Journal of Controlled Release 183 (2014) 43–50

Contents lists available at ScienceDirect

Journal of Controlled Release journal homepage: www.elsevier.com/locate/jconrel

Polymeric penetration enhancers promote humoral immune responses to mucosal vaccines Katja Klein, Jamie F.S. Mann, Paul Rogers, Robin J. Shattock ⁎ Imperial College London, Department of Infectious Diseases, Division of Medicine, Norfolk Place, London W2 1PG, UK

a r t i c l e

i n f o

Article history: Received 22 January 2014 Accepted 10 March 2014 Available online 20 March 2014 Keywords: Mucosal Vaccination Penetration Delivery Insulin Immunity

a b s t r a c t Protective mucosal immune responses are thought best induced by trans-mucosal vaccination, providing greater potential to generate potent local immune responses than conventional parenteral vaccination. However, poor trans-mucosal permeability of large macromolecular antigens limits bioavailability to local inductive immune cells. This study explores the utility of polymeric penetration enhancers to promote trans-mucosal bioavailability of insulin, as a biomarker of mucosal absorption, and two vaccine candidates: recombinant HIV-1 envelope glycoprotein (CN54gp140) and tetanus toxoid (TT). Responses to vaccinating antigens were assessed by measurement of serum and the vaginal humoral responses. Polyethyleneimine (PEI), Dimethyl-β-cyclodextrin (DM-βCD) and Chitosan enhanced the bioavailability of insulin following intranasal (IN), sublingual (SL), intravaginal (I.Vag) and intrarectal (IR) administration. The same penetration enhancers also increased antigen-specific IgG and IgA antibody responses to the model vaccine antigens in serum and vaginal secretions following IN and SL application. Co-delivery of both antigens with PEI or Chitosan showed the highest increase in systemic IgG and IgA responses following IN or SL administration. However the highest IgA titres in vaginal secretions were achieved after IN immunisations with PEI and Chitosan. None of the penetration enhancers were able to increase antibody responses to gp140 after I.Vag immunisations, while in contrast PEI and Chitosan were able to induce TT-specific systemic IgG levels following I.Vag administration. In summary, we present supporting data that suggest appropriate co-formulation of vaccine antigens with excipients known to influence mucosal barrier functions can increase the bioavailability of mucosally applied antigens promoting the induction of mucosal and systemic antibody responses. © 2014 Elsevier B.V. All rights reserved.

1. Introduction Trans-epithelial penetration of pharmaceutically relevant drugs has been widely studied. Here pharmaceutical excipients able to influence epithelial penetration have long been utilised to increase trans-dermal drug absorption [1]. However such approaches have not been fully exploited for the delivery of biopharmaceuticals across mucosal surfaces, providing a needle-free delivery strategy that could be easily administered with minimal training. In this respect, trans-mucosal delivery provides an attractive approach for regular repeat dosing of biologics. This may have particular relevance for mucosal vaccination, where gastrointestinal, respiratory and genital mucosal surfaces are the major sites of entry for most pathogens. For this reason elicitation of both systemic and mucosal immune responses is a highly desirable property for preventative vaccines and may increase their capability for controlling infection at the mucosal portals of entry. This is especially ⁎ Corresponding author at: Mucosal Infection & Immunity Group, Section of Infectious Diseases, Imperial College London, St Mary's Campus, London W2 1PG, UK. Tel.: +44 20 7594 5206. E-mail address: [email protected] (R.J. Shattock).

http://dx.doi.org/10.1016/j.jconrel.2014.03.018 0168-3659/© 2014 Elsevier B.V. All rights reserved.

true since it is believed that mucosal immune responses are more efficiently activated after direct mucosal application, compared to parenteral routes of vaccination [2]. This has driven development and commercialisation of a number of vaccines against mucosally associated pathogens [3]. Mucosally delivered vaccines are thought to function by localised imprinting of antigen-activated T and B cells facilitating their re-homing to tissues that are the source of original vaccine/antigen insult. However, a major obstacle to effective trans-mucosal vaccination is thought to be the effectiveness of mucosal epithelial barriers, limiting local bioavailability of recombinant vaccine antigens for sampling by professional antigen presenting cells (APCs) [2]. To test this hypothesis and provide deeper insight into the importance of penetration properties in relation to the efficiency of mucosal vaccination, we evaluate side-by-side the activity of three commonly used penetration enhancers with varying immune activating properties, namely Polyethyleneimine (PEI), Dimethyl-β-cyclodextrin (DM-β-CD) and chitosan glutamate. First we assess their differential ability to alter tight junction integrity in model mucosal epithelial transwell systems. Next to determine in vivo relevance, we evaluate their impact on the bioavailability of insulin, as an in vivo marker of mucosal permeability. Finally we determine their differential impact on antigen-specific systemic and mucosal

44

K. Klein et al. / Journal of Controlled Release 183 (2014) 43–50

humoral immune responses to two model vaccine candidates: recombinant HIV-1 envelope glycoprotein (CN54gp140) and tetanus toxoid (TT). Our data provide important insight as to how these penetration enhancers differentially influence the effectiveness of mucosal vaccination according to route of administration. 2. Materials and methods 2.1. Polycationic compounds and vaccine antigens Polyethyleneimine (MW 25 kDa, branched), Dimethyl-β-cyclodextrin (MW 1.33 kDa) and nonoxynol-9 (Tergitol®NP-9) were obtained from Sigma-Aldrich (UK). Chitosan glutamate (PROTASAN™ UP G 213, MW 200–600 kDa, degree of deacetylation 85%) was obtained from Novamatrix, FMC Biopolymer (Norway). PEI and N-9 were dissolved in sterile water at a concentration of 2% (v/v) while Chitosan and DM-βCD were diluted in sterile water to 2% (w/v) and 10% (w/v) respectively. TT was obtained from the Statens Serum Institute (Denmark). Trimeric recombinant CN54gp140 and demannosylated CN54gp140 were a kind gift from Dietmar Katinger (Polymun Scientific, Austria). 2.2. Assessment of compound mediated cytotoxicity using colorimetric MTT (tetrazolium) assay The MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphethyl tetrazolium bromide) assay was performed as described before [4] with some modifications. A detailed description of the protocol can be found in the Supplementary materials and methods. 2.3. Binding interactions of compounds to gp140 by interferometry Recombinant CN54gp140 at a concentration of 50 μg/ml was immobilised on ForteBio Second-Generation amine reactive biosensor probes (ARG2) using N-hydroxysuccinimide/1-ethyl-3(3dimethylaminopropyl)carbodiimide (NHS/EDC) linkage. Probes were blocked using 5 mg/ml BSA in PBS (pH 7.6) containing 0.01% Tween-20. PEI, Chitosan and monoclonal antibody 5F3 IgG were used at 25 μg/ml, DM-β-CD was used at 100 μg/ml and anti-gp140 polyclonal rabbit serum was diluted to a final concentration of 5 mg/ml of total protein. Samples were loaded into a ForteBio BLITz machine (Fortebio, Inc., Menlo Park, CA) to evaluate binding interactions. Binding is measured as a wavelength shift in nanometers [5]. Analysis was performed using Graphpad Prism (version 4 for Mac). The average rates of change were calculated using the 25% and 75% signal intensities.

(Pharmacia Limited, UK) prior to administration. All animals in the test groups were administered glucose at a dose of 1 g/kg bodyweight intraperitoneally (IP) prior to application. Recombinant human insulin (hINs) (MW 5.8 kDa) (Sigma-Aldrich, UK) was administered at 1 IU/kg bodyweight alone or in combination with penetration enhancers. Penetration enhancers were used at a concentration of 100 μg except from PEI, which was used at 40 μg for nasal and sublingual administration due to toxic side effects at higher concentrations at these routes. One group of animals received only IP PBS to obtain baseline glucose levels. For IN, I.Vag and IR application mice were anaesthetised using isoflurane. For IN application 20 μl of formulation was put onto both nostrils. For I.Vag and IR application 20 μl of formulation was applied into the entrance of the vagina or the rectal cavity respectively. For SL application the mice were heavily anaesthetised using isoflurane. A volume of 15 μl of formulation was delivered with a pipette underneath the tongue. The mice were maintained anaesthetised with their head positioned in ante-flexion for a further 10 min to avoid swallowing. Blood samples were taken to assess glucose levels at 0, 30, 60, 90 and 120 min after formulation administration using a OneTouch®UltraEasy® blood glucose monitor and OneTouch®UltraEasy® Test Strips (LifeScan, UK). At the end of each experiment mice were humanly killed by cervical dislocation. 2.6. Immunisations and sampling Female BALB/c mice were immunised three times at two-week intervals by IN, I.Vag and SL routes as described above. Serum and vaginal wash samples were taken prior to the first immunisation and two weeks after each immunisation. Serum was collected from blood from the tail vein, which was allowed to clot for one hour at room temperature before centrifugation at 400 ×g for 10 min and then frozen until analysis. Vaginal lavage samples were collected by gently flushing 3 times 25 μl of sterile PBS in the vagina using a positive displacement pipette with round-ended tips. Washes were placed on ice and supplemented with 4 μl of protease inhibitor cocktail (Roche Diagnostics, UK) for 30 min before centrifugation at 400 ×g for 20 min and freezing. At the end of each experiment mice were humanely killed by cervical dislocation. 2.7. Antigen-specific IgG, IgG1, IgG2a and IgA ELISA Serum and vaginal samples were analysed for antigen specific IgG, IgG1, IgG2a and IgA using an in-house ELISA protocol. A full description of the methods used is shown in the Supplementary methods. 2.8. Statistical analysis

2.4. Polycationic compound mediated changes in tight junction permeability The influence of PEI, DM-β-CD and Chitosan on tight junction integrity was tested using HEC-1A or Caco-2 cells as model epithelial cell layers in a transwell assay. For full details see Supplementary materials and methods.

Statistical analysis was performed using a Mann–Whitney test with a confidence interval of 95% using Graphpad Prism (version 4 for Mac) software. 3. Results

2.5. In vivo assessment of mucosal permeation using insulin as a model delivery agent

3.1. Impact of polymeric penetration enhancers on mucosal epithelial tight junction integrity

Female BALB/c mice at 6 to 8 weeks of age were obtained from Harlan Laboratories (UK) Ltd. All mouse procedures were performed under the appropriate project licence (PPL 70/6613) in accordance with the UK Home Office Animals (Scientific Procedures) Act 1986 and were ethically approved by the animal ethics committee of St. George's University of London. Mice were maintained in conditions conforming to UK Home Office guidelines to ameliorate suffering and were euthanised by cervical dislocation. The average weight of the mice was 20 g ± 2 g. Prior to application all animals were fasted overnight. Water was provided ad libitum. Mice receiving I.Vag application were treated with 2 mg of DepoProvera

We first sought to determine the effects of the various polymers on epithelial cell viability using a colorimetric MTT assay. After a 4 hour exposure using the model colorectal epithelial cell line Caco-2, 100 μg PEI mediated high levels of cytotoxicity, reducing cell viability by 79% (±4.8%) (Fig. 1A). This contrasted significantly to DM-β-CD, which resulted in a 24.3% (±23.7%) reduction in cell viability, and to Chitosan which reduced cell viability by 10% (± 11.5%). Nonoxynol-9 (N-9), a compound known to display high levels of cytotoxicity, was used as a control and resulted in 83.3% (±2.5%) reduced cell viability (Fig. 1A). Next bilayer interferometry was used to evaluate potential binding interactions with gp140 as a model vaccine antigen. Binding to the

K. Klein et al. / Journal of Controlled Release 183 (2014) 43–50

45

gp140, but with significantly faster dissociation rates than specific antibody. The ability of polymeric compounds to modify epithelial integrity was subsequently assessed using the endometrial cell line HEC-1A and Caco-2 cells in a transwell system. The apical addition of 100 μg PEI to the HEC-1A cells resulted in a dramatic reduction in tight junction integrity as measured by a drop in trans-epithelial resistance (TEER) to 64.3% (±1.9%), within the first 10 min (Fig. 1C). This resistance further decreased to 24.9% (± 3.1%) at 90 min before starting to rise, reaching 94.9% (± 1.7%) after 48 h of incubation. Addition of 100 μg DM-β-CD and 100 μg Chitosan had little impact on TEER in the first 60 min post application. However, after 240 min both DM-β-CD and Chitosan reduced TEER, to 60.4% (±1.5%) and 58.1% (±2.9%) respectively, which then returned to near baseline measurements by 48 h post application (Fig. 1C). Experiments with Caco-2 cells displayed a similar pattern (Fig. 1C). N-9 was used as a positive control for disruption of cell layers. TEER of both Hec-1A and Caco-2 monolayers, was reduced within 10 min of N-9 application. After 40 min, TEER reached 15.1% and 13.7% of baseline for both cell lines respectively and remained at these levels for the remainder of the experiment (48 h later), indicating that cells could not recover after N-9 application (Fig. 1D). Untreated cells maintained constant resistance values throughout the experiments. These results show that PEI, Chitosan and DM-β-CD had the ability to transiently alter epithelial barrier effects with PEI acting rapidly and to a greater effect than Chitosan and DM-β-CD. 3.2. Effect of penetration enhancers on trans-mucosal insulin delivery

Fig. 1. Influence of penetration enhancer on epithelial cells and interaction with vaccine antigens. A) Caco-2 cells were exposed to 100 μg of PEI, DM-β-CD or Chitosan for 4 h before analysis of cell viability (MTT). N-9 was used as a control for cell toxicity, while an untreated control represented 100% viability. The results are shown as mean % cell viability + SD of 2 independent experiments, where each was tested in duplicate. B) Binding interactions of PEI, DM-β-CD or Chitosan to gp140 was assessed using bilayer interferometry. The dotted line indicates the time point at which the condition changed from association to dissociation. Gp140-specific polyclonal sera and 5F3 IgG were used as a binding control. C) Effect of PEI, DM-β-CD and Chitosan on TEER of HEC-1A or Caco-2 cells. Penetration enhancers at a concentration of 100 μg were added and the change in TEER was measured over time. The results are shown as mean % change of TEER relative to the untreated control ± SD from 2 independent experiments, where each condition was tested in duplicate.

biosensor is measured as a shift in signal wavelength (nm) over time (s). Both PEI and Chitosan were able to physically interact with gp140 achieving peak intensities of 2.26 nm and 0.85 nm by 329.6 s and 330 s respectively. There was no detectable interaction between the DM-β-CD and gp140. As assay controls, the gp140 trimer-specific monoclonal antibody 5F3 and anti-gp140 rabbit polyclonal sera were used. Positive binding by the polyclonal sera (7 × 10−3 nm s−1) had a markedly reduced rate of change compared to the PEI (2.33 × 10−1 nm s−1) and Chitosan (1.2 × 10−2 nm s−1) with a peak intensity of 1.91 nm at 330 s, greater in magnitude than that of Chitosan, but below that of PEI (Fig. 1B). Monoclonal 5F3 (25 μg/ml) that binds to a single site on gp140, displayed maximum signal intensity at 329.8 s of 0.44 nm, with a change in signal intensity of 5 × 10−3 nm s−1. Taken together these results show that both PEI and Chitosan physically interact with the vaccine antigen

Modulation of serum glucose levels by human recombinant insulin (hINs) (5.8 kDa) was used as an in vivo biomarker of mucosal permeability. Based on previously published permeability studies PEI [6,7], DM-β-CD [8] and Chitosan [9–11] were screened for their ability to increase penetration of hINs across nasal, vaginal, sublingual and rectal mucosal surfaces in mice. The average baseline blood glucose levels of PBS treated mice were between 5.2 and 6.4 mmol/L throughout the experiments. After IP injection of 1 g glucose/kg bodyweight, mean blood glucose levels significantly increased within the first 30 min to 10.9 ± 0.97 mmol/L, decreased rapidly over the next 30 min to 7.04 mmol/L and then more gradually to near baseline levels, 6.42 ± 0.7 mmol/L, by 90 min. IN application of hINs alone significantly lowered blood glucose levels by 30 min compared to the treatment group receiving glucose alone and was further reduced by 60 min, with a serum glucose level of 4.52 ± 2.5 mmol/L, after which no further significant reduction in blood glucose was detected. Comparison of hINs co-delivered with the different penetration enhancers and administered via the IN route revealed that PEI, DM-β-CD and Chitosan each facilitated a significant drop in blood glucose concentrations compared to IN delivered hINs alone. hINs-PEI significantly reduced serum glucose concentrations by 90 and 120 min post application, while hINs–DM-β-CD and hINs– Chitosan significantly reduced serum glucose at the 30 min sample point (Fig. 2A). SL application of hINs alone had no effect on serum glucose levels relative to the glucose alone group. However, hINs–PEI and hINs–DMβ-CD were able to significantly reduce glucose concentrations between 30 and 90 min post application for PEI and over 30–120 min for DM-βCD. hINs–Chitosan administered via the SL route, failed to significantly reduce blood glucose concentration relative to hINs alone (Fig. 2B). I.Vag delivery of hINs in the absence of penetration enhancers had no impact on serum glucose levels. However, co-delivery of hINs with PEI, DM-β-CD or Chitosan resulted in significantly reduced serum glucose concentrations at 30 min post application for hINs–PEI and hINs–DM-β-CD, and 30–120 min post application for hINs–Chitosan (Fig. 2C). The IR administration of hINs in the absence of penetration enhancers had no impact on serum glucose concentration. Furthermore, while IR administration of hINs in conjunction with PEI, DM-β-CD or

46

K. Klein et al. / Journal of Controlled Release 183 (2014) 43–50

change in total blood glucose levels over time compared to hINs administered alone (Fig. 2E). These data suggest that while the kinetics of uptake can be modified by penetration enhancers, application of hINs alone could induce similar reductions in blood glucose levels over time. However after SL application co-delivery of hINs with PEI or DM-β-CD significant changes in total blood glucose levels were observed, reflective of the poor absorption of hINs alone (Fig. 2E). Likewise for both, I.Vag or IR application of hINs co-delivered with PEI, DM-β-CD or Chitosan total blood glucose levels changed significantly compared to the application of hINs alone (Fig. 2E). Taken together, these results demonstrate that PEI, DM-β-CD and Chitosan were able to increase the penetration of a functional protein across the nasal, sublingual, vaginal and rectal mucosa. However the impact of the different penetration enhancers on insulin uptake varied by route of administration. 3.3. Impact of polymeric penetration enhancers on mucosal vaccination with HIV-1 gp140

Fig. 2. Effect of PEI, DM-β-CD and Chitosan on blood glucose levels after IN, SL, I.Vag or IR administration of insulin. BALB/C mice were fasted overnight before glucose was administered at 1 g/kg bodyweight via the i.p. route. Subsequently 1 IU/kg bodyweight of hINs alone or coformulated with PEI, DM-β-CD or Chitosan was administered via the A) IN, B) SL, C) I.Vag or D) IR route. Blood glucose levels were monitored over time using a blood glucose monitor. Results are shown as mean blood glucose levels in mmol/L ± SEM (n = 5). Stars indicate significant differences between Glc + hINs (green line) and Glc + hINs + penetration enhancer (dark blue line) **p =0.008, *p = 0.016. E) Total change of blood glucose levels over 120 min after IN, SL, I.Vag or IR application of hINs alone or co-formulated with PEI, DM-β-CD or Chitosan. Means ± SEM (n = 5) are shown.

Chitosan significantly reduced serum glucose levels between 30 min and 60 min post application (Fig. 2D), the levels of reduction were modest relative to other routes of administration (IN and SL). Although IN uptake of hINs was promoted by each of the penetration enhancers, analysis of the area under the curve revealed no significant

Next we sought to examine whether penetration enhancers could significantly augment immune responses after topical mucosal vaccination. Polymers were administered to the nasal, sublingual or vaginal mucosa together with the model HIV-1 CN54gp140 env antigen, expressed as a predominate trimer (420 kDa). IR administration was not explored based on the observed poor absorption of insulin by this route. Administration of gp140 alone, to the various surfaces resulted in antigen-specific IgG (Fig. 3A) and IgA (Fig. 3B) levels at, or below, the sensitivity of the assay. The use of penetration enhancers via the IN route resulted in significantly elevated systemic antigen-specific IgG levels: gp140–PEI (58531 ± 18789), gp140–DM-β-CD (9376 ± 3523) and gp140–Chitosan (141303 ± 32736). Similar responses were seen for SL gp140–PEI (43483 ± 10224), gp140–DM-β-CD (5238 ± 1817) and gp140–Chitosan (5693 ± 417). No antigen-specific IgG responses were elicited after co-delivery of gp140 and any of the penetration enhancers via the I.Vag route (Fig. 3A). Gp140–PEI generated greater systemic antigen specific responses compared to gp140–DM-βCD after both IN and SL vaccination and greater than gp140–Chitosan when applied by the SL route (**p = 0.008). There were no significant differences between gp140–PEI and gp140–Chitosan when administered IN (Fig. 3A). Topical IN vaccination with gp140–PEI (2019 ± 223) and gp140– Chitosan (14593 ± 6346), also significantly increased specific systemic IgA responses (Fig. 3B). No significant increase in response was measured for the gp140–DM-β-CD treatment group, which contained non-responders. This contrasted to the SL route of delivery where only gp140–PEI (1150 ± 351) generated significantly elevated IgA responses when compared to the gp140 control group (Fig. 3B). Interestingly, gp140–Chitosan elicited significantly more antigenspecific IgA than both gp140–DM-β-CD and gp140-PEI when delivered via the IN route. By contrast PEI–gp140 generated significantly more specific IgA by the SL route when compared to gp140–DM-β-CD and gp140–Chitosan. As expected, vaccine elicited IgA was below the sensitivity of the assay when penetration enhancers were co-delivered with antigen via the I.Vag route (Fig. 3B). IgG subclass analysis indicated that gp140 administered with PEI, DM-β-CD and Chitosan significantly increased IgG1 and IgG2a titres compared to gp140 alone after IN application (Fig. S1a and b). The IgG1/IgG2a ratio suggests a significant Th2 biased response with all three penetration enhancers compared to gp140 alone (Fig. S1c). SL application of gp140 co-delivered with PEI and Chitosan also significantly increased serum IgG1 but not IgG2a levels (Fig. S1a and b) indicating a significant Th2 biased response (Fig. S1c). As a model mucosal surface relevant to HIV transmission, we also assessed the induction of vaginal antigen-specific IgG (Fig. 3C) and IgA (Fig. 3D) responses after vaccination. We detected no significant elevation in antigen-specific IgG antibody responses to gp140 co-delivered

K. Klein et al. / Journal of Controlled Release 183 (2014) 43–50

47

of gp140–PEI (491.8 ± 143) or gp140–Chitosan (189.2 ± 63.7), relative to antigen alone (Fig. 3D). These were also significantly augmented relative to the gp140–DM-β-CD treatment regime. However, when antigen was delivered via the SL route, only gp140–PEI was able to generate significant IgA antibody responses (Fig. 3d). Vaginal administration of gp140 with PEI, DM-β-CD or Chitosan failed to generate detectable IgA responses in vaginal lavage. These results suggest that both PEI and Chitosan promote antigen specific antibody responses in vagina when delivered via the IN route, but that only PEI was effective at promoting vaginal specific IgA by the SL route. Taken together these data indicate that the utility of the different penetration enhancers in promoting systemic and mucosal responses to gp140 differed by the mucosal route of delivery. 3.4. Impact of the glycan shield of HIV-1 on mucosal vaccination The glycan shield of HIV-1 gp140 env contains numerous mannose residues, previously associated with immunosuppressive activity following parenteral vaccination [12], however their impact on mucosal vaccination has not been fully studied. We hypothesised that such immunosuppressive activity might offer one potential explanation for the observed lack of responsiveness following vaginal vaccination. Therefore we evaluated whether an enzymatically demannosylated form of gp140 (D-gp140) could increase or modulate immunogenicity when delivered mucosally with penetration enhancers. We found that removal of gp140 associated mannose residues had no impact on antigen-specific immune responses in both serum and mucosal compartments (Fig. S2a and b) and did not alleviate the lack of responsiveness following I.Vag vaccination. Furthermore their removal had no discernable impact on the Th2 biasing of IgG1:IgG2 ratios (Fig. S3a) following IN (Fig. S3a) and SL vaccination (Fig. S3b). These data indicate that the glycan shield had minimal impact on the immunogenicity of the gp140 antigen when administered mucosally. 3.5. Mucosally applied tetanus toxoid is immunogenic when co-delivered with penetration enhancers

Fig. 3. Impact of penetration enhancers on serum and vaginal IgG and IgA antibody responses to gp140 following IN, SL or I.Vag administrations. Female BALB/c mice (n = 5) were immunised three times with 10 μg gp140 with or without PEI, DM-β-CD or Chitosan. Serum and vaginal lavage samples were taken 14 days after the last immunisation. Serum gp140-specific A) IgG, B) IgA and vaginal gp140-specific C) IgG and D) IgA reciprocal endpoint titres for each individual mouse (horizontal bars indicate mean values) are shown. Antibody titres smaller than 100 and vaginal antibody titres smaller than 10 are considered negative. Statistical significance was determined using a Mann Whitney U test (**p = 0.008).

with any of the penetration enhancers. This was regardless of the route of administration (Fig. 3C). Interestingly we were able to detect significantly elevated antigen-specific IgA responses following IN application

To determine whether our findings using gp140 co-delivered with penetration enhancers were broadly applicable to other vaccine immunogens we utilised tetanus toxoid (TT) as an alternative model antigen. TT is expressed as a predominate 150 kDa monomer. TT co-delivered IN with PEI (267033 ± 40409), DM-β-CD (9849 ± 10904) and Chitosan (266755 ± 35461) resulted in significantly elevated anti-TT IgG responses (Fig. 4A). This was also evident in co-delivery via the SL route of TT–PEI (151740 ± 25576), TT–DM-β-CD (26234 ± 5625) and TT– Chitosan (48978 ± 12799) respectively. Interestingly, specific IgG responses were significantly elevated on I.Vag delivery of TT–PEI (29188 ± 13929) or TT–Chitosan (35757 ± 32332) (Fig. 4A), this contrasted to the lack of response to I.Vag gp140. The evaluation of vaccine elicited IgA after IN application of the various penetration enhancers revealed that both TT–PEI (19528 ± 5932) and TT–Chitosan (14029 ± 1614) significantly elevated antigen-specific IgA in serum compared to TT alone (Fig. 4B). A similar result was seen after SL delivery with TT–PEI (8324 ± 1835) and TT–Chitosan (2045 ± 528.4) generating significantly elevated responses relative to TT alone (103.4 ± 3.4). No significant TT-specific IgA responses were detected in serum samples after I.Vag administration, regardless of the penetration enhancer used (Fig. 4B). Vaginal responses to TT were significantly augmented following IN administration of TT–PEI and TT–Chitosan with IgG titres of 166.1 (± 45.6) and 132.1 (± 56.2) respectively, and IgA titres of 242.6 (± 92) and 4288 (± 1487) (Fig. 4C and D). Although no consistent antigen-specific IgG responses were detected in vaginal lavage after SL application, TT–PEI (3107 ± 1735), TT–DM-β-CD (166 ± 87) and TT– Chitosan (413 ± 154) generated significantly elevated antigen-specific IgA relative to the TT alone. No significant vaginal specific IgG or IgA

48

K. Klein et al. / Journal of Controlled Release 183 (2014) 43–50

ratio (IgG1:IgG2a) were significantly enhanced for TT–PEI and TT– Chitosan compared to TT alone (Fig. S4c). Significantly higher IgG1 and IgG2a serum titres could also be seen after SL application of TT with PEI, DM-β-CD or Chitosan. Analysed as ratio TT–PEI, TT–DM-βCD or TT–Chitosan generated Th2 biased responses (Fig. S4c). I.Vag administration significantly enhanced mean IgG1 titres for TT–PEI and TT–Chitosan but not for TT–DM-β-CD (Fig. S4a). There was no significant increase in IgG2a titres after I.Vag application (Fig. S4b) and only a significantly elevated Th2 response was observed for TT–PEI (Fig. S4c). 4. Discussion

Fig. 4. TT-specific antibody responses in serum and vaginal lavage following IN, I.Vag or SL applications. Female BALB/c mice (n = 5) were immunised three times with 10 μg TT with or without PEI, DM-β-CD or Chitosan. Serum and vaginal lavage samples were collected two weeks after the last immunisation. A) Serum IgG, B) serum IgA, C) vaginal IgG and D) vaginal IgA reciprocal endpoint titres for each individual mouse (horizontal bars indicate mean values) are shown. Serum antibody titres smaller than 100 and vaginal antibody titres smaller than 10 are considered negative. Statistical significance was determined using a Mann Whitney U test (**p =0.008, *p = 0.016).

were detected after I.Vag application of TT and penetration enhancers (Fig. 4C and D). Serum specific IgG subclass analysis revealed significantly increased IgG1 and IgG2a titres after IN application of TT–PEI, TT–DM-βCD or TT–Chitosan (Fig. S4a and b). TT–PEI, TT–DM-b-CD and TT–Chitosan generated Th2 biased immune responses, which when analysed as a

The effectiveness of mucosal epithelial barriers is widely considered a major obstacle to trans-mucosal vaccination, limiting local bioavailability of recombinant vaccine antigens for sampling by professional antigen presenting cells (APCs) [2]. To test this hypothesis and provide deeper insight into the importance of penetration properties in relation to the efficiency of mucosal vaccination, we evaluated side-by-side the activity of three commonly used penetration enhancers with varying immune activating properties [8–11,13–15]. Initially we assessed the differential impact of the three penetration enhancers on tight junction integrity (assessed by TEER) and cell viability using two mucosal epithelial cell lines (HEC-1A and Caco-2) in a transwell system. PEI demonstrated an immediate effect on transepithelial resistance, while DM-β-CD and Chitosan showed slower and more modest changes. The very rapid impact of PEI on the tight junction integrity of both HEC-1A and Caco-2 cells likely reflects the observed high level of cellular cytotoxicity. This was similar to that of the surfactant N-9, known to disrupt epithelial integrity [16–18]. However, unlike N9, cells were able to recover after the application of PEI, regaining their full barrier function. PEI itself is known to have a high degree of cytotoxicity in vitro [19] but has been shown to be safe when complexed with proteins or plasmids in vivo [20,21]. These data confirmed that all three compounds influenced tight junction integrity of model mucosal epithelial barriers. To confirm the biological relevance of these in vitro observations we went on to assess their differential ability to increase trans-mucosal bioavailability of insulin across nasal, sublingual, vaginal, and rectal mucosa in mice, measuring blood glucose levels as an in vivo biomarker of mucosal permeability. IN application of insulin alone effectively lowered total blood glucose levels where the kinetics of reduction was enhanced by each of the three penetration enhancers. SL administration of insulin alone had minimal impact on blood glucose levels, however co-delivery with PEI or DM-β-CD led to significant reductions that approximated those seen with IN delivery, while Chitosan had no effect. Interestingly Chitosan promoted optimal insulin delivery via the vaginal route with reduced blood glucose levels that also approximated those seen with intranasal delivery, and was superior to hINs–DM-β-CD and hINs–PEI. IR administration of insulin was the least efficient route of administration, and although equally promoted by all three compounds, the impact on blood glucose levels was modest. These data highlight the differential effects of the individual penetration enhancers according to route of administration that were not predicted by in vitro cell models. The influence of DM-β-CD and Chitosan on insulin uptake is in agreement with previous data obtained in rat, sheep and rabbit models [15,22–25]. However the observation that PEI could enhance mucosal delivery of insulin has not previously been reported. Having determined the differential effects of the three compounds on mucosal bioavailability of insulin, used as an in vivo marker of mucosal permeability, we next assessed their impact on mucosal vaccination using two model antigens: HIV CN54gp140 and TT. Administration of either antigen alone by IN, SL or I.Vag routes induced humoral responses that were low or below the level of detection confirming the effectiveness of epithelial barriers in limiting antigen responsiveness in the absence of adjuvant [26–29]. PEI and Chitosan both significantly enhanced systemic and vaginal IgG and IgA responses to gp140 and TT

K. Klein et al. / Journal of Controlled Release 183 (2014) 43–50

after IN administrations. However, DM-β-CD was far less effective in inducing humoral responses to both antigens administered IN. These data are at variance with those observed for insulin delivery where DM-β-CD was equally as effective as PEI and chitosan in promoting reduced blood glucose levels. PEI was also superior with respect to SL immunisation relative to chitosan and DM-β-CD despite chitosan displaying similar potential to PEI for reducing blood glucose levels. The disconnect between changes in permeability and induced immune responsiveness were further highlighted by the lack of response following I.Vag applications of gp140 alone or in combination with penetration enhancers. This was particularly apparent for chitosan which promoted efficient uptake of insulin by the I.Vag route but failed to promote humoral responses to gp140. The observed lack of responsiveness to gp140 concords with that previously seen in mice [30] and humans [31] but contrast to studies performed in rabbits [32,33]. The lack of immune induction with gp140 via the vaginal route may reflect previous findings suggesting the vagina to be a poor site of immune induction [34], able to promote tolerance to applied antigens [35]. We hypothesised that the lack of response to I.Vag immunisation could be due to the high level of mannose residues [36] on gp140, which have been previously shown to exhibit immunosuppressive activity [12]. Furthermore, as mannose binds to C-type lectins on intraepithelial Langerhans cells (LCs), this may have impaired presentation to the humoral immune system [37]. In this respect, approaches to enhance the immunogenicity of HIV-1 Env by deglycosylation have been previously explored by several groups for parenteral immunisation [38–41]. Therefore, in an attempt to improve the bioavailability and immunogenicity of gp140, especially via the I.Vag route we evaluated a deglycosylated version of gp140. However, we observed no enhancement in antibody titre relative to native gp140 when combined with any of the penetration enhancers and administered by IN, SL or I.Vag routes (Fig. S2). In contrast to gp140, I.Vag immunisation with TT co-delivered with PEI or Chitosan induced significant systemic IgG titres and detectable levels of vaginal IgA, although less efficient than IN or SL routes of delivery. The ability to induce specific antibody responses to TT but not to gp140 after I.Vag immunisations is in agreement with a recent study performed in mice using TLR adjuvants [30]. An explanation for these conflicting results could be that TT is a more immunogenic antigen than gp140. Indeed systemic titres of TT specific IgG were higher than those for gp140 irrespective of the route of administration. Furthermore, when molar concentrations of the antigens are taken into consideration there were 2.8 times more TT molecules per immunisation. This could, in theory, stimulate more APCs and therefore enhance immune responses. However related studies performed using up to 50 μg of gp140 for topical vaginal administration in mice were without success (data not shown). Finally the smaller size of TT, predominantly a 150 kDa monomer may have increased the efficiency of trans-mucosal penetration relative to the larger trimeric gp140 protein at 420 kDa. Overall our data suggest that transient enhancement of epithelial permeability does not appear to be the sole determinant of immune responsiveness. Indeed DM-β-CD was the least effective at facilitating mucosal vaccination despite displaying equivalence to PEI and chitosan for intranasal delivery of insulin. Such variance may reflect differences in the interaction of the three biopolymers with the different proteins and/or additional adjuvant properties. Cyclodextrins can form inclusion complexes with drugs [42] that may influence mucosal permeability [43]. Interestingly however, only Chitosan and PEI were observed to physically interact with gp140 by bilayer interferometry (Fig. 1B). Furthermore both PEI and Chitosan have reported adjuvant properties over and above their capacity to modify trans-mucosal uptake [21,44]. In this respect both PEI and Chitosan displayed comparable influence on the induction of humoral responses to gp140 and TT after topical application, where the highest impact on serum and vaginal IgG and IgA was achieved after IN and the least after I.Vag application. These results suggest that the route of immunisation is very important for the generation of potent systemic and vaginal antibody responses. It has been shown that the

49

administration of vaccines to mucosal sites can induce effective immune responses at other mucosal compartments [45,46], via immunological linkage. A series of studies have shown that IN or SL immunisations can give rise to humoral responses in the genital-vaginal tract [29,47–51]. Differences in immune responsiveness to vaginal administration of gp140 and TT suggest that the nature of the antigen provides an additional variable to the efficiency of mucosal vaccination. In conclusion our data suggest that modulation of epithelial permeability alone is not the primary determinant of immune responsiveness to topically applied antigens and is influenced by the intrinsic immune activating potential of each penetration enhancer, the route of administration and the antigen itself. This study supports further investigation of the use of penetration enhancers for promoting mucosal vaccination in non-human primates and humans, where additional gains in immune response are predicted on co-formulation with both vaccine antigens and adjuvants. Acknowledgements Research in this publication was supported by the FP-6-funded EUROPRISE, EC grant LSHP-CT-2006-037611. We gratefully acknowledge an equipment grant from Dormeur Investment Service Ltd that provided funding to purchase equipment used in these studies. The authors would like to thank Prof. Hans Wolf and Prof. Ralf Wagner (GENEART AG, Germany) for access to the CN54gp140 sequence, Dr. Simon Jeffs (Imperial College, London) and Dr. Dietmar Katinger (Polymun Scientific, AUT) for the production of recombinant CN54gp140 and Dr Martina Chang for the generation of demannosylated CN54gp140. Support for the animal work in these studies was provided by the St George's University of London Biological Research Facility. Appendix A. Supplementary data Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.jconrel.2014.03.018. References [1] M.R. Prausnitz, R. Langer, Transdermal drug delivery, Nat. Biotechnol. 26 (11) (2008) 1261–1268. [2] M.R. Neutra, P.A. Kozlowski, Mucosal vaccines: the promise and the challenge, Nat. Rev. Immunol. 6 (2) (2006) 148–158. [3] N. Lycke, Recent progress in mucosal vaccine development: potential and limitations, Nat. Rev. Immunol. 12 (8) (2012) 592–605. [4] T. Mosmann, Rapid colorimetric assay for cellular growth and survival: application to proliferation and cytotoxicity assays, J. Immunol. Methods 65 (1–2) (1983) 55–63. [5] P.J. Carney, A.S. Lipatov, A.S. Monto, R.O. Donis, J. Stevens, Flexible label-free quantitative assay for antibodies to influenza virus hemagglutinins, Clin. Vaccine Immunol. 17 (9) (2010) 1407–1416. [6] I.M. Helander, H.L. Alakomi, K. Latva-Kala, P. Koski, Polyethyleneimine is an effective permeabilizer of gram-negative bacteria, Microbiology 143 (Pt 10) (1997) 3193–3199. [7] I.M. Helander, K. Latva-Kala, K. Lounatmaa, Permeabilizing action of polyethyleneimine on Salmonella typhimurium involves disruption of the outer membrane and interactions with lipopolysaccharide, Microbiology 144 (Pt 2) (1998) 385–390. [8] E. Marttin, J.C. Verhoef, F.W. Merkus, Efficacy, safety and mechanism of cyclodextrins as absorption enhancers in nasal delivery of peptide and protein drugs, J. Drug Target. 6 (1) (1998) 17–36. [9] I. Jabbal-Gill, A.N. Fisher, R. Rappuoli, S.S. Davis, L. Illum, Stimulation of mucosal and systemic antibody responses against Bordetella pertussis filamentous haemagglutinin and recombinant pertussis toxin after nasal administration with chitosan in mice, Vaccine 16 (20) (1998) 2039–2046. [10] E.A. McNeela, D. O'Connor, I. Jabbal-Gill, L. Illum, S.S. Davis, M. Pizza, S. Peppoloni, R. Rappuoli, K.H. Mills, A mucosal vaccine against diphtheria: formulation of cross reacting material (CRM(197)) of diphtheria toxin with chitosan enhances local and systemic antibody and Th2 responses following nasal delivery, Vaccine 19 (9–10) (2000) 1188–1198. [11] R.C. Read, S.C. Naylor, C.W. Potter, J. Bond, I. Jabbal-Gill, A. Fisher, L. Illum, R. Jennings, Effective nasal influenza vaccine delivery using chitosan, Vaccine 23 (35) (2005) 4367–4374. [12] M. Shan, P.J. Klasse, K. Banerjee, A.K. Dey, S.P. Iyer, R. Dionisio, D. Charles, L. Campbell-Gardener, W.C. Olson, R.W. Sanders, J.P. Moore, HIV-1 gp120 mannoses induce immunosuppressive responses from dendritic cells, PLoS Pathog. 3 (11) (2007) e169.

50

K. Klein et al. / Journal of Controlled Release 183 (2014) 43–50

[13] S. Hong, P.R. Leroueil, E.K. Janus, J.L. Peters, M.M. Kober, M.T. Islam, B.G. Orr, J.R. Baker Jr., M.M. Banaszak Holl, Interaction of polycationic polymers with supported lipid bilayers and cells: nanoscale hole formation and enhanced membrane permeability, Bioconjug. Chem. 17 (3) (2006) 728–734. [14] L. Hovgaard, H. Brondsted, Drug delivery studies in Caco-2 monolayers. IV. Absorption enhancer effects of cyclodextrins, Pharm. Res. 12 (9) (1995) 1328–1332. [15] L. Illum, N.F. Farraj, S.S. Davis, Chitosan as a novel nasal delivery system for peptide drugs, Pharm. Res. 11 (8) (1994) 1186–1189. [16] S.L. Hillier, T. Moench, R. Shattock, R. Black, P. Reichelderfer, F. Veronese, In vitro and in vivo: the story of nonoxynol 9, J. Acquir. Immune Defic. Syndr. 39 (1) (2005) 1–8. [17] D.L. Patton, Y.T. Cosgrove Sweeney, L.K. Rabe, S.L. Hillier, Rectal applications of nonoxynol-9 cause tissue disruption in a monkey model, Sex. Transm. Dis. 29 (10) (2002) 581–587. [18] D.L. Patton, G.G. Kidder, Y.C. Sweeney, L.K. Rabe, S.L. Hillier, Effects of multiple applications of benzalkonium chloride and nonoxynol 9 on the vaginal epithelium in the pigtailed macaque (Macaca nemestrina), Am. J. Obstet. Gynecol. 180 (5) (1999) 1080–1087. [19] A.C. Hunter, Molecular hurdles in polyfectin design and mechanistic background to polycation induced cytotoxicity, Adv. Drug Deliv. Rev. 58 (14) (2006) 1523–1531. [20] K. Regnstrom, E.G. Ragnarsson, M. Koping-Hoggard, E. Torstensson, H. Nyblom, P. Artursson, PEI — a potent, but not harmless, mucosal immuno-stimulator of mixed T-helper cell response and FasL-mediated cell death in mice, Gene Ther. 10 (18) (2003) 1575–1583. [21] F. Wegmann, K.H. Gartlan, A.M. Harandi, S.A. Brinckmann, M. Coccia, W.R. Hillson, W.L. Kok, S. Cole, L.P. Ho, T. Lambe, M. Puthia, C. Svanborg, E.M. Scherer, G. Krashias, A. Williams, J.N. Blattman, P.D. Greenberg, R.A. Flavell, A.E. Moghaddam, N.C. Sheppard, Q.J. Sattentau, Polyethyleneimine is a potent mucosal adjuvant for viral glycoprotein antigens, Nat. Biotechnol. 30 (9) (2012) 883–888. [22] C.Y. Cui, W.L. Lu, L. Xiao, S.Q. Zhang, Y.B. Huang, S.L. Li, R.J. Zhang, G.L. Wang, X. Zhang, Q. Zhang, Sublingual delivery of insulin: effects of enhancers on the mucosal lipid fluidity and protein conformation, transport, and in vivo hypoglycemic activity, Biol. Pharm. Bull. 28 (12) (2005) 2279–2288. [23] Z. Degim, T. Degim, F. Acarturk, D. Erdogan, C. Ozogul, M. Koksal, Rectal and vaginal administration of insulin-chitosan formulations: an experimental study in rabbits, J. Drug Target. 13 (10) (2005) 563–572. [24] F.W. Merkus, J.C. Verhoef, S.G. Romeijn, N.G. Schipper, Absorption enhancing effect of cyclodextrins on intranasally administered insulin in rats, Pharm. Res. 8 (5) (1991) 588–592. [25] A.M. Dyer, M. Hinchcliffe, P. Watts, J. Castile, I. Jabbal-Gill, R. Nankervis, A. Smith, L. Illum, Nasal delivery of insulin using novel chitosan based formulations: a comparative study in two animal models between simple chitosan formulations and chitosan nanoparticles, Pharm. Res. 19 (7) (2002) 998–1008. [26] D.I. Albu, A. Jones-Trower, A.M. Woron, K. Stellrecht, C.C. Broder, D.W. Metzger, Intranasal vaccination using interleukin-12 and cholera toxin subunit B as adjuvants to enhance mucosal and systemic immunity to human immunodeficiency virus type 1 glycoproteins, J. Virol. 77 (10) (2003) 5589–5597. [27] G.A. Van Roey, M.A. Arias, J.S. Tregoning, G. Rowe, R.J. Shattock, Thymic stromal lymphopoietin (TSLP) acts as a potent mucosal adjuvant for HIV-1 gp140 vaccination in mice, Eur. J. Immunol. 42 (2) (2012) 353–363. [28] J.F. Mann, P.F. McKay, S. Arokiasamy, R.K. Patel, J.S. Tregoning, R.J. Shattock, Mucosal application of gp140 encoding DNA polyplexes to different tissues results in altered immunological outcomes in mice, PLoS One 8 (6) (2013) e67412. [29] C. Hervouet, C. Luci, N. Cuburu, M. Cremel, S. Bekri, L. Vimeux, C. Maranon, C. Czerkinsky, A. Hosmalin, F. Anjuere, Sublingual immunization with an HIV subunit vaccine induces antibodies and cytotoxic T cells in the mouse female genital tract, Vaccine 28 (34) (2010) 5582–5590. [30] V. Buffa, K. Klein, L. Fischetti, R.J. Shattock, Evaluation of TLR agonists as potential mucosal adjuvants for HIV gp140 and tetanus toxoid in mice, PLoS One 7 (12) (2012) e50529. [31] D.J. Lewis, C.A. Fraser, A.N. Mahmoud, R.C. Wiggins, M. Woodrow, A. Cope, C. Cai, R. Giemza, S.A. Jeffs, M. Manoussaka, T. Cole, M.P. Cranage, R.J. Shattock, C.J. Lacey, Phase I randomised clinical trial of an HIV-1(CN54), Clade C, trimeric envelope vaccine candidate delivered vaginally, PLoS One 6 (9) (2011) e25165. [32] M.P. Cranage, C.A. Fraser, Z. Stevens, J. Huting, M. Chang, S.A. Jeffs, M.S. Seaman, A. Cope, T. Cole, R.J. Shattock, Repeated vaginal administration of trimeric HIV-1

[33]

[34] [35]

[36]

[37]

[38]

[39]

[40]

[41]

[42] [43] [44]

[45] [46]

[47]

[48]

[49]

[50]

[51]

clade C gp140 induces serum and mucosal antibody responses, Mucosal Immunol 3 (1) (2010) 57–68. F. Wegmann, G. Krashias, K. Luhn, K. Laamanen, S. Vieira, S.A. Jeffs, R.J. Shattock, Q.J. Sattentau, A novel strategy for inducing enhanced mucosal HIV-1 antibody responses in an anti-inflammatory environment, PLoS One 6 (1) (2011) e15861. M.W. Russell, Immunization for protection of the reproductive tract: a review, Am. J. Reprod. Immunol. 47 (5) (2002) 265–268. C.A. Black, L.C. Rohan, M. Cost, S.C. Watkins, R. Draviam, S. Alber, R.P. Edwards, Vaginal mucosa serves as an inductive site for tolerance, J. Immunol. 165 (9) (2000) 5077–5083. K.J. Doores, C. Bonomelli, D.J. Harvey, S. Vasiljevic, R.A. Dwek, D.R. Burton, M. Crispin, C.N. Scanlan, Envelope glycans of immunodeficiency virions are almost entirely oligomannose antigens, Proc. Natl. Acad. Sci. U. S. A. 107 (31) (2010) 13800–13805. L. de Witte, A. Nabatov, M. Pion, D. Fluitsma, M.A. de Jong, T. de Gruijl, V. Piguet, Y. van Kooyk, T.B. Geijtenbeek, Langerin is a natural barrier to HIV-1 transmission by Langerhans cells, Nat. Med. 13 (3) (2007) 367–371. K. Banerjee, S. Andjelic, P.J. Klasse, Y. Kang, R.W. Sanders, E. Michael, R.J. Durso, T. J. Ketas, W.C. Olson, J.P. Moore, Enzymatic removal of mannose moieties can increase the immune response to HIV-1 gp120 in vivo, Virology 389 (1–2) (2009) 108–121. M. Koch, M. Pancera, P.D. Kwong, P. Kolchinsky, C. Grundner, L. Wang, W.A. Hendrickson, J. Sodroski, R. Wyatt, Structure-based, targeted deglycosylation of HIV-1 gp120 and effects on neutralization sensitivity and antibody recognition, Virology 313 (2) (2003) 387–400. Y. Li, B. Cleveland, I. Klots, B. Travis, B.A. Richardson, D. Anderson, D. Montefiori, P. Polacino, S.L. Hu, Removal of a single N-linked glycan in human immunodeficiency virus type 1 gp120 results in an enhanced ability to induce neutralizing antibody responses, J. Virol. 82 (2) (2008) 638–651. Y. Wan, L. Liu, L. Wu, X. Huang, L. Ma, J. Xu, Deglycosylation or partial removal of HIV-1 CN54 gp140 V1/V2 domain enhances env-specific T cells, AIDS Res Hum Retroviruses 25 (6) (2009) 607–617. M.E. Davis, M.E. Brewster, Cyclodextrin-based pharmaceutics: past, present and future, Nat. Rev. Drug Discov. 3 (12) (2004) 1023–1035. K. Uekama, Design and evaluation of cyclodextrin-based drug formulation, Chem Pharm Bull (Tokyo) 52 (8) (2004) 900–915. R. Scherliess, S. Buske, K. Young, B. Weber, T. Rades, S. Hook, In vivo evaluation of chitosan as an adjuvant in subcutaneous vaccine formulations, Vaccine 31 (42) (2013) 4812–4819. J. Holmgren, C. Czerkinsky, Mucosal immunity and vaccines, Nat. Med. 11 (Suppl. 4) (2005) S45–S53. A. Kantele, M. Hakkinen, Z. Moldoveanu, A. Lu, E. Savilahti, R.D. Alvarez, S. Michalek, J. Mestecky, Differences in immune responses induced by oral and rectal immunizations with Salmonella typhi Ty21a: evidence for compartmentalization within the common mucosal immune system in humans, Infect. Immun. 66 (12) (1998) 5630–5635. C. Balmelli, R. Roden, A. Potts, J. Schiller, P. De Grandi, D. Nardelli-Haefliger, Nasal immunization of mice with human papillomavirus type 16 virus-like particles elicits neutralizing antibodies in mucosal secretions, J. Virol. 72 (10) (1998) 8220–8229. E.L. Johansson, L. Wassen, J. Holmgren, M. Jertborn, A. Rudin, Nasal and vaginal vaccinations have differential effects on antibody responses in vaginal and cervical secretions in humans, Infect. Immun. 69 (12) (2001) 7481–7486. C. Dupuy, D. Buzoni-Gatel, A. Touze, D. Bout, P. Coursaget, Nasal immunization of mice with human papillomavirus type 16 (HPV-16) virus-like particles or with the HPV-16 L1 gene elicits specific cytotoxic T lymphocytes in vaginal draining lymph nodes, J. Virol. 73 (11) (1999) 9063–9071. W.S. Gallichan, K.L. Rosenthal, Specific secretory immune responses in the female genital tract following intranasal immunization with a recombinant adenovirus expressing glycoprotein B of herpes simplex virus, Vaccine 13 (16) (1995) 1589–1595. K. Imaoka, C.J. Miller, M. Kubota, M.B. McChesney, B. Lohman, M. Yamamoto, K. Fujihashi, K. Someya, M. Honda, J.R. McGhee, H. Kiyono, Nasal immunization of nonhuman primates with simian immunodeficiency virus p55gag and cholera toxin adjuvant induces Th1/Th2 help for virus-specific immune responses in reproductive tissues, J. Immunol. 161 (11) (1998) 5952–5958.