ORIGINAL ARTICLE

A nanomedicine to treat ocular surface inflammation: performanceon an experimental dry eye murine modelL Contreras-Ruiz1,2,5, GK Zorzi3,5, D Hileeto1, A Lopez-Garcıa1,2, M Calonge1,2, B Seijo3,4, A Sanchez3,4 and Y Diebold1,2

MUC5AC is a glycoprotein with gel-forming properties, whose altered expression has been implicated in the pathogenesis of dryeye disease. The aim of our study was to achieve an efficient in vivo transfection of MUC5AC, restore its normal levels in an inflamedocular surface and determine whether restored MUC5AC levels improve ocular surface inflammation. Cationized gelatin-basednanoparticles (NPs) loaded with a plasmid coding a modified MUC5AC protein (pMUC5AC) were instilled in healthy andexperimental dry eye (EDE) mice. MUC5AC expression, clinical signs, corneal fluorescein staining and tear production wereevaluated. Ocular specimens were processed for histopathologic evaluation, including goblet cell count and CD4 immunostaining.Neither ocular discomfort nor irritation was observed in vivo after NP treatment. Expression of modified MUC5AC was significantlyhigher in ocular surface tissue of pMUC5AC-NP-treated animals than that of controls. In healthy mice, pMUC5AC-NPs had no effecton fluorescein staining or tear production. In EDE mice, both parameters significantly improved after pMUC5AC-NP treatment.Anterior eye segment of treated mice showed normal architecture and morphology with lack of remarkable inflammatory changes,and a decrease in CD4þ T-cell infiltration. Thus, pMUC5AC-NPs were well tolerated and able to induce the expression of modifiedMUC5A in ocular surface tissue, leading to reduction of the inflammation and, consequently improving the associated clinicalparameters, such as tear production and fluorescein staining. These results identify a potential application of pMUC5AC-NPsas a new therapeutic modality for the treatment of dry eye disease.

Gene Therapy (2013) 20, 467–477; doi:10.1038/gt.2012.56; published online 19 July 2012

Keywords: nanoparticle; MUC5AC; dry eye; ocular surface; gene delivery

INTRODUCTIONThe ocular surface represents the interface between the eye andthe outer world. It is defined as an anatomical, histological andphysiological unit composed of the tear film and the corneal,limbal and conjunctival epithelia and adjacent connective tissues.The ocular surface along with the lacrimal glands and theinterconnecting nervous, immune and endocrine systems formthe lacrimal functional unit.1 All components of the lacrimalfunctional unit are necessary to maintain the ocular surfacehomeostasis.

Dry eye syndrome is defined as a multifactorial disease of thetears and ocular surface that results in symptoms of discomfort,visual disturbance and tear film instability with potential damageto the ocular surface. It is accompanied by increased osmolarity ofthe tear film and inflammation of the ocular surface.2 Damage toany component of the lacrimal functional unit can result in dryeye, resulting in tear hyperosmolarity and tear film instability.Hyperosmolarity might cause damage to the ocular surfaceepithelium by activating a cascade of inflammatory events.Epithelial damage involves loss of goblet cells (GCs) andreduction in mucin secretion, and leads to tear film instability.This instability exacerbates ocular surface hyperosmolarity,thereby creating a vicious circle.2

Mucins are a heterogeneous group of large glycoproteins foundas major components in all mucous secretions. In the ocular

surface they act as lubricants of the corneal and conjunctivalepithelial surfaces during eyelid blinking, as stabilizers of thepreocular tear film and as physical barrier to pathogen pene-trance.3 Inflammation can result in decreased mucin production.4

Among mucins, MUC5AC is one of the most clinically relevant.MUC5AC is a highly glycosylated gel-forming mucin secreted byspecialized epithelial cells of the conjunctiva, namely the GCs,which has a key role in tear homeostasis.5 MUC5AC expressionis decreased in several ocular diseases, such as keratoconjuncti-vitis sicca, Sjogren syndrome, pemphigoid or Stevens-Johnsonsyndrome.4,6,7 Dry eye syndrome is also accompanied by decreasedMUC5AC expression and exhibits an abnormal pattern of proteinglycosylation in GCs.8 Because of its important role in the stabilityof the lacrimal fluid and its decreased expression in several ocularinflammatory conditions, MUC5AC could be a potential target fora new therapy for dry eye.

The eye is a promising candidate to benefit from gene therapy.Main reasons are its well-defined anatomy and accessibility.Promising results in the area of ocular gene therapy have alreadybeen reported. Acland et al performed successful RPE65 genereplacement therapy in natural canine models of Leber’s congenitalamaurosis that resulted in restoration of vision.9 Adeno-associatedvirus-mediated transfer of soluble VEGF receptor 1 gene hasbeen shown to reduce retinal neovascularization by 50%.10

However, ocular gene therapy has remarkable limitations.

1Ocular Surface Group-IOBA, University of Valladolid, Valladolid, Spain; 2Networking Research Center on Bioengineering, Biomaterials and Nanomedicine (CIBER BBN), Valladolid,Spain; 3NANOBIOFAR Group, Department of Pharmaceutical Technology, University of Santiago de Compostela, Santiago de Compostela, Spain and 4Molecular Imagen Group,IDIS, Complejo Hospitalario Universitario de Santiago, Santiago de Compostela, Spain. Correspondence: Dr Y Diebold, Ocular Surface Group-IOBA, University of Valladolid, EdificioIOBA, Campus Miguel Delibes, Paseo de Belen 17, E-47011 Valladolid, Spain.E-mail: yol@ioba.med.uva.es5These authors contributed equally to the work.Received 5 March 2012; revised 1 June 2012; accepted 18 June 2012; published online 19 July 2012

Gene Therapy (2013) 20, 467–477& 2013 Macmillan Publishers Limited All rights reserved 0969-7128/13

www.nature.com/gt

To achieve efficient delivery of DNA to ocular surface cells, severalbarriers must be overcome, such as corneal and conjunctivalepithelia and tear film, along with physiologic mechanism such asblinking or tear clearance, limiting the penetration of DNA into theeye.11,12

The use of nanoparticles (NPs) as carriers of therapeutic geneticmaterial for delivery to target tissues and treatment of a widerange of ocular diseases has become popular in recent years.13,14

We recently described the use of NPs based on cationized gelatine(CG) and chondroitin sulphate (CS) to efficiently transfect cornealand conjunctival epithelial cells (PCT/ES2011/07078, priority date14 November 2011).15 These NPs were able to protect the plasmidDNA from degradation and have been shown to successfullyinduce plasmid internalization in ocular surface cells, therebyrepresenting promising new carriers for gene delivery.15

The purpose of this work was to achieve an efficient in vivotransfection of MUC5AC using cationized gelatine (CG)-based NPs,to restore MUC5AC levels in an inflamed ocular surface. Ourhypothesis was that the normalization of the mucin levels wouldreduce the inflammation and break the mutually perpetuatingpathogenetic mechanisms, which sustain the inflammatory stateof the ocular surface in dry eye.

RESULTSpMUC5AC-NPs induced the expression of modified MUC5AC inocular surface tissues of healthy miceHybrid NPs based on gelatine and CS were prepared by ionicgelation technique. The physicochemical properties of the blankNPs were 143±10 nm and þ 34±2 mV (for size and zetapotential, respectively), and for the pMUC5AC-NPs 128±6 nmand þ 37±1 mV. MUC5AC mRNA expression was measured byreal-time reverse-transcription PCR (RT2-PCR) to evaluate theability of pMUC5AC-loaded NPs to produce an efficient transfec-tion of ocular tissues in healthy mice. Mice treated with 2 mg perday of pMUC5AC (both naked pMUC5AC and pMUC5AC-NPs) didnot show MUC5AC expression in either corneal or conjunctivalepithelia (data not shown). Plasmid concentration was thenincreased to 4 mg per day, and significant expression of MUC5ACmRNA was detected in conjunctiva after 5 days of pMUC5AC-NPstreatment, whereas the cornea remained negative (Figure 1).

pMUC5AC-NPs were well tolerated in ocular surface structuresof healthy miceClinical signs in all animals at all time points and with the twodifferent plasmid concentrations were scored as grade 0 as noneof the animals showed any ocular discomfort, irritation or swelling.There were also no macroscopically detectable alterations of theocular surface.

To evaluate the tear physiology, tear production was measuredin all animals before (baseline) and after the treatment (post).Although there were no statistically significant differences, thereseemed to be a tendency toward an increase in the tearproduction after treatment with pMUC5AC-loaded NPs, at all timepoints and plasmid concentrations (Figure 2a).

Corneal fluorescein staining was used as measurement of theeffect of the different treatments on the corneal epithelial barrierfunction (Figure 2b). No staining was observed in any animal.There were no differences between baseline and post-treatmentfluorescein staining in any experimental group and amonggroups.

After all treatments, histopathology evaluation showed thatcorneal and conjunctival epithelia displayed normal cell layers andmorphology (Figure 3a). Conjunctival GCs were abundant,especially in the superior tarsal conjunctiva. The lacrimal glandsalso displayed normal architecture and morphology. No differ-ences were observed among the different experimental groups.

All treated and control groups demonstrated complete lackof inflammatory changes, or focal mild alterations that werewithin the normal range and variations expected for healthysubjects.

Periodic acid–Schiff (PAS) staining showed showed a normalnumber of GCs along the conjunctiva in control mice (Figures 3band c). No significant differences were detected in mice treatedwith naked plasmid and blank NPs. However, GC density wassignificantly higher in the pMUC5AC-NPs-treated mice comparedwith control mice, after both 3 and 5 day treatment. Therefore,pMUC5AC-NPs treatment increased GC numbers in healthy mice.

pMUC5AC-NPs induced the expression of modified MUC5AC inconjunctiva of EDE miceRT2-PCR revealed an increase in MUC5AC mRNA expression afterpMUC5AC-NPs treatment for 5 days in both cornea andconjunctiva (Figure 4). This increase was statistically significantin conjunctiva. No MUC5AC expression was detected in healthymice, experimental dry eye (EDE) mice and EDE mice treated withblank NPs, whereas naked pMUC5AC-treated mice showedincreased MUC5AC expression, although at significantly lowerlevel than after pMUC5AC-NPs treatment.

pMUC5AC-NPs improved the inflammatory parameters associatedwith EDENo ocular discomfort, irritation or swelling was detected in anypMUC5AC-NPs-exposed animal. Therefore, the clinical signs in allof them were scored as grade 0. No edema, redness or cornealvascularization was observed in treated eyes. In addition, therewere no visible macroscopic alterations of the ocular surface.

Tear production was measured in all animals, before EDEinduction (baseline) and before (post EDE) and after the treatment

Figure 1. Expression of modified MUC5AC mRNA evaluated byRT2-PCR, in healthy mice cornea (a) and conjunctiva (b) treated withpMUC5AC-NPs for 3 and 5 days (4 mg pDNA per day). Controlsincluded non-treated mice (control), mice treated with nakedplasmid (naked pMUC5AC) and with blank NPs (blank NPs). Micedid not show MUC5AC expression in cornea after 3 and 5 days oftreatment (a). In conjunctiva, significant amounts of MUC5AC mRNAwere detected after 5 days of pMUC5AC-NPs treatment (b). *Po0.05.

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

468

Gene Therapy (2013) 467 – 477 & 2013 Macmillan Publishers Limited

(post). EDE produced a significant reduction in tear production inall animals. Controls without EDE did not display reduction in tearproduction. There were no significant differences between postEDE and post-treatment tear production in healthy mice, controlEDE mice, EDE mice treated with naked plasmid and with blankNPs. However, after pMUC5AC-loaded NPs treatment, EDE micetear production returned to baseline levels (Figure 5a). Therefore,the treatment with pMUC5AC-loaded NPs increased the tearproduction in EDE mice leading to normal baseline levels.

Corneal fluorescein staining was performed before EDE (base-line) and before (post EDE) and after the treatment (post).Punctate and diffuse corneal fluorescein staining were observed inEDE animals, but not in the healthy controls. Whereas there wereno changes after treatment in healthy mice, control EDE mice andEDE mice treated with naked plasmid or with blank NPs, thestaining in mouse corneas treated with pMUC5AC-loaded NPsseemed to decrease (Figure 5b).

To further evaluate the corneal epithelial barrier condition inthe different experimental groups, the expression of two proteinsimplicated in tight junction complexes, zonula occludens protein 1(ZO-1) and zonula occludens protein 2 (ZO-2), was analyzedby immunofluorescence (Figure 5c). In healthy control mice,ZO-1 and ZO-2 were located in the plasma membrane at cell–cellcontacts in the most superficial layers of corneal epithelium.After EDE induction, ZO-1 and ZO-2 showed a predominantcytosolic localization in corneal epithelial cells. However, afterpMUC5AC treatment, both proteins returned to the membrane

localization, pointing out a reorganization of tight junctioncomplexes.

In addition, phalloidin staining revealed changes in thedistribution of actin cytoskeleton in corneal epithelium. In healthymice, actin showed a well-structured corneal epithelium,whereas after EDE induction, the outer most epithelial layersshowed an altered architecture (Figure 5c). pMUC5AC-NPstreatment seems to be able to reorganize the actin cytoskeletonof EDE mice, recovering the healthy mice corneal epitheliumcondition.

There were no significant differences between pro-inflamma-tory cytokine levels in healthy control mice and pMUC5AC-NPs-treated mice, as determined by the cytokine multiplex assay (datanot shown). This result suggests that pMUC5AC do not elicit anacute inflammatory response based on cytokine release. However,pMUC5AC-NPs-treated EDE mice showed similar cytokine levelsthan EDE control mice, without significant differences. Specifically,IFNg, IL-1b, and IL-2 levels were similar in both conditions, whileIP-10 and TNFa showed a slight decrease in pMUC5AC-NPs-treated EDE mice (IP-10: 9.78±1.25 vs 4.28±0.64 pg ml� 1 andTNFa: 27.93±1.10 vs 14.29±1.60 pg ml� 1), although the differ-ences were not significant. Some cytokines were not detected inthe collected tears, such as IL-6, IL-17 or RANTES, due to thesensitivity of the assay. All together, those data would indicatethat pMUC5AC-NPs had no direct influence in short-term cytokinerelease and did neither improve nor exacerbate the inflammatorycondition of EDE mice.

Figure 2. Condition of ocular surface before (baseline) and after pMUC5AC-NPs treatment (post-treatment). Controls included non-treatedmice (control), mice treated with naked plasmid (naked pMUC5AC) and with blank NPs (blank NPs). Neither tear production (a) nor fluoresceinstaining (b) changed after pMUC5AC-NP treatment for 3 and 5 days.

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

469

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 467 – 477

In the histopathology evaluation all specimens displayed ananterior eye segment with normal architecture and morphology(Figure 6a). The histopathologic changes observed ranged fromcomplete lack of alterations to presence of focal mild and isolatedinflammatory changes, which fell into the focal mild category. Noanimals from any experimental group displayed moderate orsevere inflammatory changes.

GC numbers were normal along the conjunctiva in healthymice. The distribution of the GC population within the conjunctivais regional, with the highest density in the fornices, where thesecells form clusters. In our study, GC density showed tendencytoward decrease after EDE induction. Both blank NPs andpMUC5AC-NPs seemed to increase the GC numbers, butthe detected differences did not reach statistical significance(Figures 6b and c).

To further evaluate the inflammatory condition of the ocularsurface tissue in the different experimental groups, CD4 imnu-nostaining was performed to identify Th cells (Figure 7). Nosignificant immunoreactivity for CD4 antibody (Ab) was detectedin the conjunctiva of control healthy mice. However, EDE mice

Figure 3. (a) Light micrographs of control, naked pMUC5AC, blank NPs and pMUC5AC-NPs-treated H/E-stained tissue sections from healthymice. Representative corneal, conjunctival and lacrimal gland hematoxylin and eosin (HE)-stained sections showed normal cell layers, withoutalterations in epithelial morphology and complete lack of inflammatory changes. (b) Goblet cell (GC) density in conjunctiva in healthy micetreated with pMUC5AC-NPs. Representative light micrographs of conjunctival Alcian blue–PAS-stained tissue sections after 5 days. (c) GCcount bar charts showed a significant increased in GC density in the pMUC5AC-NPs-treated mice compared with control mice, after both 3and 5 days-treatment. Scale bar: 50 mm. *Po0.05. Arrowhead: goblet cells. ep, epithelium; st, stroma.

Figure 4. Expression of modified MUC5AC mRNA evaluated byRT2-PCR, in EDE mice cornea and conjunctiva treated withpMUC5AC-NPs (EDEþpMUC5AC-NPs) for 5 days. Controls includenon-treated healthy mice (control), control EDE mice (EDE), EDEmice treated with naked plasmid (EDEþnaked pMUC5AC) and withblank NPs (EDEþblank NPs). EDE mice treated with pMUC5AC-NPsshowed significant MUC5AC expression in cornea and conjunctiva.*Po0.05.

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

470

Gene Therapy (2013) 467 – 477 & 2013 Macmillan Publishers Limited

showed remarkable CD4þ T-cell infiltration of the conjunctivaltissue. CD4þ cells were apparent in the substantia propria and inthe epithelium. After pMUC5AC-NPs treatment, no CD4þ T cellswere identified in mice conjunctiva. In order to refine the study,CD4þ T-cell density was determined counting the positive CD4cells along the whole conjunctiva. CD4þ T-cell density signifi-cantly increased in EDE mice compared with healthy mice.pMUC5AC-NPs-treated mice showed a significant decrease inCD4þ T cells, compared with EDE control mice, while neithernaked pMUC5AC nor blank NPs had significant effect. Nosignificant differences were detected between healthy mice andEDE mice treated with pMUC5AC-NPs.

DISCUSSIONIn this study we evaluated the in vivo transfection capability of anew generation of cationized-based NPs loaded with a modifiedMUC5AC encoding novel plasmid (pMUC5AC-NPs) in ocular

tissues, in order to normalize mucin levels in an inflamed ocularsurface. We also studied whether the mucin level normalizationmight lead to improvement in the clinical signs associated withocular inflammation and ultimately alleviate the inflammatoryprocess. We found that pMUC5AC-NPs were well tolerated andinduced the expression of modified MUC5AC in ocular surfacestructures, both in healthy and EDE mice. In addition, the mucinlevel normalization led to the abolishment of the inflammationand to a remarkable improvement of the associated clinical signs.

CG-based NPs were used as plasmid carriers. NPs are recognizedas promising genetic delivery systems. There are several studies inwhich NPs have been used for ocular gene therapy.16 However,most of them reported significant difficulties associated withtoxicity, an invasive way of administration, or low ability to achievetarget cells. The hybrid NPs composed of CG and the naturalpolyanion CS, associated with the modified MUC5AC encodingplasmid have been previously tested in vitro and reported assuitable systems for ocular cells gene therapy.11,12 The small size

Figure 5. Condition of ocular surface before EDE (baseline) and before (post EDE) and after pMUC5ac-NPs treatment (post). Controls includenon-treated healthy mice (control), control EDE mice (EDE), EDE mice treated with naked plasmid (EDEþnaked pMUC5AC) and with blank NPs(EDEþblank NPs). (a) EDE produced a significant reduction in tear production in all animals. After pMUC5AC-loaded NPs treatment, tearproduction significantly increased and returned to baseline level (*Po0.05 vs baseline; #Po0.05 vs post EDE). (b) Diffuse corneal fluoresceinstaining was observed in EDE mice (arrow), while fluorescein staining in mouse corneas treated with pMUC5AC-loaded NPs seemed todecrease. (c) Reorganization of actin cytoskeleton in corneal epithelium under basal conditions, EDE induction and after pMUC5AC-NPstreatment. Actin cytoskeleton was labeled with rhodamin-conjugated phalloidin. Scale bar: 50 mm.

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

471

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 467 – 477

(o150 nm) and the positive zeta potential (approximatelyþ 30 mV) of NPs has been described to improve its uptake byepithelial cells17 and increase the interactions with the ocularepithelia18 achieving an efficient in vitro MUC5AC transfectionwithout compromising the cell viability.15 MUC5AC was chosen as

a potential target for this therapy due to its important role in thestability of the lacrimal fluid and its decreased expression inseveral ocular inflammatory conditions.8

Our study results demonstrated that treatment with 4 mg perday of pMUC5AC-NPs for 5 days was able to induce the expression

Figure 6. (a) Light micrographs of hematoxylin and eosin (HE)-stained tissue sections from healthy control mice, EDE mice and EDE micetreated with naked pMUC5AC, blank NPs and pMUC5AC-NPs-treated tissue sections. Representative corneal, conjunctival and lacrimal glandsections showed normal cell layers, no alterations in epithelial morphology and focal mild inflammatory changes in the EDE groups, whichcompletely disappeared after MUC5AC-NPs treatment. Original magnification � 20. (b) GC density along conjunctiva in EDE mice treated withpMUC5AC-NPs. Representative light micrographs of conjunctival Alcian blue–PAS-stained sections (b) and GC count bar charts (c) showeda decrease in GC density after EDE induction and a trend toward an increase in GC density after pMUC5AC-NPs. Scale bar: 50mm. *Po0.05.Arrowhead: goblet cells. ep, epithelium; st, stroma.

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

472

Gene Therapy (2013) 467 – 477 & 2013 Macmillan Publishers Limited

of MUC5AC in the conjunctiva of healthy mice, but not in cornea.This observation is in agreement with our previous in vitroexperience with these nanosystems.15

It is important to point out that this is one of the first in vivoworks that achieved effective transfection of ocular epithelia,with topical treatment of a therapeutic plasmid using non-viralvectors. Indeed, previous works have been reported successfultransfection in corneal and conjunctival epithelial cells usingmodel, non therapeutic plasmids.19,20 In contrast, pMUC5AC-NPsled to a significant increase in the expression of a functionalhuman protein, the modified MUC5AC protein, in mice ocularsurface epithelia.

Neither ocular discomfort nor irritation was observed in vivoafter pMUC5AC-NP treatment. No significant effects were detectedon fluorescein staining or tear production either. Besides, treatedmice showed normal morphology and a complete lack ofinflammatory changes. Furthermore, pMUC5AC-NPs seem toimprove some clinical and histologic parameters: pMUC5AC-NPstreatment significantly increased the GC numbers in healthy miceand there seemed to be a tendency toward an increase in tearproduction after treatment with the pMUC5AC-loaded NPs.Based upon the absence of histopathologic and functionalalterations of the ocular surface, pMUC5AC-NPs seem to benonhazardous and have excellent tolerance in all oculartissues. These results were expected as each administration wasbelow the toxic in vitro concentration attributed to these NPsaccording to studies in human corneal epithelial cells.21 Otherevidences of the safe application of NPs to the ocular surfacehave been reported.22

Once pMUC5AC-NPs were shown to be well tolerated, and toinduce the expression of modified MUC5AC in the ocular surfacetissues of healthy mice, pMUC5AC-NPs were tested in a well-established murine model of ocular inflammation (EDE). EDE wasinduced in mice through exposure to desiccating stress andscopolamine injections. EDE animals have been described todevelop ocular surface lesions that mimic those found in dry eye

patients: corneal fluorescein staining, reduced tear production,decreased conjunctival GC density and CD4þ T-cell infiltration ofthe conjunctiva.23 In this work, EDE mice demonstrated: (1) weakfluorescein staining attributable to a slight disturbance in thecorneal epithelial barrier; (2) a significant decrease in tearproduction; (3) no significant change in GC density, althoughsome apparent tendency toward decrease was observed; and(4) a significant increase in the CD4þ T cells in the conjunctiva.These findings agree with those previously reported, but in ourstudy they were less pronounced. It is to be noted that EDEinduction was performed for only 10 days, which resulted inthe EDE mice being in a very early stage of the disease. Theexperimental animals would certainly require a more prolongedperiod of follow-up to manifest to a full extent the whole range ofocular surface lesions characteristic of dry eye.

The histopathologic evaluation of the EDE mice revealed normalmorphology and focal mild inflammatory changes. These resultsare compatible with the expected morphological changes in achronic inflammatory condition in a very early stage. In thecontext of our histopathologic findings, it is to be noted thatthe dry eye syndrome is a chronic condition, which develops overa long period of time, and it could manifest clinically initiallywithout the presence of pronounced inflammatory changes.The development of some of the other histopathologic featurescharacteristic of dry eye, such as stromal fibrosis and epithelialkeratinization, would require a more prolonged period of follow-up to become morphologically apparent.

pMUC5AC-NPs were able to induce MUC5AC mRNA expressionin EDE mice. This expression was significantly higher than inuntreated control animals in both cornea and conjunctiva. Cornealtissue of EDE mice expressed MUC5AC, whereas in healthy miceMUC5AC expression was not detected. The alteration of thecorneal barrier integrity caused by the inflammatory process,which has been previously reported (manuscript submitted),might ease the NP internalization and, therefore, increase thetransfection efficiency and MUC5AC expression.

Figure 7. CD4þ T-cell infiltration in EDE mice treated with pMUC5AC-NPs. (a) CD4þ immunostaining showed remarkable CD4þ T-cellinfiltration in conjunctival tissue (arrowheads), whereas in healthy mice and EDE mice treated with pMUC5AC-NPs, no CD4þ T cells wereidentified in conjunctiva. (b) CD4þ T-cell density significantly increased in EDE mice compared with healthy mice. pMUC5AC-NPs treatmentled to a significant decrease in CD4þ T cells, achieving healthy mice levels. *Po0.05 vs control; #Po0.05 vs treatments. ep, epithelium;sp, substantia propia; st, stroma.

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

473

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 467 – 477

Regarding the condition of the ocular surface, in EDE mice,pMUC5AC-NP treatment was able to: (1) improve the integrity ofthe corneal epithelial barrier; (2) increase tear production; (3) showa tendency toward increase in the GC numbers; and (4) decreasethe CD4þ T-cell infiltration in the conjunctiva.

Fluorescein staining is a routine clinical test used to determinecorneal integrity in patients. When corneal epithelial cell–celljunctions are disrupted and/or corneal epithelial cells aredamaged, a fluorescein uptake occurs resulting in a visiblestaining.24 That disruption renders the ocular surface susceptibleto environmental insults. The staining in EDE mouse corneastreated with pMUC5AC-NPs seemed to decrease. These resultssuggest that pMUC5AC-NPs treatment might be responsible forthe improvement of the integrity of the ocular epithelial barrier inEDE mice. The mechanisms by which MUC5AC preserves cornealbarrier function remains to be fully determined, although ourpreviously reported studies may have provided some clues. In ourearlier experiments, simulated inflammatory conditions in vitro,similar to those in EDE mice, led to alterations of the cornealbarrier by modulating tight junctions protein composition.25 InEDE mice, the tight junction proteins, ZO-1 and ZO-2, in additionof actin cytoskeleton were altered, too. The anti-inflammatorycytokine IL-10 was able to reverse these conditions and reinstatethe barrier integrity.25 In a similar manner, pMUC5AC-NPs werealso able to decrease the inflammatory reaction and possiblyacting through the same mechanism of alleviating the inflamma-tion, they successfully restored the corneal barrier function andintegrity. Improvements in this barrier property are also observedin other classical dry eye anti-inflammatory therapy with corticoidsand cyclosporine A, and seem to have an important role in thereestablishment of normal lacrimal function.26

Measurement of tear secretion is one of the parameters morewidely used in the diagnosis of dry eye. The treatment withpMUC5AC-NPs increased the tear production in EDE mice tonormal baseline levels. The fact that blank NPs and free plasmiddid not increase tear production suggests that the changeobserved for the pMUC5AC-NPs was due to expression ofMUC5AC protein, and not due an intrinsic irritating process inthe eye. There is no reliable evidence to indicate how MUC5ACincreases tear production in the eye, but it might be related to ahigher tear quality. Owing to their hydrophilic character, mucinsparticipate in the prevention of the desiccation of the ocularsurface through water retention.24 Therefore, an increasedexpression of MUC5AC could have led to an improved tear film

quality with a subsequent decrease in evaporation. Currently,there is only one formulation approved by the Food and DrugAdministration able to increase natural tear production (0.05%cyclosporine emulsion; Restasis, Allergan Inc., Irvine, CA, USA).

GC density is another critical parameter that reflect the overallhealth of the ocular surface.27 Cytologic studies have indicated thatdecreased GC density is a consistent finding in dry eye in humans.28

Although GC density did not significantly change after pMUC5AC-NPs, there was an apparent tendency toward an increase in theGC numbers. Taking into account that EDE mice were in a veryearly stage of the disease, more prolonged EDE model might berequired to observe significant changes in GC density, both afterEDE induction and after pMUC5AC-NPs treatment.

No major histopathology alterations and only focal mildinflammatory changes were detected in EDE mice, before or afterpMUC5AC-NPs treatment. However, EDE mice demonstratedremarkable CD4þ T-cell infiltration in conjunctiva. pMUC5AC-NPs-treated mice demonstrated a statistically significant decreasein CD4þ T-cell numbers, comparable to those of healthy mice.CD4þ T cells are the most important inflammatory cellsimplicated in the development of dry eye syndrome.29 Thisimportant fact partially explains the ability of pMUC5AC-NPs toeffectively reduce the inflammation in EDE mice. The exactmechanism by which MUC5AC achieved the abolishment of theinflammatory response is not clear. Our hypothesis in this regard isthat an increase in MUC5AC expression might have potential rolein the process of hiding of key epitopes needed to trigger theinflammatory process. Indeed, it is reported that some secretedmucin can block the interaction among antigen epitopes and Absbecause of esteric hindrance.30,31

On the basis of these data, we propose that normalization ofMUC5AC levels leads to effective reduction of the inflammationand to break the vicious cycle of dry eye (Figure 8). NormalizedMUC5AC level has positive influence on the tear film quality andproperties. Through increasing water retention, MUC5AC wouldreduce the tear evaporation and would improve the tear filmstability and osmolarity. Maintaining normal osmolarity is animportant part of sustaining the healthy state of the ocular surfaceas it impedes and prevents the activation of new inflammatoryevents. Therefore, the proposed strategy could be considered asthe first effective treatment for dry eye that is not based in aclassical anti-inflammatory molecule.

In conclusion, pMUC5AC-NPs were well tolerated and able toinduce the expression of modified MUC5A in murine ocular

Figure 8. Proposed mechanism implicated in the reduction of the inflammation and the rupture of the vicious cycle of dry eye afterpMUC5AC-NPs treatment. The normalization of MUC5AC levels would improve the tear film quality, reduce the tear evaporation, increase thetear film stability and regulate tear osmolarity. A normal osmolarity might not activate new inflammatory events in the ocular surface.

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

474

Gene Therapy (2013) 467 – 477 & 2013 Macmillan Publishers Limited

surface tissue. This is a proof-of-concept of a safe topicaltransfection of ocular surface cells using non-viral recentlydesigned nanocarriers. Moreover, expression of the novelMUC5AC plasmid in diseased EDE mice correlated with significantreduction of the inflammation and, consequently, to remarkableimprovement of the clinical parameters associated with thiscondition including improvement of the integrity of the cornealepithelial barrier, increasing the tear production and GC numbers,and reducing the CD4þ T-cells infiltration in the conjunctiva.These results provide strong evidence in support of the potentialuse of pMUC5AC-NPs as a new therapeutic modality for dry eyedisease.

MATERIALS AND METHODSMiceAll studies adhered the Association for Research in Vision andOphthalmology statement for the Use of Animals in Ophthalmic andVision Research. Protocols were approved by the Institutional Board andsupervised by an expert in animal health care. Female C57BL/6 mice(Jackson Laboratory, Sacramento, CA, USA), 6–8 weeks old, were used inthese studies. Both eyes were used to reduce numbers of animals.

MaterialsType A gelatine (137 kDa) was generously gifted from Nitta Gelatin(Ontario, Canada). N-(3-Dimethylaminopropyl)-N0-ethylcarbodiimide hydro-chloride, tripolyphosphate, and spermine hydrochloride were purchasedfrom Sigma-Aldrich (Madrid, Spain). CS was purchased from Calbiochem(Darmstadt, Germany). The plasmid pMUC5AC, which codifies a modifiedMUC5AC protein and, optionally also, the green fluorescent protein as amarker, was designed by our group (PCT/ES2011/07078, priority date 14November 2011) and supplied by Biomedal (Sevilla, Spain)

pMUC5AC-loaded NPs (pMUC5AC-NPs)NPs were formed using the ionic gelation technique.32 The gelatine wascationized with spermine trough a carbodiimine reaction as previouslydescribed.33 CG was dissolved in water at a concentration of 1 mg ml� 1,CS at a concentration of 0.125 mg ml� 1 and tripolyphosphate ata concentration of 0.125 mg ml� 1. All solutions were prepared in waterand sterilized by filtration (0.22mm, MillexGV, Millipore, Billerica, MA, USA).The pMUC5AC plasmid and CS solution were incorporated into thetripolyphosphate solution. NPs were obtained by adding the resulting mixto the CG solution with magnetic stirring at room temperature (RT) andisolated by centrifugation (Beckman CR412, Beckman Coulter, Brea, CA,USA) at 10 000 r.c.f. for 30 min at 4 1C with 0.1% (v/v) glycerol. NPs werethen resuspended in water at a final plasmid concentration of 0.2 mg ml� 1

(PCT/ES2011/07078). The mean particle size was determined by photoncorrelation spectroscopy and the zeta potential was obtained by laserDoppler anemometry measuring the mean electrophoretic mobility. Bothassays were performed in using a Zetasizer Nano ZS (Malvern, UK).

Tolerance studyTo determine the tolerance of mice ocular surface structures to theexposure of pMUC5AC-NPs, healthy animals (n¼ 20, 5 per group) receivedinstillations (5 ml) of: (1) naked plasmid; (2) blank NPs; or (3) pMUC5AC-loaded NPs in both eyes four times a day during 3 or 5 days. Two differentconcentration of pMUC5AC were tested: 2 and 4 mg per day. Controlanimals, receiving no instillations, were included in the study as additionalcontrols. Three independent experiments were performed.

Forty-eight hours after the last instillation, MUC5AC mRNA expressionwas measured in cornea and conjunctiva by RT2-PCR. The ocular toleranceto the different treatments were evaluated using: (a) macroscopicevaluation of clinical signs; (b) fluorescein staining of the ocular surface(comparing pre- and post-treatment); (c) tear production (pre- and post-treatment); (d) tear collection for cytokine study; and (e) histopathologicevaluation of ocular tissues.

Murine EDEEDE was induced in C57BL/6 mice by injecting scopolamine (0.5 mg in0.2 ml) (Sigma-Aldrich) subcutaneously three times a day in alternativeflank for 10 days as previously described.23 Mice cages were placed in frontof a fan directed to blow across the wired screen side of the cage 24 h aday with controlled humidity (o40%).

EDE animals (n¼ 20; 5 per group) were distributed in four experimentalgroups: (1) control: no topical treatment received; (2) naked plasmid:0.2mg ml� 1 of naked pMUC5AC (total of 4 mg of pMUC5AC per day);(3) blank NPs; (4) pMUC5AC-loaded NPs group: 0.2mg ml� 1 of pMUC5AC-loaded NPs (total of 4 mg of pMUC5AC per day). In all groups, mice received5 ml per eye bilaterally four times a day. A healthy mice group (n¼ 5) wasadded as an additional control (no topical treatment received).

After 5 days under desiccating conditions and scopolamine injections,the different treatments were started in all experimental groups for 5 days,keeping the dry eye conditions at the same time. 48 h later, tears andtissues were collected. In addition, tear production and fluorescein stainingwere performed three times at different time points: (1) 24 h beforestarting the dry eye induction (baseline measurement); (2) at day 5, justafter starting the different treatments (dry eye measurement); and (3) andat day 10, the end of treatments (after treatment measurement). Twoindependent experiments were performed.

Clinical signs evaluationA macroscopic evaluation of the ocular structures to determine thepresence of clinical signs, including ocular discomfort, presence of cornealand/or conjunctival alterations, mucous discharge and lid swelling, wasperformed daily, using a grading scale of 0–2, following the criteriadescribed in Table 1.22

Tear productionTear production was measured with sterile cotton threads (ZoneQuickThread, Oasis Medical, Glendale, CA, USA). Threads, which turn red whenabsorbance of tears occurs, were held with forceps into the inferiorconjunctival sac for 30 s. Afterwards, the threads were removed and thedistance of redness was measured in millimeters.

Corneal fluorescein stainingCorneal fluorescein staining was evaluated and photographed, inisoflurane sedated mice, with a slit lamp biomicroscope using a cobaltblue light 1 min after application of 1ml of 10mg ml� 1 sodium fluorescein(Sigma-Aldrich).

Tear collection and cytokine measurementA volume of 0.5ml of Beadlyte Assay Buffer (Millipore) were applied to theocular surface. Tear fluid (2ml) was then collected by placing a glasscapillary tube in the tear meniscus at the canthus of both eyes. Teninflammatory cytokines and chemokines (IFNg, TNFa, RANTES, IP-10, IL-1b,IL-2, IL-6, IL-10, IL12p70 and IL-17) were evaluated in tear fluid with a10X-multiplex immunobead-based assay (MPXMCYTO-70K MILLIPLEX MAPMouse Cytokine/Chemokine Panel, Millipore). Cytokine multiplex beadassay was done following the manufacturer’s protocol. Briefly, 10ml of eachdiluted (1/5 in assay buffer) sample were incubated overnight on a plate

Table 1. Criteria for macroscopic evaluation

Grade Discomfort Cornea Conjunctiva Discharge Lids

0 No reaction No alterations No alterations No discharge No swelling1 Blinking Mild opacity Mild hyperemia/mild edema Mild discharge without

moistened hairMild swelling

2 Enhanced blinking/intensetearing/vocalizations

Intense opacity Intense hyperemia/intenseedema/hemorrhage

Intense discharge withmoistened hair

Obvious swelling

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

475

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 467 – 477

shaker in the dark at 4 1C with Beadlyteanti-mouse multicytokine beads(Millipore). After this, beads were washed and incubated with BeadlyteBiotinylated secondary Abs for 1 h, followed by 30-min incubation withstreptavidin–phycoerythrin solution. Finally, beads were washed, resus-pended and read on a LuminexTM100-IS instrument (Luminex Corporation,Austin, TX, USA). Data were analyzed with the BeadViewTM Software(Upstate, UK).

Histopathology studyMice were euthanized by CO2 inhalation and their eye balls with attachedeyelids and lacrimal glands were excised. Eye tissues were embedded inoptimal cutting temperature compound, frozen and sectioned with acryostat (7 mm). Tissue samples were also fixed in 4% paraformaldhehyde,embedded in paraffin and sectioned (4mm). Sections were then stainedwith hematoxylin and eosin and Alcian blue–PAS. Conjunctival and cornealepithelial morphology were evaluated. Histopathology evaluation of thespecimens was carried out by a pathologist (co-author DH) in a maskedmanner, including cornea, conjunctiva, anterior eye segment, orbital softtissue and lacrimal gland. Conjunctival GC numbers were counted in amasked manner by two independent trained observers, in the superior andinferior bulbar and tarsal conjunctivas, using an � 20 objective. Sectionsfrom both eye balls of three experimental animals from each group(in three different sets of experiments) were examined and photographed.

RNA isolation and RT2-PCRTotal RNA from corneal and conjunctival epithelia collected and pooledfrom each group (six eyes per group per experiment) were isolatedfollowing the guanidium thiocyanate–phenol–chloroform extractionmethod (Trizol, Invitrogen, Carlsbad, CA, USA). RNA concentration wasmeasured by its absorbance at 260 nm and stored at � 80 1C before used.cDNA was generated form 1 mg of total RNA with the SuperScript VilocDNAkit (Invitrogen), according to the manufacturer’s protocol.

RT2-PCR was performed on the 7500 Real Time System (AppliedBiosystem, Carlsbad, CA, USA) using SYBR Green (Applied Biosystem) toquantify levels of MUC5AC mRNA in the different treatment groups.MUC5AC primers (50-CCCACAGAACCCAGTACAA-30 and 50-AATGTGTAGCCCTCGTCT-30) and mouse 18s rRNA primers (SAbioscience, Frederick,MD, USA) were used. The cycle profile was 50 1C for 2 min, 95 1C for 2 min,40 cycles at 95 1C for 20 s, 59 1 for 30 s and 72 1C for 40 s. To verify thespecificity of the amplification reaction, a melting curve analysis wasperformed. Relative quantification of the signals was done by normalizingthe signal of MUC5AC mRNA with the mouse 18S rRNA signal. Thedifference in expression for each target in the treated mice was referredto the amount in the control group.

Immunofluorescence assaysOptimal cutting temperature-embedded eye balls with attached eyelidswere cryosectioned at 7 mm and fixed in � 20 1C acetone for 10 min. Theywere incubated at RT for 50 min with blocking buffer composed ofphosphate-buffered saline (PBS) with 4% donkey serum, 0.3% Triton X-100and 1% bovine serum albumin (all from Sigma-Aldrich) to block non-specific binding. Afterwards, they were incubated with the primary AbsZO-1 (2.5 mg ml� 1) and protein 2 (ZO-2; 4 mg ml� 1) (both from Invitrogen)for 1 h at RT. Alexa Fluor 488-conjugated secondary Ab (20mg ml� 1;Invitrogen) were used and after washing, sections were mounted andcoversliped for microscopy examination. Each experiment was performedthree times and negative controls included the omission of primary Abs.

In addition, tissue sections were stained with rhodamine-conjugatedphalloidin (Invitrogen), which targets cytoskeletal actin network. Briefly,cells were fixed in ice-cold methanol, permeabilized in 5% Triton X-100 andstained with 10 U ml� 1 phalloidin. After 30 min of incubation, cellswere washed in PBS before microscopy examination. Fluorescence wasvisualized with a Leica TCS SPE confocal laser microscope equippedwith a solid state laser with a 488-nm excitation filter (Leica Microsystems,Wetzlar, Germany). Cells were observed under a dry � 40 objectivewith a pinhole size of 1 (optical slice thickness was 1.7mm). Eachexperiment was performed three times and negative controls includedomission of primary Abs.

CD4 immunostainingOptimal cutting temperature-embedded eye balls with attached eyelidswere cryosectioned at 7mm and fixed in � 20 1C acetone for 10 min. Slides

were treated with 0.3% hydrogen peroxide for 10 min, air dried at RT for1 h and rinsed in PBS. Sections were then blocked with 20% normal rabbitserum (Dako, Carpinteria, CA, USA) in PBS for 1 h at RT, rinsed in PBS andincubated with primary Ab Rat anti-Mouse CD4 (BD Pharmingen, Sparks,MD, USA) at RT for 1 h to determine the presence of Th cells. Subsequently,sections were incubated with the biotinylated polyclonal secondary Abanti-Rat Ig (BD Pharmigen) for 30 min at RT. Vectastain Elite ABC reagents(Vector Laboratories, Burlingame, CA, USA) were applied to the tissue forantigen localization. Sections were counterstained with hematoxylin.Negative controls were performed without primary Ab. In addition, anisotope control (Rat IgGak, BD Pharmigen) was performed. CD4 positivecell numbers in conjunctiva were counted in a masked manner by twoindependent trained observers, using an � 40 objective. Three sectionsfrom the eye balls of three animals from each group (in three different setsof experiments) were examined and photographed.

Statistical analysisRepeated measures analysis of variance (ANOVA) was performed for tearproduction data. Two way ANOVA was performed for the remaining data.Results were expressed as mean±s.e.m. Differences were considered to besignificant when *Po0.05.

CONFLICT OF INTERESTThe authors declare no conflict of interest.

ACKNOWLEDGEMENTSWe thank Michael E Stern, PhD, (Allergan Inc., CA, USA) for kindly providing facilitiesfor animal experiments and for expert advice; Sagrario Callejo, PhD, and the ConfocalMicroscopy Service from the University of Valladolid, for their technical support;Amalia Enriquez de Salamanca, PhD, and Maria Jesus Benito, MSc, for technicalsupport in Luminex experiments; and Itziar Fernandez, MSc, for statistical analysisadvice. This work was supported by grants from the Spanish Ministry of Science andTechnology (MAT2007-64626-C01-01/C02-02 and FPU Scholarship Program), andProgramme AlBan, the European Union Programme of High Level Scholarshipsfor Latin America (scholarship no. E07D402978BR).

REFERENCES1 Stern ME, Pflugfelder SC. Inflammation in dry eye. Ocul Surf 2004; 2: 124–130.2 The definition and classification of dry eye disease: report of the Definition

and Classification Subcommittee of the International Dry Eye Work Shop 2007.Ocul Surf 5: 75–92.

3 Gipson IK, Inatomi T. Cellular origin of mucins of the ocular surface tear film.Adv Exp Med Biol 1998; 438: 221–227.

4 Corrales RM, Narayanan S, Fernandez I, Mayo A, Galarreta DJ, Fuentes-Paez G et al.Ocular mucin gene expression levels as biomarkers for the diagnosis of dry eyesyndrome. Invest Ophthalmol Vis Sci 2011; 52: 8363–8369.

5 Gipson IK. The ocular surface: the challenge to enable and protect vision: theFriedenwald lecture. Invest Ophthalmol Vis Sci 2007; 48: 4390–4398.

6 Argueso P, Balaram M, Spurr-Michaud S, Keutmann HT, Dana MR, Gipson IK.Decreased levels of the goblet cell mucin MUC5AC in tears of patients withSjogren syndrome. Invest Ophthalmol Vis Sci 2002; 43: 1004–1011.

7 Kunert KS, Keane-Myers AM, Spurr-Michaud S, Tisdale AS, Gipson IK. Alteration ingoblet cell numbers and mucin gene expression in a mouse model of allergicconjunctivitis. Invest Ophthalmol Vis Sci 2001; 42: 2483–2489.

8 Danjo Y, Watanabe H, Tisdale AS, George M, Tsumura T, Abelson MB et al.Alteration of mucin in human conjunctival epithelia in dry eye. Invest OphthalmolVis Sci 1998; 39: 2602–2609.

9 Bainbridge JW, Smith AJ, Barker SS, Robbie S, Henderson R, Balaggan K et al. Effectof gene therapy on visual function in Leber’s congenital amaurosis. N Engl J Med2008; 358: 2231–2239.

10 Rota R, Riccioni T, Zaccarini M, Lamartina S, Gallo AD, Fusco A et al. Markedinhibition of retinal neovascularization in rats following soluble-flt-1 gene transfer.J Gene Med 2004; 6: 992–1002.

11 de la, Fuente M, Ravina M, Paolicelli P, Sanchez A, Seijo B, Alonso MJ. Chitosan-based nanostructures: a delivery platform for ocular therapeutics. Adv Drug DelivRev 2010; 62: 100–117.

12 Paolicelli P, de la, Fuente M, Sanchez A, Seijo B, Alonso MJ. Chitosan nanoparticlesfor drug delivery to the eye. Expert Opin Drug Deliv 2009; 6: 239–253.

13 Diebold Y, Calonge M. Applications of nanoparticles in ophthalmology. Prog RetinEye Res 2010; 29: 596–609.

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

476

Gene Therapy (2013) 467 – 477 & 2013 Macmillan Publishers Limited

14 Cai X, Conley S, Naash M. Nanoparticle applications in ocular gene therapy.Vision Res 2008; 48: 319–324.

15 Konat Zorzi G, Contreras-Ruiz L, Parraga JE, Lopez-Garcia A, Romero Bello R,Diebold Y et al. Expression of MUC5AC in ocular surface epithelial cells usingcationized gelatin nanoparticles. Mol Pharm 2011; 8: 1783–1788.

16 Sanchez A, Alonso MJ. Nanoparticular carriers for ocualr drug delivery.Torchilin VP (ed). Nanoparticuales as Drug Carriers, 2009.

17 Rejman J, Oberle V, Zuhorn IS, Hoekstra D. Size-dependent internalization ofparticles via the pathways of clathrin- and caveolae-mediated endocytosis.Biochem J 2004; 377: 159–169.

18 Rabinovich-Guilatt L, Couvreur P, Lambert G, Dubernet C. Cationic vectors inocular drug delivery. J Drug Target 2004; 12: 623–633.

19 de la, Fuente M, Seijo B, Alonso MJ. Bioadhesive hyaluronan-chitosan nano-particles can transport genes across the ocular mucosa and transfect ocular tissue.Gene Therapy 2008; 15: 668–676.

20 Toropainen E, Hornof M, Kaarniranta K, Johansson P, Urtti A. Corneal epithelium asa platform for secretion of transgene products after transfection with liposomalgene eyedrops. J Gene Med 2007; 9: 208–216.

21 Zorzi GK, Parraga JE, Seijo B, Sanchez A. Hybrid nanoparticle design based oncationized gelatin and the polyanions dextran sulfate and chondroitin sulfate forocular gene therapy. Macromol Biosci 2011; 11: 905–913.

22 Contreras-Ruiz L, de la Fuente M, Garcia-Vazquez C, Saez V, Seijo B, Alonso MJet al. Ocular tolerance to a topical formulation of hyaluronic acid and chitosan-based nanoparticles. Cornea 2010; 29: 550–558.

23 Dursun D, Wang M, Monroy D, Li DQ, Lokeshwar BL, Stern ME et al. A mousemodel of keratoconjunctivitis sicca. Invest Ophthalmol Vis Sci 2002; 43: 632–638.

24 Barabino S, Chen W, Dana MR. Tear film and ocular surface tests in animal modelsof dry eye: uses and limitations. Exp Eye Res 2004; 79: 613–621.

25 Contreras-Ruiz L, Schulze U, Garcıa-Posadas L, Arranz-Valsero I, Lopez-Garcıa A,Paulsen F et al. Structural and functional alteration of corneal epithelial barrierunder inflammatory conditions. Curr Eye Res 2012, e-pub ahead of print 27 June2012; doi:10.3109/02713683.2012.700756.

26 Lemp MA. Management of dry eye disease. Am J Manag Care 2008; 14: S88–101.27 Kinoshita S, Kiorpes TC, Friend J, Thoft RA. Goblet cell density in ocular surface

disease. A better indicator than tear mucin. Arch Ophthalmol 1983; 101: 1284–1287.28 Nelson JD, Wright JC. Conjunctival goblet cell densities in ocular surface disease.

Arch Ophthalmol 1984; 102: 1049–1051.29 Niederkorn JY, Stern ME, Pflugfelder SC, De Paiva CS, Corrales RM, Gao J et al.

Desiccating stress induces T cell-mediated Sjogren’s Syndrome-like lacrimalkeratoconjunctivitis. J Immunol 176: 3950–3957.

30 Nahta R, Yu D, Hung MC, Hortobagyi GN, Esteva FJ. Mechanisms of disease:understanding resistance to HER2-targeted therapy in human breast cancer.Nat Clin Pract Oncol 2006; 3: 269–280.

31 Price-Schiavi SA, Jepson S, Li P, Arango M, Rudland PS, Yee L et al. Rat Muc4(sialomucin complex) reduces binding of anti-ErbB2 antibodies to tumor cellsurfaces, a potential mechanism for herceptin resistance. Int J Cancer 2002; 99:783–791.

32 Parraga JE, Zorzi GK, Seijo B, Sanchez A. Gelatin nanoparticles for ocular genedelivery. Hum Gene Ther 2009; 20: 1545.

33 Lopez-Cebral R, Martin-Pastor M, Parraga JE, Zorzi GK, Seijo B, Sanchez A.Chemically modified gelatin as biomaterial in the design of new nanomedicines.Med Chem 2011; 7: 145–154.

pMUC5AC-loaded NPs for the treatment of EDEL Contreras et al

477

& 2013 Macmillan Publishers Limited Gene Therapy (2013) 467 – 477