Effect of protein release rates from tablet formulations on the immune response after sublingual immunization

European Journal of Pharmaceutical Sciences 47 (2012) 695–700

Contents lists available at SciVerse ScienceDirect

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

Effect of protein release rates from tablet formulations on the immune response after sublingual immunization Annika Borde a,⇑, Annelie Ekman b, Jan Holmgren b, Anette Larsson a a

Pharmaceutical Technology, Chemical and Biological Engineering, Chalmers University of Technology, SE-41296 Gothenburg, Sweden University of Gothenburg Vaccine Research Institute (GUVAX) and Department of Microbiology and Immunology, Sahlgrenska Academy of University of Gothenburg, SE-40530 Gothenburg, Sweden b

a r t i c l e

i n f o

Article history: Received 20 March 2012 Received in revised form 30 July 2012 Accepted 21 August 2012 Available online 29 August 2012 Keywords: Protein delivery Sublingual immunization Fast release Extended release Hydrophilic matrix tablet Ovalbumin

a b s t r a c t Dry vaccine formulations for sublingual administration would provide great advantages for public health use, especially in developing countries, since they are easy to administer and might also have improved stability properties. This study investigates the influence of protein release rate from mucoadhesive twolayer tablets on the elicited antibody responses after sublingual immunization. Two fast release tablets, one based on a mixture of lactose and microcrystalline cellulose (MCC) and one protein coated ethylcellulose (EC) tablet, and three hydrophilic matrix tablets with extended release (ER) properties based on HPMC 90 SH 100000 or CarbopolÒ 974-P NF were tested. The in vitro release profiles of the model protein ovalbumin (OVA) from these tablets were characterized and correlated to the in vivo potential of the tablets to induce an immune response after sublingual immunization in BALB/c mice. It could be concluded that a tablet with fast protein release elicits antibody titres not significantly different from titres obtained with OVA in solution, whereas low immune responses were observed with a slow release of OVA from the ER formulations. Thus, an ER tablet seems not favorable for vaccine delivery to the sublingual mucosa. Thus, we can present a fast releasing tablet formulation with attractive features for sublingual immunization, whereas the use of ER formulations for sublingual vaccination has to be investigated more in detail. Ó 2012 Elsevier B.V. All rights reserved.

1. Introduction Most vaccines are still administered by injection and provide protection against infections or diseases by initiating systemic immune responses. Due to the possibility of additionally stimulating the mucosal immune system, mucosal immunization has become a widely researched field. The generation of secretory IgA (sIgA) antibodies can stop many microbial pathogens already at the mucosal surface, their main portal of entry (Holmgren and Czerkinsky, 2005). Furthermore, non-invasive mucosal vaccine delivery also provides advantages in terms of health care costs and mass vaccination campaigns in developing countries. The needle-free administration reduces the risk of transmitting infections and specialists are not required to administer the vaccine which provides the possibility of self-administration. Among the mucosal surfaces, the sublingual (s.l.) mucosa has recently attracted interest as an advantageous alternative to, for example, the peroral, buccal or nasal

⇑ Corresponding author. Address: Kemivägen 10, SE-41296 Gothenburg, Sweden. Tel.: +46 (0)31 772 3420; fax: +46 (0)31 772 3418. E-mail addresses: [email protected] (A. Borde), annelie.ekman@micro bio.gu.se (A. Ekman), [email protected] (J. Holmgren), anette.larsson@ chalmers.se (A. Larsson). 0928-0987/$ - see front matter Ó 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.ejps.2012.08.014

route of administration. Contrary to oral vaccination, the stomach’s harsh environment is avoided, and no special delivery devices are required as in the case of nasal vaccination. Compared to the buccal route, the thin, non-keratinized epithelium covering the floor of the mouth and under the tongue is more permeable than the other epithelia of the oral cavity (Mathiowitz et al., 1999). Earlier studies with the model protein ovalbumin (OVA) in combination with the adjuvant cholera toxin (CT) have shown that the s.l. mucosa is an effective site for the simultaneous induction of both systemic and widely disseminated mucosal immune responses (Çuburu et al., 2007). However, in spite of all the advantages of mucosal immunization, further research is required to make the use of such vaccines practically possible. The development of a suitable formulation is one key issue. It would be greatly beneficial to find a vaccine formulation with a needle-free, easy and safe route of administration and without the need for cold chain maintenance. The use of a solid mucoadhesive tablet formulation that can deliver antigens to the sublingual mucosa would combine the advantages of mucosal immunization with those of mucoadhesive drug delivery and solid formulations. A mucoadhesive dosage form can keep the formulation at the administration site, which might facilitate the contact of the antigen to the absorbing mucosa. A solid formulation offers

696

A. Borde et al. / European Journal of Pharmaceutical Sciences 47 (2012) 695–700

great benefits in terms of stability and logistic aspects compared to a liquid formulation, especially in developing countries since much smaller package volumes have to be transported, and no cold chain maintenance is required. Thus, it is not only the discovery of new antigens, but also the development of a suitable vaccine formulation, that is challenging in the design of new vaccines. Compressed tablets are still the dominating solid dosage form on the pharmaceutical market. When developing a tablet formulation, the release rate of the therapeutic agent from the tablet is of importance since it will affect the bioavailability and thus the in vivo efficacy. For some drugs, a fast drug release is required for an immediate pharmacological effect. In other cases, the best effect may be achieved by a solid dosage form with a modified and prolonged release profile. Tablets with an extended release (ER) of the drug can provide several advantages. For therapeutic drugs, an improved patient compliance is obtained due to a reduction of the administration frequency. For vaccines there might be other advantages of an extended release: the antigen can be presented to the immune system over a longer period of time which might lead to a better stimulation and an extremely slow release might even avoid booster immunization. The latter is however more of interest for other administration routes than the sublingual, such as for example the subcutaneous. For sublingual immunization it might be acceptable to keep a tablet under the tongue for some hours. The choice of excipients used in the tablet formulation controls the release rate of the drug/therapeutic agent and the optimal release kinetics are dependent on the drug itself as well as the desired effect. Thus, it is advantageous to be able to follow and study the release of the drug and to correlate it to the in vivo effect. In this study, we wanted to investigate if a prolonged release of antigen from a tablet, and thus a longer exposure of the antigen to the immune system, might have an advantageous effect on the elicited immune response after sublingual immunization, or if a fast release is more effective. For this purpose, tablets with different release rates of the model protein ovalbumin (OVA) were designed. Typical ER formulations are tablets based on gelling hydrophilic polymers, so called hydrophilic matrix tablets. They form a gel layer around the tablet upon contact with water and the overall swelling and erosion of the matrix as well as the properties of the incorporated therapeutic agent control the release profile (Tajarobi et al., 2009). Hydrophilic matrix tablets have been investigated for the delivery of different low molecular weight drugs to a variety of mucosal administration sites (Ìkinci et al., 2004; Kast et al., 2002; Miyazaki et al., 2006), but to our knowledge the delivery of proteins and antigens from this ER principle is a topic that has not been discussed in the literature. Therefore, ER tablets were prepared using two different hydrophilic gelling polymers, hydroxypropyl methylcellulose (HPMC) and CarbopolÒ. Both polymers are also widely investigated as mucoadhesives (Smart, 2005). Since it has previously been shown that an addition of highly soluble mannitol to HPMC tablets increases the rate of water transport into the tablet and thus the dissolution rate of the tablet (Tajarobi et al., 2009), a preparation containing HPMC and mannitol was also investigated in order to design three ER formulations with different protein release rates. All ER formulations were compared with two types of tablets that were expected to show a fast release of OVA. The first fast release tablet was composed of highly water soluble lactose and an addition of microcrystalline cellulose (MCC) as disintegrant. The second fast release tablet was made by dropping a protein solution on a water insoluble ethylcellulose (EC) layer, so that a dried protein layer on top of the tablet surface was obtained. The release profiles of OVA from the different tablet formulations were investigated in vitro and, subsequently, all formulations were tested in sublingual immunization studies in mice with the aim to correlate the release profiles with the in vivo immune responses to OVA.

2. Materials and methods 2.1. Materials 2.1.1. Chemicals The Coomassie Plus (Bradford) Assay Kit containing a Coomassie assay reagent was obtained from Pierce Biotechnology (USA). Sodium dihydrogen phosphate monohydrate and disodium hydrogen phosphate dihydrate from Fluka Chemie GmbH (Switzerland) were used for the preparation of the phosphate buffer solution. Pilocarpine hydrochloride (pilocarpine-HCl) and phosphate buffered saline (PBS, 0.01 M, pH 7.4) were purchased from Sigma–AldrichÒ (Germany). Isoflurane was obtained from IsobaÒ vet. Schering-Plough Animal Health (Sweden). CarbopolÒ 974-P NF (cross-linked polyacrylic acid, CarbopolÒ) was kindly provided by Noveon (USA), HPMC 90SH-100000SR (US Pharmacopeia type 2208, HPMC) by Shin-Etsu Chemical (Tokyo, Japan), and Ethocel™ (ethyl cellulose NF, EC) by The Dow Chemical Company (USA). Mannitol 35Ò (mannitol) was kindly provided by Roquette (France), AvicelÒ (Microcrystalline Cellulose N.F., Ph.Eur., MCC) by FMC Biopolymer (Ireland), and lactose anhydrous for direct tabletting (lactose) by Quest International, Sheffield Products (USA). Tween 20 was purchased from Merck KGaA (Germany). For the development of all enzyme-linked immunosorbent assays (ELISAs) ortho-phenylenediamine (OPD) from Sigma–AldrichÒ (USA) and 30% hydrogen peroxide from Merck KGaA (Germany) were used. Trinatriumcitrate-trihydrate from Merck KGaA (Germany) was used for the preparation of the citrate buffer and sulfuric acid (95–97%, reagent grade) from Scharlau Chemie (Spain) for the preparation of 0.5 M sulfuric acid. All chemicals were of analytical grade and were used as received. 2.1.2. Antigens, adjuvants and antibodies Albumin from chicken egg white (OVA), grade V and VI, was obtained as a crystallized and lyophilized powder from Sigma–AldrichÒ (USA) with 98% purity, determined by gel electrophoresis by the manufacturer. Cholera toxin (CT), azide-free, was purchased from List Biological Laboratories, Inc. (USA). Horseradish peroxidase (HRP)-conjugated goat anti-mouse IgG antibody was obtained from Jackson ImmunoResearch Europe Ltd. (UK). Goat anti-mouse IgA, HRP-conjugated goat anti-mouse IgA, and purified mouse IgA were purchased from SouthernBiotech (USA). 2.1.3. Animals BALB/c mice were purchased from Charles River Breeding Laboratories (Sulzfeld, Germany). For all experiments, 6–8 week old female mice were used. 2.2. Methods 2.2.1. Tablet preparation Initial adhesion tests of pure CarbopolÒ and HPMC tablets to the s.l. mucosa in mice showed that it was almost impossible to adhere any of the tablets to the floor of the mouth. This was interpreted to be due to the slippery mucilage of the mucosa hindering adhesion (Harris and Robinson, 1992). In contrast, CarbopolÒ tablets had a strong tendency to adhere to the ventral side of the tongue. Since the desirable administration site for the formulations was the s.l. mucosa and the adhesion to the floor of the mouth was low, two-layer tablets with a mucoadhesive pure CarbopolÒ layer and a second controlled release layer were prepared for the immunization studies to enhance directed transport to the s.l. mucosa. The mucoadhesive layer could adhere to the ventral side of the tongue so that the tablet was kept in place and OVA was released to the favorable s.l. area of the floor of the mouth.

A. Borde et al. / European Journal of Pharmaceutical Sciences 47 (2012) 695–700 Table 1 Compositions of mucoadhesive tablets containing OVA (grade V) as a model protein. Layer 1 Tablet Tablet Tablet Tablet Tablet

1 2 3 4 5

Layer 2 Ò

Carbopol CarbopolÒ CarbopolÒ CarbopolÒ CarbopolÒ

16.7% OVA, 78.3% lactose, 5% MCC 16.7% OVA, 83.3% HPMC 16.7% OVA, 43.3% HPMC, 40% mannitol 16.7% OVA, 83.3% CarbopolÒ 100% EC, OVA solution (100 lg/ll)

The components for each tablet mixture were weighed using an analytical balance (Sartorius CP224S, Germany), transferred into a mortar, and carefully blended. Mucoadhesive tablets were then prepared with a TA-HDi Texture Analyzer (Stable Micro Systems Ltd., UK) by direct compression at 130 ± 5 MPa using flat-faced dies. The walls of the die were lubricated with magnesium stearate dissolved in ethanol prior to use. For the in vivo immunizations, two-layer tablets with equal weights of each layer were prepared. 1.5 mg CarbopolÒ powder was first compressed into 2 mm tablets, i.e. the first layer. In the second step, 1.5 mg of the blend containing OVA according to Table 1 was compressed onto this layer, resulting in two-layer tablets with a mucoadhesive side and a layer with an OVA dose of 250 lg. For the preparation of tablet 5, a solution of OVA (100 lg/ll) in phosphate buffered saline (PBS) (I = 0.1 M, pH 7.4) was dropped onto the second layer consisting only of EC, and allowed to dry at room temperature. For the release studies, 45 mg of each OVA blend was compressed to 6 mm diameter tablets resulting in tablets with an OVA dose of 7.5 mg. Since the mucoadhesive CarbopolÒ layer is not relevant for the protein release studies, all tablets were prepared without this layer. 2.2.2. Protein release The release of protein from the prepared tablets was measured using a USP dissolution apparatus (Dissolution Tester, Prolabo, France) equipped with a standard USP II paddle. The tablets were fixed in baskets sized 3 cm  2.5 cm  1 cm with a mesh size of 2.5 mm  2.5 mm, which were placed 1 cm above the paddle and 3 cm from the center of the paddle. The release medium, was 500 ml phosphate buffer (I = 0.1 M, pH 7.4) and the temperature was 37 °C. The paddle speed was set to 100 rpm. Aliquots of 1 ml were taken from the release medium at regular predetermined intervals and replaced by the same amount of release medium. All experiments were carried out in triplicates. 2.2.3. Bradford determination of protein concentration The concentration of OVA at each sampling time was determined by the Bradford method (Bradford, 1976) using the Bradford Assay Kit through the Micro Test Tube Protocol (Bradford, 1976). Briefly, the Coomassie Blue (CB) assay reagent was mixed with the sample at a 1:1 volume ratio, incubated for 10 min at room temperature and the absorbance at 595 nm was recorded using a Cintra 40 UV/vis (ultraviolet/visible) spectrophotometer (GBC Scientific Equipment, Australia) with a 100 mM phosphate buffer (pH 7.4) as baseline. The protein concentration was calculated using a standard calibration curve at a wavelength of 595 nm. Since it has previously been shown by Carlsson et al. that CarbopolÒ causes perturbations at the monitoring wavelength (595 nm) of the Bradford method, samples containing CarbopolÒ were also measured at 850 nm and the absorbance for the calculation of the protein concentration was corrected according to their proposed method (Carlsson et al., 2011). All other excipients used in this study, except for HPMC which has proved to leave the Bradford method unaffected, were investigated for possible perturbations on the Bradford absorbance measurements at 595 nm. The

697

spectra of all investigated excipient solutions were found to be virtually identical with pure buffer solution, i.e. there are no detectable effects of any excipient on the CB absorption spectrum under Bradford assay conditions. 2.2.4. Sublingual immunizations Mice were immunized sublingually (s.l.) by two rounds of immunizations. The first dose was given twice on two consecutive days, and the second dose after 15 days. The mice were slightly anaesthetized with isoflurane prior to immunization. Since water transport into hydrophilic matrix tablets has been shown to be an important factor for a strong mucoadhesive effect (Borde et al., 2010; Smart, 1999; Smart et al., 1991; Wilson, 1990), and the mouse’s mouth becomes very dry upon anaesthetization, a dose of 50 lg pilocarpine-HCl (500 lg/ml in PBS) was given subcutaneously (s.c.) to all animals prior to the immunizations in order to stimulate salivation thus providing better swelling and adhesion of the tablets. All tested formulations had a dose of 250 lg OVA and were given in co-administration with 7.5 lg cholera toxin (CT) adjuvant in PBS (1 lg/ll). 10 ll of a solution of OVA and CT in PBS (pH 7.4) with the same doses was given to each mouse in the reference group. The tablets were administered by pressing the uncoated CarbopolÒ side onto the ventral side of the tongue so that the antigen was released to the s.l. mucosa of the floor of the mouth. The animals were maintained for 30 min without food and water after immunization. The retention of the slow release tablets was evaluated in a control group of five mice 30 min and 2 h after administration and it was found that all tablets were gelled and still attached to the tongue. This ensured that our formulations were located at the administration site for at least 2 h. Samples were taken 12 days after the second immunization to evaluate the antibody responses. For blood sampling, mice were anaesthetized, blood was taken from the subclavian vein and sera were separated by centrifugation prior to storage at 20 °C. Immediately after sampling, the mice were euthanized. For the estimation of intestinal IgA responses, intestinal tissue extracts were prepared using a modified version of the perfusion–extraction technique (PERFEXT) (Johansson et al., 1998). Each mouse was perfused with at least 20 ml heparin (0.1% solution of 5000 IU/ml, Lövens Kemiske Fabrik, Denmark) in PBS before removing the small intestine. All tissue samples were stored overnight at 20 °C in 450 ll PBS containing 2 mM PMSF, 0.1 mg/ml STI, and 0.05 mM EDTA, and then thawed and permeabilized by adding saponin (Riedel-de Haën, Germany) to a concentration of 2% (w/ v). After incubation at 4 °C overnight, the samples were centrifuged (15.5 k  g, 10 min, RT) and supernatants were collected and frozen at 20 °C. All animals in this study were housed under specific pathogen free conditions and the experiments were approved by the Ethical Committee for Laboratory Animals in Gothenburg. 2.2.5. Time dependence of CT administration in relation to sublingual immunization with OVA To examine the effect of different time points of CT administration in relation to the administration of OVA the same mouse model and immunization scheme was used. Doses of 100 lg OVA in PBS were given s.l. and 5 lg CT adjuvant was administered at different time intervals ( 3 h, 1 h, ±0 h, +1 h, +3 h). One week after the second round of immunization, samples were taken in the same manner as described above. 2.2.6. Analysis of antibody responses Serum and mucosal anti-OVA antibody responses were determined by enzyme-linked immunosorbent assay (ELISA) according to a method described by Anjuère et al., 2003 with the following

A. Borde et al. / European Journal of Pharmaceutical Sciences 47 (2012) 695–700

modifications: high binding ELISA plates (Greiner, Germany) were coated with OVA grade VI (20 lg/ml in PBS) and kept overnight at 4 °C. After three washes with PBS, all plates were blocked with 0.1% BSA in PBS. Samples and a standard of known activity included on each plate were titrated in 5-fold falling dilutions. Plates were incubated for 90 min at room temperature and then washed twice with 0.05% (v/v) Tween 20 in PBS and once with PBS. Goat antimouse IgG-HRP was added to the plates with serum samples, and goat anti-mouse IgA-HRP to the plates with small intestine extracts. All plates were incubated at 4 °C overnight and finally developed with a solution containing 1 mg/ml OPD and 0.012% hydrogenperoxide in 0.1 M citrate buffer with pH 4.5. The reaction was stopped after 15 min with 0.5 M sulfuric acid. The absorptions of all samples were measured in a plate reader (BioTek Power Wave XS, BioTek Instrumentals, USA) at 490 nm and analyzed using the associated software Gen5, version 1.09 (BioTek Instruments, USA). Intestinal IgA-antibody responses were standardized as ELISA units (EUs) per mg of total IgA according to a method described by Vindurampulle and Attridge (Vindurampulle and Attridge, 2003), but with the following modifications in the total IgA determination: high binding ELISA plates were coated with 1 lg/ml goat anti-mouse IgA in PBS; samples and purified mouse IgA as standard were titrated in 3-fold falling dilutions; and goat anti-mouse IgA-HRP conjugate was used as above. All results were presented as geometric means ± standard error of the mean (GM ± SEM). 2.2.7. Statistical analyses For all statistical analyses the Prism software system GraphPad 5 (GraphPad Software Inc., San Diego) was used. Before calculations, all values were log10 transformed in order to obtain approximate normality and to equalize variances between the populations. Geometric means were calculated and multigroup comparisons were performed using one-way ANOVA with Tukey post-test. Two-sided P-values <0.05 were considered significant. 3. Results and discussion Poor retention on mucosal surfaces is an often mentioned problem in mucosal delivery of proteins, and a mucoadhesive formulation with an extended release might be advantageous in such cases. The aim of this study was therefore to investigate the influence of the release rate of therapeutic proteins or antigens from tablet formulations on the elicited immune responses after s.l. immunization, i.e. whether an extended release and thus a prolonged antigen presentation is advantageous or not. For this purpose, tablet formulations with different release rates were compared. Three gelling hydrophilic matrix tablets with an expected extended release of protein were prepared with OVA as a model protein and compared with two tablets with an expected fast protein release. The release of OVA from all formulations was characterized and then correlated to the systemic and mucosal immune responses in mice after s.l. immunization. 3.1. Characterization of protein release rates Protein release of OVA from five different tablet formulations was investigated using a modified USP II paddle method with a paddle speed of 100 rpm where the tablets were fixed in baskets. The different release profiles of the tested formulations are presented in Fig. 1 and show that the formulations can, as expected, be grouped into two different release types. Both the tablet composed of lactose and MCC and the tablet with an OVA coating on an EC layer showed a fast release; both had released all protein during the first 5 min of the experiment. Both the HPMC tablets

120 100 OVA release (%)

698

80 60 40 20 0 0

10

20 Time (h)

30

Fig. 1. Release of OVA at 100 rpm paddle stirring rate from tablet formulations containing a dose of 7.5 mg OVA. (D) Lactose/MCC, (d) OVA coated EC layer, (X) CarbopolÒ 974 P-NF, (N) HPMC 90 SH 100000 and (h) HPMC 90 SH 100000 and mannitol. Each data point represents the mean (±SD) of three replicate measurements.

and the CarbopolÒ tablet were expected to release the protein during a prolonged period of time, but with different release rates. The HPMC formulation containing 40% mannitol was expected to show the fastest release of the three ER formulations since it has been found that mannitol increases the rate of water transport into HPMC tablets and thus the dissolution rate of the tablet (Tajarobi et al., 2009). This expectation could be confirmed; the HPMC/mannitol tablet had released all protein after 22 h, whereas the HPMC tablet without mannitol had released all protein after 26 h. The CarbopolÒ tablet showed the slowest release rate of all tested formulations; only about 60% OVA was released from this tablet at the time when all protein was released from the HPMC tablet. The protein release data of all three ER tablets were analyzed using the empirical Power-law equation by Ritger and Peppas (Ritger and Peppas, 1987a,b). The fitted n values for the tested formulations in this study were 0.76, 0.82, and 0.89 for the HPMC/ mannitol, HPMC, and CarbopolÒ tablets respectively, and thus indicate that an erosion controlled protein release is the dominating mechanism for all three ER tablets. 3.2. Influence of protein release rate on the immune response in sublingual immunized mice The five tablet formulations with different release rates were tested in vivo as potential delivery systems for s.l. administered vaccines with the aim to evaluate the influence of the in vitro protein release rate on the elicited in vivo immune response. After s.c. administration of pilocarpine-HCl for stimulation of salivation, mice were immunized s.l. with the different formulations together with CT as adjuvant. A reference group received a solution of OVA in PBS. Serum IgG and small intestine IgA antibody responses to OVA were examined using ELISA. The results are shown in Fig. 2 and it can be seen that all investigated formulations were able to elicit both systemic and mucosal immune responses and the antibody profiles were the same for both, which shows that the formulations had no impact on the type of then immune response. The two formulations with fast release were expected to elicit immune responses similar to the reference solution of OVA in PBS. No significant differences in neither serum nor intestinal antibody responses were found between the reference group and the group that was immunized with the lactose/MCC tablet (P > 0.05). However, the group of mice immunized with the OVA coated EC tablet developed significantly lower antibody titres compared to the reference (P < 0.01), but the levels were still higher than those

699

A. Borde et al. / European Journal of Pharmaceutical Sciences 47 (2012) 695–700

(a) 1000000

(b) 100000 **

10000

***

1000

*** ***

100

***

SI standardized anti-OVA IgA titre (EU/mg tot IgA)

Serum anti-OVA IgG titre

ns

100000

10000

1000

**

**

**

***

100

10

H PM C

+

R

H

ef

er

en

ce PM m an C ni t EC Ca ol r bo w it p La h c ol ct oat C ose ing on + tr M ol C C (p re se ra )

ef er en ce H PM H PM C + C m an ni to l C ar EC bo w po ith l co La a ct tin os g e+ M C C

10

R

ns

Fig. 2. Serum and intestinal anti-OVA antibody responses in mice immunized sublingually with ovalbumin (OVA) and cholera toxin as adjuvant. Doses of 250 lg OVA and 7.5 lg CT were given on days 0 + 1 and 16. Five groups of five mice each received tablet formulations. The reference group was immunized with a solution of OVA in PBS. (a) Serum IgG responses and (b) standardized small intestine (SI) IgA responses were measured against OVA by ELISA and are expressed as log10 antibody titres (GM ± SEM). Asterisks denote significant differences (P < 0.01, P < 0.001, ns = non-significant) compared to the reference group.

obtained with the ER tablets. Mice that were immunized with OVA/PBS had surprisingly high antibody titres. All mice that were immunized with ER formulations elicited serum and small intestine antibody responses that were significantly lower than in the reference group (P < 0.01 and P < 0.001). When comparing the three different ER tablets, slight differences in the antibody responses can be seen that correspond to the different release rates of OVA from the tablets. The tablet based on HPMC/mannitol elicited the highest antibody titre, followed by the pure HPMC tablet and the CarbopolÒ tablet. However, these differences were not statistically significant (P > 0.05). The in vivo differences between fast releasing and extended release formulations were not as distinct as they were in vitro. This is not unexpected since one has to take into account that, firstly, no reliable in vitro model has been developed for the prediction of s.l. release yet, and secondly, the dose response correlation is not expected to be linear. From earlier studies in our laboratory it is known that a certain concentration of antigen is required to trigger an immune response. An increase of antigen leads to an increase of the immune response, but this effect levels out at a certain point. In addition, the correlation between in vitro and in vivo erosion is often not 1:1 for gelling tablets even for other administration routes such as the oral route since it is difficult to imitate the in vivo situation. Apart from the lactose/MCC tablet, no tested formulation was able to elicit antibody responses as high as the reference solution

(a)

10 6

SI standardized anti-OVA IgA titre (EU/mg tot IgA)

10 5

10 4

10 3

10 2

(b)

10 5

10 4

10 3

3h +

1h +

h 0

-1 h

3h +

1h +

h 0

-1 h

-3 h

10 2 -3 h

Serum anti-OVA IgG titre

10 6

which to some extent can depend on two possible circumstances. Firstly, the uptake area for protein is decreased when OVA is administered in a tablet formulation. The mucoadhesive CarbopolÒ layer without OVA was pressed onto the ventral side of the tongue and OVA was released unidirectionally and could thus only be taken up via the s.l. mucosa of the floor of the mouth. When OVA is administered in solution the uptake is not restricted to the floor of the mouth but can also occur via the ventral side of the tongue so that a larger area for antigen uptake is available. Secondly, the administration of pilocarpine-HCl for better salivation and thus better swelling and dissolution of the tablets might, in addition to the desired effect also have had a negative effect, especially on the two formulations with fast protein release. The induced salivation was clearly visible and might have caused some released protein as well as adjuvant to slaver out. The poor effect and correlation of release rate and immune response of all ER formulations can depend on several factors that might influence the elicited immune response. We propose two main explanations for this phenomenon. From an immunological point of view, the continuous release of the protein results in a presentation of a very low dose to the immune system each time which might not be sufficient for eliciting a strong response. This could be overcome by increasing the total dose of the antigen, however, if much higher doses should be released each time, the corresponding, several times larger amount of protein might be

Fig. 3. Serum and intestinal anti-OVA antibody responses in mice immunized sublingually with ovalbumin (OVA) and cholera toxin as adjuvant on days 0 + 1 and 16. Five groups of three mice each were immunized with 100 lg OVA and 5 lg CT adjuvant was administered 3 or 1 h before, together with, or 1 or 3 h after the OVA administration ( 3 h, 1 h, ±0 h, +1 h, +3 h). (a) Serum IgG responses and (b) standardized small intestine (SI) IgA responses were measured against OVA using ELISA and are expressed as log10 antibody titres (GM ± SEM). No significant differences were found between the groups.

700

A. Borde et al. / European Journal of Pharmaceutical Sciences 47 (2012) 695–700

impossible to incorporate in a suitable sublingual tablet formulation. In addition to these explanations, it must also be considered that it is difficult to predict the in vivo behavior of the sublingual tablets in mice. Both the amount of salivation and chewing movements will affect the dissolution of the tablets. Although we had tested the retention of the tablets in a control group, there is still a risk that the tablets loosen from the tongue before all protein is released due to mechanical stress or a high degree of swelling after a longer time, so that the tablets spit out or are swallowed and the remaining protein cannot be taken up sublingually. Another aspect that has to be taken into account as an explanation for the poor immune responses to the ER formulations is that the adjuvant CT is not delivered simultaneously since CT was not incorporated in the tablets. Oral immunization studies with keyhole limpet haemocyanin (KLH) by Lycke and Holmgren have shown that a simultaneous administration of CT is essential for an enhanced anti-KLH response (Lycke and Holmgren, 1986) and other experiments in our laboratory have shown that much lower anti-OVA titres are obtained when OVA is given s.l. without CT. Thus, it is thinkable that the time of administration of CT in relation to OVA in sublingual immunizations is critical for the adjuvant effect. A time dependence study was therefore performed, where the time intervals of the CT administration were varied between 3 h before and 3 h after the administration of OVA in PBS. Fig. 3 shows an adjuvant effect of CT both on the systemic and mucosal anti-OVA responses in all groups of mice immunized with OVA and CT, where the highest antibody levels and thus the strongest adjuvant effect were observed in the group of mice that had received OVA and CT simultaneously. Slightly lower antibody titres were obtained in the groups with a time difference between CT and OVA administration, regardless if CT was administered before or after OVA. However, no significant differences were found between the different groups. Thus, the time effect is not regarded as a main explanation for the low antibody responses to the ER formulations. Summarizing, the results of the investigated model system show that a sufficient immune response comparable to protein solution could be achieved with a fast releasing tablet formulation. Since the data from all three ER formulations could not show any benefits in the present study we assume that a slow release is not preferable over a fast release, however, more studies are necessary in order to reliably answer this question. 4. Conclusions This is the first study that investigates the effect of protein release on the elicited in vivo immune responses in sublingually immunized mice. No benefits could be found with the use of an ER formulation with the present model system and the data slightly indicate that a decreasing release rate might lead to decreasing immune responses. This might indicate that a prolonged release of antigen is not favorable in sublingual immunization. Since the used in vivo model might though fail to detect potential benefits of ER formulations, further studies are required to draw a reliable conclusion on this formulation type. The results from the fast releasing tablets show, however, that it is possible to deliver protein from a tablet formulation to the sublingual mucosa and to elicit antibody responses comparable to that of a solution. With the tested formulation composed of highly soluble lactose and the disintegrant MCC both strong systemic and mucosal immune responses were obtained that did not significantly differ from the effect obtained with protein solution. The design of a fast releasing vaccine tablet that could furthermore provide mucoadhesive properties, and thus deliver the antigen locally to the desired administration site, would be greatly

beneficial for mass vaccinations and might also be of interest for sublingual immunotherapy (SLIT). Acknowledgements The authors acknowledge the Chalmers Bioscience Program funded by a strategic initiative of Chalmers University of Technology Foundation and the Swedish Research Council for funding this work. Thanks also to Erik Nygren at the Institute of Biomedicine at Gothenburg University for his help with the statistical analyses. References Anjuère, F., George-Chandy, A., Audant, F., Rousseau, D., Holmgren, J., Czerkinsky, C., 2003. Transcutaneous immunization with cholera toxin B subunit adjuvant suppresses IgE antibody responses via selective induction of th1 immune responses. J. Immun. 170, 1586–1592. Borde, A., Bergstrand, A., Gunnarsson, C., Larsson, A., 2010. Osmotic-driven mass transport of water: impact on the adhesiveness of hydrophilic polymers. J. Colloid Interf. Sci. 341, 255–260. Bradford, M.M., 1976. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein–dye binding. Anal. Biochem. 72, 248–254. Carlsson, N., Borde, A., Wölfel, S., Åkerman, B., Larsson, A., 2011. Quantification of protein concentration by the Bradford method in the presence of pharmaceutical polymers. Anal. Biochem. 411, 116–121. Coomassie Plus (Bradford) Assay Kit, Product No. 23236. Thermo Scientific (Pierce Biotechnology), Rockford, IL. Çuburu, N., Kweon, M.-N., Song, J.-H., Hervouet, C., Luci, C., Sun, J.-B., Hofman, P., Holmgren, J., Anjuère, F., Czerkinsky, C., 2007. Sublingual immunization induces broad-based systemic and mucosal immune responses in mice. Vaccine 25, 8598–8610. Harris, D., Robinson, J.R., 1992. Drug delivery via the mucous membranes of the oral cavity. J. Pharm. Sci. 81, 1–10. Holmgren, J., Czerkinsky, C., 2005. Mucosal immunity and vaccines. Nat. Med. Ìkinci, G., Senel, S., Wilson, C.G., Sumnu, M., 2004. Development of a buccal bioadhesive nicotine tablet formulation for smoking cessation. Int. J. Pharm. 277, 173–178. Johansson, E.-L., Rask, C., Fredriksson, M., Eriksson, K., Czerkinsky, C., Holmgren, J., 1998. Antibodies and antibody-secreting cells in the female genital tract after vaginal or intranasal immunization with cholera toxin B subunit or conjugates. Infect. Immun. 66, 514–520. Kast, C.E., Valenta, C., Leopold, M., Bernkop-Schnürch, A., 2002. Design and in vitro evaluation of a novel bioadhesive vaginal drug delivery system for clotrimazole. J. Controlled Release 81, 347–354. Lycke, N., Holmgren, J., 1986. Strong adjuvant properties of cholera toxin on gut mucosal immune responses to orally presented antigens. Immunology 59, 301– 308. Mathiowitz, E., Chickering, D.E., Lehr, C.-M., 1999. Bioadhesive Drug Delivery Systems: Fundamentals, Novel Approaches and Development. Marcel Dekker, New York. Miyazaki, Y., Tanaka, Y., Yakou, S., Takayama, K., 2006. In vivo drug release from hydrophilic dextran tablets capable of forming polyion complex. J. Controlled Release 114, 47–52. Ritger, P.L., Peppas, N.A., 1987a. A simple equation for description of solute release I. Fickian and non-fickian release from non-swellable devices in the form of slabs, spheres, cylinders or discs. J. Controlled Release 5, 23–36. Ritger, P.L., Peppas, N.A., 1987b. A simple equation for description of solute release II. Fickian and anomalous release from swellable devices. J. Controlled Release 5, 37–42. Smart, J.D., 1999. The role of water movement and polymer hydration in mucoadhesion. In: Mathiowitz, E., Chickering, D.E., Lehr, C.-M. (Eds.), Bioadhesive Drug Delivery Systems: Fundamentals, Novel Approaches and Development. Marcel Dekker, Inc., New York, pp. 11–23. Smart, J.D., 2005. The basics and underlying mechanisms of mucoadhesion. Adv. Drug Deliv. Rev. 57, 1556–1568. Smart, J.D., Carpenter, B., Mortazavi, S.A., 1991. Is mucus dehydration an important factor in mucoadhesion?, in: Kellaway, I.W. (Ed.), Proc. Program Int. Symp. Controlled Release Bioact. Mater., 18th. Controlled Release Society, Inc., pp. 629–630. Tajarobi, F., Abrahmsen-Alami, S., Hansen, M., Larsson, A., 2009. The impact of dose and solubility of additives on the release from HPMC matrix tablets-identifying critical conditions. Pharm. Res. 26, 1496–1503. Vindurampulle, C.J., Attridge, S.R., 2003. Impact of vector priming on the immunogenicity of recombinant salmonella vaccines. Infect. Immun. 71, 287– 297. Wilson, C.G., 1990. In vivo testing of bioadhesion. In: Gurny, R., Junginger, H.E. (Eds.), Bioadhesion-Possibilities and Future Trends. Wissenschaftliche Verlagsgesellschaft, Stuttgart, pp. 93–108.