dissolved in a DCl/D2O co-solvent system before analysis. 2.2.7. Fabrication of nanoparticles Nanoparticles were prepared based on the procedures reported by Calvo et al. (1997) with modification. Briefly, nanoparticles were obtained spontaneously via a modified ionic gelation method involving the addition of aqueous TPP solution drop-wisely to COL or FA–PEG–COL aqueous solution (62 °C) at a weight ratio of 1:7.1 (TPP:COL) and 1:8.1 (TPP:FA–PEG–COL) under constant agitation (tube mixer, model 2280, Wakenyaku, Co., Ltd., Kyoto, Japan). The formed nanoparticles were continuously agitated for 30 min in the dark at ambient temperature and the nanoparticles were further incubated at ambient temperature for 30 min before use or analysis. 2.2.8. Preparation of siRNA loaded COL and FA–PEG–COL nanoparticles Short interfering RNA loaded TPP-cross linked nanoparticles were prepared based on the procedures reported by Katas and Alpar (2006) with modification. For encapsulation of siRNA into COL or FA–PEG–COL nanoparticles, 18.2 ll of siRNA (1.38 lg/ll) in DEPC treated water was added to TPP solution (500 ll, 0.74 mg/ ml for COL and 0.70 mg/ml for FA–PEG–COL) before adding dropwisely to pre-warmed (62 °C) COL or FA–PEG–COL solution (500 ll, 6 mg/ml), yielding a TPP to COL or FA–PEG–COL weight ratio of 1:8.1 and 1:8.6, respectively. The nanoparticles were shaken in the dark for 30 min and were incubated at ambient temperature for 30 min before use or analysis. 2.2.9. Dynamic light scattering and zeta potential Particle size and polydispersity index (PDI) of the different formulations were determined using a dynamic light scatting method. Samples were re-suspend in Milli-Q water, the hydrodynamic diameter of the particles were measured with an dynamic light scattering particle size analyser (ELS-8000, Photal, Osaka, Japan) employing a He–Ne laser light scattering at an angle of 90° at 25 °C. The zeta potential of the different formulations were determined by a laser Doppler microelectrophoresis Zetasizer (ELS 21, Photal, Osaka, Japan) at an angle of 22° at 25 °C.

2.2.4. MALDI-TOF mass spectrometry of FA–PEG conjugate Conjugation between FA and NH2–PEG–COOH was verified with MALDI-TOF-MS (Applied Biosystems SCIEX TOF/TOF 5800). Freezedried FA–PEG was dissolved in DMSO. a-Cyano-4 hydroxycinnamic acid (5 mg/ml) dissolved in acetonitrile solution was freshly prepared prior to experimentation as the matrix. An aliquot of 1 ll of sample to matrix ratio 1:9 was spotted onto a MALDI-TOF-MS plate and allowed to air dry in the dark before analysis.

2.2.10. Determination of siRNA encapsulation efficiency The encapsulation efficiency of siRNA (%) in COL and FA–PEG– COL nanoparticles (formulated as per Section 2.2.8) were determined by absorbance measurement at 257 nm of the free siRNA concentration in the supernatant recovered after nanoparticles collection by centrifugation (13,000g, 15 min). Supernatant recovered from unloaded COL or FA–PEG–COL nanoparticles were used as the blank. The siRNA encapsulation efficiency (%) was calculated as the difference between the total amount of siRNA added and the amount of siRNA remaining in the supernatant after centrifugation to the total amount of siRNA added. Concentrations were obtained from absorbance using standard curve prepared with standard siRNA solution.

2.2.5. Fourier transform infrared spectroscopy (FT-IR) of FA, NH2– PEG–COOH and FA–PEG conjugate Fourier transform infrared spectroscopy (FT/IR-460, JASCO Corporation, Tokyo, Japan) was used to further confirm FA–PEG conjugation. One milligram of FA, NH2–PEG–COOH or FA–PEG conjugate was mixed with 200 mg previously dried potassium bromide (KBr) and the mixed powders were compressed to form a disc. Background scan using pure KBr disc was performed before samples were analyzed. Spectra were obtained from 500 to 4000 cm1.

2.2.11. In vitro release profile The in vitro release study was carried out as follows: 100 ll, equivalent to 13.75 lg of siRNA was loaded into either COL or FA–PEG–COL nanoparticles. siRNA loaded nanoparticles were resuspended in 1.3 ml of 1 x phosphate buffered saline (PBS) (pH 7.4) and placed into a 1.5 ml tube and into a water bath with shaker at 37 °C. Four hundred ll of sample was taken out at pre-determined time point (3, 9, 12, 24, 36, 48, 60, 84 and 108 h) and was transferred into another 1.5 ml tube. The tube with the 400 ll of

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

sample was centrifuged (13,000g, 10 min). Two hundred ll of the supernatant was removed and the absorbance taken at 257 nm. The remainder of the 400 ll of supernatant was discarded leaving the nanoparticles at the bottom of the tube. Four hundred ll of fresh pre-warmed PBS was used to re-suspend the nanoparticles. The re-suspended nanoparticles were then returned to the original tube and returned to the water bath. The cumulative amount of siRNA released and percentage of total siRNA released were integrated from the first measurement. Concentrations were obtained from absorbance using standard curve prepared with standard siRNA solutions. 2.2.12. Serum stability of siRNA-loaded nanoparticles 2.2.12.1. siRNA recovery from formulation. The siRNA recovery efficiency from formulation using heparin as an extraction solvent was assessed using 4% agarose gel (containing 5 lg/ml ethidium bromide (EtBr)). siRNA (100 ll, equivalent to 13.75 lg of siRNA) was loaded into COL nanoparticles. The nanoparticles were centrifuged (13,000g, 15 min) and the supernatant discarded. 200 ll of heparin (10,000 U/ml) (Ajinomoto, Tokyo, Japan) was added and shaken on a tube shaker for 15 min at full speed to displace the loaded siRNA from the nanoparticles. The nanoparticles were centrifuged (13,000g, 15 min) and sample of the supernatant was mixed with loading dye (6x Blue/Orange Loading dye, Wako Pure chemical Industries) and added to the gel. Unformulated siRNA of known concentrations were loaded and ran on the same gel for the purpose of determining the response (concentration versus signal). Electrophoresis was carried out at a constant voltage of 100 V (Mupid-2plus, Takara, Japan) in TBE buffer (45 mM Tris–borate and 1 mM EDTA). siRNA bands were visualized under a UV transilluminator (Fluo_Link, Vilber Lourmat, France) equipped with a camera (Bioinstrument, Atto, Japan) and an image saver (AE6905H image saver HR, Atto, Japan). The resulting image was processed with Image J, (U.S. National Institutes of Health, Maryland, USA). Recovery method efficiency was tested with COL nanoparticles formulation only since the electrostatic interaction between COL and siRNA was stronger than FA–PEG–COL. Therefore, if acceptable recovery of siRNA from COL nanoparticles were obtained (85–115% with a % CV of less than 5% (US FDA, 2001)), the recovery method was deemed applicable for use with FA–PEG– COL formulation. 2.2.12.2. Serum stability of siRNA-loaded nanoparticles. Serum stability of siRNA-loaded nanoparticles was based on the procedures reported by Katas and Alpar (2006) with modification. Chitosan oligosaccharide lactate and FA–PEG–COL nanoparticles loaded with siRNA (100 ll, equivalent to 13.75 lg of siRNA) were incubated at 37 °C with equal volume of DMEM supplemented with 50% final concentration of FBS. Serum was used without heat inactivation. At each pre-determined time point (0, 30 min, 1, 3, 6, 9, 12, 24 and 36 h) samples were removed and stored at 20 °C until gel electrophoresis was performed. To terminate serum activity, samples were incubated in a water bath at 80 °C for 10 min and then centrifuged (13,000g, 15 min). The supernatant was removed and 200 ll of heparin (1000 U/ml) was added to each sample then shaken and centrifuged. The integrity of the siRNA was analyzed by 4% agarose gel (containing 5 lg/ml EtBr). Electrophoresis was carried out at a constant voltage of 100 V in TBE buffer (45 mM Tris– borate and 1 mM EDTA). The siRNA bands were visualized under a UV transilluminator equipped with a camera and an image saver. 2.2.13. Formulation blood compatibility assays 2.2.13.1. RBCs aggregation assay. Freeze-dried COL or FA–PEG–COL nanoparticles were re-suspended in PBS at the following concentrations of 0.1, 1.0, 2.0, 5.0, 10.0 and 15 mg/ml.

51

Mouse blood was collected by cardiac puncture. RBCs were collected by centrifugation (1500g, 10 min at 4 °C) and were diluted 100 times with pre-chilled PBS. 250 ll of RBCs and 250 ll of formulation were added to a 1.5 ml tube and shaken gently for 1 min. The samples were allowed to incubate at 37 °C for 60 min. Each sample was mixed gently by pipetting. Five ll of each sample was placed on a glass slide and cover slip on top. The samples were then viewed with a phase contrasting microscope (IMT-2, Olympus corporation, Tokyo, Japan) equipped with a camera (DP12, Olympus corporation, Tokyo, Japan). 2.2.13.2. Hemolysis assay. Freeze-dried COL nanoparticles or FA– PEG–COL were re-suspended in PBS at the following concentrations of 0.1, 1.0, 2.0, 5.0, 10.0 and 15 mg/ml. Mouse blood was collected and prepared as described in the RBCs aggregation assay. Two hundred and fifty ll of RBCs and 250 ll of formulation were added to a 1.5 ml tube and shaken gently for 1 min. The samples were allowed to incubate at 37 °C for 60 min. The samples were centrifuged (1500g, 15 min at 4 °C) and the absorbance of the supernatants was taken at 541 nm. PBS was used as the negative control and Triton X-100 (2% w/v) was used as 100% lysis positive control. The relative hemolytic activities of FA–PEG–COL and COL nanoparticles were calculated by:

Relative hemolytic activity ð%Þ   ½Abssample  ½AbsPBS  100 ¼ ½AbsTriton  ½AbsPBS

ð1Þ

2.2.14. Effect of formulation on cell proliferation (cytotoxicity assay) The WST-1 proliferation assay (Takara Bio Inc., Shiga, Japan) was used to evaluate the cytotoxicity of COL and FA–PEG–COL nanoparticles. Mouse RAW 264.7 mouse macrophage cells were seeded in 96well plates at a density of 5000 cells per well in 50 ll of DMEM supplemented with 10% FBS and placed into a cell culture incubator (95% air, 5% CO2 at 37 °C). The cells were allowed to adhere overnight. Freeze-dried COL or FA–PEG–COL nanoparticles were re-suspended in PBS and diluted to 0.1, 1.0, 2.0, 5.0, 10.0 and 15 mg/ml. Forty ll of each formulation at the respective concentrations were added to the wells and allowed to incubate for 16 h, 10 ll of WST-1 premix was added to each well and the cells were incubated for a further 4 h before the absorbance was taken at 450 nm (600 nm was used as the reference) with a micro-plate reader (Molecular Devices, Sunnyvale, CA, USA). 2.2.15. Cellular uptake profile of formulations 2.2.15.1. Flow cytometry. OVK18#2 human ovarian cancer cells were cultured in 35 mm plates in 2 ml of FA-free RPMI 1640 at 5  105 cells per plate and placed into a cell culture incubator (95% air, 5% CO2 at 37 °C). Cells were allowed to adhere overnight. 500 ll of FITC–COL or FITC–FA–PEG–COL nanoparticles (formulated from TPP (500 ll) and FITC–COL or FITC–FA–PEG–COL solution (500 ll, 6 mg/ml) at a weight ratio of 1:8.1 and 1:8.6, respectively) was added and allowed to incubate for 3, 6, 9, 12, 24, 36 and 48 h. At each time point, samples were removed and washed 3 times with ice cold PBS, cells were then trypzinated, re-suspended in PBS and filtered through a 40 lm nylon mesh screen filter. Quantitative cellular uptake of formulations was performed using a FACS Calibur flow cytometer (Becton–Dickinson, San Jose, CA, USA). 2.2.16. In vitro gene knockdown efficiency (HIF-1a) In vitro transfection and gene knockdown studies were performed in human ovarian cancer OVK18#2 cells. Cells were seeded

52

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

in 6 well plate at a density of 1.75  106 cells per well in 2.0 ml FAfree RPMI 1640 containing 10% FBS and was placed into a cell culture incubator (95% air, 5% CO2 at 37 °C) and incubate for 8 h before changing to fresh FA-free RPMI 1640 containing CoCl2 (150 lM final concentration). HIF1-a siRNA (60 nM final concentration) was transfected with COL or FA–PEG–COL nanoparticles. Briefly, HIF1a siRNA (150 pmol) was loaded into COL or FA–PEG–COL nanoparticles. Nanoparticles were collected by centrifugation (13,000g, 15 min) and re-suspended in 500 ll of Opti-MEM I reduced serum medium (Gibco, Invitrogen Corporation, NY, USA) before addition to the cells. Lipofectamine 2000 was loaded with HIF1-a siRNA according to the manufacturer’s instructions for gene knockdown comparisons. Non-coding siRNA was loaded into each of the formulations as negative control (indicating knockdown specificity). Transfection/gene knockdown was allowed to proceed for 24 h before analysis. HIF1-a gene knockdown were assessed at two levels, at the protein level using SDS–PAGE followed by Western blotting and at the mRNA level using Real Time quantitatively-PCR (RT-qPCR). 2.2.16.1. Western blot analysis. Cells were washed twice with icecold PBS, total proteins were extracted with RIPA buffer (50 mM Tris, HCl pH 7.4, 1% Triton X-100, 0.5% sodium doxycholate, 0.1% SDS, 150 mM NaCl, 2 mM EDTA and complete protease inhibitor cocktail (Roche Diagnostics, Mannheim, Germany). Cells were scraped and agitated. The lysate was allowed to stand on ice for 30 min with vortex every 10 min. The lysate was sonicated for 1 min and centrifuged (14,000g, 15 min at 4 °C). Protein concentrations of the lysates were measured with a bicinchoninic acid (BCA) protein assay kit (Thermos scientific, Rockford, IL, USA). Twentyfive lg of proteins was mixed with 2x loading buffer and heated for 5 min at 98 °C before loading. Proteins were resolved on a 8% sodium polyacrylamide–SDS gel and was transferred to a polyvinylidene fluoride (PVDF) membrane (Millipore Corporation, Billerica, MA, USA) followed by blocking in 5% non-fat milk in TBS-T at ambient temperature for 1 h. The membrane was probed with a goat anti-human HIF-1a primary antibody (R&D system, Minneapolis, MN, USA) in blocking buffer (1:2500) overnight at 4 °C followed by 3 washes of 10 min in TBS-T. The membrane was incubated with a horseradish peroxidise (HRP)-conjugated donkey anti-goat secondary antibody (Santa Cruz Biotechnology, Santa Cruz, CA, USA) in blocking buffer (1:50,000) for 1 h at ambient temperature followed by 3 washes of 10 min in TBS-T. Immuncomplexes were visualized using ECL (Millipore Corporation, Billercia, MA, USA). Antibodies were stripped with stripping buffer (64.5 mM Tris–HCl (pH 6.8), 0.1 M 2-Mercaptoethanol and 2% SDS in Milli-Q water) at 50 °C for 30 min. The membrane was washed 3 times in TBS-T for 10 min each, followed by blocking in 5% skim milk in TBS-T for 1 h in ambient temperature. The membrane was probed with rabbit anti-b-Actin primary antibody (Cell Signalling Technology Inc., USA) in blocking buffer (1:40,000) overnight at 4 °C followed by 3 washes of 10 min in TBS-T. The membrane was incubated with a HRP-conjugated goat anti-rabbit secondary antibody (Promega, Madison, WI, USA) in blocking buffer (1:50,000) for 1 h at ambient temperature followed by 3 washes of 10 min in TBS-T. Immuncomplexes were visualized using ECL as described previously. 2.3.16.2. Total RNA extraction and RT-qPCR analysis. Total RNA was isolated using the GenElute Mammalian total RNA Miniprep kit (Sigma Aldrich Co., St. Louis, MO, USA) along with an on-Column DNase I Digestion kit (Sigma Aldrich Co., St. Louis, MO, USA) to minimize DNA contamination. The quality and quantities of the RNAs were determined by absorbance measurement at 260 nm (Nanodrop Technology, Wilmington, DE, USA). One mg of the total RNA was reverse transcribed into cDNA using a High-Capacity

cDNA Reverse Transcription Kit (Applied Biosystems, Foster City, CA, USA). RT-qPCR reaction was carried out in a final volume of 10 ll using Power SYBR Green Master Mix (Applied Biosystems, Warrington, UK). Reaction conditions were as follows: 95 °C for 10 min, 40 cycles at 95 °C for 30 s, 60 °C for 30 s, and 72 °C for 30 s. Melt curve analysis was performed after every amplification program to verify specificity of the target and the absence of primer dimmers. No-template control (NTC) was included in each assay to verify that the PCR master mixes was free from contamination. The following primers were used: HIF-1a forward 50 -GTACCCTAACTAGCCGAGG-30 , HIF-1a reverse 50 -GTGAATGTGGCCTGTGCAG-30 (Eurofins MWG Operon, Tokyo, Japan) at a final concentration of 300 nM forward and reverse, and 18S forward 50 -GTAACCCGTTGAACCCCATT-30 , 18S reverse 50 CCATCCAATCGGTAGTAGCG-30 at a final concentration of 150 nM forward and reverse. The 18S housekeeper gene was validated against qPCR Human Reference Total RNA (Clonetech Laboratories, Mountain View, CA, USA). The quantity in each samples were determined from the experimental threshold cycles on a standard curve of the data from a series of serial dilutions of the mixture of generated cDNA. The mRNA level of HIF-1a was normalized using 18S as an endogenous control. 2.2.17. Biological activity of siRNA (HIF1-a) loaded formulation on ovarian cancer cells proliferation OVK18#2 human ovarian cancer cells were seeded in 96-well plate at a density of 25,000 cells per well in 100 ll of FA-free RPMI 1640 containing 10% FBS and was placed into a cell culture incubator (95% air, 5% CO2 at 37 °C) and was allowed to incubate and adhere for 8 h before changing to fresh FA-free RPMI 1640 containing CoCl2 (150 lM final concentration). HIF1-a siRNA (60 nM final concentration) was transfected with COL nanoparticles or FA–PEG– COL nanoparticles. Briefly, siRNA (HIF1-a) was loaded into COL nanoparticles or FA–PEG–COL nanoparticles. The nanoparticles were collected by centrifugation (13,000g, 15 min) and re-suspended in 50 ll of Opti-MEM I Reduced serum medium before addition to the cells. Lipofectamine 2000 was loaded with HIF1-a siRNA for comparisons. Cells were returned back into the incubator and allowed to incubate for 20 h. Fifteen ll of WST-1 premix solution was added into each well and the cells were incubated for an additional 4 h before the absorbance was taken at 450 nm with reference absorbance at 600 nm on a micro-plate reader and the effect of gene knockdown on cell proliferations determined. 2.2.18. In vivo ovarian cancer xeno-graft accumulation efficiency Six weeks old BALB nu/nu mice week were implanted with 28  106 OVK18#2 human ovarian cancer cells. The xeno-graft was allowed to grow for 4 weeks and mice were fed an alfalfa free chow. Nanoparticles were formulated from TPP (500 ll) and Alexa Fluor 647–COL or Alexa Fluor 647–FA–PEG–COL (500 ll, 6 mg/ml) at a weight ratio of 1:8.1 and 1:8.6, respectively. The nanoparticles were collected by centrifugation (13,000g, 15 min) and re-suspended in 400 ll of PBS. Mice were given via the tail vein a 400 ll bolus injection of either, Alexa Fluor 647–COL nanoparticles, Alexa Fluor 647–FA–PEG–COL nanoparticles or PBS. In vivo nearÒ infrared fluorescence (NIRF) images were captured using an IVIS Spectrum imaging system (Caliper LifeScience, Hopkinton, MA, USA) at 3, 6, 12 and 24 h with filters set at (kexc 640 nm, kemi 680 nm). The data were analyzed by the software package supplied by the system’s manufacturer. 2.2.19. Statistical analysis Data are presented as the means ± standard deviations (SD). Statistical analysis was performed using GraphPad Prism Version 6.0 software.

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

3. Results

53

cles, requiring approximately 48 h compared to 30 h for 50% of the loaded siRNA to be released, respectively.

3.1. Synthesis and characterization of FA–PEG–COL 3.5. Serum stability of siRNA-loaded nanoparticles Folic acid (FA) was linked to COL using a hetero-bifunctional PEG. The synthesis of the tri-polymer was completed in two stages: (1) FA conjugation to NH2–PEG–COOH and, (2) FA–PEG–COOH conjugation to COL. MALDI-TOF-MS and FT-IR were used to verify the conjugation of FA to PEG. The mass/charge (m/z) value of NH2–PEG–COOH increased from 3294 to 3762 after conjugation with FA (Fig. 1A). The difference in molecular weight approximately corresponded to that of FA. FT-IR was further used to verify the chemical structure of FA–PEG conjugate (Fig. 1B). The spectrum of FA had the characteristic peaks at 1695 cm1 and 1606 cm1 which corresponded to the stretching vibration of (C@O) of the carboxyl group and (C@C) bond of the aromatic phenyl ring, respectively. The spectrum of PEG had a strong peak at 2886 cm1 for the stretching of (RACH2) of the alkyl group and at 1116 cm1 for the stretching of the ether group. The spectrum of FA–PEG showed dissociated peaks around 2900 cm1 derived from the alkyl group of PEG along with strong peaks at 1689 cm1 and 1627 cm1 derived from FA. The chemical structure of FA–PEG–COL conjugate was characterized by 1H NMR. Spectra were obtained in 35 wt% DCl in D2O for the de-protonation of the amine group. The peak at 7.5 ppm was attributed to the aromatic proton of FA (Fig. 2). The peak at 4.7, 3.5 and 2.7 ppm corresponded to the ACH of COL, the ACH2 of PEG and the NHACO bond between COL and PEG, respectively. The characteristic peaks specific to each of the components verified the successful conjugation between FA, PEG and COL. 3.2. Fluorescein isothiocyanate (FITC) and Alexa Fluor 647 conjugation of COL and FA–PEG–COL conjugate The labelling efficiency of FITC of COL and FA–PEG–COL were 4.7% and 4.5%, respectively, and was consistence with previous reports (Onishi and Machida, 1999). The Alexa fluor 647 content in labelled COL and FA–PEG–COL were 0.96% and 0.93%, respectively. 3.3. Formation of nanoparticles by ionic gelation and their physical properties Nanoparticles were prepared using a mild ionic gelation method. The formation of nanoparticles occurred spontaneously upon the addition of TPP aqueous solution into COL aqueous solution. The physical characteristics of the formulations are summarized in Table 1. The particle sizes and zeta potentials in the different formulations are graphically shown in Fig. 3. The particle sizes ranged from 130 to 200 nm and was dependent on the level of modification to COL. When siRNA loaded, the particle size increased and the average particle size remained below 250 nm (Fig. 3). The zeta potentials ranged from +8 to +45 mV. The positive charge of COL was lowered by conjugation with other components as well as by loading siRNA. One hundred % siRNA loading efficiency was obtained for both COL and FA–PEG–COL nanoparticles at the tested concentrations. This indicated that the modification with FA–PEG did not significantly affect the availability of reactive primary amine groups on COL for siRNA interactions. 3.4. siRNA in vitro release profile from formulations The in vitro release profiles of siRNA from COL and FA–PEG–COL nanoparticles in PBS were examined. The siRNA release rate was slower in COL nanoparticles compared to FA–PEG–COL nanoparti-

After exposure of siRNA-loaded nanoparticles for various times, remaining siRNA on the nanoparticles was recovered by heparin. The average recovery efficiency was 93 ± 3.0% with % CV of 3.2% (n = 3). The linearity of standards for the recovery method was r = 0.9989. The recovery efficiency satisfied the requirements set by the FDA for bio-analytical method recovery (results not shown) (US FDA, 2001). The serum stability of naked siRNA and siRNA loaded into the 2 formulations were tested in 50% serum-containing media for 36 h. Rapid degradation of the unformulated siRNA was observed after 1 h (Fig. 5). Some siRNA remained intact for both formulations over 36 h. COL nanoparticles provided a greater protection against serum degradation compared to FA–PEG–COL nanoparticles (Fig. 5). This coincided with the in vitro release profile (Fig. 4). 3.6. Formulation blood compatibility Effects of COL nanoparticles and FA–PEG–COL nanoparticles on RBCs were tested (Fig. 6A). RBCs severely aggregated by incubation with 2 mg/ml of COL while they did not with PBS alone. In contrast, the aggregation hardly occurred with 2 mg/ml of FA–PEG–COL and it became apparent only after 10 mg/ml. As the concentrations of both formulations increased, hemolysis level was elevated (Fig. 6B). However, a considerable difference in the hemolytic activity between COL and FA–PEG–COL nanoparticles was seen at all tested concentrations (P < 0.05, Prism multiple t-test). At the concentration of 15 mg/ml, a 4-fold greater hemolytic activity was observed with COL nanoparticles compared to FA–PEG–COL nanoparticles. The result of FA–PEG–COL nanoparticles at 15 mg/ml was similar to that of COL nanoparticles at 2 mg/ml (4.9% versus 6.1%). The hemolysis was consistent with the RBC aggregation and the marked reduction in the RBC aggregation and haemolysis with FA–PEG–COL as compared to COL suggests the effectiveness of PEGylation. 3.7. Cytotoxicity of formulations To investigate cytotoxicity of COL and FA–PEG–COL nanoparticles, their effects on cell viability of mouse RAW 264.7 macrophage cells that actively engulf the formulations were investigated by the WST-1 assay (Fig. 7). There was an opposite non-linear correlation between the concentration of COL nanoparticles and cytotoxicity (P > 0.05, Prism non-parametric correlation test). When compared to the control group, FA–PEG–COL nanoparticles showed no significant difference at all concentrations tested (Fig. 7: P = 0.9 for all samples, Prism one-way ANOVA with Dunnett’s multiple comparisons test). In contrast, when compared to the control, COL nanoparticles showed a significantly lower proliferation rate at 10 (P < 0.05) and 15 mg/ml (P < 0.001) indicating toxicity. These results indicated that FA–PEG–COL nanoparticles were significantly less toxic compared to COL nanoparticles. 3.8. Cellular uptake profile of formulations Cellular uptake of FITC–COL and FITC–FA–PEG–COL nanoparticles in ovarian cancer cells were quantified by flow cytometry (see Supplementary data for selected time points of flow cytometry data). The FA receptor-mediated uptake of the FITC–FA–PEG–COL nanoparticles peaked at 9 h with a subsequent drop in mean fluorescent intensity (MFI) (Fig. 8). The gradual reduction after 9 h reflects the proliferation of the cells that took in the

54

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

Fig. 1. Characterization of FA–PEG–COOH. (A) MALDI-TOF-MS analysis of (a) NH2–PEG–COOH and (b) FA–PEG–COOH. (B) FTIR spectra of (a) FA, (b) NH2–PEG–COOH and (c) FA–PEG–COOH.

FITC–FA–PEG–COL nanoparticles. The uptake of FITC–COL nanoparticles was slower than that of FITC–FA–PEG–COL nanoparticles and still in progress after 24 h. These results verified the advantage of employing receptor mediated uptake over non-specific uptake in the cellular uptake rate of formulations. 3.9. In vitro gene silencing of HIF-1a The HIF-1a gene was targeted by siRNA using COL nanoparticles, FA–PEG–COL nanoparticles and Lipofectamine 2000. The knockdown efficiency was determined at the protein level and mRNA level. Lipofectamine 2000 provided the greatest inhibitory effect on protein level followed by FA–PEG–COL nanoparticles and COL nanoparticles as detected by Western blotting (Fig. 9A).

Unformulated siRNA showed minimal effect in the suppression of protein activity. HIF-1a band was absent in cells not treated with CoCl2, suggesting that HIF-1a was rapidly degraded. The results of mRNA expression evidenced by RT-qPCR (Fig. 9B) coincided with protein expression. When compared to 150 lM CoCl2 treated cells (100%), unformulated siRNA reduced mRNA levels of HIF-1a to 94%, siRNA delivered by COL nanoparticles to 38% and Lipofectamine 2000 to 21%. mRNA expression of siRNA delivered by both COL and Lipofectamine 2000 reduced when compared to the CoCl2 treated group (P < 0.0001, Prism one-way ANOVA with Tukey’s multiple comparison test). There was a reduction in the mRNA level by siRNA delivered by FA–PEG–COL nanoparticles to 24% that was comparable to Lipofectamine 2000 (P = 0.99). Non-targeting siRNA with Lipofectamine, FA–PEG–COL and COL

55

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

Fig. 2. 1H NMR analysis of FA–PEG–COL tri-polymer. The peak labelled as ‘‘A’’ at 7.5 ppm was attributed to the aromatic proton of FA. The peak labelled as ‘‘B’’ at 4.7 ppm corresponded to the ACH of COL and the major peak at 3.5 ppm labelled as ‘‘C’’ corresponded to the ACH2 of PEG block.

Table 1 Physico-chemical characterization of formulations. Results are expressed as mean ± SD of at least 2 experiments (n = 3). Formulation

siRNA loaded

COL:TPP weight ratio

Particle size (nm)

Zeta potential (mV)

PDI

COL NPs COL NPs FA–PEG–COL NPs FA–PEG–COL NPs FITC–COL NPs Alexa fluor 647–COL NPs FITC–FA–PEG–COL NPs Alexa fluor 647–FA–PEG–COL NPs

No Yes No Yes No No No No

7.1:1 8.1:1 8.1:1 8.6:1 8.1:1 8.1:1 8.6:1 8.6:1

130.7 ± 9.7 161.1 ± 10.7 154.5 ± 3.8 200.3 ± 4.6 151.3 ± 6.4 189.0 ± 7.0 173.7 ± 24.2 200.0 ± 22.8

45.0 ± 5.2 30.5 ± 4.3 17.3 ± 3.8 8.4 ± 1.9 35.3 ± 4.6 27.7 ± 2.1 9.7 ± 2.2 7.9 ± 1.4

0.12 ± 0.03 0.16 ± 0.01 0.24 ± 0.06 0.20 ± 0.07 0.14 ± 0.08 0.17 ± 0.10 0.42 ± 0.11 0.36 ± 0.08

NPs denotes nanoparticles.

Fig. 3. Particle size and zeta potential of COL NPs and FA–PEG–COL NPs and the effect of siRNA loading.

nanoparticles showed no inhibitory effect on HIF-1a mRNA levels when compared to the CoCl2 treated group, indicating that the formulations themselves do not contribute to HIF-1a gene knockdown. 3.10. In vitro HIF-1a gene knock down and the effects on cell proliferations WST-1 assay was used to determine the effect of HIF-1a knockdown by siRNA on proliferation of OVK18#2 ovarian cancer cells.

Cells treated with 150 lM of CoCl2 had an approximately 5-fold increased proliferation rates when compared to untreated cells (Fig. 10). HIF-1a siRNA delivered by all tested formulations resulted in significant difference in suppression of proliferations (P < 0.0001, Prism one-way ANOVA with Tukey’s multiple comparison test) when compared with the CoCl2 treated group. The relative proliferation rate of siRNA delivered by Lipofectamine 2000, FA–PEG–COL and COL nanoparticles compared to the CoCl2 treated group was 26%, 29% and 57%, respectively. There was a significant difference in the suppression of proliferation rates between FA–

56

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

100

Cumulave siRNA released (%)

90 80

*

70 60 50 40 30 20 10 0 0

10

20

30

40

50 60 Time (Hr)

FA-PEG-COL nanoparcles

70

80

90

100

110

COL nanoparcles

Fig. 4. In vitro release profile of siRNA from formulations. Results are expressed as mean ± SD of at least 2 separate experiments (n = 6). Significant difference was observed between the formulations at the 108 h final siRNA cumulative release time-point (*P < 0.05, Prism, unpaired t test with Welch’s correction).

Ex vivo images of mouse organs further confirmed the presence of Alexa fluor 647–COL nanoparticles at the 24 h time point (Fig. 11D). However, only a low amount was seen, suggesting that majority of Alexa fluor 647–COL nanoparticles were processed by the liver at an earlier time and the remnants were persistently taken up by the liver and excreted. Uptake of Alexa fluor 647–FA– PEG–COL by liver was not seen in the in vivo and ex vivo images, indicating the effectiveness of PEGylation and rapid active targeting ability with FA. Fig. 5. Assay for serum stability: 4% agarose gel electrophoresis analysis of unformulated siRNA and siRNA released from formulations following incubation with 50% FBS. (A) unformulated siRNA; (B) siRNA from COL nanoparticles; and (C) siRNA from FA–PEG–COL nanoparticles. Cont denotes control.

PEG–COL and COL nanoparticles (P < 0.001), suggesting that the small difference in the mRNA suppression leads to an enormous difference at the end point effect. 3.11. In vivo tumor targeting efficiency The near infrared fluorescent (NIRF) intensity of Alexa fluor 647–COL nanoparticles and Alexa fluor 647–FA–PEG–COL nanoparticles were very similar with only a 0.7% difference, 5.249  1012 and 5.212  1012 (p/s/cm2/sr)/(lW/cm2), respectively before administration as determined by the in vivo imager. Fig. 11 shows the in vivo time dependent tumor accumulation profile of Alexa fluor 647–COL, Alexa fluor 647–FA–PEG–COL nanoparticles and PBS, after intravenous injections into nude mice bearing OVK18#2 human ovarian cancer cells. Both Alexa fluor 647–COL and Alexa fluor 647–FA–PEG–COL nanoparticles presented a strong NIRF signal throughout the entire body within 3 h post injection, indicating that nanoparticles were rapidly circulated via the bloodstream. Significant accumulation at the tumor site can be seen at 3 h post injection with subsequent increase at the 12 and 24 h time points with the active targeting of Alexa fluor 647–FA–PEG–COL nanoparticles. Small amounts of accumulation at the tumor site were observed at 12 h and with a slight increase at 24 h for the passive ERP effect-based targeting of Alexa fluor 647–COL nanoparticles. Significant liver accumulation was observed from 3 h through to 6 h with Alexa fluor 647–COL nanoparticles (Fig. 11A; yellow arrows), suggesting opsonization and macrophage-associated transportation to the liver reticular endothelial system (RES).

4. Discussion In this study, low molecular weight COL was chosen over traditional chitosan because of high solubility in water with low viscosity (6 cps) and high degree of deacetylation (>90%) (Sigma– Aldrich product information, 2009). The former allows uniform distribution of drugs within the particle, while the latter provides more amino groups that enable a stronger interaction with siRNA, resulting in a greater protection from degradation and a sustained release profile (Fan et al., 2012). In solution, the negatively charged TPP interacts with the protonated amino group of COL. This crosslinking results in a larger and less soluble ionic complex, which precipitates out of solution. Particle size of the formulation is a critical factor that determines cellular uptake rate. The DLS analysis showed that all the formulations in the present study achieved a sub-250 nm particle size. The average size of siRNA/COL nanoparticles was approximately 160 nm and increased to approximately 200 nm with the siRNA/FA–PEG–COL nanoparticles, indicating the successful conjugation of FA, PEG and COL. Modulation of particle size was able by adjustment of the COL and TPP ratio, which depends on availability of amino groups of COL for the interaction with TPP. To achieve the smallest possible particle size in this study, the temperature of FA– PEG–COL and COL solution was kept at 62 °C before the addition of TPP. It has been reported that an increase in temperature up to 60 °C results in a decrease in intrinsic viscosity that facilitates the formation of finer particles (Fan et al., 2012; Chen and Tsaih, 1998). Zeta potential is another important factor that can determine the fate of the formulation during transit to the target site. Excessive positive charge on the surface of the particles leads to interaction with RBCs causing aggregation or even hemolysis (Rekha and

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

57

A

B

Fig. 6. Formulation blood compatibility. (A) Effect of formulations on RBCs aggregation. (a) RBCs in PBS (pH 7.4), (b) RBCs with COL nanoparticles (2 mg/ml); (c) RBCs with FA–PEG–COL nanoparticles (2 mg/ml); (d) RBCs with FA–PEG–COL nanoparticles (5 mg/ml); (e) RBCs with FA–PEG–COL nanoparticles (10 mg/ml) and (f) RBCs with FA–PEG– COL nanoparticles (15 mg/ml). (B) Formulations hemolytic activities. Results are expressed as mean ± SD of at least 3 separate experiments (n = 3). Significant differences between the two formulations were observed at each concentration (*P < 0.05, Prism multiple t-test).

Sharma, 2009; Chein, 1975). Moreover, excess positive charge leads to opsonization with plasma proteins resulting in interaction with cells of the RES and deposition to the liver, and subsequently destruction of the delivery system. In the current study, PEGylation significantly reduced the Zeta potential of the nanoparticles (Table 1 and Fig. 3) and improved the blood compatibility (Fig. 6). The interactions with RBCs are often overlooked during formulation development (Yang et al., 2010). The RBCs aggregation is manifested by a combination of two physico-chemical mechanisms: one being neutralization of the negatively charged glycocalyx on the surface of RBCs by positively-charged particles, and the other being reduction in repulsion forces between individual RBCs allowing stable formation of physical bridge through the formula-

tion (Creda-Cristerna et al., 2011; Moreau et al., 2002, 2000). In the current study, the RBCs aggregation was considerably higher with the COL nanoparticles than the FA–PEG–COL nanoparticles, implying the role of PEGylation in prevention of the RBCs aggregation. Hemolysis of RBCs as a result of interaction with polycation formulations results in anemia, jaundice and reticulocytosis (Yang et al., 2010). The hemolytic activitiy is attributed to the electrostatic interactions between the cationic groups on the formulations and the anionic glycoproteins on the RBC membrane. This can lead to membrane destabilization with rapid redistribution of phosphatidylserine, resulting in membrane curvature and rupture (Moreau et al., 2000; Pribush et al., 2007; Fischer et al., 2003; Carreno-Gomez and Duncan, 1997). In the present study,

58

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

Fig. 7. Cytotoxicity of formulations. Results are expressed as mean ± SD of at least 2 separate experiments (n = 6). There is a negative non-linear correlation between the concentration of COL nanoparticles and cytotoxicity (P > 0.05, Prism non-parametric correlation test). Significant difference was observed when the control group was compared to COL nanoparticles at 10 mg/ml (*P < 0.05) and 15 mg/ml (***P < 0.001). (Prism one-way ANOVA with Dunnett’s multiple comparisons test). One set of control result was used to compare the two formulations.

Fig. 8. Flow cytometry analysis of OVK-18#2 cell uptake of formulations. Results are expressed as mean ± SD of at least 3 separate experiments (n = 3).

hemolysis increased as the concentration rose in both formulations, coinciding with the findings that a greater hemolytic activity accompanies an increase in the charge potential of the formulation (Creda-Cristerna et al., 2011; Pribush et al., 2007; Fischer et al., 2003; Carreno-Gomez and Duncan, 1997). Hemolysis was suppressed with the FA–PEG–COL nanoparticles as compared to the COL particles, implying the role of PEGylation in prevention of hemolysis. Cytotoxicity of RAW 264.7 cells increased as the concentration of the COL and FA–PEG–COL nanoparticles rose (Fig. 7). Although chitosan and its derivatives are known for their good biocompatibility but the current study showed that their cationic property affects the cell viability. This is particularly profound in the COL nanoparticles compared to the FA–PEG–COL nanoparticles; the effect on the cell viability was 100 fold greater in the COL nanoparticles compared to the FA–PEG–COL nanoparticles. This observation highlighted the advantages of neutralization and masking of positive charge on the surface of the nanoparticles with PEGylation, coinciding with the view that cationic formulations can be cytotoxic through interactions with negatively-charged

cellular components, interfering with cell adhesion (Lappalainen et al., 1994) as well as cell division (Litzinger and Huang, 1992). PEGylation is also effective in prolonging the blood circulating half-life but reduces cellular uptake. The reduction of cellular uptake is due to its interference in the interaction of the positivelycharged nanoparticles with the negatively-charged cell membrane (Park et al., 2010; Jokerst et al., 2011). PEGylated nanoparticles have a better chance of reaching the tumor site via the passive targeting because of their prolonged blood circulating half-life. However, the delay in tumor accumulation via the passive targeting (Fig. 11) can lead to loss of drug (siRNA) caused by release from nanoparticles (Fig. 4) and destruction (Fig. 5) in the blood circulation. The active targeting mediated by the FA ligand can compensate for these drawbacks of PEGylation by shortening the time required for reaching the target site (Fig. 11) and enhancing the internalization of drug (siRNA) (Fig. 8). In the present study, HIF-1a was chosen as the target gene for in vitro evaluation of transfection efficiency and effects of gene knock down mediated by siRNA delivered by the FA–PEG–COL nanoparticles. Administration of siRNA for HIF-1a delivered by

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

59

A

B

Fig. 9. Gene targeting of HIF-1a by siRNA transfected with formulations in OVK18#2 ovarian cancer cells. (A) Western blot analysis showing protein expression of HIF-1a after treatment with siRNA transfected with different formulations. (1) No CoCl2; (2) CoCl2 treated (150 lM); (3) Lipofectamine + siRNA (non-coding); (4) FA–PEG–COL nanoparticles + siRNA (non-coding); (5) COL nanoparticles + siRNA (non-coding); (6) Lipofectamine + siRNA (HIF-1a); (7) FA–PEG–COL nanoparticles + siRNA (HIF-1a); (8) COL nanoparticles + siRNA (HIF-1a); (9) unformulated siRNA (HIF-1a). b-actin was used as the loading control. (B) RT-qPCR analysis of HIF-1a knockdown by siRNA transfected with different formulations. 150 lM CoCl2 was used for induction of HIF-1a. Results are expressed as mean ± SD of at least 3 separate experiments (n = 3). **** Denotes significant differences from CoCl2 treated group (P < 0.0001) (Prism one-way ANOVA with Tukey’s multiple comparison). NPs denote nanoparticles.

Fig. 10. Effect of HIF-1a knockdown by siRNA on proliferation of OVK18#2 ovarian cancer cells. 150 lM CoCl2 was used for induction of HIF-1a. Each point represents the mean ± SD of 3 separate experiments (n = 3). **** and ààà Denotes significant differences when compared to the CoCl2 treated group (P < 0.0001) and COL and FA–PEG–COL nanoparticles (P < 0.001), respectively (Prism one-way ANOVA with Tukey’s multiple comparison). NPs denote nanoparticles.

the COL and FA–PEG–COL nanoparticles significantly reduced both protein (Fig. 9A) and mRNA (Fig. 9B) levels of HIF-1a, leading to a strong suppression of cell proliferation in human ovarian cancer cells (Fig. 10). Furthermore, the delivery of siRNA by the FA– PEG–COL nanoparticles showed a significantly greater inhibition of proliferation compared to the COL nanoparticles (Fig. 10). These results suggest that active targeting and receptor-mediated endocytosis of siRNA-loaded nanoparticles bring an advantage for overall gene knockdown effects of siRNA. Finally, RNA interference takes place in the cytosole, and hence efficient release of siRNA

from the endosomes where it is entrapped after endocytosis is an important issue (Gilleron et al., 2013; Varkouhi et al., 2011). It would be desired to design a formulation that facilitate the endosomal release for improvement of siRNA delivery systems. 5. Conclusions In conclusion, we have shown the effectiveness of using a FA–PEG–COL nanoparticles system for delivery of siRNA for the potential anti-HIF-1a treatment of ovarian cancer. High

60

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61

Fig. 11. In vivo NIR fluorescence images of nude mice bearing OVK-18#2 tumor after injection of formulation, (A) Alexa fluor 647–COL nanoparticles; (B) Alexa fluor 647–FA– PEG–COL nanoparticles; (C) PBS. White arrow indicates location of tumor xenograft. (kexc 640 nm, kemi 680 nm), exposure time 1 s. Yellow arrow indicates liver accumulation. (D) Ex vivo images of the sacrificed organs in mice at 24 h post-injection. (1) Heart; (2) kidneys; (3) spleen and (4) liver. (kexc 640 nm, kemi 680 nm) exposure time 5 s.

encapsulating efficiency and good protection of siRNA from serum degradation with appropriate particle size were obtained with this system. Moreover, FA–PEG–COL nanoparticles have excellent tumor targeting ability with minimum uptake by the liver, implying that it has great potential not only for targeting the primary cancer site but also metastatic sites. The results shown here indicated that FA–PEG–COL nanoparticles containing siRNA may have good potential in use along with traditional chemotherapy and radiotherapy to achieve maximum tumoricidal activities. Acknowledgement Tony Shing Chau Li’s PhD scholarship was provided by the Japan Ministry of Education, Culture, Sports, Science and Technology (MEXT). Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.ejps.2013.10.011. References Calvo, P., Remunan-Lopez, C., Vila-Jato, J.L., Alonso, M.J., 1997. Novel hydrophilic chitosan–polyethylene oxide nano particles as protein carriers. Journal of Applied Polymer Science 63 (1), 125–132. Carreno-Gomez, B., Duncan, R., 1997. Evaluation of the biological properties of soluble chitosan and chitosan microspheres. International Journal of Pharmaceutics 148 (2), 231–240. Chae, S.Y., Jang, M.K., Nah, J.W., 2005. Influence of molecular weight on oral absorption of water soluble chitosans. Journal of Controlled Release 102 (2), 383–394. Chan, P., Kurisawa, M., Chung, J.E., Yang, Y.Y., 2007. Synthesis and characterization of chitosan-g-poly(ethylene glycol)-folate as a non-viral carrier for tumortargeted gene delivery. Biomaterials 28 (3), 540–549. Chein, S., 1975. Biophysical behaviours of red blood cells in suspensions. In: Surgenor, DM. (Ed.), Red Blood Cell. Academic Press, New York, pp. 1032–1121.

Chen, R.H., Tsaih, M.L., 1998. Effect of temperature on the intrinsic viscosity and conformation of chitosan in dilute HCl solution. International Journal of Biological Macromolecules 23 (2), 135–141. Chiavarina, B., Whitaker-Menezes, D., Migneco, G., Martinez-Outschoorn, E.U., Pavlides, S., Howell, A., Tanowitz, H.B., Casimiro, M.C., Wang, C., Pestell, R.G., Grieshaber, P., Caro, J., Sotgia, F., Lisanti, M.P., 2010. HIF-alpha functions as a tumor promoter in cancer associated fibroblasts, and as a tumor suppressor in breast cancer cells: Autophagy drives compartment-specific oncogenesis. Cell Cycle 9 (17), 3534–3551. Cho, I.S., Kim, J., Lim, D.H., Ahn, H.C., Kim, H., Lee, K.B., Lee, Y.S., 2008. Improved serum stability and biophysical properties of siRNAs following chemical modification. Biotechnology Letters 30 (11), 1901–1908. Creda-Cristerna, B.I., Flores, H., Pozos-Guillen, A., Perez, E., Sevrin, C., Grandfils, C., 2011. Hemocompatibilty assessment of poly(2-dimethylamino ethylmethacrylate) (PDMAEMA)-based polymer. Journal of Controlled Release 153 (3), 269–277. Creusat, G., Thomann, J.S., Maglott, A., Pons, B., Dontenwill, M., Guerin, E., Frisch, B., Zuber, G., 2012. Pyridylthiourea-grafted polyethylenimine offers an effective assistance to siRNA-mediated gene silencing in vitro and in vivo. Journal of Controlled Release 157, 418–426. Davis, M.E., Zuckerman, J.E., Choi, C.H.J., Seligson, D., Tolcher, A., Alabi, C.A., Yen, Y., Heidel, J.D., Ribas, A., 2010. Evidence of RNAi in humans from systemically administered siRNA via targeted nanoparticles. Nature 464, 1067–1070. Dykxhoorn, D.M., Lieberman, J., 2006. The silent treatment: siRNAs as small molecule drugs. Gene Therapy 13 (6), 541–552. Fan, W., Yan, W., Xu, Z., Ni, H., 2012. Formation mechanism of monodisperse, low molecular weight chitosan nanoparticles by ionic gelation technique. Colloids and surfaces B: Biointerfaces 90, 21–27. Fischer, D., Li, Y., Ahlemeyer, B., Krieglstein, J., Kissel, T., 2003. In vitro cytotoxicity testing of polycations: influence of polymer structure on cell viability and hemolysis. Biomaterials 24 (7), 1121–1131. Gilleron, J., Querbes, W., Zeigerer, A., Borodovsky, A., Marsico, G., Schubert, U., Manygoats, K., Seifert, S., Andree, C., Stöter, M., Epstein-Barash, H., Zhang, L., Koteliansky, V., Fitzgerald, K., Fava, E., Bickle, M., Kalaidzidis, Y., Akinc, A., Maier, M., Zerial, M., 2013. Image-based analysis of lipid nanoparticle-mediated siRNA delivery, intracellular trafficking and endosomal escape. Nature Biotechnology 31 (7), 638–646. Hirota, K., Semenza, G.L., 2006. Regulation of angiogenesis by hypoxia-inducible factor 1. Critical Reviews in Oncology/Hematology 59 (1), 15–26. Hua, K., Din, J., Cao, Q., Feng, W., Zhang, Y., Yao, L., Huang, Y., Zhao, Y., Feng, Y., 2009. Estrogen and progestin regulate HIF-1a expression in ovarian cell lines via the activation of Akt signalling transduction pathway. Oncology Reports 21 (4), 893–898. Huang, M., Ma, Z., Khor, E., Lim, L.-Y., 2002. Uptake of FITC–chitosan nanoparticles by A549 cells. Pharmaceuticals Research 19 (10), 1488–1494.

T.S.C. Li et al. / European Journal of Pharmaceutical Sciences 52 (2014) 48–61 Jokerst, J.V., Lobovkina, T., Zare, R.N., Gambhir, S.S., 2011. Nanoparticle PEGylation for imaging and therapy. Nanomedicine 6 (4), 715–728. Kalli, K.R., Oberg, A.L., Keeney, G.L., Christianson, T.J., Low, P.S., Knutson, K.L., Hartmann, L.C., 2008. Folate receptor alpha as a tumor target in epithelial ovarian cancer. Gynecologic Oncology 108 (3), 619–626. Katas, H., Alpar, H.O., 2006. Development and characterisation of chitosan nanoparticles for siRNA delivery. Journal of Controlled Release 115 (2), 216– 225. Lappalainen, K., Jaaskelinen, I., Syrjanen, K., Urtti, A., Syrjanen, S., 1994. Comparison of cell proliferation and toxicity assays using two cationic liposomes. Pharmaceutical Research 11 (8), 1127–1131. Liang, B., He, M.L., Chan, Y.C., Chen, Y.C., Li, X.P., Zheng, D., Lin, M.C., Kung, H.F., Shuai, X.T., Peng, Y., 2009. The use of folate-PEG-grafted-hybranched-PEI nonviral vector for the inhibition of glioma in the rat. Biomaterials 30 (23– 24), 4014–4020. Litzinger, D.C., Huang, L., 1992. Phosphatidylethanolamine liposomes: drug delivery, gene transfer and immunodiagnostic applications. Biochimica et Biophysica Acta 1113 (2), 201–227. Miyagishi, M., Sumimoto, H., Miyoshi, H., Kawakami, Y., Taira, K., 2004. Optimization of an siRNA-expression system with an improved hairpin and its significant suppressive effects in mammalian cells. The Journal of Gene Medicine 6, 715–723. Moreau, E., Ferrari, I., Drochon, A., Chapon, P., Vert, M., Domurado, D., 2000. Interactions between reb blood cells and a lethal, partly quaternized tertiary polyamine. Journal of Controlled Release 64 (1–2), 115–128. Moreau, E., Domurado, M., Chapon, P., Vert, M., Domurado, D., 2002. In vitro agglutination and lysis of red blood cells and in vivo toxicity. Journal of Drug Targeting 10 (2), 161–173. Onishi, H., Machida, Y., 1999. Biodegradation and distribution of water-soluble chitosan in mice. Biomaterials 20 (2), 175–182. Park, Y., Kwon, O.J., Kwon, O.J., Hwang, T., Park, H., Lee, J.M., Kim, J.H., Yun, C.O., 2010. Ionically crosslinked Ad/chitosan nanocomplexes processed by electrospinning for targeted cancer gene therapy. Journal of Controlled Release 148 (1), 75–82. Parker, N., Turk, M.J., Westrick, E., Lewis, J.D., Low, P.S., Leamon, C.P., 2005. Folate receptor expression in carcinomas and normal tissues determined by a quantitative radioligand binding assay. Analytical Biochemistry 338 (2), 284– 293. Pribush, A., Zilberman-Kravits, D., Meyerstein, N., 2007. The mechanism of the dextran-induced red blood cell aggregation. European Biophysics Journal 36 (2), 85–94.

61

Rekha, M.R., Sharma, C.P., 2009. Blood compatibility and in vitro transfection studies on cationically modified pullulan for liver cell targeted gene delivery. Biomaterials 30 (34), 6655–6664. Salmaso, S., Semenzato, A., Caliceti, P., Hoebeke, J., Sonvico, F., Dubernet, C., Couvreur, P., 2004. Specific antitumor targetable beta-cyclodetrinpoly(ethylene glycol)–folic acid drug delivery bioconjugate. Bioconjugate Chemistry 15 (5), 997–1004. Semenza, S.L., 2001. Hypoxia-inducible factor 1: oxygen homeostasis and disease pathophysiology. Trends in Molecular Medicine 7 (8), 345–350. Sigma–Aldrich, 2009. Product Information: Chitosan-A Technologically Important Biomaterial. Sigma–Aldrich, Milwaukee, WI, USA. Sioud, M., Sorensen, D.R., 2003. Cationic liposome-mediated delivery of siRNAs in adult mice. Biochemical and Biophysical Research Communications 312, 1220– 1225. Sorensen, D.R., Leirdal, M., Sioud, M., 2003. Gene silencing by systemic delivery of synthetic siRNAs in adult mice. Journal of Molecular Biology 327, 761–766. Tagawa, T., Morgan, R., Yen, Y., Mortimer, J., 2012. Ovarian cancer: opportunity for targeted therapy. Journal of Oncology (Article ID 682480). US FDA, 2001. Guidance for Industry: Bioanalytical method validation. U.S. Department of Health and Human Services Food and Drug Administration Center for Drug Evaluation and Research (CDER), Center for Veterinary Medicine (CVM). Varkouhi, A.K., Scholte, M., Storm, G., Haisma, H.J., 2011. Endosomal escape pathways for delivery of biologicals. Journal of Controlled Release 151 (3), 220–228. Werner, M.E., Karve, S., Sukumar, R., Cummings, N.D., Copp, J.A., Chen, R.C., Zhang, T., Wang, A.Z., 2011. Folate-targeted nanoparticle delivery of chemo- and radiotherapeutics for the treatment of ovarian cancer peritoneal metastasis. Biomaterials 32 (33), 8548–8554. Yang, S.J., Lin, F.H., Tsai, K.C., Wei, M.F., Tsai, H.M., Wong, M.J., Shieh, M.J., 2010. Folic acid-conjugated chitosan nanoparticles enhanced protoporphyrin IX accumulation in colorectal cancer cells. Bioconjugate Chemistry 21 (4), 679– 689. Yeo, E.J., Chun, Y.S., Park, J.W., 2004. New anticancer strategies targeting HIF-1. Biochemical Pharmacology 68 (6), 1061–1069. Ziello, J.E., Jovin, I.S., Huang, Y., 2007. Hypoxia-inducible factor (HIF)-1 regulatory pathway and its potential for therapeutic intervention in malignancy and ischemia. Yale Journal of Biology and Medicine 80, 51–60.