International Journal of Pharmaceutics 465 (2014) 413–426

Contents lists available at ScienceDirect

International Journal of Pharmaceutics

journal homepage: www.elsevier .com/ locate / i jpharm

A novel gene therapy vector based on hyaluronic acid and solid lipidnanoparticles for ocular diseases

Paola Stephanie Apaolaza, Diego Delgado, Ana del Pozo-Rodríguez,Alicia Rodríguez Gascón, M.Ángeles Solinís *Pharmacokinetic, Nanotechnology and Gene Therapy Group (PharmaNanoGene), Faculty of Pharmacy, Centro de investigación Lascaray ikergunea, University ofthe Basque Country UPV/EHU, Vitoria-Gasteiz, Spain

A R T I C L E I N F O

Article history:

Received 7 January 2014Received in revised form 17 February 2014Accepted 20 February 2014Available online 24 February 2014

Keywords:Gene therapySolid lipid nanoparticlesHyaluronic acidProtamineTransfectionX linked juvenile retinoschisis

* Corresponding author at: PharmacyandFaculty of Pharmacy, University of the BUniversidad no. 7. 01006, Vitoria-Gasteiz, Sp013040.

E-mail address: marian.solinis@ehu.es (

http://dx.doi.org/10.1016/j.ijpharm.2014.020378-5173/ã 2014 Elsevier B.V. All rights r

Pharmaceasque Couain. Tel.: +

M.Á. Soliní

.038eserved.

A B S T R A C T

The introduction of therapeutic genes in target tissues is considered as a novel tool for the treatment ofseveral diseases.We have developed nanoparticles consisting on SLNs, protamine (P) and hyaluronic acid(HA) as carrier for gene therapy. Stable complexes positively charged andwith a particle size ranging from240nm to 340nm were obtained. Transfection studies in ARPE-19 and HEK-293 cells showed theversatility of vectors to efficiently transfect cells with different division rate, widening the potentialapplications of SLN-based vectors. In ARPE 19 cells, the incorporation of P and HA induced almost a 7-foldincrease in the transfection capacity of SLNs. The CD44 inhibition studies suggested the participation ofthis receptor in the internalization of the vectors in this cell line. The intracellular disposition of DNAshowed that the HA is able to modulate the high degree of condensation of DNA due to the protamineinside the cells; an important fact, if the vector is uptaken via non-degradative endocytosis. Besides, thetherapeutic plasmid which encodes the protein retinoschisin was employed achieving a positivetransfection in ARPE-19 cells, showing a promising application of this new non-viral system for thetreatment of X-linked juvenile retinoschisis by gene therapy.

ã 2014 Elsevier B.V. All rights reserved.

1. Introduction

The introduction of therapeutic genes in the target tissues isconsidered as a promising alternative to conventional drugproducts for the treatment of several chronic diseases. However,despite the recognizedvalue of DNA-based therapeutics, importantproblems, such as poor cellular uptake, a rapid in vivo degradation,limited transport to the target, and low effective delivery of thegenetic material to the cell nucleus, need to be solved before thistherapeutic strategy becomes viable (de la Fuente et al., 2008a).Therefore, a key challenge to realizing the broad potential of DNA-based therapeutics is the need for safety and effective deliverymethods. A broad diversity of materials is under exploration toaddress the challenges of delivery, including, viral vectors,inorganic particles, polymeric-, cationic lipid-, and peptide-basedvectors (Gascón et al., 2013). Despite the delivery successes knownby some of these carriers, advances are necessary to allow the

utical Technology Laboratory,ntry UPV/EHU, Paseo de la34 945 013469; fax: +34 945

s).

fullest application of DNA in the clinic. In this regard, non-viraldelivery systemspossess features that result advantageous for theiruse in gene therapy: safety, low-cost production, high-reproduc-ibility and no limit size of DNA to transport (del Pozo-Rodríguezet al., 2011).

Solid lipid nanoparticles (SLNs), areapromisingnon-viral vectorfor gene therapy due to their capacity to condense and protect thegenetic material, besides they show efficacy in the cell internaliza-tion, and once inside, release the genetic material (Gascón et al.,2012a, 2012b), as well as the ability of SLNs for transfection in vitroin several cell lines, and in vivo after intravenous and ocularadministration as it has been previously demonstrated (del Pozo-Rodríguez et al., 2010; Delgado et al., 2012a,2012b). In order toimprove the capacity of SLNs to transfect, several ligands have beenincorporated on the nanoparticle surface; for instance, cellpenetration peptides (del Pozo-Rodríguez et al., 2009), protamine(Delgado et al., 2011), dextran (Delgado et al., 2012b) oroligochitosans (Delgado et al., 2013), among others.

Hyaluronic acid (HA), a high molecular linear glycosaminogly-can composed of repeated disaccharide units ofb-1,4-D-glucuronicacid-b-1,3-N-acetyl-D-glucosamine, is an attractive polymer in thefield of pharmaceutical technology due to its biocompatibility,biodegradability andmucoadhesive character (Aragona, 2004). HA

414 P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426

presents several advantages from a biological point of view, since itpromotes the adhesion andproliferation inmammalian cells, and itis involved in biological processes like cell signalling (de la Fuenteet al., 2008b). Furthermore, as it is a hydrophilic polymer, it couldprevent opsonin adsorption by steric repulsion allowing mononu-clear phagocyte systemuptake to be reduced (Wojcicki et al., 2012).Another property of HA is thewell-known capacity to interact withtheCD44 receptor,which is expressed indifferent tissues, includinghuman cornea and conjunctive, andparticipates in awide varietyofcellular functions, like the receptor-mediated internalizationbetween others. Therefore, the interaction of HA with CD44 andother HA-specific receptors facilitates cell internalization ofdifferent systems (Ruponen et al., 2001); moreover, its capacityto favour the nuclear entry is well known, and the role as atranscription activator has also been suggested (de la Fuente et al.,2008b). All these properties make HA a very useful compound fortransfection approaches (Yamada et al., 2013).

With this premise in mind, we developed nanoparticlesconsisting on SLNs, protamine and HA as a carrier for gene therapy.We prepared different complex compositions, with regard to themolecular weight of HA and HA–protamine–SLN mass ratios, andstudied transfection efficiency, cell uptake and intracellulartrafficking pathways in two cell models: ARPE-19 and HEK-293.Finally, it was also studied the capacity of the vectors to transfectARPE-19 cells with the therapeutic plasmid that encodes theprotein retinoschisin,which is responsible of a retinal degenerativedisorder, the X-linked juvenile retinoschisis.

2. Material and methods

2.1. Preparation of the hyaluronic acid–protamine–DNA–SLN vectors(HA–P–DNA–SLN)

The SLNs were produced by a previously described solventemulsification-evaporation technique (del Pozo-Rodríguez et al.,2007). The solid matrix of the particles was composed by the lipidPrecirol1ATO 5whichwas kindly gifted byGateffossé. The cationiclipid N-(1-(2,3-Dioleoyloxy) propyl)-N,N,N trimethyl ammoniummethyl sulfate (DOTAP), purchased from Avanti Polar Lipids, wasused to form the external phase of the nanoparticles, and thesurfactant Tween 80 (Panreac) was selected to obtain the initialemulsion. To prepare the SLNs the lipid Precirol1 ATO 5 wasdissolved in the organic solvent dichloromethane (Panreac) (5%, w/v) and emulsified by sonication in the aqueous phase containingDOTAP (0.4%,w/v) and Tween 80 (0.1%, w/v). Dichloromethanewasthen evaporated by magnetic stirring followed by vacuumconditions, so that the inner lipid Precirol1 ATO 5 precipitatedand nanoparticles were obtained.

To prepare the HA–P–DNA–SLN vectors, firstly a solution ofprotamine (P) was mixed with an aqueous solution of pCMS–EGFPplasmid (0.2mg/ml), which encodes the enhanced green fluores-cent protein (EGFP), at a fixed ratio of 2:1 (w:w) during 5min. Then,an aqueous solution of HA was added to form the HA–P–DNAcomplex at thedesired ratios. After 15min, theHA–P–DNAcomplexwas mixed with a suspension of SLNs during 20min at roomtemperature, and electrostatic interactions between the HAPDNAcomplexes and SLNs led to the formation of HA–P–DNA–SLNvectors. All the vectors were brought up to a final volume of 300mlwithHBS buffer. HAof three differentmolecularweights purchasedfrom Sigma–Aldrich were used for the preparation of the vectors:150kDa (HA150), 500kDa (HA500), and 1630kDa (HA1630). ThepCMS–EGFP plasmid was purchased from BD Biosciences Clontech(Palo Alto, California, USA) and amplified by Dro Biosystems S.L.(San Sebastian, Spain).

As control, a DNA–SLN vector was prepared by mixing 50ml ofan aqueous solution of pCMS–EGFP plasmid (0.2mg/ml) together

with 150ml of the SLNs suspension (1mg/ml). The incubation timefor the electrostatic bindingwas 20min at room temperature. In allcases, the SLN to DNA ratio, expressed as the ratio of DOTAP to DNA(w:w), was fixed at 5:1.

Finally, vectors bearing the pCEP4-RS1 plasmid were preparedas explained above. This plasmid encodes retinoschisinprotein, andwas obtained from the Institute of Human Genetics of theUniversity of Regensburg (Germany). The pCEP4-RS1 plasmidwas formulated in HA–P–DNA–SLN vectors prepared with HA150

and HA500 at different w:w ratios.

2.2. Studies with the pCMS–EGFP plasmid

2.2.1. Size and Zeta potential measurementsSizes of SLNs, DNA–SLN and HA–P–DNA–SLN vectors were

determined by photon correlation spectroscopy (PCS). Zetapotentials were measured by Laser Doppler Velocimetry (LDV).Both measurements were performed on a Zetasizer Nano series-Nano ZS (Malvern Instruments, Worcestershire, UK). All sampleswere diluted in Milli-QTM water.

2.2.2. Electrophoresis on agarose gelBinding efficiency of DNA by HA–P–DNA–SLN complex at

different HA ratios, protection from DNAse I (Sigma–Aldrich)digestion and DNA plasmid release from the vectors wereperformed using a 0.7% agarose gel (Sigma) containing Gel RedTM

(Biotium) for visualization during 30min at 120V. The bands wereobserved with an Uvitec Uvidoc D-55-LCD-20M Auto transillumi-nator.

Complexes were diluted in Milli-QTM water up to a finalconcentration of 0.03mg DNA/mL. A concentration of 1U DNase I/2.5mg DNA was added to DNA–SLN and HA–P–DNA–SLN vectors,and the mixtures were then incubated at 37 �C for 30min;afterwards, 4% SDS solution was added to the samples to a finalconcentration of 1% to release the DNA from the SLNs. Finally, theintegrity of the DNA in each sample was compared to a control ofuntreated DNA.

2.2.3. Preparation of cell culturesIn vitro assays were performed with human retinal pigment

epithelial (ARPE-19) cells and human embrionic kidney (HEK-293)cells, both obtained from the American Type Culture Collection(ATCC).

ARPE-19 cells and HEK-93 cells were cultured in Dulbecco’sModified Eagle’s Medium–Han’s Nutrient Mixture F-12 (1:1)medium (D-MEM/F-12, Gibco) and in Eagle’s Minimal Essentialwith Earle’s BSS and 2mM L-glutamine (EMEM, LGC Promochem),respectively, supplemented with 10% heat-inactivated fetal bovineserum (FBS) and 1% NormocinTM antibiotic solution (Invivogen).Cells were incubated at 37 �C in 5% CO2 atmosphere andsubcultured every 2–3 days using trypsin–EDTA (Lonza).

ARPE-19 cells were seeded on 12 well plates (3�104 cells/well)and HEK-293 cells were seeded on 24 well plates (1.5�105 cells/well). Before the treatment with vectors both cell lines wereallowed to adhere overnight. Vectors were diluted in HEPES BufferSolution (HBS) andadded to the cell cultures. In all cases, anamountof vectors equivalent to 2.5mg of DNAwas added to eachwell. Cellswere incubated in presence of the vectors for 4h at 37 �C, afterwhich the medium containing the complexes was refreshed with1mL of complete medium. Transfection efficacy was quantified at24h, 48h, 72 h, 96h and one week.

2.2.4. Transfection evaluation by flowcytometryand fluorimetric assayAt the established times, the cells were washed once with

300mL of Phosphate Buffered Saline (PBS), and then detachedwith300mL of 0.05% trypsin–EDTA. The cells were then centrifuged at

[(Fig._1)TD$FIG]

Fig. 1. Example of a calibration curve for the HA150–P–DNA–SLN vector.

P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426 415

1500� g, and the supernatantwas discarded. After resuspension inPBS, cells were directly introduced into a GalliosTM flow cytometer(Beckman Coulter, FLorida, Miami, US). For each sample, 1�104

events were collected. Transfection efficacy was quantified bymeasuring the fluorescence of EGFP at 525nm (FL1) emitted by thecells.

We also measured the emission signal of the intracellularprotein EGFP at each time by a fluorometric assay. For this purpose,the cells were lysed with Reporter Lysis Buffer (RLB, Promega);afterwards the samples were centrifuged at 12,000� g, and the celllysate was measured in a GlomaxTM Multi detection System(Promega). The quantity of EGFP was expressed as RelativeFluorescent Units (RFU) per milligram of total protein. The totalamount of proteinwas quantified byMicro BCATM Protein Assay kit(Thermo Scientific).

2.2.5. Viability evaluation by fluorimetric assayThe assay CellTiter-FluorTM Cell Viability Assay (Promega) was

employed. Following the protocol supplied by the manufacturer,ARPE-19 and HEK-293 cells were seeded on 96 black well plates(1�104 cells/well) and the assay was performed at 24h and 96h.The fluorescent reagent (100mL) was added to each sample, andafter 2h of incubation at 37 �C, the fluorescence corresponding toliving cellswasmeasured at 505nm. Cell viabilitywas calculated asa percentage with regard to control cells.

2.2.6. Cellular uptake of non-viral vectorsEntry of vectors to the ARPE-19 and HEK-293 cells was studied

qualitativelyandquantitativelyby flowcytometryand fluorometrictechniques, respectively. For this purpose, SLNs were labelled withthe fluorescent dyeNile Red (l =590nm) as described in a previousstudy (del Pozo-Rodríguez et al., 2008). Two hours after theaddition of the labelled vectors, the cells were washed three timeswith PBS and detached from plates. Cells incorporating eitherDNA–SLN or HA–P–DNA–SLN vectors were analysed by flowcytometryat 650nm(FL3).1�104eventsper samplewere collectedfor the analysis.

A quantitative uptake study was performed by a fluorometricassay. Both cell lines were cultured in Minimal Essential Mediumwith reduction of FBS and no phenol red (Opti-MEM). The samecellulardensityand the sameamountof vectors than in transfectionassays were used. At different times after the addition of thevectors, the medium containing the non-uptaken vectors wascollected and poured into 96 black well plates. The fluorescencesignal was measured in a GlomaxTM Multi detection System (lex

525nm and lem 580–640nm). In contrast to Nile Red loaded SLNs,free Nile Red in suspension did not change the intrinsic cellfluorescence of the cells when compared to untreated controls.Therefore, Nile Red is uptaken only when loaded in SLNs andprovided a good assessment of SLNs internalization. In order toquantify the cellular uptake, calibration curves with increasingamounts of complexes in culture medium were performed. Therange for the calibration curve was from 3.5mg to 15mgcorresponding to the total amount of Nile Red in the complexes.The determination coefficient was always over 0.99, and theindividual error of every point in the calibration curve was alwayslower than5%. The sampleswere showntobe stable for at least 24h.As an example, Fig. 1 displays the calibration curve for theHA150–P–DNA–SLN vector.

2.2.7. Detection of CD44 expression by immunocytochemistryIn order to study the presence of the specific HA receptor

CD44, firstly the ARPE-19 cell line (6�104 cells/ml) and HEK-293cell line (1.5�105 cells/ml) were seeded and incubate overnightin 24 well-plates. The next day the cells were fixed with 4% PFA(Panreac) during 10min. Afterwards the cells were blocked and

permeabilized during 30min at room temperature with asolution containing PB buffer, 0.3% Triton X-100 (SigmaAldrich)and 10% goat serum (Sigma–Aldrich). After the blocking proce-dure an antibody solution containing PB buffer, 2.5% goat serumand 0.1% triton X-100 with the rat monoclonal CD44 antibody(Abcam) was added keeping in contact with the cells overnight.The samples were washed with PBS and dyed with Alexa Fluor488-conjugated goat anti-rat IgG (Abcam) for 1 h in the dark. Thenuclei were dyed with the mounting media DAPI-Fluoromount-G(SouthernBiotech). The images were captured with an AxioObserver Inverted Microscope Z.1 with Apotome (Zeiss) using63� magnification.

2.2.8. Evaluation of the CD44 receptor influence in the uptakeTo analyze the effect of the CD44 receptor when HA is merged

into the vector P–DNA–SLN, theuptakeof the vectorswas evaluatedby flow cytometry (FL3). In order to accomplish this target, SLNslabelled with the Nile Red dye (l =590nm) were employed after apreincubation of the cells with a saturated solution of hyaluronicacid (3mg/ml) (Evanko and Wight, 1999). One hour before theaddition of the vectors, the solution of hyaluronan mixed with thecompletemedium corresponding to each cell linewas added to thewell-plates. Before the addition of the vectors the cells werewashed with PBS and refreshed with newmedium. Afterwards, 2 hafter the cellswere detached fromplates beforewashing themwithPBS. The cells able to uptake the HA–P–DNA–SLN vectors wereanalyzed by flow cytometry at 650nm (FL3). 1�104 events persample were collected for the analysis. The vector DNA–SLN wasalso evaluated.

2.2.9. Internalization mechanismThe endocytic processes involved in the internalization of the

non-viral systems in both cell lines were analyzed qualitatively bycolocalization studies with AlexaFluor488-Cholera Toxin andAlexaFluor488-Transferrin (Molecular Probes), which are markersof caveolae/lipid raft-mediated endocytosis and clathrin-mediatedendocytosis, respectively. Confocal laser scanning microscopy(CLSM) was used.

Cellswereseededincoverslipcontainingplatesandco-incubatedfor 2h with Nile Red labelled vectors and either AlexaFluor488-Cholera Toxin (10mg/mL) or AlexaFluor488-Transferrin (50mg/mL).After incubation, the mediumwas removed and cells were washedwith PBS and fixed with paraformaldehyde (PFA) 3.7–4% (Panreac).PFA had beenpreviously tested to not interactwith the fluorescencesignal of either Nile Red. Preparations were mounted withFluoromount GTM (Southern Biotech) and then, air-drying imageswere obtained with an Olympus Fluoview FV500 confocal micro-scope, using sequential acquisition to avoid overlapping of fluores-cent emission spectra. CLSM images were captured in the GeneralServiceofAnalyticalMicroscopyandHighResolution inBiomedicine

416 P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426

of the University of the Basque Country (UPV/EHU, Bizkaia, Spain).Colocalization resultswere estimatedbymeansofMander’sOverlapCoefficient (R), where 0.6�R�1.0 indicates colocalization (overlapof the signals)(Zinchuk and Zinchuk, 2008).

2.2.10. Detection of intracellular EMA-labelled DNA by fluorescencemicroscopy

In order to study the intracellular disposition of DNA into thecytoplasm, the cells were seeded in coverslip containing plates andincubated with vectors containing EMA-labelled-DNA (DroBiosys-tems). Nuclei were labelled with DAPI-fluoromount-GTM (South-ernBiotech). Images were captured at 4h and 12h with an AxioObserver Inverted Microscope Z.1 with ApoTome (Zeiss) using 63�magnification.

2.3. Studies with the pCEP4-RS1 plasmid

2.3.1. Size and zeta potential measurementsSizes of HA–P–DNA–SLN vectors were determined as explained

for the plasmid pCMS–EGFP.

2.3.2. Electrophoresis on agarose gelBindingefficiencyof thepCEP4-Rs1plasmidbyHA–P–DNA–SLN

vectors at differentHAratios,protection fromDNAse I digestionandDNAplasmid release from the vectorswere performed as explainedabove.

2.3.3. Quantification of secreted retinoschisinARPE-19 cells were seeded on 12 well plates (3�104 cells/well)

and allowed to adhere overnight. Meanwhile, the pCEP4-RS1containing vectors were prepared and diluted in HBS. A total of2.5mg of plasmid was added to the culture medium. Cells wereincubated with the vectors at 37 �C, and after 4h, the mediumcontaining the complexes in thewellswas refreshedwith 500mL ofcomplete medium.

Forty-eight and seventy-two hours after the addition of thevectors, the levels of retinoschisin produced by the cells weredetermined in the culturemedium by Enzyme-linked Immunosor-bent Assay (ELISA) Kit (USCN1).

2.3.4. Immunochemical detection of retinoschisinARPE-19 cells were seeded in coverslips and transfected with

the plasmid pCEP4-RS1 containing vectors as described above.Seventy-two hours after the addition of the vectors cells wereimmunolabelled as previously reported (Delgado et al., 2012a).AftermountingwithDAPI-fluoromount-GTM imageswere capturedwith an inverted microscopy equipped with an attachment forfluorescent observation (EclipseTE2000-S, Nikon).

2.4. Statistical analysis

Statistical analysis was performed with PASW1 Statistic 18(Chicago, USA). Normal distribution of sampleswas assessed by theShapiro–Wilk test, andhomogeneityof variance, by the Levene test.The different formulations were compared with ANOVA andstudent’s t test, whereby differences were considered statisticallysignificant at p<0.05.

3. Results

3.1. Characterization of the vectors with pCMS–EGFP

Hyaluronic acid (HA) of three different molecular weights,protamine (P), pCMS–EGFP plasmid and SLNs were combined atratios of 0.1/0.3/0.5/0.7:2:1:5, respectively, and the vectors weresubjected to electrophoresis on agarose gels. The gel illustrates that

the plasmid was fully bound in all vectors, regardless of themolecular weight and proportion of HA (Fig. 2A).

The protection capacity of the vectors and their release of DNAwere studied by analyzing the integrity of DNA in agarose gelelectrophoresis after treatmentwith DNase I and SDS, respectively.The formulations assayed were able to preserve the pCMS–EGFPplasmid they transport (Fig. 2B), although the band correspondingto the supercoiledDNA topoisoform(lower band)wasmore intensewith thevectors containingHAandprotamine (lanes 4 to15).Whenvectors were treated with SDS (Fig. 2C) all of them were able torelease DNA.

Table 1 shows the particle size, the polydispersity index and thesurface charge of the vectors. Particle size ranged from 240nm to340nm, without significant differences between formulations(p>0.05). The polydispersity index was always lower than 0.4. Thesurface charge varied from +30mV to +40mV for all formulations.

The labelling of the vectors with Nile Red or EMA did not inducechanges in the particle size, polydispersity index or zeta potential(p>0.05)

3.2. In vitro transfection with the plasmid pCMS–EGFP

3.2.1. ARPE-19 cellsPrior to theevaluationof the efficacyofHA–P–DNA–SLNvectors,

we showed that HA–P–DNA complexes were not able to transfectARPE-19 cells. Fig. 3 shows the percent of transfected ARPE-19 cells72h after the addition of the vectors. The vectors containingHAandprotamine induced significantly higher transfection than theDNA–SLN vector, regardless of the molecular weight and theproportion of HA. When HA150 and HA500 were used, thetransfection efficacy increased as the HA:DNA ratio increases,reaching the highest transfection level with the 0.5:1 ratio (36%EGFP positive cells); higher amount of HA in the vector did notincrease transfection. However, whenHA1630was used, the percentof transfected cells did not vary at HA to DNA ratios of 0.3:1, 0.5:1and 0.7:1, and in a lower proportion (0.1:1), the transfection wassignificantly lower.

Fig. 4A shows the transfection level induced by the vectorDNA–SLN over time and those composed by HA and P at an HA toDNA ratio of 0.5:1. Twenty-four hours after treatment, thepercentage of positive cells was similar with all the vectors (about10%); however, after48h, transfectionwas significantlyhigherwiththe vectors bearing HA and P. With these vectors, transfectionincreased over time, reaching a maximum at 96h (60–70% EGFPpositive cells). Only at 96h and 7 days, small but significantdifferenceswere detected among the vectors preparedwith the HAof different molecular weights.

In order to study the production of the EGFP, we measured theRFUs emitted by the EGFP permilligramof total protein at differenttimes. The results showed that the quantity of protein producedwas significantly higher (p<0.05) with the HA–P–DNA–SLNvectors than with the DNA–SLN vectors at all time points(Fig. 4B). Significant differences due to the molecular weight ofthe HA were only detected at 72h between the formulationprepared with HA150 and HA500.

3.2.2. HEK-293 cellsPrior to the evaluation of the efficacy of HA–P–DNA–SLN

vectors, we showed that HA–P–DNA complexes were not able totransfect HEK-293 cells. Fig. 4C shows the level of transfectioninduced by the DNA–SLN vector and those composed by HA and Pat a HA to DNA ratio of 0.5:1. At 24h and 48h the highesttransfection levels were achieved with the vector preparedwithout HA and protamine. At 72h, vectors prepared with HA150

and HA500 induced significantly higher transfection than thevector prepared without HA and P, and the vector prepared with

[(Fig._2)TD$FIG]

Fig. 2. pCMS–EGFP binding to SLNs, protection against DNAse I and SDS-induced release of DNA fromHA–P–DNA–SLNvectorswithHAof differentmolecularweights. (A)DNAbinding. Lane 1= free DNA; lane 2=DNA–SLN; lanes 3 to 6 = 0.1/0.3/0.5/0.7(HA150):2:1:5; lanes 7 to 10 =0.1/0.3/0.5/0.7(HA500):2:1:5; lanes 11 to 14 =0.1/0.3/0.5/0.7(HA1630):2:1:5. (B)DNAprotectionagainst DNAse I. Lane1= freeDNA; lane2 = freeDNAtreatedwithDNAse I; lane3 =DNA–SLN; lanes 4 to7 = 0.1/0.3/0.5/0.7(HA150):2:1:5; lanes8 to 11=0.1/0.3/0.5/0.7(HA500):2:1:5; lanes 12 to 15=0.1/0.3/0.5/0.7(HA1630):2:1:5. (C) DNA release. Lane 1= free DNA; lane 2=DNA–SLN; lanes 3 to 6 =0.1/0.3/0.5/0.7(HA150):2:1:5; lanes 7 to 10 =0.1/0.3/0.5/0.7(HA500):2:1:5; lanes 11 to 14 =0.1/0.3/0.5/0.7(HA1630):2:1:5.

P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426 417

HA1630. At day 7, we also found significant differences in thetransfection levels induced by the vectors.

Fig. 4D shows the amount of EGFP produced by the HEK-293cells after being transfected by the vectors (DNA–SLN andHA–P–DNA–SLN at a HA to DNA ratio of 0.5:1). The amount ofprotein increased over timewith all the formulations, regardless ofthe presence of HA and P. No statistically significant differences(p>0.05) were observed in the amount of protein produced by thecells after transfectionwith the different vectors at any time point.

3.3. Viability evaluation

Cell viability was over 90% at 24 h and over 80% at 96 h in bothcell lines with all formulations. It was significantly higher with

Table 1Physical characterization of SLNs-based vectors bearing the plasmid pCMS–EGFP. Particsuperficial charge represent mean� SD (n =3).

Vector HA–protamine–DNA–SLN ratio Size

DNA–SLN 0:0:1:5 262HA150 0.1:2:1:5 316

0.3:2:1:5 3400.5:2:1:5 2920.7:2:1:5 302

HA500 0.1:2:1:5 2640.3:2:1:5 2650.5:2:1:5 2590.7:2:1:5 243

HA1630 0.1:2:1:5 3350.3:2:1:5 2810.5:2:1:5 2710.7:2:1:5 264

the HA–P–DNA–SLN vectors than with DNA–SLN vector(p<0.05).

3.4. Cellular uptake

Cellular uptake in both cell lines over time was quantifiedindirectly by measuring the fluorescence of Nile Red-labelledvectors remaining in the culture medium. Fig. 5A and B feature thepercentage uptake over time. The uptake in ARPE-19 from6h to 8hwas significantly higher (p<0.05) for the vectors preparedwithHAand P than for the DNA–SLN vector (around 50% vs. 30%). The samepattern was observed for HEK-293 cells (about 60% vs. 38%). Theseresults were confirmed by flow cytometry (Fig. 5C and D).

le size (nm); Polydispersity index (PDI); and Zeta potential (mV). The size, PDI and

(nm) PDI Zeta potential (mV)

.28�16.52 0.24�0.03 +36.17�2.35

.80�18.50 0.39�0.02 +32.73�1.25

.07�42.75 0.36�0.00 +30.00�0.28.10�51.36 0.33�0.00 +33.00�1.18.67�57.46 0.34�0.02 +29.70�1.80.03�9.57 0.36�0.00 +40.60�0.30.77�4.10 0.36�0.00 +37.70�0.40.13�16.51 0.33�0.00 +36.77�0.35.50�5.11 0.31�0.03 +33.60�2.54.23�36.43 0.28�0.01 +33.33�1.60.57�47.23 0.25�0.00 +30.43�0.84.93�17.11 0.26�0.02 +30.73�1.20.50�27.02 0.25�0.00 +29.73�0.31

[(Fig._3)TD$FIG]

Fig. 3. Evaluation of EGFP positive ARPE-19 cells 72h after treatment withHA–P–DNA–SLN vectors with HA of different molecular weights, 150 kDa, 500 kDaand 1630kDa at different w:w ratios. Error bars represent SD (n =3). *p<0.05respect to DNA–SLN; **p<0.01 respect to DNA–SLN.

418 P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426

3.5. Detection of CD44 expression by immunocytochemistry

Fig. 6 shows the fluorescence images of ARPE-19 cells (Fig. 6A)andHEK-293 cells (Fig. 6B) after the treatmentwith the specific HAreceptor CD44. We observed in green colour the qualitativeexpression in both cells lines. As it is shown, ARPE-19 cells presenta higher expressionof the receptorCD44 compare to the expressionin HEK-293 cells.

[(Fig._4)TD$FIG]

Fig. 4. Transfection in ARPE-19 and HEK-293 cells over time, after treatment with HA–1630kDa at 0.5:2:1:5weight ratio. (A) Percentage of EGFP positive ARPE-19 cells analyzed19cells (C) Percentageof EGFPpositiveHEK-293 cells analyzedby flowcytometry. (D)QuaSD (n =3). *p<0.05 respect toDNA–SLN; **p<0.01 respect toDNA–SLN; ap<0.05 respectto HA150 based vectors.

3.6. Evaluation of the CD44 receptor influence in the uptake

Cellular uptake in both cells lines after the pre-treatmentwith asaturated solution of HAwas quantified bymeasuring the Nile Red-labelled vectorswhichwere inside the cells 2 h after the addition ofHA–P–DNA–SLN vectors.

3.6.1. ARPE-19 cellsFig. 7A shows the flow cytometry results of cells without

previous incubation (blue line) and treated with the solution of HA(yellow filled curve). The mean relative fluorescence signalobtained was 111.41�15.36 for non-treated cells with the HAsolution and 91.21�2.17 for the cells treated previously with theHA solution. The values obtained for DNA–SLN vectors were363.64�42.91 for the cells without a previous saturation and176.60�15.81 for the cells previously treated. Statistically signifi-cant differences were observed (p<0.05).

3.6.2. HEK-293 cellsIn the Fig. 7B is shown the flow cytometry analysis of cells

without previous incubation (blue line) and treated with thesolution of HA (yellow filled curve). The mean relativefluorescence signal obtained was 46.38�1.43 for non-treatedcells with the HA solution and 40.90�1.54 for the cells treatedpreviously with the HA solution. The values obtained forDNA–SLN vectors were 59.31�2.67 for the cells without aprevious saturation and 51.71�0.86 for the cells previouslytreated. Statistically significant differences were observed(p<0.05).

P–DNA–SLN vectors with HA of different molecular weights, 150 kDa, 500 kDa andby flowcytometry. (B) Quantification of EGFP protein by fluorimetric assay in ARPE-ntificationof EGFPproteinby fluorimetric assay inARPE-19 cells. Errorbars representtoHA1630 based vectors; aap<0.01 respect toHA1630 based vectors; yyp<0.01 respect

[(Fig._5)TD$FIG]

Fig. 5. Cellular uptake of HA–P–DNA–SLN vectors with HA of different molecular weights, 150kDa, 500 kDa and 1630kDa at 0.5:2:1:5 weight ratio. (A, B) Percentage ofnanoparticle uptaking over time in ARPE-19 and HEK-293 cells. Error bars represent SD (n =3). *p<0.05 respect to DNA–SLN vectors; yp<0.05 respect to HA150 based vectors.(C, D) Flow cytometry histograms of ARPE-19 cells and HEK-293 cells treated with Nile Red-labelled vectors. Grey filled curve =Untreated cells; Black curve=DNA–SLN; Redcurve=HA150–P–DNA–SLN; Blue curve=HA500–P–DNA–SLN; Green curveHA1630–P–DNA–SLN. (For interpretation of the references to color in this figure legend, the reader isreferred to the web version of the article).

P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426 419

3.7. Internalization mechanism

3.7.1. ARPE-19 cellsFig. 8 features the confocal images of ARPE-19 cells after the

treatment with the endocytosis markers (Transferrin or CholeraToxin) andwith the vectors preparedwithout andwith HA and P atan HA:DNA ratio of 0.5:1. We observed colocalization between thecholera toxin and the vectors containingHA (yellow in Fig. 8C, E andG). This result was confirmed by the Mander’s Overlap coefficient;the highest value of this coefficient (0.74�0.14) was achievedwiththe HA150–P–DNA–SLN vectors and cholera toxin (values higher

[(Fig._6)TD$FIG]

Fig. 6. Fluorescence microscopy images of Alexa Fluor 488-CD44 receptor (green). ARPEImages are at 63� magnification. (For interpretation of the references to color in this fi

than 0.6 indicate colocalization). Overlap coefficients were lowerthan 0.6 when the cells were treated with transferrin, the clathrinmarker, and the vectors, regardless of whether they were preparedwith or without HA. These results indicate that the endocytosis viacaveolae/lipid raft is themainmechanism of cell internalization forall the vectors in ARPE-19 cells.

3.7.2. HEK-293 cellsWhen colocalization studieswere analyzed in HEK-293 cells, no

differenceswere detected in themechanismof endocytosis used bythe DNA–SLN vector (Fig. 9A and B) and the HA–P–DNA–SLN vector

-19 cells (A) and HEK-293 cells (B) were treated with DAPI-fluoromount G1 (blue).gure legend, the reader is referred to the web version of the article).

[(Fig._7)TD$FIG]

Fig. 7. Flow cytometry histograms of ARPE-19 cells (A) and HEK-293 cells (B) treated with Nile Red-labelled vectors. Grey filled curve=Untreated cells; Bluecurve =HA–P–DNA–SLN without previous addition of a saturated HA solution; Yellow filled curve=HA–P–DNA–SLN with previous addition of a saturated HA solution. (Forinterpretation of the references to color in this figure legend, the reader is referred to the web version of the article).

420 P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426

(Fig. 9C–H). As depicted in the Fig. 9, all the Mander’s Overlapcoefficients ranged from 0.60 to 0.66 for all the vectors, indicatingthat both mechanisms of internalization are used in a similarmanner in HEK-293 cells.

3.8. Intracellular distribution of EMA-labelled DNA

Figs. 10 and 11 present images captured by fluorescencemicroscopy of the ARPE-19 andHEK-293 cells, respectively, treatedwith thevectorspreparedwithEMA-labelledpCMS–EGFP (red). Forthis study, different formulations were prepared: EMA–SLN,P–EMA–SLN and HA150–P–EMA–SLN.

In ARPE-19 cells (Fig. 10), after the addition of the EMA-SLNvector, the plasmid (red) was detected poorly condensed in thecytoplasm, both at 4h and 12h. Nevertheless, when the vectorswere prepared with P and without HA the DNA appears highlycondensed and near the nucleus. When the cells were treatedwiththe vector bearing HA and P, a higher amount of DNAwas detectedin the cytoplasm in comparison with the other two vectors;moreover, the DNA appeared condensed and dispersed all over the

[(Fig._8)TD$FIG]

Fig. 8. CLSM images of Nile Red labelled SLNs and AlexaFluor 488-Cholera toxin (gremagnification. Presence of yellow color represents the overlay of the endocytic pathwacolocalization. (A–B)DNA–SLNvectors at 1:5weight ratio. (C–D)HA150 (0.5):2:1:5 vectorsof the references to color in this figure legend, the reader is referred to the web version

cytoplasm and the perinuclear area at 4h, whereas at 12h part ofDNA was uncondensed.

Fig. 11 shows the fluorescence microscopy pictures of HEK-293cells after the treatment with the vectors bearing EMA-labelledpCMS–EGFP. Like in ARPE-19 cells, when the DNA was formulatedin the P–EMA–SLN vector, it appeared highly condensed and closeto the nucleus. When the cells were treated with the HAPDNASLNvector, a higher amount of DNA was detected inside the cells,mostly condensed at 4h, and in both forms, condensed anduncondensed at 12h.

3.9. Characterization of the vectors with pCEP4-RS1

As observed in Fig. 12, the vectors composed by SLNs anddifferent proportions of protamine andHA150 or HA500were able toprotect the plasmid pCEP4-RS1 from DNase I and to release it.Regarding size and zeta potential (Table 2), vectors prepared withHA500 showed bigger size, higher PDI and lower zeta potential thanvectors prepared with HA150.

en) and AlexaFluor 488-Transferrin (green) in ARPE-19 cells. Images are at 63�y marker and the vectors. Mander’s Overlap coefficient (R) values over 0.6 indicate. (E–F)HA500 (0.5):2:1:5 vectors. (G–H)HA1630 (0.5):2:1:5 vectors. (For interpretationof the article).

[(Fig._9)TD$FIG]

Fig. 9. CLSM images of Nile Red labelled SLNs and AlexaFluor 488-Cholera toxin (green) and AlexaFluor 488-Transferrin (green) in HEK-293 cells. Images are at 63�magnification. Presence of yellow color represents the overlay of the endocytic pathway marker and the vectors. Mander’s Overlap coefficient (R) values over 0.6 indicatecolocalization. (A–B)DNA–SLNvectors at 1:5weight ratio. (C–D)HA150 (0.5):2:1:5 vectors. (E–F)HA500 (0.5):2:1:5 vectors. (G–H)HA1630 (0.5):2:1:5 vectors. (For interpretationof the references to color in this figure legend, the reader is referred to the web version of the article).

[(Fig._10)TD$FIG]

Fig.10. Fluorescencemicroscopy images ofARPE-19 cells at 4 hand12hafteradditionofDNA–SLN; P–DNA–SLNandHA–P–DNA–SLNvectors. Imagesareat63�magnification.Cells were treated with DAPI-fluoromount G1 (blue) and vectors containing EMA-labelled pCMS–EGFP plasmid (red). The ratio (w/w) were 1:5 for DNA–SLN vector; 2:1:5 forP–DNA–SLN vector and 0.5:2:1:5 for HA based vector. (For interpretation of the references to color in this figure legend, the reader is referred to thewebversion of the article).

P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426 421

[(Fig._11)TD$FIG]

Fig.11. Fluorescencemicroscopy imagesofHEK-293cells at 4hand12hafteradditionofDNA–SLN;P–DNA–SLNandHA–P–DNA–SLNvectors. Imagesareat63�magnification.Cells were treatedwith DAPI-fluoromount G1 (blue) and vectors containing EMA-labelled pCMS–EGFP plasmid (red). The ratio (w/w) were 1:5 for DNA–SLN vector; 2:1:5 forP–DNA–SLN vector and 0.5:2:1:5 for HA based vector. (For interpretation of the references to color in this figure legend, the reader is referred to thewebversion of the article).

[(Fig._12)TD$FIG]

Fig.12. pCEP4-RS1 binding to SLNs, protection against DNAse I and SDS-induced release of DNA fromHA–P–DNA–SLN vectorswithHA150 andHA500 at 0.5:2:1:2 and 0.5:3:1:3ratios.

422 P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426

Table 2Physical characterization of SLNs-based vectors bearing the plasmid pCEP4-RS1. Particle size (nm); Polydispersity index (PDI); and Zeta potential (mV). The size, PDI andsuperficial charge represent mean� SD (n=3).

Vector HA–protamine–DNA–SLN ratio Size (nm) PDI Zeta potential (mV)

HA150 0.5:2:1:2 261.50�23.77 0.25�0.02 +32.50�1.250.5:3:1:3 280.77�25.82 0.24�0.01 +33.70�2.52

HA500 0.5:2:1:2 333.80�82.32 0.47�0.04 +20.30�0.520.5:3:1:3 355.73�32.21 0.36�0.01 +22.60�0.62

[(Fig._13)TD$FIG]

Fig.13. Transfectionof theplasmidpCEP4-RS1 inARPE-19cellswithHA150–P–DNA–SLNvectors at 0.5:2:1:2and0.5:3:1:3 ratios. Photographsshowimagesof the retinsochisin(red)produced inARPE-19 cells treatedwithDAPI-fluoromountG1 (blue). Thegraphicdepicts the concentrationof retinoschisin in the culturemediumofARPE-19 cells 48 and72h after addition of vectors. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of the article).

P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426 423

3.10. In vitro transfection with the plasmid pCEP4-RS1

The retinoschisin protein expressed in ARPE-19 cells treatedwith the vectors HA150:P:DNA:SLN containing the plasmid pCEPS-RS1 is shown in the images in Fig.13 (red). The protein secretedwasdetected at least 72h after addition of vectors (graphic in Fig. 13).

4. Discussion

Non-viral gene therapy based on solid lipid nanoparticles (SLNs)is apromising strategy for the treatmentof severaldiseases (Gascónet al., 2012a, 2012b). As non-viral vectors, their main advantage istheir safety profile, but the low efficiency of transfection of SLNsrenders them far from an ideal vector. At present, the improvementof the efficiency of transfection of non-viral vectors in general, andSLNs in particular, is a challenge for real progresses in gene therapy.With this goal inmind, in thepresent studywehavedesigned anewvector composed by SLNs, protamine (P) and hyaluronic acid (HA)and its potential utility for gene therapy has been studied.

Stable complexeswithnet positive chargewere obtained thanksto the electrostatic interaction between the positive charge of theprotamine and SLN, and the negative charge of HA and DNA. Theparticle size, ranging from 240nm to 340nm, is conditioned by thebalance between both, the ability of the components to condensetheDNA,whichwould implya reduction in theparticle size, and thespace these components take up (Delgado et al., 2012b; del Pozo-Rodríguez et al., 2009).

All vectors, regardless of the content and molecular weight ofHA, were able to fully bind the pCMS–EGFP plasmid, to protect itagainst the action ofDNAse, and to release it inpresence of SDS. Theprotamine, due its high capacity to condense DNA (Ye et al., 2008;Sorgi et al.,1997), is at least inpart, responsible for the protection ofthe plasmid against enzymatic degradation. It is important to takeinto account that a balance between the capacity to condense DNA,to protect the geneticmaterial, and the ability of DNA to be releasedfrom the vector and to reach the nucleus, is needed for a successfultransfection process (del Pozo-Rodríguez et al., 2008).

In order to know the influence of the proportion of HA and itsmolecular weight on transfection efficacy, we transfected ARPE-19

424 P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426

cellwith thevectors andmeasured theEGFPexpressionat 72h.As itcanbe seen in Fig. 3, regardless of the amount andmolecularweightof HA, transfection was higher with the vectors containing thepolyanion. Molecular weight of HA did not induce significantdifferences in transfection;however, themoreHA150 andHA500wasincluded in the vector, thehigher thepercentageof transfectedcellsobtained up to 0.5:1 ratio. Based on these results, we selected thevectors prepared with the three HA in a proportion of 0.5:1 for therest of the study.

One of the most critical steps for transfection is the entry of theDNA into the nucleus, which is favoured during cell division. In thecase of slow or non-dividing cells, whose nuclear membrane ismuch more difficult to cross, transfection is particularly problem-atic (Zabner et al.,1995). In order to better knowthepotential of ourvectors, we evaluated the transfection capacity of our vectors inARPE-19 cells, which divide slowly and are a goodmodel for retinaldiseases, and in HEK-293 cells, that divide rapidly and arefrequently used as a cell model to evaluate novel strategies forgene therapy.

In both cell lines, transfection studies over time showed that allthe new vectors prepared with protamine and HA presented highcapacity to transfect without compromising the cell viability. InARPE 19 cells, the highest level of transfectionwas achieved at 96h(72.8�1.3% EGFP positive cells); the incorporation of P and HAinduced almost a 7-fold increase in the transfection capacity of theSLNs. HEK-293 cellswere also efficiently transfected by the vectors;contrary to ARPE-19 cell, protamine and HA hardly affected thecapacity of transfection of SLN (Fig. 4).We had previously (Delgadoet al., 2011) showed that protaminedrastically reduced the capacityof SLNs to transfect HEK-293 cells, but HA incorporated in the SLNconfers to this new vector the ability to transfect them. Therefore,HA–P–DNA–SLN formulations possess a high capacity to efficientlytransfect cells with different division rate, allowing us broaden therange of target cells.

In order to obtain insights into the mechanisms by which HAinfluences the transfection in both cell lines,we studied the cellularuptake, the endocytosis mechanism and the intracellular distribu-tion of the vectors. For this purpose, we labelled the vectors withNile Red. All the vectors were detected into both cell lines by flowcytometry (in Fig. 5C and D, rightward displacement of thehistograms in comparisonwith non-treated cells), and the cellularuptakeprofileover timewas similar inboth cell lines (Fig. 5AandB).From6h the uptake of the vectors preparedwith protamine andHAwas significantly higher than the uptake of the DNA–SLN vector,being this difference higher in HEK-293. Differences in cell uptakebetween the vector DNA–SLN and the vectors HA–P–DNA–SLNcould be due to differences in the capacity to bind to cell surface.Internalization efficiency over time depends on the total number ofbinding sites/receptors per cell, the rate of receptor recycling andthe effect of the cargo on the receptor up/down regulation (Duncanand Richardson, 2012). As mentioned above, it is well-known thecapacity of HA to interact with CD44 and other HA-specificreceptors, that facilitates cell internalization (Ruponen et al., 2001).This could justify the higher uptake of the vectors bearing HA.However, differences in cellular uptake do not justify the differ-ences in cell transfection obtained with the vectors.

Endocytosis has beenpostulated as themainmechanismof entryfor non-viral systems. While certain endocytic pathways areubiquitous to all cells, others are cell-specific or play an enhancedrole in certain cell types. This is an important consideration whennew nanomedicines are developed to act within a particular celltype/diseased tissue (Duncan and Richardson, 2012). In ARPE-19cells,wherebothclathrinandcaveolae/lipid raftendocyticprocessesare equally active (Delgado et al., 2012b), the HA–protamine-DNA–SLN vectors showed higher colocalization with the caveolae/lipid raft endocytic marker, whereas the DNA–SLN complexes have

previously shown colocalization with both markers (Delgado et al.,2011). InHEK-293 cells the vectorsusedboth clathrin- and caveoale/lipidraft-mediatedendocytosis, regardlessof thepresenceofHAandprotamine on the surface of the nanoparticle. However, it isimportant take into account that caveoale/lipid raft-mediatedendocytosis is quantitatively much more active in HEK-293 cellsthan clathrin-mediated endocytosis (Delgado et al., 2012b), whichmeans that the contribution of caveolae/lipid raft-mediatedendocytosis to the cell uptake of the complexeswill bemuch higherthan clathrin-mediated endocytosis.

Since caveolae-mediated endocytosis has the ability to bypasslysosomes (though there are few exceptions), efficient transfectionof HEK-293 is thought to occur if the release of the DNA is notdependent of the action of lysosomal enzymes. In fact, it has beenpreviously shown that when SLN are prepared with protamine butwithoutHA, theyareunable to transfect this cell line (Delgado et al.,2011) because of the high degree of condensation of theDNAdue tothe protamine. Moreover, in other cells in which vectors bearingprotamine enter by clathrin-mediated endocytosis, efficienttransfection was achieved (Delgado et al., 2011; Ruiz de Garibayet al., 2012), since the lysosomal activity associated to this uptakemechanism facilitates theDNA release from the complex. However,in the present work the formulations bearing protamine and HAenter HEK-293 cells via caveola/lipid raft and are able to efficientlytransfect. These results seem to indicate that HA facilitates therelease of the DNA in absence of lysosomal enzymes.

SinceourvectorsarepreparedwithHAand this polyanion is ableto interact with CD44, this receptor may be involved in theinternalization process of the formulations. Actually, CD44receptors have been shown to be involved in the cell uptake ofdifferent nanocarriers (Wojcicki et al., 2012; Lu et al., 2011; de laFuente et al., 2008a).

Immunochemistry showed a high expression of CD44 receptorsin ARPE-19 cells (Fig. 6A), as previously reported (de Kozak et al.,2004); in contrast, this receptor was detected in HEK 293 cells in alesser extension (Fig. 6B).When CD44was inhibited byan excess ofHA, hardlyaffect cell internalizationof the formulations inHEK-293cells. However a meaningful decrease in the uptake of the vectorswas observed only in ARPE-19 cells; this result suggests theparticipation of this receptor in the cell internalization of thevectors. The decreaseof theHA–P–DNA–SLNvectors uptakemaybedue to the high affinity of HA to CD44 receptors of the surface ofmany mammalian cells (Arpicco et al., 2013). In the case ofDNA–SLN vectors, which mainly enter by caveolae/lipid raft-mediated endocytosis, the decrease of uptake due to saturationwith HA could be related with the fact that caveolin-mediatedendocytosishas been shown tobe activatedwhenHA interactswithCD44, andbothcaveolinandCD44arepresent in lipid rafts domains(Wojcicki et al., 2012). However, it has to be taken inmind that dueto the nature of biological systems, several dynamic processesmight take place in parallel which might turn in competewith oneanother, which difficulties the interpretation of quantitativeresults.

In the study of the intracellular disposition of the vectors, inorder to know the effect of protamine and HA separately, a vectorprepared with protamine but without HA (P–DNA–SLN) was alsoassayed in both cell lines. Condensation of theDNAwas very high inthe cells treated with P–DNA–SLN, lower in the cells treated withHA–P–DNA–SLN, andeven lower in the cells treatedwithDNA–SLN.These results indicate that the HA is able to modulate the highdegree of condensation of the DNA due to the protamine. In ouropinion, the interaction of HA with the intracellular matrix couldchange the binding among the components of the formulation; asresult DNA loosening occurs, which favours the approach oftranscription factors (Ito et al., 2006). Actually, this DNAdecondensation seems to be related with the transfection capacity

P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426 425

in HEK-293 cells, since, as mentioned above, P–DNA–SLN is unableto transfect (Delgado et al., 2011), and HA–P–DNA–SLN transfects,but slower than DNA–SLN. These results confirm the importance ofthe capacity of the formulations to control the release of the DNAinside the cell when the internalization is mediated by caveolae/lipid raft.

In order to explore the potential application of our vectors forthe treatment of ocular diseases, we incorporate in the vectors theplasmid that encodes retinoschisin (pCEP4-RS1), a deficient proteinin the X-linked juvenile retinoschisis. This degenerative disorder ofthe retina is a common cause of juvenile blindness in males, with aprevalence of 1:5000 to 1:25,000 (Mooy et al., 2002). Theapplication of gene replacement therapy in this disease has beenconsidered a promising therapeutic approach. In this part of thestudy AH150 and HA500 were used to prepare vectors. Although thevector prepared with HA1630 was able to efficiently induce theexpression of the EGFP, increasing the size of HA also increases theinherent viscosity, making the vector sticky. pCEP4-RS1 has ahigher molecular mass than pCMS–EGFP; therefore, thew:w ratiosused to prepare the new vectors were modified to ensure DNAbinding, protection and later release, as observed in Fig. 12. Thevectors prepared with HA500 and pCEP4-RS1 presented higherparticle size and PDI, and lower zeta potential than the vectorsbearing the pCMS–EGFP plasmid (Table 2); following transfectionstudies were only carried out with the vectors bearing HA150.

The vectors prepared with the pCEP4-RS1 were able to inducethe production of retinoschisin in ARPE-19 cell (Fig.13). Comparingwith a previous study, these vectors induced a higher amount ofretinoschisin (up to0.7 ng/mL) than avector preparedwith SLNanddextran (0.2 ng/mL) (Delgado et al., 2012a). The ocular administra-tion to rats of the vector bearing dextran showed capacity totransfect in vivo ocular tissues; therefore, the vector prepared withHA150might induce even higher response in vivo than that obtainedwith the vector preparedwith dextran. In a previous study (Li et al.,2013), the authors also used shielded liponanoparticles with HA toachieve RPE-targeted distribution and prolonged intraocularresidence due to the capacity of HA to interact with CD44, highlypresent in ARPE-19 cells.

5. Conclusions

In conclusion, the incorporation of HA and protamine to theSLNs allowed us to obtain a versatile vector, able to efficientlytransfect cells with different rate of cellular division, widening thepotential applications of SLN-based vectors. HA is able tomodulatethe high degree of condensation of the DNA due to the protamine;this is important if the vector is uptaken mainly by caveoale/lipidraft-mediated endocytosis, since by means of optimization of theformulationwecan influence thedecondensationof theDNA insidethe cell, and thus, the transfection rate. Moreover, the vectorprepared with the plasmid that encodes the retinoschisinwas ableto efficiently transfect ARPE-19 cells. Hence, we have shown thepotential application of protamine, HA and SLN-based vectors forthe treatment of X-linked juvenile retinoschisis by gene therapy.

Acknowledgments

This work was supported by the Basque Government’sDepartment of Education, Universities and Investigation (IT-341-10) and by the Spanish Ministry of Economy and Competi-tiveness (SAF2010-19862).. We thank the University of Basquecountry (UPV-EHU) for the grant awarded to Paola S. Apaolaza.The authors acknowledge the General Service (SGIker) ofAnalytical Microscopy and High Resolution in Biomedicine ofthe University of the Basque Country (UPV-EHU) for technicaladvice on confocal microscopy as well.

References

Aragona, P., 2004. Hyaluronan in the treatment of ocular surface disorders. In: Garg,H.G., Hales, C.A. (Eds.), Chemistry and Biology of Hyaluronan. Elsevier Ltd,Oxford, pp. 529–551.

Arpicco, S., De Rosa, G., Fattal, E., 2013. Lipid-based nanovectors for targeting ofCD44-overexpressing tumor cells. Journal of Drug Delivery ID 860780, 8 pages.

deKozak, Y., Andrieux,K., Villarroya,H., Klein, C., Thillaye-Goldenberg, B., Naud,M.C.,et al., 2004. Intraocular injection of tamoxifen-loaded nanoparticles: a newtreatment of experimental autoimmune uveoretinitis. European Journal ofImmunology 34 (37), 02–12.

de la Fuente, M., Seijo, B., Alonso, M.J., 2008a. Bioadhesive hyaluronan-chitosannanoparticles can transport genes across the ocularmucosa and transfect oculartissue. Gene Therapy 15, 668–676.

de la Fuente, M., Seijo, B., Alonso, M.J., 2008b. Novel hyaluronic acid-chitosannanoparticles for ocular gene therapy. Investigative Ophthalmology and VisualScience 49, 2016–2024.

del Pozo-Rodríguez, A., Delgado,D., Solinís,M.A., Gascón,A.R., Pedraz, J.L., 2007. Solidlipid nanoparticles: formulation factors affecting cell transfection capacity.International Journal of Pharmaceutics 339, 261–268.

del Pozo-Rodríguez, A., Delgado, D., Solinís, M.A., Gascón, A.R., Pedraz, J.L., 2008.Solid lipid nanoparticles for retinal gene therapy: transfection and intracellulartrafficking in RPE cells. International Journal of Pharmaceutics 360, 177–183.

del Pozo-Rodríguez, A., Pujals, S., Delgado, D., Solinís, M.A., Gascón, A.R., Giralt, E., etal., 2009. A proline-rich peptide improves cell transfection of solid lipidnanoparticle-based non-viral vectors. Journal of Controlled Release 133, 52–59.

del Pozo-Rodríguez, A., Delgado, D., Solinís, M.A., Pedraz, J.L., Echevarría, E.,Rodríguez, J.M., et al., 2010. Solid lipid nanoparticles as potential tools for genetherapy: in vivo protein expression after intravenous administration. Interna-tional Journal of Pharmaceutics 385, 157–162.

del Pozo-Rodríguez, A., Delgado, D., Solinís, M.A., Gascón, A.R., 2011. Lipidnanoparticles as vehicles for macromolecules: nucleic acids and peptides.Recent Patents on Drug Delivery & Formulation 5, 214–226.

Delgado, D., del Pozo-Rodríguez, A., Solinís, M.A., Gascón, A.R., 2011. Understandingthe mechanism of protamine in solid lipid nanoparticle-based lipofection: theimportance of the entry pathway. European Journal of Pharmaceutics andBiopharmaceutics 79, 495–502.

Delgado, D., Del Pozo-Rodríguez, A., Solinís, M.A., Aviles-Triqueros, M., Weber, B.H.,Fernández, E., et al., 2012.Dextranandprotamine-basedsolid lipidnanoparticlesas potential vectors for the treatment of x-linked juvenile retinoschisis. HumanGene Therapy 23, 345–355.

Delgado, D., Gascón, A.R., del Pozo-Rodríguez, A., Echevarría, E., Ruiz de Garibay, A.P.,Rodríguez, J.M., et al., 2012. Dextran–protamine–solid lipid nanoparticles as anon-viral vector for gene therapy: in vitro characterization and in vivotransfection after intravenous administration to mice. International Journal ofPharmaceutics 425, 35–43.

Delgado, D., del Pozo-Rodríguez, A., Solinís, M.A., Bartkowiak, A., Gascón, A.R., 2013.Newgenedelivery systembasedonoligochitosan and solid lipid nanoparticles: ‘.European Journal of Pharmaceutical Sciences 50 (3-4), 484–491.

Duncan, R., Richardson, S.C.W., 2012. Endocytosis and intracellular trafficking asgateways for nanomedicine delivery: opportunities and challenges. MolecularPharmaceutics 9, 2380–2402.

Evanko, S.P., Wight, T.N., 1999. Intracellular localization of hyaluronan inproliferating cells. Journal of Histochemistry and Cytochemistry 47 (10),1331–1341.

Gascón, A.R., Solinís, M.A., del Pozo-Rodríguez, A., Delgado, D., Pedraz, J.L. Inventors.Universityof the BasqueCountryUPV/EHU. Lipid nanoparticles for gene therapy.US 20120183589 A1. 2012.

Gascón, A.R., Solinís, M.A., del Pozo-Rodríguez, A., Delgado, D., Fernández, E.Lipidnanoparticles for treatingoculardiseases.Universityof theBasqueCountryUPV/EHU and Universidad Miguel Hernández de Elche. WO 2012085318 A1. 2012.

Gascón, A.R., del Pozo-Rodríguez, A., Solinís, M.A., 2013. Non-viral delivery systemsin gene therapy. In: Martin, F. (Ed.), Gene Therapy – Tools and PotentialApplications, pp. 3–33.

Ito, T., Iida-Tanaka, N., Niidome, T., Kawano, T., Kubo, K., Yoshikawa, K., et al., 2006.Hyaluronic acid and its derivative as a multi-functional gene expressionenhancer: protection from non-specific interactions, adhesion to targeted cells,and transcriptional activation. Journal of Controlled Release 112, 382–388.

Mooy, C.M., Van Den Born, L.I., Baarsma, S., Paridaens, D.A., Kraaijenbrink, T., Bergen,A., et al., 2002. Hereditary X-linked juvenile retinoschisis: a review of the role ofMüller cells. Archives of Ophthalmology 120, 979–984.

Li, G., Jing, W., Yanan, Z., Dan, C., Chunliu, Z., Jianping, L., Yong, G., 2013. Hyaluronan-modified core–shell liponanoparticles targeting CD44-positive retinal pigmentepithelium cells via intravitreal injection. Biomaterials 34, 5978–5987.

Lu, H.D., Zhao, H.Q., Wang, K., Lv, L.L., 2011. Novel hyaluronic acid–chitosannanoparticles as non-viral gene delivery vectors targeting osteoarthritis.International Journal of Pharmaceutics 420, 358–365.

RuizdeGaribay,A.P.,Delgado,D., del Pozo-Rodríguez,A., Solinís,M.A.,Gascón,A.R., 2012.Multicomponentnanoparticlesasnonviralvectors forthetreatmentofFabrydiseaseby gene therapy. Drug Design, Development and Therapy 6, 303–310.

Ruponen, M., Ronkko, S., Honkakoski, P., Pelkonen, J., Tammi, M., Urtti, A., 2001.Extracellular glycosaminoglycans modify cellular trafficking of lipoplexes andpolyplexes. Journal of Biological Chemistry 276, 33875–33880.

Sorgi, F.L., Bhattacharya, S., Huang, L., 1997. Protamine sulfate enhances lipid-mediated gene transfer. Gene Therapy 4, 961–968.

Wo

Yam

Ye,

Zab

Zin

426 P. Apaolaza et al. / International Journal of Pharmaceutics 465 (2014) 413–426

jcicki, A.D., Hillaireau, H., Nascimento, T.L., Arpicco, S., Taverna,M., Ribes, S., et al.,2012. Hyaluronic acid-bearing lipoplexes: physico-chemical characterizationand in vitro targeting of the CD44 receptor. Journal of Controlled Release 162,545–552.ada, Y.,Hashida,M., Y,Hayashi, Y., Tabata,M.,Hyodo,M., Ara,M.N., et al., 2013. Anapproach to transgene expression in liver endothelial cells using a liposome-based gene vector coated with hyaluronic acid. Journal of PharmaceuticalSciences . doi:10.1002/jps.23480.

J., Wang, A., Liu, C., Chen, Z., Zhang, N., 2008. Anionic solid lipid nanoparticlessupported on protamine/DNA complexes. Nanotechnology 19, 285708.ner, J., Fasbender, A.J.,Moninger, T., Poellinger, K.A.,Welsh,M.J.,1995. Cellular andmolecular barriers to gene transfer by a cationic lipid. Journal of BiologicalChemistry 270 (32), 18997–19007.chuk, V., Zinchuk, O., 2008. Quantitative colocalization analysis of confocalfluorescence microscopy images. Current Protocols in Stem Cell Biology 4.19,4.19.1–4.19.16.