lable at ScienceDirect

Biomaterials 30 (2009) 879–891

Contents lists avai

Biomaterials

journal homepage: www.elsevier .com/locate/biomateria ls

The enhancement of the immune response against S. equi antigens through theintranasal administration of poly-3-caprolactone-based nanoparticles

H.F. Florindo a,b, S. Pandit b, L. Lacerda b, L.M.D. Gonçalves a,c, H.O. Alpar b, A.J. Almeida a,*

a iMED, Faculdade de Farmacia da Universidade de Lisboa, Avenida Professor Gama Pinto, 1649-003 Lisbon, Portugalb CDDR, The School of Pharmacy, University of London, 29-39 Brunswick square, London WC1N 1AX, UKc Unidade Piloto, Instituto de Biologia Experimental e Tecnologica, 2781-901 Oeiras, Portugal

a r t i c l e i n f o

Article history:Received 10 September 2008Accepted 21 October 2008Available online 22 November 2008

Keywords:Poly-3-caprolactoneNanospheresAdjuvantsStreptococcus equi,Mucosal immunisation

* Corresponding author. Tel.: þ351 217 946 409; faE-mail address: aalmeida@ff.ul.pt (A.J. Almeida).

0142-9612/$ – see front matter � 2008 Elsevier Ltd.doi:10.1016/j.biomaterials.2008.10.035

a b s t r a c t

Strangles is a bacterial infection of the Equidae family that affects the nasopharynx and draining lymphnodes, caused by Streptococcus equi subspecies equi. This agent is responsible for 30% of all worldwideequine infections and is quite sensitive to penicillin and other antibiotics. However, prevention is still thebest option because the current antibiotic therapy and vaccination is often ineffective. As S. equi inducesvery strong systemic and mucosal responses in convalescent horses, an effective and economic stranglesvaccine is still a priority. In this study the humoral, cellular and mucosal immune responses to S. equiantigens encapsulated or adsorbed onto poly-3-caprolactone nanospheres were evaluated in mice.Particles were produced by a double (w/o/w) emulsion solvent evaporation technique and containedmucoadhesive polymers (alginate or chitosan) and absorption enhancers (spermine, oleic acid). Theirintranasal administration, particularly those constituted by the mucoadhesive polymers, increased theimmunogenicity and mucosal immune responses (SIgA) to the antigen. The inclusion of cholera toxin Bsubunit in the formulations successfully further activated the paths leading to Th1 and Th2 cells.Therefore, those PCL nanospheres are potential carriers for the delivery of S.equi antigens to protectanimals against strangles.

� 2008 Elsevier Ltd. All rights reserved.

1. Introduction

Streptococcus equi subsp. equi (S. equi) is the aetiological agent ofstrangles, one of the most costly, commonly diagnosed and wide-spread infectious diseases of Equidae worldwide, that has lead todevastating epidemics in stables where horses are housed [1,2].This is an acute, contagious and deadly respiratory tract disease,which typical signs of infection include pyrexia, suppurative,mucopurulent nasal discharge, lymphadenitis and abscessation,often in the lymph nodes of head and neck [1–3]. It occasionallyaffects other lymph nodes and organs, resulting in a severe stage ofthe disease called bastard strangles. Because S. equi persists in theenvironment for only a few weeks, the most important factor forthe maintenance of infection is the infected horse [3]. Even ifgenerally after 4–6 weeks the infected animals recover fromdisease eliminating S. equi from their organism, 10% will constitutelong-term S. equi carriers, harbouring the microorganism formonths. The presence of the pathogen is not detectable in S. equilong-term carriers and animals do not show any clinical signs ofdisease [1–3]. On the other hand, although S. equi is sensitive in

x: þ351 217 937 703.

All rights reserved.

vitro to some antibiotics, its use is not consensual as most of thetreatments are ineffective when external signs of disease arealready detectable [3,4].

Animals that recover from a S. equi natural infection showa protective immunity, mostly against one of the most virulentfactor of this bacteria, the anti-phagocytic cell wall-associatedS. equi M-like protein (SeM), that seems to remain in 75% of theanimals for five years, although the protection mechanism is notyet fully understood [3]. In fact, convalescent horses present IgGand IgA against SeM, both in serum and nasal secretions, whichencourages the development of efficient vaccines. As a result, theinclusion of animals in regular vaccination programmes, simulta-neously with an effective diagnostic test, constitute the key tocombat large strangles outbreaks [1,2]. Since the 1980s, bacterinsand adjuvant extracts have been widely used in the field; howeverthose vaccines, which specifically target the SeM protein, did notshow useful protection in horses [1,3,5]. Secretory IgA (SIgA) anti-bodies locally produced in the respiratory tract are extremelyimportant for prevention and control of infection at mucosal sites.Therefore, it appears that an effective strangles vaccine must beable to stimulate a nasopharyngeal immune response [6].

Biodegradable particles are of particular interest as technologyplatforms for the generation of both protective local and systemicimmunity. These carriers increase the level concentration of

Table 1Design of intranasal immunisation study.

Groups Formulation

1 Adsorbed PCL-PVAads.2 PCL-GCSads.3 PCL-ALGads.4 PCL-PVAads. þ CTB5 Encapsulated PCL-PVA6 PCL-GCS7 PCL-ALG8 PCL-SP9 PCL-OA10 PCL-PVA þ CTB11 FREE12 FREE þ CTB

PCL-PVA, polycaprolactone–polyvinyl alcohol nanospheres; PCL-GCS, poly-caprolactone–glycol-chitosan nanospheres; PCL-ALG, polycaprolactone–alginatenanospheres; CTB-B, subunit of cholera toxin; PCL-SP, polycaprolactone–sperminenanospheres; PCL-OA, polycaprolactone–oleic acid nanospheres; FREE, S. equienzymatic extract in solution; FREE þ CTB, S. equi enzymatic extract admixed Bsubunit of cholera toxin.

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891880

antigen over an extended time and uptake by lymph nodes,promoting the interaction between the antigen and specific antigenpresenting cells (APC); protect the antigen from proteolyticenzymes and have a great ability to gain access to the mucosal-associated lymphoid tissue (MALT), when compared to free anti-gens, thus providing a means of target vaccine delivery [7,8]. Thesedelivery systems have the capacity to be modified in order toachieve a prolonged and pulsatile release of encapsulated antigen,mimicking the primary dose and boosting commonly used in theconventional vaccination programs, reducing therefore the numberof doses needed to be administered in order to maintain an even-tual immune response for a long period, as was already shown inseveral publications [9–11]. The efficacy of polymeric carriers ininducing effective immune responses was reviewed by Lemoineet al. [12].

Polymeric particles induce a mucosal response after intranasal(i.n.) administration in lymphoid tissues regardless the target site,and the response is paralleled by the appearance of antibodies insecretions of glands distant from the site of immunisation due tothe role of the common mucosal immune system (CMIS), whichhighlights the immunoadjuvant properties of these carriers [13].The modification of polymeric and biodegradable particles withmucoadhesive compounds such as chitosan (CS) [11,14,15] andalginate (ALG [16]) to extend residence time, and absorptionenhancers as spermine (SP [17]) and oleic acid (OA [18]), constitutesan interesting approach in order to overcome the nasal barriers andimprove the delivery of antigens administered by mucosal routes[18]. Accordingly, in the present study SP was used as a potentialadjuvant to improve S. equi absorption through the nasal epithe-lium. In order to overcome CS poor solubility in physiological pH,a glycolchitosan (GCS) was used as it is soluble and active as anabsorption enhancer at pH 7.4 [19].

Antigens adsorbed and encapsulated in particulate systems havebeen able to induce strong immune responses. In a previous work,our group has shown that non-aggregated S. equi enzymatic extractprotein-adsorbed PCL microspheres enhanced serum specific IgG,IgG1 and IgG2a antibody responses and elicited an elevated type Iimmune response (Th1) cytokines, even 300 days after a singlesubcutaneous dose administration [20].

The purpose of this study is to use PCL nanospheres, modified bydifferent adjuvants (GCS, ALG, SP and OA), as potential carriers for S.equi surface proteins, and compare their ability to induce bothsystemic and local protective immunity after mucosal administra-tion in a mouse model. As cholera toxin B subunit (CTB) is one of themost potent mucosal adjuvant [7,9], both PCL loaded nanospheresand free S. equi antigens were admixed with CTB, and the resultantimmunogenicity was evaluated and compared with the afore-mentioned formulations.

2. Materials and methods

2.1. Materials

Poly-3-caprolactone (PCL, average molecular weight (MW) 42.5 kDa), polyvinylalcohol (PVA, MW 13–23 kDa, 87–89% hydrolyzed), alginate low viscosity (ALG), SP,OA, GCS, sucrose, CTB, Bicinchoninic acid (BCA) kit, trypsin–chymotrypsin inhibitor,EDTA sodium salt, iodoacetic acid and phenylmethanosulphonyl fluoride solution(PMSF) were supplied by Sigma Aldrich Co., UK. Micro BCA� Protein assay kit wassupplied by GOSS Scientific Ltd and Pierce UK. Dichloromethane (DCM) wasobtained from BDH Laboratory Supplies, UK. Streptococcus equi subsp. equi (strainLEX) ATCC 53186 were a kind gift from Professor J.F. Timoney (University of Ken-tucky, USA).

2.2. Animals

Each group of experimentation used in the in vivo studies was composed of fourfemale, 6–8-week-old BALB/c mice with food and drink provided ad libitum.Experiments were performed in strict accordance with the UK 1986 Animals(Scientific Procedures) Act.

2.3. Preparation of PCL nanospheres

PCL nanospheres were prepared aseptically at room temperature by a modifi-cation of a double emulsion (w/o/w) solvent evaporation method previouslyreported [20]. In brief, polymer was dissolved in 6 ml DCM and emulsified byhomogenisation using an Ultra-Turrax T25 (Janke & Kunkel, IKA-Labortechnik) for3 min at 24,000 rpm, with a 10% (w/v) PVA or a 1% (w/v) GCS solution, whichcontained the S. equi antigens (10.0 mg) in the formulation of nanospheres loaded byencapsulation. To prepare the S. equi antigens, a bacteria culture was harvested andtreated with N-acetyl muraminidase and lysozyme (Sigma Aldrich Co., UK) to extractthe cell wall proteins as previously described [20]. Absorption enhancers (10 mg),such as SP and OA were also used, being the first one dissolved in the PVA solution,while the second was previously molecularly dispersed in 1 ml of ethanol and thusmixed with DCM. The (w/o) emulsion was then added dropwise into 30 ml of a 1.25%(w/v) PVA or 0.75% (w/v) ALG solution, and homogenised for 7 min at 10,000 rpmwith the Silverson homogeniser (Silverson model L4RT, UK). The resultant (w/o/w)emulsion was magnetically stirred at room temperature for 4 h to evaporate theorganic solvent. The PCL nanospheres were harvested by centrifugation(20,000 rpm, 45 min, 15 �C; Beckman J2-21 High speed centrifuge), washed threetimes with 0.02% (w/v) sucrose solution and subsequently freeze-dried (Virtis, UK)to obtain a fine, free-flowing dry powder of nanospheres.

To adsorb the S. equi extract proteins, 20 mg of plain particles (PCL-PVA, PCL-GCSand PCL-ALG) were weighed and dispersed in 2 ml of a 1250 mg/ml proteins solutionin a LoBind� Eppendorf tube (Eppendorf, UK). Particles were then incubated for 1 hin a water bath at 37 �C, under agitation (100 rpm). After incubation, particlessuspension was centrifuged at 7,500 rpm for 10 min (IEC Micromax Eppendorfcentrifuge, UK) the pellet washed twice and then allowed to dry in a desiccator,while the supernatant and the washes were kept frozen at �20 �C until futureanalysis.

For i.n. immunisations, 50 ml of a particles suspension in PBS (pH 7.4) containingadsorbed or encapsulated S. equi antigens equivalent to 10 mg of SeM were admin-istered per mouse. Plain particles and free antigen were used as controls (Table 1).

2.4. Physicochemical characterisation of PCL nanospheres

The size and surface charge of nanospheres were determined by MalvernZetaSizer (Malvern Instruments, UK). For zeta potential measurement, a 10 mM

potassium chloride solution (Sigma Aldrich Co., UK) was used to disperse particles.Zeta potential data (mV) were obtained from the average of six measurements witha standard deviation of �5%.

The surface morphology of nanospheres was analysed by scanning electronmicroscopy (SEM, Phillips/FEI XL30 SEM) as previously described [20].

2.5. Determination of antigen loading

The total amount (% w/w) of protein entrapped per unit weight of nanospheres(loading capacity, L.C.) was directly measured using the MicroBCA� protein assayafter digestion of 10 mg of particles with 2.5 ml of a 5% (w/v) sodium dodecylsulphate (SDS, BDH UK) in 0.1 M sodium hydroxide (NaOH, Fisher Co. UK) solutionand maintained under magnetic stirring at 37 �C, until a clear solution was obtained.Moreover, the total amount of protein adsorbed onto particles was determined by anindirect method, assessing the protein concentration in the supernatants using theBCA protein assay. The integrity of protein structure, after adsorption or encapsu-lation, was studied by sodium dodecyl sulphate–polyacrylamide gel electrophoresis(SDS–PAGE).

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891 881

2.6. In vitro release studies

Antigen-containing nanospheres were accurately weighed (10 mg), placed inone LoBind� Eppendorf tube per time point and dispersed in 2 ml of a PBS buffer (pH7.4), containing 5% (w/v) of SDS and 0.02% (w/v) of sodium azide. Tubes wereincubated in a shaking water-bath at 37 �C. At predetermined intervals, threeEppendorf tubes were collected, centrifuged at 7500 rpm for 10 min (IEC MicromaxEppendorf centrifuge, UK) and the supernatants were analysed for protein contentand integrity by BCA, Micro BCA� and SDS–PAGE, respectively.

2.7. Structural integrity of S. equi antigens

Dried particles (10 mg) were suspended in 500 ml of a PBS solution, containing10% (w/v) SDS, and incubated for 2 h at 37 �C in an orbital shaker. S. equi proteinsthus extracted from nanospheres, as well as the supernatants obtained after antigenadsorption were loaded onto 10% (w/v) polyacrylamide (Bio-Rad, UK) mini-gel andrun at a constant voltage of 100 V for 120 min using a Bio-Rad 300 Power PackElectrophoresis system (Bio-Rad, Hercules, CA, USA). Protein bands were observedafter staining with SimplyBlue� SafeStain solution (Invitrogen, USA).

2.8. Western blotting

The antigenicity of entrapped and adsorbed S. equi antigens was assessed byWestern blotting [21]. Samples were transferred from acrylamide gel onto the PVDFmembrane using Bio-Rad mini Trans-Blot Electrophoretic Transfer Cell for 1 h. Themembrane was then washed and blocked by its incubation with 10% (w/v) skimmedmilk powder (Merck KGaA, UK) dissolved in PBS containing 0.02% (v/v) of Tween 20(PBST; Sigma Aldrich, Co., UK), for 1 h under constant agitation in a orbital shaker(100 rpm). The membrane was further incubated for 1 h at room temperature withrabbit monoclonal anti-SeM diluted in the blocking buffer (1:200), under constantagitation. After washing, the blot was incubated with a goat anti-rabbit IgG conju-gated to phosphatase alkaline (Sigma Aldrich Co., UK), diluted 1:1000 in blockingbuffer for 1 h at room temperature. The capacity of anti-serum to recognise S. equiantigens was revealed colorimetrically using nitro blue tetrazolinum (NBT) and5-bromo-4-chloro-3-indolyl-1-phosphate (BCIP) (Pierce, USA).

2.9. In vitro cytotoxicity

BALB/c monocyte macrophages (J774A.1 cells line, American Type CultureCollection; ATCC#TIB-67) were used to study the cytotoxicity of PCL nanospheres.Cell viability was evaluated by measuring the reduction of 3-(4,5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide (MTT) by mitochondrial dehydrogenase ofliving cells [22]. Cells were treated with samples, in triplicate, at different concen-trations and incubated for 4 h. MTT reagent was added and, after incubation, thecomplete medium was removed and the absorbance measured at 570 nm usinga Dynex MRX Microplate Reader (Dynex, UK), after crystals dissolution with DMSO.

The relative cell viability related to control wells containing non-treated cellswas determined by the following equation:

% Cell viability ¼ASampales

AControl� 100

where ASamples is the absorbance value obtained for cells treated with differentparticle formulations, while AControl was the reading obtained when cells were justincubated with Dulbecco’s modified Eagle’s medium (DMEM, Sigma Aldrich Co., UK).

2.10. Cellular uptake studies

Bovine serum albumin fluorescein isothiocyanate (FITC-BSA) was used to assessthe binding and uptake of PCL nanospheres by BALB/c monocyte macrophages(J774A.1). To prepare the dye-entrapped nanospheres, FITC-BSA protein was previ-ously dissolved in the internal phase of the first w/o emulsion. The concentration ofthe FITC-BSA was optimised to 2% (w/w) of the polymer weight based on prelimi-nary studies. An efficient loading in the PCL nanospheres was achieved, with noleaching of protein during the incubation in the cell culture medium. Cells werepreviously cultured on a 35 mm glass bottom culture dishes with a cover-slip coatedwith poly-D-lysine (5 � 104 cells/well; P35GC-0-10C, MatTek Corporation, USA) inDMEM (Sigma Aldrich Co., UK) supplemented with 4 mM glutamine (Sigma AldrichCo., UK) plus 10% foetal bovine serum (FBS; Gibco BRL, UK), 100 U/ml penicillin and100 mg/ml streptomycin (Sigma, Poole, Dorset, UK). BALB/c cells were sub-culturedduring the experimental period and frequently checked for viability using Trypanblue exclusion assay. For qualitative uptake studies, the culture medium wasremoved and serum-free DMEM was added to cells. Freeze-dried S. equi loadedparticles (50 mg/ml; 250 mg/ml) suspended in DMEM medium were added to cellsand left for 60 min at 37 �C in 5% CO2 and 90% relative humidity. The dishes wereafterwards observed with a Zeiss LSM 510 confocal laser scanning microscope (CarlZeiss Microscope Systems, Germany) with Zeiss LSM Image Browser� 4.0 softwarefor the uptake and distribution of the particles. Differential interference contrast andfluorescence images were obtained and processed using Adobe Photoshop�

software.

2.11. Immunisation schedule

Twelve groups of female BALB/c mice (25 g; n ¼ 4/group) were immunised byi.n. route on day 1 and boosted on day 21, by instillation using a micropipette tip toadminister 50 ml of sample (25 ml in each nostril) containing S. equi antigen equiv-alent to 10 mg of SeM, either free, encapsulated or adsorbed onto nanospheres (Table1). Before each administration, mice were lightly anesthetised with 3% of isofluranein oxygen (300 cm3 min�1). The formulations were delivered slowly onto nares, sothat the mice could inhale it in. Controls consisted of plain particles and 10 mg of CTB(Sigma Aldrich Co., UK) associated to the free antigen or S. equi loaded nanospheres.All formulations were freshly and aseptically prepared by particles dispersion in PBSpH 7.4 in a biological safety cabinet, immediately prior dosing.

2.12. Collection of samples

Blood samples were collected in heparinised capillary tubes (Vitrex, Denmark)by tail vein bleeding. The blood was incubated overnight at 4 �C and the serum wasseparated by centrifugation at 15,000 rpm for 20 min at 4 �C and stored at �20 �Cuntil further analysis.

In order to assess the mucosal immune response, lung and gut secretions werecollected from ethically sacrificed animal following a method previously reported[23]. Following the culling of animals, the small intestines were aseptically isolated,longitudinally sectioned and transferred into a Petri dish which contained 4 ml ofice-cold enzyme inhibitor mix (1 mM iodoacetic acid, 0.1% (v/v) trypsin inhibitorsoybean type 1 and 10 mM EDTA sodium salt). These were scraped to expose themucus and the liquid was transferred into a centrifuge tube and stored in acid-bath.Next, the tube was sonicated for 30 s to release protein from mucin, and centrifugedat 20,000 � g for 30 min, at 4 �C. The supernatants were removed and centrifugationwas repeated in order to obtain a clear sample, which was then stored at �70 �Covernight and lyophilised (Virtis Advantage freeze dryer, UK) to obtain a dry powder.These samples were then reconstituted with 500 ml of PBS immediately before SIgAquantification by ELISA. To collect lung secretions, the trachea and lungs ofeuthanised animals were exposed and 5 ml of lavage solution (1 mM PMSF, 0.9% (w/v) sodium chloride, 0.5% (v/v) Tween 20 and 0.1% (w/v) sodium azide) were addedinto the trachea to inflate the lungs. The liquid was immediately removed andtreated as aforementioned for gut washes.

2.13. Quantification of systemic (serum) and mucosal antibody responses

The level of anti-S. equi specific IgG, IgG1, IgG2a antibodies in serum and IgAS. equi specific antibodies in lung and gut washes were assessed by indirect ELISAusing Immulon 2 flat-bottom plates (Dynatech, UK) [24]. Each well was coatedovernight at 4 �C, with 1.5 mg/ml S. equi M proteins in PBS solution (pH 7.4). Afterwashing, non-specific protein-binding sites were blocked with 2% (w/v) bovineserum albumin solution (BSA, Sigma Aldrich Co., UK) in PBS and, after 1 h incubationat 37 �C, plates were washed with PBS containing 0.05% of Tween 20 and 0.1% of BSA,fraction V (washing buffer). Serum samples, controls, lung and gut washes wereadded to the ELISA plates at a 32-fold dilution in PBS (100 ml), and serially dilutedtwo-fold with the same buffer. After incubating the plates at 37 �C for 2 h, these werewashed before being incubated for 90 min at 37 �C with horseradish peroxidaseconjugate goat antibody (Sigma, Poole, Dorset, UK) against mouse IgG (Serotec, UK)diluted to 1:1000 in PBS, IgG subclass 1 (IgG1; Serotec, UK) and IgG subclass 2a(IgG2a; Serotec, UK) or IgA (Serotec, UK) all diluted to 1:2000. After washing, boundantibodies were detected by a colorimetric reaction developed by 2-azino-bis(3-ethylbenzthiazoline-6-sulphonic acid) [ABTS] (Sigma Aldrich Co., UK) in citratebuffer and optical densities were read at 405 nm.

The serum antibody titres were expressed as the average (mean � s.d., n ¼ 4) ofthe reciprocal of the dilution at which the optical density (OD) at 405 nm was 5%higher than the strongest negative control reading. The IgA titres were presented asthe absorbance value directly obtained from the optical densities resulting from IgAquantification in lung washes.

2.14. Splenocyte culture studies

The spleens were aseptically removed from euthanised animals and a spleencells suspension was prepared as mentioned elsewhere [20]. These cell homoge-nates (100 ml) obtained from the four mice of each group were plated individually in96-well tissue culture plates (Fisher, UK) in triplicate, along with 100 ml of RPMI 1640(Gibco, UK) containing soluble S. equi enzymatic extract at the concentration of2.5 mg/ml. DuoSet� ELISA Development kit (R&D Systems Europe, UK) was used toquantify, according to the manufacturer’s instructions, the interleukin concentra-tions (IL-2, IL-4, IL-6 and interferon (IFN)-g) in the cell supernatants obtained after48 h of splenocyte stimulation with the soluble antigen at 37 �C, in a humidifiedincubator at 5% CO2 environment.

2.15. Statistical analysis

Differences of significance between groups of in vitro and in vivo studies weredetermined by analysis of variance (ANOVA) linear model with SPSS software(Version 13, Microsoft), with significance set at P � 0.05. This analysis was further

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891882

complemented by a multi-comparison LSD post hoc test in order to state thedifference identified between the groups in which ANOVA analysis revealeda P � 0.05.

3. Results and discussion

3.1. PCL nanosphere physicochemical characteristics

Several studies have demonstrated the potential of PCL micro-particles as a vaccine delivery system which, along with its lowtoxicity, low cost and in vitro stability supported the choice of PCL asthe polymer used for the formulation of these delivery systems.Moreover, PCL degradation does not decrease the pH particles innermicroenvironment, which is of vital importance for the mainte-nance of antigen properties and, therefore, for the induction ofcellular and antibody-mediated responses [25].

Nanospheres presented a spherical shape, a surface free of anypores or cracks (Fig. 1) and a narrow particle size distribution,

Fig. 1. Scanning electron micrographs of the (A) PCL-PVA, (B) PCL

particularly PCL-GCS and PCL-OA (Table 2). PCL particle size rangedfrom 248.4 to 450.2 nm for antigen adsorbed particles and 242.57to 422.83 nm when S. equi antigens were entrapped on PCL nano-spheres. When GCS and SP were added as adjuvants and dissolvedin the internal phase of the first (w/o) emulsion, the size of nano-spheres increased. The smallest particles were formed when PVA,ALG and OA were used as adjuvants. GCS 1% (w/v) solution pre-sented a higher viscosity compared to PVA, which can thus explainthe increase of nanosphere size. On the other hand, the polyamineSP forms relatively low viscosity solutions after dissolution in thePVA 10% (w/v) solution and therefore an increase of particles sizewas not predictable. In fact, PVA is an excellent colloidal polymericstabiliser. As is well accepted, amphiphiles having the property toalign at the droplet surface spontaneously form micelles in anaqueous media, due to intra- and/or intermolecular interactionswith hydrophobic moieties, so as to promote stability by loweringthe free energy at the interface between two phases and resistingcoalescence and flocculation of the emulsion droplets, resulting in

-GCS, (C) PCL-ALG, (D) PCL-SP and (E) PCL-OA nanospheres.

Table 2Particles size, zeta potential and loading capacity (L.C.) of PCL nanospheres(mean � s.d.; n ¼ 6).

Formulation VMDa

(nm)Zeta potential (mV) L.C.d

(% w/w)BPc APc

Adsb PCL-PVA 248.4 � 55.80 �30.7 � 6.90 �32.7 � 7.10 10.75 � 1.12PCL-GCS 450.2 � 32.65 þ38.7 � 7.20 þ5.40 � 7.50 11.38 � 2.50PCL-ALG 287.1 � 21.19 �51.3 � 6.72 �52.30 � 7.60 10.13 � 1.73

Encapsulated PCL-PVA 242.57 � 64.19 �32.8 � 6.20 �34.30 � 6.72 5.19 � 0.98PCL-GCS 422.83 � 14.46 þ31.7 � 7.43 þ21.10 � 7.65 8.42 � 1.81PCL-ALG 264.67 � 69.26 �53.1 � 5. 30 �54.90 � 5.26 5.04 � 1.13PCL-SP 348.37 � 23.36 �27.8 � 6.50 �23.10 � 5.46 4.92 � 0.76PCL-OA 243.40� 10.60 �31.60 � 3.70 �31.73 � 2.52 6.50 � 1.07

a VMD, volume mean diameter.b Ads, adsorbed.c PCL nanospheres surface charge, before (BP) and after (AP) protein association

(mean � s.d., n ¼ 6).d Loading capacity (L.C.) represents the mean percentage of actual amount S. equi

extract proteins content � s.d. (n ¼ 6).

Marker(kDa)

210125101

56

35.8

29

211 2 3 4 5 6 7 8

Marker(kDa)

210

125101

56

35.8

29

211 2 3 4 5 6 7

Fig. 2. (A) SDS–PAGE (10% gel) of S. equi enzymatic extract solutions before and afteradsorption onto PCL (MWT 40 kDa) nanospheres. Lanes: (1) standard molecularweight markers; S. equi proteins extracted from adsorbed particles (2) PCL-PVA, (3)PCL-GCS, (4) PCL-ALG; (5) S. equi enzymatic extract standard solution at 1250 mg/ml;supernatants obtained after particles centrifugation: (6) PCL-PVA; (7) PCL-GCS; (8)PCL-ALG. (B) SDS–PAGE (10% gel) of S. equi enzymatic extract solutions before and afterencapsulation in PCL (MWT 42.5 kDa) nanospheres. Lanes: (1) standard molecularweight marker; (2) S. equi enzymatic extract standard solution at 1250 mg/ml; and S.equi proteins extracted from (3) PCL-PVA, (4) PCL-GCS, (5) PCL-OA, (6) PCL-SP and (7)PCL-ALG nanospheres.

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891 883

smaller particle diameters and polydispersity [26]. The resultantcarriers are composed by a hydrophobic core that can be used asdrugs reservoir. Moreover, nanosphere surface modification by theuse of a hydrophilic polymer such as alginate will improve thestability of the second emulsion, as it limits the exchange betweenthe external phase and the organic phase of the first emulsion.

Nanosphere production was previously optimised using OVA asa model drug (results not shown). Therefore, the amounts andconcentration of polymers and stabilisers used in the (w/o/w)double emulsion solvent evaporation technique were chosen inorder to produce particles that would fit within the nanometricrange, as well as with a low polydispersity index. As a result,despite particle composition, formulations presented a size lowerthan 1 mm, except for PCL-SP nanospheres, which means that theymay be suitable for use by intranasal administration and for theadjuvant effect, as they would be more readily taken up bymacrophages and dendritic cells [27].

For nanospheres formulated with SP and OA, the zeta potentialis slightly more negative compared to the control, non-modifiedPCL nanospheres, whereas it became positive when GCS was useddue to its cationic polymeric nature. PCL-ALG nanospheres pre-sented a highly negative charge which was even higher afterprotein adsorption (Table 2).

CS has been widely used in the pharmaceutical field due to itsunique properties as it is a non-toxic, biodegradable and positivelycharged polysaccharide. However, its low solubility at physiologicalpH constitutes one of its most important limitations. As a result,a biocompatible CS derivative (GCS) was chosen for particleproduction as its good solubility in a broad range of pH aqueoussystems makes possible its inclusion in the internal phase of (w/o)first emulsion. Actually, as already reported in the literature, GCS hasindeed the ability to protect proteins not only during particleformation, but also throughout release in physiological media [28].The increase in viscosity of the inner phase not only reduces theinteraction between the encapsulated proteins and organic phase inthe interface during the homogenisation of the first (w/o) emulsion,but also reduces leakage from the nanospheres, resulting in a higherloading efficiency [28]. In fact, the encapsulation efficiency (E.E.)varied from 49.2% (PCL-SP) to 84.2% (PCL-GCS), with no compromiseof antigen molecular weight by the encapsulation process (Fig. 2B). Itwas observed that GCS substantially increased the loading capacityof S. equi antigens in the nanospheres (Table 2). Besides PCL-SP lowerloading capacity, particle suspension was very easy to disperse inwater evidently due to the high hydrophilicity of the molecule.

All PCL nanospheres loaded by adsorption presented higherloading capacities than those resulting from S. equi antigenencapsulation. The total protein amount adsorbed was 86% (w/w)

onto non-modified PCL-PVA nanoparticles, 91% (w/w) onto PCL-GCS nanospheres and 81% (w/w) onto PCL-ALG, which supports theidea that, for the mixture of proteins studied, both differences inparticle size and charge seemed not to have an influence in theadsorption phenomena.

In S. equi antigen-encapsulated PCL nanospheres, proteins wereexposed to potentially harsh conditions (shear force, contact withsurfactants and the organic solvent DCM) which can potentiatedegradation, causing alteration of its native structure [29]. There-fore, Western blotting (Fig. 3) was used to evaluate antigenicity, andSDS–PAGE (Fig. 2A and B) was undertaken in order to assess proteinintegrity after encapsulation and adsorption in the different PCLnanospheres. SeM is a 58 kDa M-like protein therefore, in Fig. 2Aand B, its band may be seen near the 56 kDa marker [4]. Proteinband patterns of migration obtained for native S. equi antigens andthose extracted from loaded nanospheres are identical, whichsuggests that protein molecular weights were not affected by theentrapment or adsorption techniques. In addition, S. equi epitopesnecessary for immune responses were as well preserved as anantiserum raised against the protein SeM still recognise encapsu-lated and adsorbed S. equi antigens (Fig. 3).

The total amount of protein adsorbed was similar in particleswith different composition, size and charge (Table 2), but the SeMband is stronger in the solution obtained after protein extractionfrom PCL-GCS nanosphere surface, which is in accordance with theloading results (Table 2).

S. equi antigens remained largely linked to the PCL nanospheresurface for 4 h, as until this time point 45% (w/w), 55% (w/w) and

0

20

40

60

80

100

0 5 10 15 20

Pro

tein

cu

mu

lative released

(%

)

Time (day)

Pro

tein

cu

mu

lative released

(%

)

0

20

40

60

80

0 3 6 9 12 15 18Time (day)

PCL-PVA PCL-G

PCL-PVA PCL-SPPCL-GCS

A

B

Pro

tein

cu

mu

lative

released

(%

)

Fig. 4. In vitro cumulative release of S. equi enzymatic extract proteins associated to differen(n ¼ 3, mean � s.d.).

Marker(kDa)

181.8115.5

82.264.248.837.125.919.414.8

6.0 1 2 3 4 5 6 1 7 8 9

Fig. 3. Western blot analysis S. equi antigens extracted from adsorbed and encapsu-lated PCL (MWT 42.5 kDa) nanospheres blotted against anti-SeM antibody as describedin Section 2. Lanes: (1) standard molecular weight marker; S. equi proteins extractedfrom (2) PCL-PVA and (3) PCL-GCS; (4) PCL-OA, (5) PCL-SP and (6) PCL-ALG nano-spheres; S. equi proteins removed from adsorbed (7) PCL-PVA and (8) PCL-GCSnanospheres; (9) S. equi enzymatic extract standard solution at 1250 mg/ml.

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891884

42% (w/w) of the adsorbed antigen remained in the surfacerespectively of PCL-PVA, PCL-GCS and PCL-ALG, under physiologicalconditions (Fig. 4A). Therefore, this system is suitable for nasaldelivery as the antigen associated to particles will not be released toa large extent before loaded PCL nanospheres are taken up by thenasal epithelium and release the antigen in the nasal cavity. CS andALG are bioadhesive polymers, therefore the PCL-GCS and PCL-ALGnanospheres may resist the mucociliary clearance in the nose moreefficiently than the PCL-PVA ones. Even 30 days after particleincubation in PBS (pH 7.4), 22% (w/w) of loaded S. equi proteinremained adsorbed onto PCL-GCS particles surface, while only10.7% (w/w) and 12.2% (w/w) were still adsorbed onto PCL-ALG andPCL-PVA particles, respectively (Fig. 4A). For this reason, in spite ofadsorbing a mixture of proteins, it seems that the positive chargepresented by the GCS particles contributed to a stronger interactionbetween S. equi proteins and PCL particles surface, as it significantlyreleased a lower amount of adsorbed antigen.

All particulate systems containing S. equi antigens presenteda burst release that varied from 26% (w/w) for PCL-GCS to 33% (w/w) for PCL-OA. After this time point, all particles presented a sus-tained release up to 30 days. At this time, 40% (w/w) for PCL-OA to55% (w/w) for PCL-GCS remained entrapped, which is in accor-dance with the degradability of PCL polymer (Fig. 4B).

25 30

21 24 27 30

CS PCL-ALG

PCL-ALGPCL-OA

0

20

40

60

80

0 1 2 3

PCL-PVAPCL-GCSPCL-ALG

Time (day)

t PCL-based nanospheres, by adsorption (A) or encapsulation (B), for a period of 30 days

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891 885

3.2. PCL nanospheres in vitro cell viability assay

The cytotoxicity of PCL nanospheres was assessed by deter-mining the viability of cells by an MTT test on microtitre plates [22].PCL nanospheres showed no cytotoxicity effect on BALB/c cells(J774A.1 cell line) in a concentration up to 2.5 mg/ml. Incubationwith PCL nanospheres led to viability as high as 117% compared tothe control cells, which was significantly higher (P < 0.001) thanthat obtained with the positive control PEI.

3.3. In vitro cellular uptake studies

Particle uptake is determined by several factors, includingparticle size, charge, hydrophobicity, but also by associated adju-vants such as GCS, ALG, SP and OA [16,17]. A positive correlation wasfound for the hydrophobicity and uptake of particles by cells, andthese cells had higher affinity for PCL particles when comparedwith PLA delivery systems. PCL nanospheres used in immunisationstudies have not been extensively reported, which can be due to thedifficulty of producing non-aggregation PCL nanospheres. However,the existing studies suggest promising results of this polymer asa vaccine adjuvant [30,31]. In addition, particle size may determinethe type of immune response induced in animals, as particles witha diameter >5 mm are taken up by Peyer’s patches and remain intheir area, and carriers with inferior diameters were identified notonly in Peyer’s patches but also in mesenteric lymph nodes andspleen, and as a result would preferably induce a systemic immuneresponse, while the first ones will preferentially elicit a mucosalimmunity as they remain in IgA inductive environment [24,32].

As can be seen in Fig. 5, cells tested do not have background inthe fluorescence emission range of FTIC. Presence of green fluo-rescent nanospheres surrounding the nucleus of several cells couldbe found in PCL-PVA, PCL-GCS and PCL-SP nanospheres. It isimportant to stress that differential interference contrast (A, B andD) and fluorescence images (C and E) were obtained at Z of 4 mm, inwhich the nucleus of the cells was focused. These observationssuggest that 1 hwas enough to verify that these PCL-GCS nano-spheres were successfully taken up by cells and therefore canfunction as an effective antigen delivery system. In fact, PCL-ALGnanospheres were only able to attach to cell membrane during the

Fig. 5. Cellular uptake of PCL nanospheres encapsulated with BSA-FITC, by mouse BALB/c mo(E) PCL-OA (n ¼ 3, mean � s.d.).

time of experiment, which can be due to its high negative charge asits size is as small as the ones presented by PCL-PVA and PCL-OA.

3.4. Systemic IgG antibody immune responses

Two i.n. administrations were conducted to confirm theimmunogenicity of the PCL nanospheres loaded with S. equi anti-gens. In addition to the encapsulation technologies believed toenhance the immune response by promoting the antigen uptake byAPC or M cells, immunoadjuvants such as CTB can be used in orderto up-regulate mucosal responses [33]. Therefore, the association ofCTB to non-modified polymeric nanospheres (PCL-PVA) bya mucosal route was expected to improve the local (SIgA) andsystemic (IgG) immune responses.

No S. equi-specific IgG antibodies were detected in serum ofanimals vaccinated with plain particles (data not shown), whichemphasised the utilisation of these PCL polymeric systems as anadjuvant for S. equi antigens, as they have the ability to change S.equi antigen immunogenicity (Fig. 6A and B).

Two weeks after animal vaccination, only adsorbed-PCL nano-sphere groups produced serum anti-S. equi IgG specific titresignificantly higher than those elicited by the free antigen, freeantigen associated with the adjuvant CTB and PCL-PVA encapsu-lated with S. equi antigens and co-administered with CTB (Fig. 6B).PCL-GCSads led to noticeably higher and more prolonged antibodyIgG levels in vaccinated animals, and the difference between thesetitres and those induced and quantified in all other animals vacci-nated with the loaded particulate systems was significant(P < 0.001). By day 28, not only PCL-PVAads and PCL-GCSadsinduced significantly higher IgG systemic levels (P < 0.001), butalso S. equi-encapsulated PCL-ALG and PCL-OA had already sero-converted and their IgG responses were significantly superior(P < 0.02) than those induced by free antigen and freeþ CTB. PCL-PVAads þ CTB markedly enhanced the immunogenicity comparedto free and freeþ CTB groups (P < 0.002) (Fig. 6A and B). PCL hada MW of 42.5 kDa which may justify the slower increase in antibodytitres of animals vaccinated with encapsulated PCL nanospheres, asit presents a slow degradation rate.

The second i.n. immunisation boosted the immune responseinduced by PCL-GCS, as after 7 weeks the IgG titres consequent to

nocyte macrophage cells (J774A.1): (A) PCL-PVA; (B) PCL-GCS; (C) PCL-ALG; (D) PCL-SP;

1

10

100

1000

10000

100000

2 4 7 12

2 4 7 12

PCL-PVA PCL-GCS PCL-ALG PCL-SPPCL-OA PCL-PVA+CTB FREE FREE+CTB

Se

ru

m Ig

G titre

s

Time (weeks)

1

10

100

1000

10000

100000

PCL-PVAads. PCL-GCSads. PCL-ALGads.PCL-PVAads.+CTB FREE FREE+CTB

Time (weeks)

Se

ru

m Ig

G titre

s

A

B

Fig. 6. Serum anti-S. equi specific IgG profile of mice immunised by i.n. route withdifferent formulations of PCL nanospheres (A) encapsulated and (B) adsorbed withS. equi antigens (n ¼ 4, mean � s.d.).

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891886

animal vaccination with these nanospheres, as well as by PCL-ALGand PCL-OA, were significantly higher than the free antigen and thefree antigen þ CTB (P < 0.02 for the first formulations referred andP < 0.008 for the PCL-OA). PCL-PVA and PCL-SP, despite inducingIgG antibody levels statistically higher than those induced by freeantigen (P < 0.05), were not able to present a significant differencewhen compared to the freeþ CTB treated group (Fig. 6A). More-over, in PCL-PVAads and PCL-GCS treated groups it was still possibleto distinguish IgG levels that were statistically different from the S.equi antigens and S. equi antigens associated with CTB. However,PCL-PVAads elicited a serum IgG immune response that, at thisperiod, was statistically higher than that induced by whatevertested formulations, with exception only when this formulationwas associated with the potent adjuvant CTB (Fig. 6B).

Twelve weeks after the beginning of the in vivo study, allformulations enhanced the antibody response in terms of periph-eral IgG levels, and the difference between these responses andthose induced by free antigen and freeþ CTB was statisticallygreater (Fig. 6A and B). It is important to mention that co-admin-istration of CTB with S. equi-encapsulated PCL-PVA nanospheresdid not result in a considerable increase of serum IgG antibodylevels when compared to the PCL-PVA group, and therefore it wasnot identified as a synergic adjuvant effect from the association ofboth adjuvants (P ¼ 0.107) (Fig. 6A). On the contrary, i.n. adminis-tration of CTB admixed with PCL-PVA nanospheres adsorbed withS. equi antigens significantly increased the IgG titres (P ¼ 0.018) to

one and half times the level induced by nanospheres in the absenceof this adjuvant (Fig. 6B). Among the animals vaccinated with PCLnanospheres, the difference between the serum anti-S. equi IgGspecific titre induced by the groups of particles was only significantwhen compared with the PCL-ALGads treated group (P < 0.004).

Over the trial period, serum IgG antibody levels of animalsvaccinated with S. equi antigens encapsulated or adsorbed onto PCLparticles, despite the type of antigen association, was significantlyhigher (P < 0.005) than those elicited by the control (plain parti-cles), free antigens, or even free antigen adjuvanted with CTB, withthe exception of S. equi adsorbed PCL-ALG nanospheres (Fig. 6A andB). The exact reason for this effect remains uncertain but antigenimmobilisation in particles are taken up and processed differently,presenting different properties in terms of uptake and trafficking ascompared with soluble proteins [34]. Despite the lower serum S.equi-specific IgG responses obtained after animal i.n. immunisationwith PCL-ALG particles, this natural polymer is known by its bio-adhesive properties and consequently might enhance a localimmune response to strangles after a nasal immunisation [35].

Even if the second i.n. immunisation generally raised the serumimmune response, the resultant antibody increase was not signif-icant, which confirms that one single immunisation of PCL micro-spheres may be enough to induce elevated immune response ina mouse model. Moreover, the antibody titre generated by i.n.administration of these different PCL nanospheres was significantlyhigher than the ones induced after a single dose of PCL micro-spheres previously studied, which highlights the importance of theroute of administration and particle size in the uptake and subse-quent presentation of the antigen [20]. This makes the PCL nano-spheres especially attractive as a nasal vaccine delivery system as itelicited strong immune responses after a single dose, while mostavailable nasal vaccines need booster vaccination to reach antibodylevels as high as the ones obtained with a parenteral administration[6]. The adjuvant activity of these PCL nanospheres of differentcomposition was found to be comparable as, excepting PCL-ALGads,all loaded PCL nanospheres tested were equally effective, whichhighlights the significant adjuvant effect of these polymeric carriersand the versatility of this approach.

3.5. IgG subtype profiling

As observed for serum specific anti-S. equi IgG immuneresponse, analysis of IgG subclasses in serum collected fromvaccinated mice 2 weeks after the prime dose with S. equi anti-genadsorbed PCL-PVA or PCL-GCS revealed that IgG1 serum titreswere significantly higher than those quantified in all other groups(P < 0.013 for PCL-PVAads and P < 0.001 for PCL-GCSads), theantibody levels induced by the latter formulation in fact beingmarkedly higher than IgG1 antibodies induced by PCL-PVAads(P < 0.001) (Figs. 7A and 8A). In addition, IgG2a immune responsesinduced by PCL-GCSads (P < 0.010) and PCL-ALGads (P < 0.035)nanospheres were statistically different from those induced by anyother vaccinated group (Figs. 7B and 8B). At this point, no efficientspecific IgG1 and IgG2a response against S. equi antigens could bedetected in mice vaccinated with the S. equi antigens encapsulatedin any of the PCL nanospheres. CTB enhanced the development ofIgG1 and IgG2a to S. equi only marginally in mice treated with freeor nanoencapsulated antigens (Fig. 7A and B). Only antigen adsor-bed onto the surface of PCL-GCS and PCL-PVA nanospheres mark-edly enhanced the immunogenicity in animals, eliciting an anti-S.equi IgG1 titre in serum statistically elevated when compared withall experimental groups (P < 0.006) except from that induced byCTB associated to adsorbed PCL-PVA nanospheres (Fig. 8A). Serumfrom mice treated with nanoencapsulated antigens in PCL-GCS andPCL-OA nanospheres contained a higher concentration of anti-S.equi IgG1 and IgG2a than the serum collected from animals

1

10

100

1000

10000

100000

2 4 7 12

PCL-PVAads. PCL-GCSads. PCL-ALGads.PCL-PVAads.+CTB FREE

Time(weeks)

Seru

m Ig

G1 titres

1

10

100

1000

10000

2 4 7 12

PCL-PVAads. PCL-GCSads. PCL-ALGads.PCL-PVAads.+CTB FREE FREE+CTB

FREE+CTB

Time (weeks)

Seru

m Ig

G2a titres

A

B

Fig. 8. Serum anti-S. equi specific (A) IgG1 and (B) IgG2a titres induced after i.n.administration of PCL nanospheres adsorbed with S. equi antigens to female BALB/cmice compared to the free antigen immunisation and free co-administered with CTB(n ¼ 4, mean � s.d.).

1

10

100

1000

10000

100000

2 4 7 12

PCL-PVA PCL-GCS PCL-ALG PCL-SPPCL-OA PCL-PVA+CTB FREE FREE+CTB

Time (weeks)

Se

ru

m Ig

G1

titre

s

1

10

100

1000

10000

2 4 7 12

PCL-PVA PCL-GCS PCL-ALG PCL-SPPCL-OA PCL-PVA+CTB FREE FREE+CTB

Time (weeks)

Seru

m Ig

G2a titres

A

B

Fig. 7. Serum anti-S. equi specific (A) IgG1 and (B) IgG2a titres induced after i.n.administration of PCL nanospheres encapsulated with S. equi antigens to female BALB/cmice compared to the free antigen immunisation and free co-administered with CTB(n ¼ 4, mean � s.d.).

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891 887

vaccinated with antigen in the free form (P < 0.033) (Fig. 7A and B).PCL-PVA adsorbed particles induced IgG2a levels statisticallyhigher than those dosed in the sera of all treated mice (P < 0.01),with the exception of specific anti-S. equi IgG2a assessed in animalsvaccinated with PCL-PVAads þ CTB nanospheres. Conversely, atthis time point, the immunisation of mice with PCL-GCSads nano-spheres induced superior IgG2a antibody titres compared withalginate nanospheres, both associated to the antigen by adsorptionand encapsulation (P < 0.038), and the free antigen, either alone orcombined with the adjuvant CTB (P < 0.05) (Figs. 7B and 8B). Afterboosting, quantification of IgG subtypes indicated a decrease ofIgG1 antibody levels and general elevated titres of IgG2a in animalsvaccinated with adsorbed polymeric particles, while no change inthese antibodylevels was obtained in mice treated with antigenfree or co-administered with CTB adjuvant (Fig. 8A and B). Asa result, PCL-PVAads and PCL-GCSads at this time point inducedIgG1 and IgG2a titres higher than those elicited by the antigen, freeor associated with CTB (P < 0.04 and P < 0.06, respectively). Theseobservations were maintained even 12 weeks after the beginningof the study (Fig. 8A and B). After 7 weeks, particulate systems co-administered with CTB still induced higher IgG2a levels than all theother experimental groups (P < 0.03), while IgG1 antibodies weremuch higher than the ones induced by PVA and ALG formulations(P < 0.027), as well as the FREE and FREE þ CTB groups (P < 0.001)(Figs. 7 and 8). It is important to mention that at the end of the

experiment, PCL-PVAads þ CTB induced IgG1 (P < 0.008) andIgG2a (P < 0.04) levels significantly higher than those induced bythe encapsulated antigens, except for the IgG2a induced by PCL-PVA and PCL-SP. In contrast, PCL-PVA þ CTB nanospheres hadmarkedly elevated the IgG subclass titres relative to the adsorbedparticulate systems (P < 0.041 for IgG2a and P < 0.033 for IgG1),including PCL-PVAads þ CTB. After boosting, encapsulated antigensinduced a continuous increase of both IgG subclass levels (Figs. 7and 8). Despite the serum of all animals treated with this nano-encapsulated systems containing IgG1 and IgG2a levels signifi-cantly elevated in comparison with those assessed in animalsvaccinated with the free form (P < 0.011 and P < 0.06), PCL-GCSnanospheres induced IgG1 antibody levels different from thoseobtained in the serum of mice immunised with PCL-PVA(P < 0.027) and PCL-ALG (P < 0.001), with the antigen associatedeither by adsorption or encapsulation. Anti-S. equi IgG1 assessed inthe serum of animals vaccinated with encapsulated PCL nano-spheres continued to increase even 12 weeks after the animal’simmunisation (Fig. 7A). Therefore, these antibody titres were stillstatistically elevated when compared with the antigen, either free(P < 0.001) or co-adjuvanted by CTB (P < 0.001). On the other hand,IgG2a titres to S. equi in the experimental groups treated withnanoencapsulated antigen in GCS, ALG and OA nanospheres hadrisen to a level that was statistically elevated relative to groups PCL-PVA (P < 0.045) and all groups of mice immunised with adsorbed

0,000

0,200

0,400

0,600

0,800

1,000

PCL-

PVAa

ds.

PCL-

GC

Sads

.PC

L-AL

Gad

sPC

L-PV

Aads

+CTB

PCL-

PVA

PCL-

GC

SPC

L-AL

GPC

L-SP

PCL-

OA

PCL-

PVA+

CTB

Free

Free

+CTB

PCL-

PVAp

lain

PCL-

GC

Spla

inPC

L-Sp

plai

nOp

tic

al d

en

sity

(4

05

n

m)

Fig. 9. Secretory IgA level in lung washes of mice intranasally immunised withdifferent formulations after 12 weeks of booster dose (n ¼ 4, mean � s.d.).

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891888

PCL carriers (P < 0.007), including the PCL-PVAads associated toCTB, and PCL-SP (P < 0.045) (Figs. 7B and 8B). Therefore, despite theimmune responses induced by antigen adsorbed onto PCL nano-spheres being higher during the first 7 weeks of the study, theantibody titres induced by the encapsulated antigen continuouslyincreased even 12 weeks after the immunisation, which supportsthe hypothesis of delayed immune responses induced by theseparticulate carriers being due to the slow release and biodegrada-tion of PCL polymer. As a result, it seems that the best system mayresult from the design of a PCL polymeric carrier containing theantigen adsorbed but also encapsulated, in order to not only inducehigh immune responses just after the prime dose, but also tomaintain the antibody levels for prolonged periods of time.

Overall, mice vaccination with S. equi antigens loaded PCLnanospheres, in the absence of CTB, maintained statisticallyelevated serum IgG subtype titres to S. equi in comparison withanimals treated with the free vaccine (P < 0.04) throughout the 12-week schedule, indicating the generation of Th1/Th2 mixedresponse. Moreover, at the end point of the experiment the ratio ofanti-S. equi IgG subclasses (IgG2a/IgG1) in serum from treatedanimals appeared not to be markedly influenced by the presence orabsence of CTB in the vaccine formulation. Serum of mice immu-nised intranasally with encapsulated or adsorbed nanospherescontaining ALG and GCS contained elevated levels of anti-S. equiIgG1 and IgG2a, suggestive of a more balanced immune response(IgG2a/IgG1 ¼ 0.35 and IgG2a/IgG1 ¼ 0.49, respectively), oppositeto those obtained with the remaining particulate systems. Until thispoint, having in mind the particles’ physicochemical characteristics,the loading capacity and these systemic in vivo results, GCS seemsto be an attractive adjuvant for nasal vaccination of mice againststrangles.

3.6. Local IgA immune responses

The mucosal immunisation used to vaccinate animals in thisstudy is known to be able to induce not only the production of SIgA,but also a systemic antibody and cell-mediated immunity, withlower amounts of antigen [36]. The systemic and local immuneresponses induced after an i.n. vaccination depend on severalfactors, namely the antigen nature, the correct selection of anadjuvant, the dose volume, the frequency of administration and thedelivery systems chosen [37]. Larger doses administered into thenose may be swallowed, leading to oral delivery. Eyles et al. [38]have observed that 40% of the dose was found in the lung after anadministration of 50 ml of labelled microspheres, while most of theparticles were retained in the nose when 10 ml was dosed [38].Moreover, the efficacy of vaccines to prevent strangles seems to bedependent on the induction of SIgA antibodies, as they may block S.equi attachment to the mucosal surfaces and stimulate a cross-protective immunity that has been observed as more efficient thanserum IgG antibody [37].

Soluble antigens need to be associated or delivered by adjuvantsas, after animal i.n. vaccination, they will cross the nasal epithelium,interact with dendritic cells, macrophages and lymphocytes, andthen be transported to posterior lymph nodes. As a result, local anddistant mucosal immune responses can be elicited as a result of thematuration of the precursors of IgA mucosal cells in regional lymphnodes or, after being transported, in the lamina propria of distantmucosal sites of the CMIS [36].

In the present work, the two i.n. immunisations induced elevatedS. equi-specific IgA antibodies in lung washes of all groups of animalstreated with S. equi-loaded PCL nanospheres compared to controls(plain particles, P < 0.002) or antigen, free (P < 0.01) or admixedwith CTB (P < 0.02) (Fig. 9). In fact, modest SIgA levels were detectedin lung washes of animals vaccinated with antigen. In addition, i.n.administrations did not induce detectable SIgA levels in intestinal

contents of animals in all groups (data not shown). PCL-SP group hadthe highest SIgA antibody response, besides not being statisticallydifferent from that assessed in antigens encapsulated in PCL-PVA þ CTB nanospheres (Fig. 9). Among the adsorbed nanospheres,the PCL-ALGads group induced the highest SIgA antibody response,which was significantly different from those induced in all the othergroups (P < 0.01), although PCL-GCSads had also highly enhancedthis antibody level. Among all particulate systems tested, only non-modified PCL-PVA formulation did not induce an extremely elevatedmucosal response after its i.n. administration (Fig. 9). These resultswere expected, as these natural polymers (GCS and ALG) are knownby their bioadhesive properties and consequently would probablyenhance a local immune response to strangles after i.n. immunisa-tion. In addition, PLGA microspheres that were modified with CSwere successfully developed to improve nasal vaccine delivery byJaganathan et al. [14]. The strong mucoadhesiveness and theparticulate nature of these carriers are important properties for itsimmunostimulatory effect. PCL-GCS nanospheres are positivelycharged and therefore are likely to establish electrostatic interac-tions with the negatively charged mucus and thereby prolong theresidence time of the antigen in the nasal cavity [9,14,22]. The factorsthat most probably contributed for the immunostimulating effect ofGCS and ALG based systems on i.n. administration are improveduptake by M cells at the nasal mucosa and efficient delivery to themucosal lymph nodes, a consequence of the prolonged exposure ofthe antigen. Therefore, it is clear from these results that the antigenloaded nanospheres studied here successfully induced not onlya systemic but also a local immune response after i.n. antigenexposure.

3.7. Splenocyte culture studies

It is generally accepted that the way the antigen is delivered toAPCs is one of the major factors vital to elicit the Th1/Th2 relation[9,39]. Cytokines released from cells play an important role in theactivation, growth and differentiation of B and T cells, whichinteract with each other and secrete factors that will stimulateother cell types and eventually themselves. Th1 cells secreteinterleukin-2 (IL-2), interferon-gamma (IFN-g) and tumournecrosis factor-beta (TNF-b), which will increase gene expression inAPC and activate B cells to differentiate and secrete IgG2a, whileTh2 produces IL-4, -5, -6, -10 and -13, IgG1 being the major isotypeelicited from B cells following activation of Th2 cells. Therefore, Th1cells are involved in the cell mediated immunity, while Th2 cells arerelated with humoral immunological response [29].

FREE

+C

TBFREE

PCL-

PVA

+ C

TB

PCL-

OA

PCL-

SP

PCL-

ALG

PCL-

GC

S

PCL-

PVA

PCL-

PVAa

ds.

+CTB

PCL-

ALG

ads.

PCL-

GC

Sads

.

PCL-

PVAa

ds.

700

600

500

400

300

200

100

0

IL

2 c

on

ce

ntra

tio

n (p

g/m

l)

FREE

+C

TBFREE

PCL-

PVA

+ C

TB

PCL-

OA

PCL-

SP

PCL-

ALG

PCL-

GC

S

PCL-

PVA

PCL-

PVAa

ds.

+CTB

PCL-

ALG

ads.

PCL-

GC

Sads

.

PCL-

PVAa

ds.

30

20

10

0

IL

4 co

ncn

etratio

n (p

g/m

l)

A

B

Fig. 10. S. equi-specific recall responses after stimulation of splenocytes derived from BALB/c mice immunised subcutaneously with S. equi enzymatic extract alone (FREE),co-administered with CTB and adsorbed or encapsulated onto PCL nanospheres. Cytokine concentration in culture supernatants was determined in splenocyte culture stimulatedwith 5 mg/ml of soluble S. equi enzymatic extract: (A) IL-2; (B) IL-4; (C) IL-6 and (D) IFN-g (n ¼ 4, mean � s.d.).

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891 889

Recently, whole killed S. equi cells or bacterial lysates encapsu-lated in PLGA microspheres, after i.n. administration in mice,induced protective immunity to a strangles murine model [34]. Asa result, the nasal administration of particles adsorbed or encap-sulated with S. equi antigens was undertaken in order to assess ifPCL, a polymer with higher hydrophobicity than PLGA, associatedwith several adjuvants, can induce an immune response after i.n.administration. From Fig. 10 it can be seen that in general theparticulate systems induced an increase of splenocyte response as

the differences observed in IL-2, IL-4, IL-6 and IFN-g stimulated bynanospheres formulations and free antigen were significant(P < 0.021). Among the antigen adsorbed particles, PCL-PVAads andPCL-GCSads induced elevated IL-2 production after spleen cellstimulation with soluble antigen (5 mg/ml), whose concentrationwas statistically different from those elicited by the remainingexperimental formulations (P < 0.011), except when comparedwith IL-2 levels as a consequence of antigen encapsulated PCL-PVAnanosphere dosing (Fig. 10A). In addition, the differences found in

FREE

+C

TBFREE

PCL-

PVA

+C

TB

PCL-

OA

PCL-

SP

PCL-

ALG

PCL-

GC

S

PCL-

PVA

PCL-

PVAa

ds.

+CTB

PCL-

ALG

ads.

PCL-

GC

Sads

.

PCL-

PVAa

ds.

500

400

300

200

100

0

IL

6 c

on

ce

ntra

tio

n (p

g/m

l)

FREE

+C

TBFREE

PCL-

PVA

+ C

TB

PCL-

OA

PCL-

SP

PCL-

ALG

PCL-

GC

S

PCL-

PVA

PCL-

PVAa

ds.

+CTB

PCL-

ALG

ads.

PCL-

GC

Sads

.

PCL-

PVAa

ds.

2500

2000

1500

1000

500

0

IF

NG

AM

MA

co

ncen

tratio

n (p

g/m

l)

C

D

Fig. 10. (continued).

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891890

the IL-2 amounts are statistically significant for all formulationstested with the exception of the IL-2 induced by PCL-ALGads whencompared with those elicited in animals vaccinated with PCL-GCS(P ¼ 0.061). Even so, PCL-ALG not only elicited the highest IL-2concentration, but also noticeable IFN-g levels, the latter beingstatistically higher than those induced by all other treated groups(P < 0.001) (Fig. 10A and D), which is in accordance with the IgAantibody levels (Fig. 9). Nevertheless, the differences in IFN-g levelsobserved in Fig. 10D are significant for all vaccinated animals(P < 0.04), except for PCL-GCS, in comparison with the onesassessed in PCL-OA and CTB associated with nanosphere treatedgroups. Indeed, both Th1 dependent cytokines and their high levelsare evidence for the strong cell-mediated immune response elicited

by ALG modified PCL-based formulations encapsulated with S. equiantigens administered intranasally. The generation of Th1 cytokineprofile is extremely important in permitting the eradication of S.equi pathogen, being therefore a useful vaccine for therapeuticimmunisation of S. equi long-term carriers. IL-4 production wassignificantly induced by PCL-SP among the particulate systemstested (P < 0.001) (Fig. 10B). Also, CTB admixed with S. equiadsorbed PCL-PVA particles elicited IL-4 amounts significantlyhigher than all formulations except PCL-SP and PCL-PVAads(P ¼ 0.23) (Fig. 10B). All antigen loaded nanospheres induced highlevels of IL-6 but those induced by PCL-GCS were statistically higherthan the IL-6 concentration assessed in all the remaining groups(P < 0.001) (Fig. 10C).

H.F. Florindo et al. / Biomaterials 30 (2009) 879–891 891

Despite the differences mentioned above, modified PCL carriersloaded with S. equi antigens gave higher lymphokine levels whenadministered intranasally when compared with the immunologicalresponses induced by non-modified carriers (PCL-PVA and PCL-PVAads), in accordance with cellular and humoral immuneresponses, compared to the free antigen, successfully activating thepaths leading to Th2 and Th1 cells; this highlights the role of thesenanospheres as an adjuvant and their use in protecting animalsagainst strangles.

Altogether, the data suggest that the quality of immuneresponse to S. equi antigens is increased by its association to the PCLpolymeric carriers produced, mainly to the ones modified by themucoadhesive polymers GCS and ALG, as not only a pronouncedsystemic but also local and mixed cellular immune responses havebeen detected after i.n. vaccination in animals.

4. Conclusions

The purpose of this study was to compare the systemic andmucosal immune responses induced after i.n. vaccination of micewith PCL nanospheres of different compositions, associated toS. equi antigens by adsorption or encapsulation methods. Theseantigens were successfully associated with PCL nanosphereswithout causing any damage in the protein structure, althoughadsorption resulted in higher loading. Even so, both adsorbed andencapsulated nanospheres induced significantly higher specificimmune responses to S. equi antigens, which were confirmed by thecytokine levels quantified. It would be interesting to combine ina future work the adsorption and encapsulation of S. equi antigensin a single PCL nanospheres system in order to obtain a combinedimmune response profile. Moreover, modified PCL nanospheres arepotential nasal delivery systems to vaccinate animals againststrangles as humoral, cellular and mucosal immune responses,important for the defence after bacterial pathogens such as S. equiinvades the host via the mucosal surface, were noticeably induced.

Acknowledgements

The authors are grateful to Dr David McCarthy from The LondonSchool of Pharmacy, UK for the SEM analysis. This work was sup-ported by Fundaçao para a Ciencia e Tecnologia, Ministerio daCiencia e da Tecnologia, Portugal (SFRH/BD/14300/2003, POCI/BIO/59147/2004 and PPCDT/BIO/59147/2004).

References

[1] Waller AS, Jolley KA. Getting a grip on strangles: recent progress towardsimproved diagnostics and vaccines. Vet J 2007;173:492–501.

[2] Sweeney CR, Timoney JF, Newton JR, Hines MT. Streptococcus equi infections inhorses: guidelines for treatment, control, and prevention of strangles. J VetIntern Med 2005;19:123–34.

[3] Sweeney CR. Strangles: Streptococcus equi infection in horses. Equine Vet Educ1996;8:317–22.

[4] Harrington DJ, Sutcliffe I, Chanter N. The molecular basis of Streptococcus equiinfection and disease. Microbes Infect 2002;4:501–10.

[5] Sheoran AS, Artiushin S, Timoney JF. Nasal mucosal immunogenicity for thehorse of a SeM peptide of Streptococcus equi genetically coupled to choleratoxin. Vaccine 2002;20:1653–9.

[6] Amidi M, Romeijn SG, Verhoef CJHE, Bungener LHA, Crommlin DJA, Jiskoot W.N-Trimethyl chitosan (TMC) nanoparticles loaded with influenza subunitantigen for intranasal vaccination: Biological properties and immunogenicityin a mouse model. Vaccine 2007;25:144–53.

[7] Aguilar JC, Rodrıguez EG. Vaccine adjuvants revisited. Vaccine 2007;25:3752–62.

[8] Jaganathan KS, Rao YU, Singh P, Prabakaran D, Gupta S, Jain A, et al.Development of a single dose tetanus toxoid formulation based on polymericmicrospheres: a comparative study of poly(D,L-lactic-co-glycolic acid) versuschitosan microspheres. Int J Pharm 2005;294:23–32.

[9] Almeida AJ, Alpar HO. Mucosal immunisation with antigen-containingmicroparticles. In: Gander B, Merkle HP, Corradi G, editors. Antigen deliverysystems: immunological and technological issues. Amsterdam: HarwoodAcademic Publishers; 1997. p. 207–26.

[10] Bramwell VW, Perrie Y. Particulate delivery systems for vaccines: what can weexpect? J Pharm Pharmacol 2006;58:717–28.

[11] Singh M, O’Hagan DT. Recent advances in vaccine adjuvants. Pharm Res2002;19:715–28.

[12] Lemoine D, Deschuyteneer M, Hogge F, Preat V. Intranasal immunizationagainst influenza virus using polymeric particles. J Biomater Sci Polym Ed1999;10:805–25.

[13] Partidos CD. Intranasal vaccines: forthcoming challenges. PSTT 2000;3:273–81.[14] Jaganathan KS, Vyas SP. Strong systemic and mucosal immune responses to

surface-modified PLGA microspheres containing recombinant hepatitis Bantigen administered intranasally. Vaccine 2006;24:4201–11.

[15] Vila A, Sanchez A, Janes K, Behrens I, Kissel T, Vila Jato JS, et al. Low molecularweight chitosan nanoparticles as new carriers for nasal vaccine delivery inmice. Eur J Pharm Biopharm 2004;57:123–31.

[16] Borges O, Borchard G, Verhoef jC Sousa A, Junginger HE. Preparation of coatednanoparticles for a new mucosal vaccine delivery system. Int J Pharm 2005;299:155–66.

[17] Sugita Y, Takao K, Toyama Y, Shirahatal A. Enhancement of intestinal absorp-tion of macromolecules by spermine in rats. Amino Acids 2007;33:253–60.

[18] Ascarateil S, Dupuis L. Surfactants in vaccine adjuvants: description andperspectives. Vaccine 2006;24S2:S2/83–85.

[19] Gogev S, Fays K, Versali M, Gautier S, Thiry E. Glycol chitosan improves theefficacy of intranasally administered replication defective human adenovirustype 5 expressing glycoprotein D of bovine herpes virus 1. Vaccine 2004;22:1946–53.

[20] Florindo HF, Pandit S, Goncalves LM, Alpar HO, Almeida AJ. S. equi antigensadsorbed onto surface modified poly-caprolactone microspheres inducehumoral and cellular specific immune responses. Vaccine 2008;26:4168–77.

[21] Towbin H, Staehlin T, Gordin J. Electrophoretic transfer of proteins frompolyacrylamide gels to nitrocellulose sheets. Procedure and some applications.Proc Natl Acad Sci USA 1979;76:4350–4.

[22] Mosmann T. Rapid colorimetric assay for cellular growth and survival:application to proliferation and cytotoxicity assays. J Immunol Methods 1983;65:55–63.

[23] Farzad Z, James K, McClelland DBL. Measurement of human and mouse anti-tetanus antibodies and isotype analysis by ELISA. J Immunol Methods1986;87:119–25.

[24] Ying M, Audran R, Thomasin C, Eberl G, Demotz S, Merkle HP, et al. MHC classI- and class II-restricted processing and presentation of microencapsulatedantigens. Vaccine 1999;17:1047–56.

[25] Baras B, Benoit MA, Gillard J. Influence of various technological parameters onthe preparation of spray-dried poly(epsilon-caprolactone) microparticlescontaining a model antigen. J Microencapsul 2000;17:485–98.

[26] Lamprecht A, Ubrich N, Hombreiro PM, Lehr C, Hoffman M, Maincent P.Biodegradable monodispersed nanoparticles prepared by pressure homoge-nization-emulsification. Int J Pharm 1999;184:97–105.

[27] Eldridge JH, Staas JK, Meulbroek JA, Tice TR, Gilley RM. Biodegradable andbiocompatible poly(dl-lactide-co-glycolide) microspheres as an adjuvant forstaphylococcal enterotoxin B toxoid which enhances the level of toxin-neutralizing antibodies. Infect Immun 1991;59:2978–86.

[28] Lee ES, Park K, Park IS, Na K. Glycol chitosan as a stabilizer for proteinencapsulated into poly(lactide-co-glycolide) microparticle. Int J Pharm 2007;338:310–6.

[29] Johansen P, Men Y, Merkle HP, Gander B. Revisiting PLA/PLGA microspheres:an analysis of their potential in parenteral vaccination. Eur J Pharm Biopharm2000;50:129–46.

[30] Benoit MA, Baras B, Gillard J. Preparation and characterization of protein-loaded poly(epsilon-caprolactone) microparticles for oral vaccine delivery. IntJ Pharm 1999;184:73–84.

[31] Singh J, Pandit S, Bramwell VW, Alpar HO. Diphtheria toxoid loaded poly-(epsilon-caprolactone) nanoparticles as mucosal vaccine delivery systems.Methods 2006;38:96–105.

[32] Alpar HO, Almeida AJ. Identification of some physicochemical characteristicsof microspheres which influence the induction of the immune responsefollowing mucosal delivery. Eur J Pharm Biopharm 1994;40:198–202.

[33] Isaka M, Yasuda Y, Taniguchi T, et al. Mucosal and systemic antibody responsesagainst an acellular pertussis vaccine in mice after intranasal co-administra-tion with recombinant cholera toxin B subunit as an adjuvant. Vaccine2003;21:1165–73.

[34] Azevedo AF, Galhardas J, Cunha A, Cruz P, Goncalves LM, Almeida AJ. Micro-encapsulation of Streptococcus equi antigens in biodegradable microspheresand preliminary immunisation studies. Eur J Pharm Biopharm 2006;64(2):131–7.

[35] Singh M, Singh A, Talwar GP. Controlled delivery of diphtheria toxoid usingbiodegradable poly(D,L-lactide) microcapsules. Pharm Res 1991;8:958–61.

[36] Davis SS. Nasal vaccines. Adv Drug Deliv Rev 2001;51:21–42.[37] Wu HY, Russel MW. Nasal lymphoid tissue, intranasal immunization and

compartmentalization of the common mucosal immune system. Immunol Res1997;16:187–201.

[38] Eyles JE, Spiers ID, Williamson ED, Alpar HO. Analysis of local and systemicimmunological responses after intra-tracheal, intra-nasal, intra-muscularadministration of microsphere co-encapsulated Yersinia pestis sub-unit vaccines.Vaccine 1998;16:2000–9.

[39] Szakal AK, Burton GF, Smith JP, Tew JG. Antigen processing and presentation invivo. In: Spriggs DR, Koff WC, editors. Topics in vaccine adjuvant research.Florida: CRC Press; 2001. p. 11–21.