ISSN 0891-4168, Molecular Genetics, Microbiology and Virology, 2009, Vol. 24, No. 1, pp. 2431.

Allerton Press, Inc., 2009. Original Russian Text V.G. Lunin, N.E. Sharapova, T.V. Tikhonova, N.N. Poletaeva, Z.M. Galushkina, E.I. Aksenova, V.I. Grabko, G.A. Velikodvorskaya, N.V. Lavrova, Yu.V. Ananina, 2009, published in Molekulyarnaya Genetika, Mikrobiologiya i Virusologiya, 2009, No. 1, pp. 2127.

In vivo Study of the Immunogenic Properties of the Recombinant Cellulose-binding Domain of Anaerocellum thermophilum
V. G. Lunina, N. E. Sharapovaa, T. V. Tikhonovab, N. N. Poletaevaa, Z. M. Galushkinaa, E. I. Aksenovaa, V. I. Grabkoa, G. A. Velikodvorskayac, N. V. Lavrovaa, and Yu. V. Ananinaa
Research Institute of Epidemiology and Microbiology, Russian Academy of Medical Sciences, ul. Gamalei 18, Moscow, 123098 Russia bAll-Russia Institute of Agricultural Biotechnology, Russian Academy of Agricultural Sciences, 127550, Moscow, Russia cInstitute of Molecular Genetics, Russian Academy of Sciences, pl. Kurchatova 2, Moscow, 123182 Russia
Received June 24, 2008
aGamaleya

AbstractDevelopment of novel technologies enabling the one-stage production of preparations of a required degree of purity is one of the most important goals of contemporary medical biotechnology. The incorporation of afnity domains capable of binding to relevant sorbents into recombinant proteins is one of the most promising approaches. The method of one-stage production of proteincellulose injectants without using chemical conjugates is based on the construction of two-component proteins containing a target antigen and a cellulose-binding domain (CBD) capable of spontaneous binding to cellulose-containing sorbents. To study the immunogenic properties of the CBD, as well as the effect of cellulose sorbents on the immune response, a recombinant CBD was obtained and puried using the one-stage method based on the CBD immobilization on amorphous cellulose; the free CBD protein was isolated. A comparative study of the immunogenic properties of the CBD, the CBDcellulose complex and the cellulose sorbent was performed using a rat model. The titers of specic antibodies, as well as the spectrum and concentrations of cytokines in the blood serum of rats induced by the obtained preparations, were determined using the BioPlex suspension array system. The effect of the CBD on the induction of cytokine synthesis observed on injection of the free protein and its immobilized form was studied. It has been demonstrated that the immobilization of CBD on the cellulose sorbent enhances the synthesis of specic antibodies. It was found that cellulose immunosorbent is not immunotolerant and induces the synthesis of the cytokines involved in the regulation of the humoral immune response. DOI: 10.3103/S0891416809010054

Recently, the use of recombinant antigens or their fragments has become the major approach to the development of a new generation of safe and low reactogenic vaccines. However, despite that these subunit vaccines offer signicant advantages over traditional ones, their low immunogenicity is a major problem. This can be attributed to the lack of certain components (bacterial DNA, lipopolysaccharide, and peptidoglycan typical of whole-cell vaccines) which can nonspecically stimulate the immune response [30, 34]. Therefore, various adjuvants are used to enhance the specic activity of subunit vaccines. One of the most effective and widely employed approaches to enhance the immunological potency of recombinant antigens is immobilization of soluble proteins on inorganic or polymeric organic adjuvants [3, 5, 9]. Among inorganic sorbents, aluminum hydroxide, calcium and aluminum phosphates, and other salts are most frequently used [1, 34]. The organic polymers used for these purposes can be arbitrarily divided into several groups, including articial polymers produced by chemical synthesis (i.e., polioksidoniy [16]) and natural ones (in most cases, various polysaccharides). The possibilities for the application of natural polysaccharide sorbents
24

to increase the antigen immunogenicity was rst demonstrated by Prof. A.E. Gurvich, who worked at the Gamaleya Research Institute of Epidemiology and Microbiology, Russian Academy of Medical Sciences from 1961 to 1986. The researchers in his laboratory were the rst to obtain proteincellulose complexes by chemical synthesis and studied their immunogenicity [2, 4, 7]. In a number of works by Gurvich et al., the high efciency of proteincellulose complexes resulting in a maximum enhancement of antigen immunogenicity as compared to the immunogenicity of soluble protein forms was discussed [46, 10]. Cellulose was found to be the most convenient and cheap immunosorbent that exhibited a high capacity and came in a wide variety of forms (amorphous, spherical, and various modications). An additional advantage of cellulose as an immunosorbent is that, in animal tissues, its polysaccharide particles do not split and cause much less damage [5] than, for instance, Freunds adjuvant [23], which is widely used in laboratory experiments for obtaining hyperimmune sera. However, the necessity of using large amounts of the studied antigen and the need for its pretreatment, as well as the complexity of the chemical process of producing anti-

IN VIVO STUDY OF THE IMMUNOGENIC PROPERTIES

25

gensorbent conjugates, are the main shortcomings of Gurvichs method [2, 3, 5, 24]. To overcome these difculties, a unique genetic approach has been developed. This approach allows the isolation and purication of recombinant proteins to be carried out in one stage by their immobilization on cellulose-containing sorbents. This involves the development of two-component recombinant proteins consisting of a target protein and a cellulose-binding domain (CBD) capable of spontaneous binding to cellulose-containing sorbents with a high binding constant [14]. This method allows one to produce proteins immobilized on cellulose-containing sorbents that can be used for various purposes [11, 12]. It should be noted that, when attached to the CBD, the target protein completely retains its conformation [13] and functional properties [12]. However, the immunogenic properties of CBD used as a carrier protein, its effect on the antibody production, and the type of induced immune response have not yet been studied. Moreover, Prof. Gurvich supposed that cellulose is immunotolerant and does not affect the development of humoral immune responses; however, this hypothesis remains unconrmed [5]. The purpose of this work was to conduct a comparative study of the immunogenic properties of the CBD in its free form and in the CBDcellulose complex, as well as to determine the cytokine prole of the immune response using the cellulose immunosorbent. MATERIALS AND METHODS The recombinant plasmid pTT10 [11, 12] containing the CBD gene [14] was used to obtain strains that produce the recombinant CBD protein. For this purpose, the cells of strain E. coli M15 were transformed by electroporation using the above-mentioned plasmid. Electrocompetent E. coli cells were obtained and transformed using a Bio-Rad electroporator (United States) according to the manufacturers instructions. Expression of the cloned gene and measurement of protein solubility in the cells of E. coli M15 [pREP4] were performed according to the protocols of QIAGEN [25, 28]. An analysis of the cell lysates of bacteria producing CBD protein was performed by electrophoresis in a 12% polyacrylamide gel under denaturing conditions using the Laemmli buffer system [29]. To obtain the CBDcellulose preparation, the cells of E. coli M15 [pREP4] transformed with the recombinant pTT10 plasmid were grown and harvested by centrifugation at 3000 g. The proteincellulose complexes were obtained by the method described in [13]. For protein immobilization, a 20% suspension of amorphous cellulose was used. The cellulose suspension was prepared according to the methods described in [2] with our modications. To remove cell lysate, cellulose with bound protein was washed tree times with a tenfold volume of buffer containing 0.5 M of sodium chloride and

0.5% Triton X-100. To obtain the pure CBD solution, the resultant protein was eluted with an 8-M urea solution; the eluted fractions were dialyzed overnight against phosphate-buffer solution (PBS) (pH 7.2) containing the following: NaCl, 140 mM; KCl, 3 mM; Na2HPO4 7H2O, 10 mM; and KH2PO4, 2 mM. To study the cytokine proles of immune response, as well as to determine the titer of specic antibodies, white outbred male rats (250300 g) were used. Preparations of free CBD protein from the CBDcellulose complex and amorphous cellulose suspension were administered subcutaneously into the dorsal region three times every two weeks. To study the cytokine proles, blood samples were taken from the middle caudal vein before the experiment and on days 1, 2, 15, 27, 29, and 58 of the experiment (a total of seven samples for each experimental group). To determine the cytokine concentrations in rat sera, a Rat Cytokine 9-Plex Panel (IL-1, IL-1, IL-2, IL-4, IL-6, IL-10, GM-CSF, IFN, and TNF) and a BioPlex Rat Serum Diluent Kit (Bio-Rad, United States) for multiplex suspension immunoassay were used. The sera were prepared according to the instruction for Rat Cytokine 9-Plex Panel. The blood samples were collected into tubes with anticoagulant (6% EDTA) in a ratio of 1 : 19. The samples were stored at 37C for 1 h. The resultant clots were centrifuged at 500 g and 4C for 15 min. The serum was then transferred to a clean tube, aliquoted into 5070 l portions, and stored frozen at 40C until analyzed. The cytokine concentrations in rat sera were determined according to the instructions for the BioPlex suspension array system and the cytokine assay kit (BioRad, United States). Before the analysis, the serum aliquots were thawed on ice and diluted twofold with sample diluent. Fourfold dilutions of cytokine standards were prepared. The wells of polystyrene plates with wet lter paper were supplemented with 50-l aliquots of the microsphere suspension (2000 microspheres per a plate well) conjugated with cytokine antibodies and with 50-l portions of prepared standards and samples. The plates were then incubated in the dark with agitation (300 rpm) at room temperature for 30 min. The microspheres were washed three times with a special buffer solution using a vacuum pump. The plate wells were then supplemented with 25-l portions of a fresh solution of biotinylated antibodies and incubated in the dark with agitation (300 rpm) at room temperature for 30 min. To remove unbound antibodies, the microspheres were washed as described above. The plate wells were supplemented with 50-l portions of a fresh solution of streptavidin conjugate with the uorescent dye phycoerythrin (16 g/ml) and incubated in the dark with agitation (300 rpm) at room temperature for 10 min. The microspheres were washed as described above, resuspended in 125 l of the buffer solution for analysis with agitation on a shaker (500 rpm) for 1 min, and analyzed using a ow-through laser BioPlex uoVol. 24 No. 1 2009

MOLECULAR GENETICS, MICROBIOLOGY AND VIROLOGY

26 1 2 3 4 5 6

LUNIN et al. 7

product of the peroxidase reaction was measured with a microplate Multiskan Ascent photometer (Thermo Scientic, United States) at 450 nm. The inverse value of the last dilution of the studied specimen whose optical density (OD450) was two times higher than the OD450 of the negative control was considered an adequate antibody titer. RESULTS AND DISCUSSION The recombinant plasmid pTT10 was used to obtain active producers of the CBD protein on the basis of strain E. coli M15 [pREP4] [25, 28]. Selection of the transformant clones containing the pTT10 plasmid was performed by counterselection on the solid agarized LB medium as described by T. Maniatis et al. [15]. The producer cells carrying the recombinant pTT10 plasmid were grown in LB medium until the logarithmic growth phase was reached; the medium was then supplemented with the inducer of target gene expression (isopropyl--D-thio-galactopyranoside, IPTG) and incubated for 4 h. The cells were harvested by lowspeed centrifugation. Analysis of cell lysates of the producer strains demonstrated that, during the induction of IPTG expression, a protein product of the expected molecular mass (22 kD) accumulates in the cells. The level of expression of recombinant proteins was approximately 1520% of the total protein (Fig. 1). Isolation and purication of recombinant CBD proteins was performed on amorphous cellulose. After disruption of the induced E. coli M15 [pTT10] cells and removal of the cell debris, cell lysate was obtained by centrifugation. The lysate containing the fraction of soluble E. coli proteins, including recombinant ones, was incubated with 20% suspension of amorphous cellulose; as a result, the specic sorption of CBD on cellulose particles occurred due to the CBD afnity properties [13]. After the threefold washing-off the unbound proteins, the suspended CBDcellulose preparation suitable for immunization was obtained. The techniques used for the isolation and purication of recombinant protein, as well as for the obtaining of the CBDcellulose preparation, are quite simple and do not require any special tools. This one-stage method of production and purication of preparations allows producing ready-to-use injections for immunization of animals. A CBDcellulose complex with protein concentration of 5 mg/ml and the degree of purity of 95% has been obtained. To obtain a free CBD protein, the solution was eluted with a 8 M urea solution; after the dialysis, the protein solution in phosphate buffer (pH 7.2) was obtained. The concentration of free protein in the preparation was measured on a spectrophotometer and calculated using Kalckars formula as follows: (mg/ml) = 1.45A280 0.74 A260, where A280 and A260 are the OD values of the protein solutions at wavelengths 280 and 260 nm, respectively.
Vol. 24 No. 1 2009

Fig. 1. Electrophoretogram of the CBD gene expression in the E. coli. cells. 1, strain M15 [pREP4], before induction; 2, strain M15 [pREP4], induction; 3, strain [pREP4, pTT10], clone 1 before induction; 4, strain [pREP4, pTT10], clone 1, induction; 5, strain [pREP4, pTT10], clone 2 before induction; 6, strain [pREP4, pTT10], clone 2, induction; 7, molecular mass, control (30 kD).

rometer (Bio-Rad, United States) according to the manufacturers recommendations. The results of the analysis were processed using the BioPlex Manager Software 4.1.1. The titers of specic antibodies to the CBD protein and the CBDcellulose complex in rat sera were determined by indirect solid-phase enzyme immunoassay. The recombinant CBD antigen dissolved in 30 mM potassium carbonate buffer (pH 9.29.6) was immobilized on the walls and bottoms of the wells of a polystyrene 96-well plate (Corning Inc., United States) at 4C with agitation for 24 h. The wells were rinsed twice with distilled water and three times with PBS containing 0.05% Tween-20 (PBST) and supplemented with the dilutions (starting from 1 : 100) of the sera to be tested prepared using PBST. The sera were titrated by twofold dilutions and incubated at 37C on a shaker for 1 h. The plate was then washed as described above. The wells were supplemented with anti-species rabbit immunoperoxidase conjugate to rat immunoglobulins (P-RAQ, series no. 30410, Imtek Ltd., Russia) at the working dilution of 1 : 10000. Incubation was carried out at 37C on the shaker. The plate was then washed as described above; the reaction was catalyzed by the addition of 100-l portions of substrate-indicator mixture containing 0.01% 3,3',5,5'-tetramethyl benzidine (Sigma, United States). Incubation with the substrate was carried out at room temperature for 15 min. The reaction was terminated by adding 50 l of 8 M H2SO4 into each well. The optical absorption of the stained

MOLECULAR GENETICS, MICROBIOLOGY AND VIROLOGY

IN VIVO STUDY OF THE IMMUNOGENIC PROPERTIES

27

The CBD concentration in the free protein preparation was 4 mg/ml. To study the immunogenic properties of the CBD cellulose complex and a preparation consisting of free CBD and cellulose sorbent, the rats were inoculated with the above preparations. The titers of specic antibodies were then determined by indirect solid-phase enzyme immunoassay; the concentrations of cytokines in the blood serum of rats were determined using the BioPlex suspension array system [31, 37]. Induction of the synthesis of cytokines and specic antibodies in response to the injection of pure preparations of recombinant CBD proteins, the CBDcellulose complex, and the amorphous cellulose suspension was studied. Laboratory animals were divided into four groups of ve rats each. The rst group was inoculated with the pure CBD preparation, while the second was immunized with the CBDcellulose complex; the third group was inoculated with the cellulose suspension, and the fourth group (serving as a negative control) was inoculated with sterile buffered saline. The inoculum dose was 10 g protein per one animal; in the case of immunization with the cellulose suspension, a 20% immunosorbent suspension was used. Sterile buffered saline (pH 7.2) was used as a diluent of protein preparations. Small nodules appeared after the injection of the CBD complex and cellulose suspension, which then disappeared within the next three or four days. On the whole, the health status of the animals during immunization was satisfactory; neither body-weight loss nor lack of appetite (or other adverse reactions) were observed. The preparations were administered to all animals three times every two weeks on days 1, 14, and 28 of the experiment without the use of adjuvants. To study the cytokine proles, blood samples were taken from the middle caudal vein before the experiment and on days 1, 2, 15, 27, 29 and 58 of the experiment (a total of seven samples for each experimental group). The sera were prepared according to the recommendations of Bio-Rad, the manufacturer of the 9-Plex, a kit used for multiplex uorescent immunoassay of nine cytokines (IL-1, IL-1, IL-2, IL-4, IL-6, IL-10, GM-CSF, IFN, and TNF). The calibration curves were obtained using the BioPlex Manager Software 4.1.1. In our study, we used calibration curves to determine cytokine concentrations within the range of 232000 pg/ml. To construct the calibration curves, fourfold dilutions of cytokine standards (320002 pg/ml) were prepared. The calibration curves were obtained by using the 5-parameter nonlinear regression equation according to the recommendations of the manufacturer and the literature data [21, 26, 37]. The primary results were represented in the form of measurement logs of the average uorescence intensity for each analysis. These logs were processed using the

BioPlex Manager Software 4.1.1. For each value, the background was subtracted, and the concentration of each analyte was calculated using the relevant calibration curve. The statistical analysis of the results obtained was carried out. For each data set, the average value (M), standard deviation, and standard error of mean (m) were calculated. The data were statistically analyzed using the Students t criterion. The table shows the measurement results of cytokine concentrations in rat sera. The concentrations of ve cytokines (IL-1, IL-1, IL-4, IL-10, and TNF) in the control serum samples were determined. The concentrations of four cytokines (IL-1, IL-1, IL-10, and TNF) were determined in the serum samples taken from the experimental group. The concentrations of other cytokines were out of the range of the method sensitivity. In the rst, second and third experimental groups (except for the fourth, i.e., control, group), the rate of IL-1 synthesis was considerably higher, especially during the rst two days of the experiment (p < 0.01), than in the control group. When cellulose and CBD preparations were injected, the IL-1 concentrations in rat serum were 916.6 39.0 and 886.2 42.1 pg/ml (on the rst day of the experiment, and 2 h after the rst injection, respectively). During the rst day of the experiment, when the CBDcellulose complex was used, the content of IL-1 in the serum samples (436.0 19.9 pg/ml) was two times lower than in the serum samples from the rst and second groups. The next day, it increased up to 709.9 31.1 pg/ml. On the day 15 of the experiment, the IL-1 concentration was similar in all three experimental groups (225.6 18.9 pg/ml in the rst group, 223.9 16.8 pg/ml in the second group, and 292.6 21.8 pg/ml in the third group), whereas the IL-1 concentration in the fourth (control) group was four times lower. On the day 58 of the experiment, the IL-1 concentration in the serum samples from the rst (CBD) group was almost the same as in the samples from the control group; the concentrations of this analyte in the samples from the second and third groups were three times higher (251.7 17.7 and 256.1 16.2 pg/ml, respectively). The following pattern was revealed by a comparison of IL-1 content in the serum samples taken from different groups. On the rst day of the experiment, the IL-1 concentrations in the samples taken from all the experimental groups were similar to the initial concentration. On the second day of the experiment, the IL-1 concentration increased markedly in the samples taken from the second and third groups (up to 511.6 16.4 and 1122.7 28.4 pg/ml, respectively). On the second day of the experiment, the IL-1 content in the samples taken from the rst group remained almost unchanged as compared to the initial value. Only during the day 15 of the experiment, a considerable increase (up to 224.5 11.4 pg/ml) in the IL-1 content in the serum samples taken from the rats immunized with CBD was observed.
Vol. 24 No. 1 2009

MOLECULAR GENETICS, MICROBIOLOGY AND VIROLOGY

28 Cytokine concentrations (pg/ml) in rat serum (M m) Experimental group 1 Preparation CBD Day of experiment Initial 1 2 15 27 29 58 Initial 1 2 15 27 29 58 Initial 1 2 5 27 29 58 Initial 1 2 15 27 29 58

LUNIN et al.

IL-1

IL-1

IL-4

IL-10 45.6 5.4 67.1 8.9 47.5 5.2 132.4 14.7 69.9 9.3 265.8 11.6 152.5 8.8 56.4 6.8 115.3 9.9 135.3 8.2 115.3 5.9 41.9 3.4 138.2 12.1 109.6 7.7 63.8 6.2 219.1 4.4 401.2 10.6 295.1 9.8 112.4 8.4 702.7 13.5 306.8 9.9 25.8 3.8 44.6 6.3 17.5 0.8 70.8 7.2 95.9 9.3 52.6 5.1 72.3 4.2

TNF ND ND 87.7 15.3 267.3 7.4 108.4 6.6 261.1 7.7 51.1 6.9 82.3 12.6 ND 129.5 6.8 78.2 7.6 ND 162.5 11.2 131.9 8.8 ND ND 558.7 21.1 134.3 13.3 54.1 4.5 387.5 16.6 122.2 5.8 41.2 3.5 ND ND 62.8 5.4 ND 113.5 6.6 76.9 3.3

Cellulose

CBDcellulose

Buffered saline

21.4 2.3 22.5 3.8 ND 886.2 42.1 14.7 2.2 ND 553.3 22.2 34.4 4.6 ND 225.6 18.9 224.5 11.4 17.2 0.9 39.1 5.8 ND ND 59.7 8.9 130.3 19.3 ND 80.3 7.6 336.0 12.3 21.8 3.6 33.6 3.6 66.4 6.9 ND 916.6 39.0 84.5 11.1 26.2 7.3 618.6 26.3 511.6 16.4 20.7 4.8 223.9 16.8 66.9 6.3 ND 13.6 0.6 ND ND 32.9 4.2 15.5 3.2 ND 251.7 17.7 16.3 5.7 ND 48.2 4.9 20.6 2.3 ND 436.0 19.9 35.7 5.1 ND 709.9 31.1 1122.7 28.4 ND 292.6 21.8 187.0 13.6 36.4 8.9 44.3 4.4 ND ND 60.9 8.0 198.1 9.4 ND 256.1 16.2 58.6 4.7 8.0 0.7 44.4 6.3 11.2 0.4 49.1 6.1 214.4 8.9 18.8 0.3 153.3 9.2 212.8 11.5 19.1 0.8 72.3 8.7 58.0 6.6 35.8 2.1 51.8 5.1 107.0 9.2 62.8 4.4 90.0 9.2 41.6 5.5 48.9 4.2 60.4 4.3 56.6 8.4 13.4 0.2 75.4 5.3

Note: ND stands for not determined.

At the same time, the IL-1 content in the samples taken from the third group was 187.0 13.6 pg/ml; in the samples taken from the second group, the IL-1 content was approximately 3 times lower (66.9 6.3 pg/ml). Within 1 day of the third injection, the rate of the IL-1 synthesis in the rats from the second group remained unchanged; in the rats from the rst and third groups, the concentrations of IL-1 were 130.3 19.3 and 198.1 9.4 pg/ml, respectively. The results obtained demonstrate the intensity of reactions, which triggered the immune response of antigen-presenting cells during the rst one or two days after injection, as well as the development of intense inammatory response [8, 18]. Throughout the whole experiment, IL-4 was hardly detected in the serum samples taken from the experimental groups; however, a small amount of this cytokine (49.1 6.1153.3 9.2 pg/ml) was detected in the control serum samples (p < 0.05). The level of IL-10 in the samples taken from the experimental groups was higher than in the samples

taken from the control group (p < 0.05). Three peak concentrations of this interleukin were obtained in the analysis of samples taken from different experimental groups: on days 2 and 15 of the experiment, i.e., within 24 h of the rst or second injections, the IL-10 concentrations were 47.47401.15 pg/ml and 115.26 295.06 pg/ml, respectively. On day 29 of the experiment (within 24 h after the third injection), the IL-10 content reached 138.16702.67 pg/ml. When comparing the levels of IL-10 in the samples taken from different experimental groups, the following patterns were observed. On the rst day of the experiment, the highest rate of the IL-10 synthesis (219.1 4.4 pg/ml) was observed in the third group, whereas, in the rst group, the concentration of this cytokine was approximately three times lower (67.1 8.9 pg/ml); in the second group, it was 1.5 times higher and reached 115.3 9.9 pg/ml. On the second day of the experiment, the concentration of IL-10 in the serum samples taken from rats immunized with the CBDcellulose complex was 401.2 10.6 pg/ml, which was 8.5 and
Vol. 24 No. 1 2009

MOLECULAR GENETICS, MICROBIOLOGY AND VIROLOGY

IN VIVO STUDY OF THE IMMUNOGENIC PROPERTIES

29

three times higher than in the rst and second groups (47.5 5.2 and 135.3 8.2 pg/ml, respectively). On day 29 of the experiment, the IL-10 content in the samples taken from the third group was 702.7 13.5 pg/ml. A comparison of the IL-10 concentrations in the serum samples taken from the rst and third groups suggests that immobilization of CBD on a cellulose sorbent enhances the efciency of the IL-10 synthesis stimulation by more than twofold. A similar dynamics of uctuation of the TNF concentrations was observed. On the rst day of the experiment, the TNF content in the serum samples taken from the experimental groups remained at the negative control level. The next day, after each injection, a reliable increase in the TNF content was observed in the serum samples taken from the rst, second, and third groups; the highest level of TNF, i.e., 558.7 21.1 pg/ml, (p < 0.05), was observed in the samples taken from the third group, which was immunized with the CBDcellulose complex. Within 24 h of the third injection (day 29 of the experiment), an increase in the TNF concentration (p < 0.05) was again detected; the peak TNF concentration was 387.5 16.6 pg/ml (third group). On the whole, cytokines that primarily regulate the humoral immune response (Th2 response) were detected in the serum samples taken from rats immunized with the recombinant CBD protein and the CBD cellulose complex [18]. Within the rst two days, all experimental groups demonstrated a high level of the IL-1 and IL-1 production, which indicates the early activation of B lymphocytes, monocytes, and macrophages. A local and systematic increase in the concentrations of proinammatory cytokines (IL-1, IL-1, and TNF) occurs in the course of the detection and presentation of antigens by macrophages and is required for the activation, proliferation, and differentiation of lymphocytes, as well as for altering the functional state of neutrophiles [8]. The stimulating effect of the CBD cellulose complexes on the synthesis of the above-mentioned proinammatory cytokines was higher than that of free protein. The results obtained may serve as a conrmation of Gurvichs hypothesis about the presentation efciency of antigens immobilized on cellulose, as well as about the activation and differentiation of lymphocytes [5, 8]. In addition, the results obtained indicate that the cellulose-containing immunosorbent is directly involved in the regulation of immune response via activation of the synthesis of proinammatory cytokines. In our experiment, injections of the cellulose suspension resulted in an earlier and more signicant increase in the concentrations of IL-1 and IL-1 in rat serum than injections of the free CBD protein. These results do not agree with Gurvichs hypothesis that cellulose is an immunotolerant carrier [5] since, in our experiment, it induced the activation of antigen-presenting cells and,

2.0 1.8 1.6 1.4 1.2 1.0 0.8 0.6 0.4 0.2 0 100 200 400 800 1600 3200 6400 12800

Fig. 2. Determination of the titers of specic antibodies to CBD in rat serum. The abscissa shows the inverse values of the serum dilutions; the ordinate shows the optical density at 450 nm. The curve with triangles denotes titration of the serum samples taken from rats immunized with the free CBD protein; the curve with circles denotes titration of the serum samples taken from rats immunized with the CBDcellulose complex; the curve with diamonds denotes titration of normal (negative) rat serum.

probably, lymphocyte proliferation; on the whole, it was responsible for the initiation of the humoral immune response [8, 18]. In the case of immunization with cellulose suspension, low concentrations of key cytokines of the humoral immune response, i.e., IL-4 and IL-10, were detected in rat serum [18]. The low intensity of the production of IL-4 and IL-10 (or weak activation of Th2 cells, major producers of the above-mentioned cytokines) can be attributed to the absence of immunocompetent Toll-like-receptor (TLR) cells on the surface, which is required for binding to cellulose-containing sorbents [27, 32]. Due to this, the effective stimulation of innate immunity was not observed. However, other polysaccharides were found to have immunostimulating properties, including lipopolysaccharide [20], peptidoglycan [27, 30, 32], -glycane [19, 22, 33, 35, 36], for which TLR were detected and that show considerable promise as immunosorbents for antigens [30, 34]. The study of the IL-2 and IL-8 concentrations in rat serum demonstrated that they were low and did not exceed the range of the method sensitivity (<2 pg/ml). On the whole, the results obtained coincide with the concentrations of IL-2 and IL-8 in healthy donors and can be attributed to the short half-life of most cytokines (several minutes) in circulation, as well as to the presence of natural inhibitors in blood serum [8, 18].
Vol. 24 No. 1 2009

MOLECULAR GENETICS, MICROBIOLOGY AND VIROLOGY

30

LUNIN et al.

The titers of specic antibodies were determined by indirect solid-phase enzyme immunoassay using the recombinant CBD as an antigen. Figure 2 shows the results of rat serum titration. On day 58 of the experiment, the OD values of the serum samples taken from the rats immunized with the free CBD protein were similar to those of the negative control. In the samples taken from the rats immunized with the CBDcellulose complex, the OD values (p < 0.01) were reliably different from those of the negative control. The inverse value of the last dilution of the studied samples whose optical density (OD450) was two times higher than the OD450 of the negative control was considered an adequate antibody titer. Hence, the maximum titer of specic antibodies to CBD in the group of animals immunized with the CBDcellulose complex was 6400 on days 29 and 58 of the experiment. In this experiment, the considerable increase in the rate of antibody formation was observed in the case of immunization with the CBDcellulose complex as compared to the results obtained when using the free antigen, which is in accordance with the results obtained by Gurvich [2, 4, 7, 9]. The increase in the titer of specic antibodies was due to the following effects reached by the use of cellulose immunosorbent: aggregation and polymerization of antigen molecules, formation of an antigen depot in the organism, and enhanced presentation efciency of immobilized antigens to immunocompetent cells [1, 4, 5]. Even in the case of immunization with small antigen doses, the use of the proteincellulose complex allow us to obtain sera with high titers of specic antibodies [13, 14]. In addition, it should be noted that the location of the antigen in the proteincellulose complex on the surface of immunosorbent is not random [17], since the location of the antigen is strictly opposite due to the CBD afnity properties. This allows us to increase the bioavailability of the antigen for immunocompetent and, most importantly, antigen-presenting cells, which results in the more effective stimulation of T-helper cells in the immune system [13]. On the whole, proteins immobilized on sorbents may be considered an effective alternative to the use of reactogenic adjuvants for obtaining hyperimmune sera for immunization of animals. This approach may be particularly efcient when using immunosorbent capable of stimulating both adaptive and innate immunity. Moreover, the technology for obtaining injection preparations of recombinant proteins may be used as a basis for the development (through genetic engineering) of safe and efcient subunit vaccines for anti-infective immunization.

REFERENCES
1. Vorobev, A.A. and Vasilev, N.N., Adyuvanty (nespetsicheskie stimulyatory immunogeneza) (Adjuvants (Nonspecic Stimulants of Immunogenesis)), Moscow, 1969. 2. Gurvich, A.E., Biokhimiya, 1957, vol. 22, issue 6, pp. 10281034. 3. Gurvich, A.E., Kuzovleva, O.B., and Tumanova, A.E., Biokhimiya, 1961, vol. 26, pp. 934942. 4. Gurvich, A.E., Kuzovleva, O.B., and Orlov, G.E., Lab. Delo, 1967, no. 12, pp. 732734. 5. Gurvich, A.E., Korukova, A.A., and Elgort, D.A., Molekulyarnye i kletochnye mekhanizmy regulyatsii immuniteta (Molecular and Cellular Mechanisms of Immunity Regulation), 1985, pp. 5662. 6. Gurvich, A.E., Immunosorbenty i ikhispolzovanie v biotekhnologii (Immunosorbents and Their Use in Biotechnology), 1986, pp. 522. 7. Gurvich, A.E., Korukova, A.A., and Elgort, D.A., Immunomodulyatory (Immunomodulators), Moscow, 1987, pp. 6776. 8. Ketlinskii, S.A., Immunologiya, 1999, vol. 4, issue 1, pp. 4652. 9. Korukova, A.A., Grigoreva, O.S., and Gurvich, A.E., Byul. Eksper. Biol., 1985, vol. 100, issue 7, pp. 4446. 10. Korukova, A.A., Elgort, D.A., and Gurvich, A.E., Byul. Eksper. Biol., 1986, vol. 102, issue 12, pp. 736738. 11. RF Patent No. 2270249 (2006). 12. RF Patent No. 2278160 (2006). 13. Lyashchuk, A.M., Lunin, V.G., Karyagina, A.S., et al., Zh. Mikrobiol. Epidemiol. Immunobiol., 2006, no. 4, pp. 6568. 14. Lyashchuk, A.M., Lunin, V.G., Karyagina, A.S., et al., Biotekhnologiya, 2006, no. 5, pp. 2332. 15. Maniatis, T., Fritsch, E.F., and Sambrook, J., Molecular Cloning: A Laboratory Manual, Cold Spring Harbor: Cold Spring Harbor Laboratory, 1982. Translated under the title Metody geneticheskoi inzhenerii. Molekulyarnoe klonirovanie, Moscow: Mir, 1984. 16. Petrov, R.V., Khaitov, R.M., Nekrasov, A.V., et al., Allergiya, Astma, Klin. Immunol., 1999, issue 3, pp. 36. 17. Skvortsov, V.T. and Gurvich, A.E., Byul. Eksper. Biol., 1984, vol. 97, issue 2, pp. 179180. 18. Yarilin, A.A., Immunologiya, 1997, no. 5, pp. 714. 19. Aderem, A. and Underhill, D.M., Annu. Rev. Immunol., 1999, vol. 17, pp. 593623. 20. Beutler, B., Curr. Opin. Microbiol., 2000, vol. 3, pp. 2328. 21. Carson, R. and Vignali, D., J. Immunol. Methods, 1999, vol. 227, pp. 4152. 22. Dillon, S., Agrawal, S., Banerjee, K., et al., J. Clin. Invest., vol. 116, no. 4, pp. 916928. 23. Freund, J., Am. J. Clin. Pathol., 1951, vol. 21, no. 7, pp. 645656. 24. Gurvich, A.E. and Drizlikh, G.I., Nature, 1964, vol. 203, no. 4945, pp. 648649. 25. Henco, K. and Crowe, J., The QIAexpressionist, 1992. 26. Jager, W. and Rijkers, G., Methods, 2006, vol. 38, pp. 294303.
Vol. 24 No. 1 2009

MOLECULAR GENETICS, MICROBIOLOGY AND VIROLOGY

IN VIVO STUDY OF THE IMMUNOGENIC PROPERTIES 27. Janeway, C. and Medzhitov, R., Ann. Rev. Immunol., 2002, vol. 20, pp. 197216. 28. Karsten, H. and Joanne, C., The QIAexpressionist, 1992. 29. Laemmli, U., Nature, 1970, vol. 227, pp. 680685. 30. McKee, A., Munks, M., and Marrack, P., Immunity, 2007, vol. 27, pp. 687690. 31. Morgan, E., Varro, R., Sepulveda, H., et al., Clin. Immunol., 2004, vol. 110, pp. 252266. 32. Pasare, C. and Medzhitov, R., Semin. Immunol., 2004, vol. 16, pp. 2326.

31

33. Sato, M., Sano, H., Iwaki, D., et al., J. Immunol., 2003, vol. 171, pp. 417425. 34. Singh, M. and OHagan, D., Pharm Res., 2002, vol. 19, pp. 715728. 35. Takeda, K. and Akura, S., Int. Immunol., 2005, vol. 17, pp. 114. 36. Underhill, D., J. Endotoxin Res., 2003, vol. 9, no. 3, pp. 176180. 37. Vignali, D., J. Immunol. Methods, 2000, vol. 243, pp. 243255.

MOLECULAR GENETICS, MICROBIOLOGY AND VIROLOGY

Vol. 24

No. 1

2009

Vaccine 21 (2003) 32003207

Cellulose beads bound to cellulose binding domain-fused recombinant proteins; an adjuvant system for parenteral vaccination of sh
Sarah Maurice a, , Mara Dekel b , Oded Shoseyov b , Arieh Gertler a
Institute of Biochemistry, Food Science and Nutrition, Faculty of Agriculture, Food and Environmental Quality Sciences, The Hebrew University of Jerusalem, Freiburg Building, Rm 10, P.O. Box 12, Rehovot 76100, Israel b The Kennedy Leigh Centre for Horticulture Research, Faculty of Agriculture, Food and Environmental Quality Sciences, The Hebrew University of Jerusalem, P.O. Box 12, Rehovot 76100, Israel Received 17 September 2002; received in revised form 18 March 2003; accepted 19 March 2003
a

Abstract A recombinant form of the outer membrane protein (A-layer protein) associated with atypical Aeromonas salmonicida was expressed, fused to a cellulose binding domain (CBD) isolated from Clostridium cellulovorans. The resultant chimerical protein was bound to either Sigmacell 20 or OrbicellTM cellulose particles. Common goldsh were injected intraperitoneally with the celluloseprotein complex and blood serum antibody levels produced against A-protein were examined weekly by means of ELISA. These titers were compared to those induced by immunization of goldsh with the same protein, with or without Freunds incomplete adjuvant, as well as to a standard bacterin-adjuvant system. Small Orbicell beads (110 M) induced antibody levels that were equal to the titers produced by the adjuvanted protein and bacterin formulae. In comparison, the larger Sigmacell particles (1020 M) proved to be poor immunopotentiators. The long-term titer elicited from a single injection of A-protein bound to Orbicell beads was equivalent to that induced by two injections. All the vaccinated sh demonstrated memory to the A-layer protein after exposure to a pathogenic load of atypical A. salmonicida with Orbicell treated sh displaying the highest titer. No direct correlation was found between the presence of anti-A-protein antibodies and protection against infection. The paper describes a simple and safe method to increase the potential immunogenicity of soluble recombinant proteins by employing relatively inexpensive cellulose particles. 2003 Elsevier Science Ltd. All rights reserved.
Keywords: Cellulose binding domain (CBD); Adjuvant; Recombinant A-protein

1. Introduction Fish are notoriously difcult candidates for vaccination regimes as they are highly susceptible to handling stress and the process itself compromises their natural immune system, predisposing them to secondary infections. Vaccines will vary in their efcacy depending on their antigen composition and accompanying adjuvant [1,2]. Adjuvants, which are often employed in order to enhance the cell-mediated and humoral immune responses, can cause deleterious secondary reactions that override the desired protective effect of the vaccine. Although injectable oil-adjuvanted vaccines have been accepted by the industry for use in sh larger than 15 g, there are still major concerns over possible side effects, which can evolve as a result of the antigens or adjuvants employed [35]. Midtlyng et al. [6] examined four adjuvant systems in Atlantic salmon: mineral oil, aluminum

Corresponding author. Fax: +972-89476189. E-mail address: maurice@agri.huji.ac.il (S. Maurice).

salts, glucan and levamisole. Long lasting immunity and protection resulted only in the mineral oil-adjuvanted vaccines, but the sh displayed signicant intra-abdominal lesions. Lillenhaug et al. [7] also observed injection site lesions, reduced appetite and growth retardation in Atlantic salmon up to three weeks post-vaccination. The need for suitable vaccines and delivery systems specically designed for sh, which entail a balance between safety and efcacy, has led researchers in the direction of molecular biotechnology. Effective synthetic peptide and protein vaccines have been prepared against a spectrum of animal viruses and human microbial diseases [8,9]. Recombinant proteins and synthetic peptides are considered to be safer potential vaccine sources than inactivated or live attenuated microorganism, but only one vaccine containing recombinant products is commercially available for aquaculture use [1012]. Studies have shown that soluble immunogens rarely induce high titers of antibodies without the use of strong adjuvants [13]. Recombinant proteins, designed to mimic specic immunogenic and pathogenic factors, can

0264-410X/03/$ see front matter 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0264-410X(03)00231-7

S. Maurice et al. / Vaccine 21 (2003) 32003207

3201

be expressed fused to built-in adjuvants or targeting components that can overcome this deciency [14,15]. Encapsulation of puried native and dened synthetic antigens in biodegradable microspheres using poly-lactide-co-glycolide (PLG) and poly-dl-lactide-co-glycolide (PLGA) microparticles is a method that has been employed to increase the immunogenicity of soluble protein preparations when employed parenterally in rodents, without causing any undesirable local reactions [1618]. Carrier material intended for recombinant protein vaccines should be biodegradable, applicable for use with assorted species of biologically active peptides and proteins, as well as simple to administer. It is imperative that the course of binding the recombinant proteins to the carrier material occurs under mild conditions in order to maintain the epitopic structure of the molecules [19]. In the following study we produced the recombinant virulent A-protein associated with atypical Aeromonas salmonicida fused to the cellulose binding domain (CBD) cloned from Clostridium cellulovorans [20]. Cellulose binding domains are discrete protein modules that possess the intrinsic ability to bind rmly to different forms of cellulose. The strong afnity between cellulose and the CBD is used for many applications in industry and science [21]. The theoretical aim was to employ cellulose beads as an adjuvant carrier material in a recombinant protein vaccine and to achieve equal or higher antibody titers and a booster effect similar to that seen with standard adjuvants while avoiding internal or external damage. 2. Materials and methods 2.1. Maintenance of sh Five-month-old Carassius auartus (weight 35 5.6 g) maintained in outdoor earthen ponds were acquired from Hazorea Aquatics, Israel in mid-summer during ambient day/night water temperatures of 2628 C. The goldsh were divided into groups of 50 sh each and housed in 200 l aquariums containing two 5 l biolters. Water temperature was maintained at 28 C for one month, then slowly cooled to 21 C and maintained at this temperature for the duration of the experiment. Fish were fed twice daily (0.5% body weight) with oating pelleted food containing 30% animal protein. The sh were closely monitored over a two-month period for signs of disease and only those which maintained a high level of health were used for the vaccination trials. 2.2. Recombinant proteins Recombinant A-protein (At) was produced according to methods previously described [22]. In order to express the recombinant CBD-A-protein (CBD-At) the At vap gene was ligated into the NcoI/BamHI restriction sites of the core plasmid pET-CBD-180 [23] in which the single cysteine moiety had been transformed to glycine (Ira Marton, unpublished

data) to form the plasmid pET-CBD-180cys -At. The original pET-CBD-180cys contained two NcoI restriction sites; one located at the 5 initiation codon and the second one at the 3 end, located in the multiple cloning site, 36 bp upstream from the BamHI site. The 5 NcoI site was eliminated by means of point mutation, in vitro site-directed mutagenesis (Stratagene, La Jolla, Ca) employing two mutagenic primers that transformed a single cytosine to adenine. CBD-At expression and purication was similar to that of At. Protein expression was carried out in the host E. coli BL21(DE3) pLysS grown to the optical density (OD600 ) of 0.9 and induction performed with isopropyl -d-thiogalactopyranoside (IPTG) to a nal concentration of 0.5 mM. Inclusion bodies were isolated from bacterial proteins, suspended in ice-cold water by sonication, refolded in Trizmaurea buffer (nal concentration 20 mM Trizma, 4.5 mM urea, pH 11.3), dialyzed against 10 mM Tris (pH 8.5) and puried on a Q-sepharose ion exchange column using step-wise concentrations of NaCl in 10 mM Tris (50200 mM NaCl). The eluted protein was further dialyzed against sodium bicarbonate to nal Na bicarbonate to protein concentration ratio of 1:4, lyophilized and maintained at 20 C. 2.3. Bacterin preparation The primary isolate of atypical A. salmonicida (At3 -96) was maintained at 80 C in Brain Heart Infusion (BHI) Broth containing 25% glycerol. For preparation of the bacterin, frozen cultures were spread on BHI agar plates containing 0.01% Coomassie Brilliant Blue (BHI-CBB) and incubated at 21 C for 3 days. A single colony from the plate was then inoculated into 200 ml BHI broth and incubated at the same temperature under static conditions. Bacterial cells were separated from the growth media by centrifugation (4000 g, 10 min), washed twice with phosphate buffered saline, pH 7.3 (PBS), dialyzed against 4% formalin in PBS for 2 days (4 C, 10 volumes, 8), followed by PBS (10 volumes, 10) for two additional days and nally suspended in the same buffer to the concentration of 109 colony forming units (cfus) per millilitre. The bacterin was examined for bacterial viability by inoculating 10 l into BHI broth, which was incubated for 5 days (21 C) and subsequently applied to BHI-CBB. In order to verify the presence of A-protein, a 10 l sample of the nal product was analyzed by 10% polyacrylamide SDS-PAGE [24] and Western blot [25] employing polyclonal antibodies directed against recombinant A-protein [22]. The bacterin was maintained at 4 C and homogenized with an adjuvant at a ratio of 1:1 (v/v) prior to application. 2.4. Recombinant protein vaccines Lyophilized At and CBD-At proteins were dissolved in double distilled water (ddw) to produce a nal At concentration of 1 and 2 mg/ml. The 2 mg/ml solutions were employed concurrently with an adjuvant at a 1:1 ratio while

3202

S. Maurice et al. / Vaccine 21 (2003) 32003207

the 1 mg/ml preparations were injected without additives. In addition, CBD-At was bound to either, 20 m Simacell 20 (SimgaAldrich, USA) cellulose particles or 13 m OrbicellTM cellulose beads (Accurate Polymers Ltd., IN, USA) in the following manner; for a single dose, 12 mg (wet weight) of pre-washed cellulose beads were suspended in 1 ml CBD-At (2 mg/ml) and maintained in suspension for 1 h with the aid of a rotamixer (Elmi, Riga Latvia). The beads were separated from the protein solution by centrifugation (4000 g, 5 min), washed once with 1 M NaCl followed by repeated washes with 20 mM Tris (pH 7.0) until protein could not be detected in the supernatant. The concentration of total protein bound to the cellulose beads was determined by the Bradford method [26] according to the protocol recorded in the following section. The nal dosage of cellulose injected was determined from the concentration of protein bound to the beads. The concentration of Sigmacell or Orbicell beads which contained a molar concentration equivalent to 200 g At was suspended in PBS and injected intraperitoneally (i.p.) without adjuvant. 2.5. Cellulose binding capacity of CBD fused protein CBD fused proteins were bound to cellulose beads as described in the previous section. In order to quantitatively determine the maximum protein-binding capacity, the two types of cellulose beads were incubated with 1, 2 or 3 mg/ml CBD-At for 30 min. The protein-binding capacity was determined by the following method: the beads were isolated by centrifugation and placed in a clean Eppendorf tube. To remove the absorbed proteins, the bead complex was incubated (rotamixer, RT) with 100 l of 4.5 M urea in 20 mM Tris base for 10 min, centrifuged and the supernatant collected for further analysis. This process was repeated until protein was no longer evident in the supernatant. The concentration of protein released into each supernatant sample was determined by the Bradford method. 2.6. Immunization regime A preliminary trial vaccination, consisting of 10 sh, was performed in order to select the optimal adjuvant. At was homogenized with Complete Freunds (FCA), Incomplete Freunds (IFA), Squalene (Sigma, Israel) or ISA 206 (Montanide, Paris France). Blood samples were taken two weeks post-vaccination and the level of antibody response was determined by means of ELISA. The adjuvant that elicited the highest response and which proved to be the least toxic was selected for the large-scale vaccination experiments. Prior to vaccination, sh were anesthetized by immersion in 50 mg/l benzocain. Each trial group (n = 50) was injected i.p. with 200 l of one of the following preparations: (1) At dissolved in PBS; (2) At emulsied with an adjuvant; (3) CBD-At dissolved in PBS; (4) CBD-At emulsied with an adjuvant; (5) CBD-At bound to Sigmacell or Orbicell beads; (6) Sigmacell or Orbicell suspended in PBS; (7) PBS emulsied in an

adjuvant; (8) atypical A. salmonicida bacterin with an adjuvant. A second injection of the same dosage was administered four weeks post-primary vaccination. Blood samples were collected weekly by caudal vein puncture (n = 5 per trial) for a period of nine weeks with the exclusion of week 7. Serum samples were maintained at 20 C. Fish were bled only once and removed from the trial group. In addition, 50 sh received only one injection of C-At bound to Orbicell beads and were bled once at the termination of the study. A random group of sh (n = 20), representing each experimental aquarium, was weighed at the initiation and summation of the trial. 2.7. Analysis of serum antibody titers by ELISA Nunc ImmunoTM plates (Nalge Nubc, Denmark) were coated (o/n, 4 C) with 100 l At (10 g/ml) in 0.5 mM carbonatebicarbonate buffer (pH 9.6). Phosphate buffered saline containing 0.05% (v/v) Tween 20 (PBS-T) was used to wash the ELISA plates three times between each stage and sodium caseinate (pH 7.0) was used as a blocking agent and for sh serum and primary and secondary antibody dilutions. At-coated plates were washed followed by blocking for 2 h at 37 C. Hundred microlitres of a four-fold serial dilution (1/400 to 1/409, 600) of each test serum were applied to the wells and incubated for 2 h (37 C). Each sample was tested in duplicate and in parallel to positive and negative controls on the same plate. Following removal of the unbound antibodies, the plates were washed and incubated further with 100 l/well rabbit anti-carp IgM (1:30,000) followed by washing and further incubation with 100 l per well (1 h, 37 C) biotinylated HRP labeled goat anti-rabbit IgG (1:35,000). The plates were washed and 100 l of TMB peroxidase substrate (KPL, Gaithersberg, MD, USA) were added to each well. The color was allowed to develop for 20 min at room temperature after which, 100 l TMB stop solution was added. Plates were analyzed with a Sunrise absorbance reader (Tecan, Switzerland) at OD 450 nm and the mean optical density determined for each duplicate with the aid of Magellan reader control software. The results seen in Figs. 2 and 3 represent the absorbance values derived from the experimental serum samples diluted 1/400. They are expressed as the ratio of the test sera OD to positive reference sera. The positive control consisted of pooled sera possessing a high anti-At titer that had been collected from large sh (600 gm) vaccinated with A-protein emulsied in Freunds complete adjuvant and boosted with the same protein in Freunds incomplete adjuvant. In addition, serum samples were collected from nave sh and used for the negative control. 2.8. Rabbit anti-carp IgM Anti-IgM was prepared according to the method described in Fish Immunology Technical Communications [27]. Briey, adult carp of the type Dor 70 (n = 10)

S. Maurice et al. / Vaccine 21 (2003) 32003207

3203

were immunized i.p. with goat IgG at weeks 0, 2, and 5. One week after the nal injection, the sh were bled and the sera pooled. Specic carp anti-goat IgM was afnity puried on Afgel-10 covalently bound with goat IgG. The resultant puried serum was employed in the production of rabbit polyclonal anti-carp IgM antibodies as previously described [22]. 2.9. Atypical A. salmonicida challenge Four months following the nal injection, 10 sh representing each experimental immunization group, were moved to alternate aquaria (20 l) where they were exposed to a total of 0.2 104 l1 unwashed, atypical A. salmonicida (At-96) bacterial cells for a period of 1 h. They were briey washed and placed in clean 50 l aquaria for observation. Bacterial cells employed in the disease model were cultured as for the bacterin and cell numbers were determined by the MilesMisra method [27]. Serum samples were collected two weeks after exposure and examined for anti-At antibodies. Each sh was examined and classied numerically according to external symptoms. The values allotted were as follows: 1: no symptoms, 2: isolated areas of clouding of the epidermis, 3: damage or hemorrhagic areas in the ns, 4: localized raised area of scales without necrosis and 5: dermal and muscular necrosis. 2.10. Statistical analysis Intra and interplate assay differences were determined prior to interpretation of the data. Serum antibody titers were examined using a one-way analysis of variance (ANOVA) with the aid of the program, Analyze-it for Microsoft Excel 2000. The Tukey post-test was used to examine the variance of titer by week in the individual treatments. The Dunnet post-test was employed for comparison of the experimental groups with the negative control. A Students t-test was performed to compare the titer elicited from one injection and two injections of protein-bound cellulose beads. In all cases P < 0.05 with a CI of 95% was considered signicant. 3. Results 3.1. CBD-At production CBD-At protein recovered from the bacterial inclusion bodies was puried by ion exchange chromatography on a Q-Sepharose column (not shown). The protein bound to the resin at 90% efciency and the majority of the monomeric protein was eluted with 50 mM NaCl. An average nal yield of 500 mg of almost 100% pure monomeric protein as evidenced by both SDS-PAGE and gel ltration under non-denaturing conditions (Fig. 1a and b) was produced from 5 l bacterial growth. The molecular mass determined

Fig. 1. (a) SDS-PAGE analysis of recombinant proteins. Electrophoresis of 10 g samples dissolved in sample buffer containing 2-mercapto-ethanol performed by Laemmli discontinuous system on 12% acrylamide gel. (Lane 1) CBD-A-protein; (lane 2) A-protein; (lane 3) molecular mass markers (top to bottom: 97.4, 66.2, 45, 31, 21.5 and 14.4); (lane 4) CBD. (b) Analysis of recombinant proteins by gel ltration chromatography on a SuperdexTM 75HR 10/30 column. Proteins were monitored by absorbance at 280 nm using a Merk-Hitachi D-200 integrator. The column was developed at 0.8 ml/min with 10 mM TrisHCl buffer (pH 8.0) containing 150 mM NaCl and calibrated with bovine serum albumin (66 kDa, RT = 10.81) and trypsin inhibitor (22.1 kDa, RT = 13.60). (A) CBD-A-protein, RT = 10.90. (B) A-protein, RT = 10.21.

by both methods was 70 kDa as predicted from the theoretical calculations. 3.2. Cellulose binding capacity of CBD fused protein CBD-At was bound to either Sigmacell or Orbicell cellulose beads as previously described in methods. The concentration of bound CBD-At was determined by stripping the protein from the beads by means of a denaturing urea buffer. When incubated with 1, 2 or 3 mg protein, Orbicell beads displayed a capacity of 12.8 2.5, 29.3 1.9 and 30.9 0.9 ug protein/mg cellulose, while Sigmacell bound 11.9 3.8, 29.8 1.4 and 30.3 1.8 ug protein/mg beads, respectively. 3.3. Adjuvants During the preliminary experiments, CFA and ISA were found to be toxic. Two days following the i.p. injection, the ISA treated sh became lethargic and displayed dispersed

3204

S. Maurice et al. / Vaccine 21 (2003) 32003207

hemorrhagic patches. All of these sh died within 5 days. Throughout the 10-day observation period, the CFA injected sh demonstrated signs of anorexia and apathy. Four out of the 10 sh died within the 10-day observation period. The mineral oil adjuvants, Squalene and IFA, were well tolerated but IFA appeared to illicit slightly higher and longer lasting antibody titers and was therefore employed throughout the entire experiment. Internal and external damage was not observed in the cellulose bead and IFA treated sh. 3.4. Antibody response (two injections) Immunization trials were repeated twice: the rst vaccination trial included all of the vaccine formulae described above. The second vaccination trial consisted of only a selected group of vaccine formulae for the following reasons; during the initial study anti-At titers elicited by At and CBD-At were similar and At homogenized with IFA induced antibody titers which were parallel to those seen with similarly adjuvanted CBD-At. Fish that were vaccinated with Orbicell beads, Sigmacell beads or PBS emulsied in IFA displayed predominately negative titers. The second vaccination trial consisted of an abridged form of the rst protocol and only the fused protein, CBD-At, was employed as the antigen. PBS emulsied with IFA represented the negative control. The results detailed in this paper represent an average of the two trials and relate only to the titers de-

rived from the i.p. injection of CBD-At, CBD-At with IFA (CBD-At/IFA), CBD-At bound to Sigmacell (CBD-At/Sig) or Orbicell cellulose beads (CBD-At/Orb), PBS in IFA (PBS/IFA) and bacterin emulsied in IFA (Bact/IFA). Blood samples were taken once a week for a period of nine weeks, excluding week 7, and serum antibody levels were assessed by ELISA. The inter-assay variance of the positive controls was 8.2% and intra-assay variance 4.1%. Fig. 2A and B summarize the weekly mean antibody response of individual sh to the various antigen formulae for a period of nine weeks. When compared with the negative PBS/IFA control, it is apparent that a primary reaction (Fig. 2A) was achieved in the goldsh after i.p. injection of the rst dose of CBD-At/IFA, Bact/IFA and CBD-At bound to either Sigmacell or Orbicell beads. A rapid appearance of antibodies occurred in the rst week after vaccination with CBD-At/IFA (0.64 0.05) and Bact/IFA (0.18 0.04) whereas antibodies associate with C-At/Sig (0.230.01) and C-At/Orb (0.35 0.09) injections appeared only during the second week. The IFA adjuvanted protein titer persisted for at least 21 days and declined to background level at the end of the fourth week while the Bact/IFA antibody titers persisted for only three weeks. CBD-At/Sig and CBD-At/Orb induced similar antibody levels, which remained constant through to the third week, but as in the other treatments, the titer of antibodies in the CBD-At/Sig sh dipped to negligible levels by the fourth week. The anti-At titer in sh

Fig. 2. A-protein antibody titers induced by six different vaccine formulae (see text). Antibody titers of vaccinated sh (n = 5) were examined weekly by means of an indirect ELISA. (A) Primary response; weeks 14 after initial injection. (B) Secondary response; weeks 59 after booster. The graphical results are expressed as the ratio of the test sera OD to a positive reference serum (mean S.E.M.; P 0.05).

S. Maurice et al. / Vaccine 21 (2003) 32003207

3205

injected with protein coated Orbicell beads increased signicantly after 28 days (0.60 0.14) and reached a similar level as the At/IFA (0.64 0.05) sh displayed in the rst week after immunization. Five weeks after the initiation of the study, i.p. injections were repeated for all sh (Fig. 2B), excluding one Orbicell group that received a single primary dose only. The consequence of a subsequent application of CBD-At/FIA and Bact/IFA was a signicant increase in antibody concentration; CBD-At/IFA (0.97 0.30) within the rst 7 days and Bact/IFA (0.58 0.10) only during the second week. In both cases, the antibody levels were boosted back to the maximum concentrations noted after the primary injections and they remained high until the end of the nine-week experimental period. CBD-At/IFA treated sh demonstrated an antibody titer (0.27 0.15) in the eighth week equivalent to that which appeared in the second week and which was still apparent after nine weeks. The booster injection of CBD-At/Orb did not affect a signicant increase of anti-At antibodies above that seen in the fourth week after the primary injection. The CBD-At/Orb treated sh maintained elevated antibody titers until the end of week 9 (0.79 0.09). PBS/IFA and CBD-At treated sh did not display signicant anti-At antibodies during the entire nine-week period. 3.5. Antibody titers after a single injection of CBD-At/Orb Serum samples from sh (n = 10) that had received one or two injections of protein-treated Orbicell beads were tested for the presence of anti-At antibodies sixteen weeks after the primary vaccination. At this time period no significant difference (P < 0.01) in the antibody concentrations was observed in the sh, which received a single injection (0.34 0.10) and the respective titer values observed in the experimental group which had been vaccinated with a primary injection and a booster after four weeks (0.35 0.10). 3.6. Challenge with atypical A. salmonicida Fish from each vaccination group (n = 10) were exposed to 2103 cfu/l of atypical A. salmonicida as described in the methods. One week after exposure, samples of their serum were tested for antibody titers (Fig. 3). All the sh developed anti-At antibodies as a result of bacterial exposure. CBD-At/Orb and Bact/IFA treated subjects displayed significantly higher levels of response than CBD-At, CBD-At/IFA and CBD-At/Sig (P < 0.01). The sh were examined and graded for gross external appearance of disease symptoms (Fig. 4). Bact/IFA formula induced the greatest level of protection (no sh with symptoms graded 35), while the CBD-At/Orb demonstrated a moderate decrease in the intensity of ulceration. No direct signicant correlation was seen between the level of antibody titers that appeared in exposed sh of each vaccine formula and the degree of disease symptoms, though in C-At/Orb

Fig. 3. Antibody titers expressed in vaccinated sh one week after challenge with atypical A. salmonicida. Three months after the booster injection, vaccinated sh were challenged with pathogenic atypical A. salmonicida. Antibody titers were examined by ELISA, one week post-exposure. The results presented here are absorbance values expressed as the ratio of the test sera OD to a positive reference serum (meanS.E.M.; P 0.05).

Fig. 4. Disease symptoms in vaccinated and challenged sh. Vaccinated sh (n = 10) were exposed to 2 103 cfu/l of atypical A. salmonicda four months after the booster injection. One week post-exposure, sh were examined for symptoms of disease and graded in the following manner: (1) no symptoms; (2) areas of clouding of the epidermis; (3) hemorrhagic areas in the ns; (4) localized raised scales without necrosis; (5) ulceration. The bullets represent individual sh and the horizontal line in each column represents the average intensity of the disease symptoms in a single group.

and Bact/IFA treatments the protective effect was more pronounced.

4. Discussion The binding of biomolecules to cellulose through a cellulose binding domain meets the requirements necessary for a carrier material. Cellulose is a chemically inert matrix that has stable physical properties as well as low afnity for non-specic protein binding. It is pharmaceutically safe

3206

S. Maurice et al. / Vaccine 21 (2003) 32003207

and relatively inexpensive. In our research, we employed the CBD tag, fused to a target protein, in order to bind micron sized cellulose beads; thus transforming the antigen from soluble to particulate form suitable for sh vaccination. The recombinant chimeric CBD-At protein was puried to homogeneity and its proper refolding was documented by a monomeric appearance in gel ltration under non-denaturative conditions. In our preliminary experiments, we found that several commercially available adjuvants such as FCA and Montanide ISA were extremely toxic to i.p. injected goldsh, whereas similar treatments with IFA or Squalene mineral oil did not result in signicant immediate or delayed post-vaccination lesions or mortalities. In contrast, both Sigmacell and Orbicell cellulose beads were well tolerated by the sh and no deleterious response reactions were apparent throughout the entire experimental period. Increased antibody titers were observed in all treatment groups which had received some form of adjuvanted vaccine, but all the titers, with the exception of CBD-At/Orb, declined to background levels by the end of the fourth week. Titers elicited from a single injection and the standard booster method were still elevated and did not differ (P < 0.01) when examined four months after the primary injection. It can be thus concluded that the antibody titer observed in the fth week was the in progress response to the initial vaccination and that the second dose had neither a positive nor negative effect. Similar results have been seen by others with microencapsulated peptide and protein vaccines. Reddin et al. [16], Brayden et al. [17] and Moynihan et al. [18] incorporated subunit vaccines into PLG/PLGA microparticles for i.p. introduction to mice. They concluded that a single dose of such a vaccine was equivalent and, in some cases, superior to standard mineral oil- and alum-adjuvanted formulae when relating to the duration of antibody response, antibody species induced and protection acquired. Immunopotentiation with these microparticles is based on entrapment of the antigen and slow release over a dened period of time. The cellulose bead formula used in this project is closer, in comparison, to alum-precipitated vaccines in which the antigen is also adsorbed onto a carrier. The CBD binds irreversibly to crystalline cellulose beads by means of multiple hydrophobic bonds [28,29]. Alum-based vaccines are less effective and must be delivered several times in order to ensure the development of adequate titers of antibody [2] while a single injection of the protein-Orbicell complex induced long-term elevation of antibody titers. Anti-A-protein titers and kinetics observed in CBD-At/IFA and CBD-At/Orb treated sh exhibited different patterns after primary and booster immunization. While injections with CBD-At/IFA resulted in elevated titers as early as one week after primary and booster vaccinations, reactions to CBD-At/Orb were delayed to the second week after treatment. This variance is indicative of the time lapse necessary for the uptake and presentation of a pure soluble protein, as compared to a

particulate antigen. A similar but less obvious response was apparent with the Bact/IFA treatment. The cellulose particles most likely display multiple mechanisms for enhancing responses to the entrapped antigen; local irritation for chemoattraction of macrophage, co-delivery of antigens into antigen-presenting cells for immuno-stimulation and the classical adjuvant depot sustained release mechanism. Although the cellulose treated sh received an equal molar concentrations of protein, the overall immune response to the 20 m Sigmacell particles was consistently weaker than that achieved with the 13 m sized Orbicell particles The fact that the smaller sized particles tend to induce a better response could be due to increased particle uptake by the peritoneal macrophages in addition to the greater total surface area available for antigen presentation. Microscopic examination of peritoneal macrophage collected from sh that had been injected i.p. with either CBD-At/Sig or CBD-At/Orb and stained with Calcauor White M2R (SigmaAldrich, USA) revealed cellulose particles within the macrophage cell bodies of Orbicell treated individuals only (results not shown). This leads us to conclude that the larger Sigmacell particles are less efcient targets for phagocytosis. The specic humoral immune response induced by the Orbicell particles, indicates that a true adjuvant effect was achieved. In order to further examine the efcacy of the various vaccine formulae, vaccinated and control sh were exposed to an infectious dose of atypical A. salmonicida. Signicant lower degrees of ulceration and infection were seen only in the sh which had been immunized with Bact/IFA vaccine while a slight decrease in the incidence of ulceration was also apparent in the CBD-At/Orb treated sh. Although the A-layer protein of A. salmonicida has been shown to be a major factor for pathogenicity of atypical A. salmonicida [3033], the literature is divided as to the importance of the presence of serum antibodies to the outer membrane protein and immunity to the infection. Using Atlantic salmon, Gudmundsdottir et al. [34,35] reported a positive interdependence between A-protein antibody levels and survival during challenged with typical A. salmonicida. In this study, the presence of serum antibodies to the puried protein did not correlate positively with resistance in all cases. Our results indicate that antibodies directed only at the monomeric form of the A-layer protein play a minor role in protection from infection with GUD. The Bact/IFA vaccine contained the native tetrameric membrane form of the A-layer protein, as well other immunogenic factors, which most likely are responsible for the decrease in disease symptoms. It has been demonstrated in this work that, intraperitoneal injection of goldsh with Orbicell cellulose beads bound to the fused recombinant protein, CBD-A-protein, can be employed as an alternative to IFA to induce a signicant adjuvant effect. It was conrmed that these specic cellulose particles were responsible for the immunopotentiating effect. The CBD preparation conferred good results with respect to the duration of the response whether as a single-step or double dosage protocol was employed. In addition, the

S. Maurice et al. / Vaccine 21 (2003) 32003207

3207

absence of any immediate or delayed post-vaccination mortalities conrms that i.p. injection of sh with cellulose beads is safe. This system can be further developed to include the incorporation of multiple dened antigens, which would broaden the potential of candidate subunit vaccines.

Acknowledgements This work was supported by grants from the Israeli Ministry of Industry and Commerce and The Hebrew University of Jerusalem (Agreement no. 0325878). The authors would like to thank David Shelach for expert maintenance of the wet-lab and the experimental sh and to Paulina Mendoza and Dr. Noah Lavid for the weekly collection of blood samples. References
[1] Anderson DP, Jeney G. Immunostimulants added to injected Aeromonas salmonicida bacterin enhance the defense mechanisms and protection in rainbow trout (Oncorhynchus mykiss). Vet Immunol Immunopathol 1992;34(34):37989. [2] Anderson DP. Adjuvants and immunostimulants for enhancing vaccine potency in sh. Dev Biol Stand 1997;90:25765. [3] Ronsholdt B, Mclean E. The effect of vaccination and vaccine components upon short-term growth and feed conversion efciency in rainbow trout. Aquaculture 1999;174(34):21321. [4] Midtlyng PJ. Vaccinated sh welfare: protection versus side-effects. Dev Biol Stand 1997;90:3719. [5] Ellis AE. Immunization with bacterial antigens: furunculosis. Dev Biol Stand 1997;90:10716. [6] Midtlyng PJ, Reitan LJ, Speilberg L. Experimental studies on the efcacy and side-effects of intraperitoneal vaccination of Atlantic salmon (Salmo salar L.) against furunculosis. Fish Shellsh Immunol 1996;6:33550. [7] Lillenhaug A, Lunder T. Field testing of adjuvanted furunculosis vaccines in Atlantic salmon Salmo salar L. J Fish Dis 1992;6:155 74. [8] Partidos CD, Stanley CM, Steward MW. Immune responses in mice following immunization with chimeric synthetic peptides representing B and T cell epitopes of measles virus proteins. J Gen Virol 1991;72:12939. [9] Borras-Cuesta F, Petit-Camurdan A, Fedon Y. Engineering of immunogenic peptides by co-linear synthesis of determinants recognized by B and T cells. Eur J Immunol 1987;17:12135. [10] Christie KE. Immunization with viral antigens: infectious pancreatic necrosis. Dev Biol Stand 1997;90:1919. [11] Grudding RA, Lillenhaug A, Ebensen D. Recent developments in sh vaccinology. Vet Immunol Immunopath 1999;72:20312. [12] Lorenzen N, LaPatra S. Immunity to rhabdoviruses in rainbow trout: the antibody response. Fish Shellsh Immunol 1999;9:34560. [13] Leong JC, Anderson E, Bootland LM, Chiou P-W, Johnson M, Kim C. Fish vaccine antigens produced or delivered by recombinant DNA technologies. In: Midtlyng PJ, Brown F, editors. Fish Vaccinology. Basel: Karger. Dev Biol Stand 1997;90:26777. [14] Mowat M, Donachie AM, Jagewall S, Schon K, Lowenadler B, Dalsgaard K, et al. CTA1-DD-immune stimulating complexes: a novel, rationally designed combined mucosal vaccine adjuvant effective with nanogram doses of antigen. Immunol Invest 2001; 15(167(6)):3398405. [15] Wiesmuller KH, Fleckenstein B, Jung G. Peptide vaccines and peptide libraries. Biol Chem 2001;382(4):5719.

[16] Reddin KM, Easterbrook TJ, Eley SM, Russel P, Mobsby VA, Jones DH, et al. Comparison of the immunological and protective responses elicited by microencapsulated formulations of the F1 antigen from Yersinia pestis. Vaccine 1998;16(8):7617. [17] Brayden DJ, Templeton L, McClean S, Barbour R, Huang J, Nguyen M, et al. Encapsulation in biodegradable microparticles enhances serum antibody response to parenterally-delivered beta-amyloid in mice. Vaccine 2001;19(30):418593. [18] Moynihan JS, Jones DH, Farrar GH, Howard CR. A novel microencapsulated peptide vaccine against hepatitis B. Vaccine 2001;19:3292300. [19] Higaki M, Asechi Y, Takase T, Igarashi R, Nagafara S, Sano A, et al. Collagen minipellet as a controlled release delivery system for tetanus and diptheria toxoid. Vaccine 2001;19:30916. [20] Shoseyov O, Doi RH. Essential 170 kDa subunit for degradation of crystalline cellulose by Clostridium cellulovorans. Proc Natl Acad Sci USA 1990;87(6):21925. [21] Levy I, Shoseyov O. Cellulose-binding domains. Biotechnological approaches. Biotech Adv 2002;20:191213. [22] Maurice S, Hadge D, Dekel M, Friedman A, Gertler A, Shoseyov O. A-protein from achromogenic atypical Aeromonas salmonicida: molecular cloning, expression, purication, and characterization. Protein Exp Purif 1999;16(3):396404. [23] Shipgel E, Goldlust A, Efroni G, Avrah A, Eshel A, Dekel M, et al. Immobilization of recombinant heparinase I fused to cellulose-binding domain. Biotech Bioeng 1999;65(1):1723. [24] Laemmli UK. Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 1970;227:6805. [25] Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Biotechnology 1992;24:1459. [26] Bradford M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of protein-dye binding. Anal Biochem 1976;72:24854. [27] Cruickshank R, Duguid J, Swain R. Counting bacteria. In: Medical microbiology. Edinburgh and London: E & S Livingston Ltd.; 1965. p. 870. [28] Tomme P, Boraston A, McLean B, Kormos J, Creagh L, Sturch K, et al. Characterization and afnity application of cellulose binding domains. J Chromatogr 1998;715:28396. [29] Goldstein MA, Takagi M, Hashida S, Shoseyov O, Doi RH, Segal IH. Characterization of the cellulose-binding domain of the Clostridium cellulorvorans cellulose-binding protein-A. J Bacteriol 1993;176:57628. [30] Chu S, Cavaignac S, Feutrier J, Phipps BM, Kostrzynska M, Kay WW, et al. Structure of the tetragonal surface virulence array protein and gene of Aeromonas salmonicida. J Biol Chem 1991;266(23):1525865. [31] Ishiguro EE, Kay WW, Ainsworth T, Chamberlain JB, Austen RA, Buckley JT, et al. Loss of virulence during culture of Aeromonas salmonicida at high temperature. J Bacteriol 1981;148(1):33340. [32] Kay WW, Buckley JT, Ishiguro EE, Phipps BM, Monette JP, Trust TJ. Purication and disposition of a surface protein associated with virulence of Aeromonas salmonicida. J Bacteriol 1981;147(3):1077 84. [33] Munn CB, Ishiguro EE, Kay WW, Trust TJ. Role of surface components in serum resistance of virulent Aeromonas salmonicida. Infect Immunol 1982;36(3):106975. [34] Gudmundsdottir BK, Magnadottir B. Protection of Atlantic salmon (Salmo salar L.) against an experimental infection of Aeromonas salmonicida ssp achromogenes. Fish Shellsh Immunol 1997;7:55 69. [35] Gudmundsdottir BK, Jonsdottir H, Steinthorsdottir V, Magnadottir B, Gudmundsdottir S. Survival and humoral antibody response of Atlantic salmon, Salmo salar L., vaccinated against Aeromonas salmonicida ssp achromogenes. J Fish Dis 1997;20:35160.

Vaccine 23 (2004) 450459

Oral immunization of Carassius auratus with modied recombinant A-layer proteins entrapped in alginate beads
Sarah Mauricea, , Amos Nussinovitcha , Nicole Jaffea , Oded Shoseyovb , Arieh Gertlera
b a Institute of Biochemistry, The Hebrew University of Jerusalem, P.O. Box 12, Rechovot 76100, Israel Faculty of Agriculture, Food and Environmental Quality Sciences, The Kennedy Leigh Centre for Horticulture Research, The Hebrew University of Jerusalem, P.O. Box 12, Rechovot 76100, Israel

Received 9 July 2003; received in revised form 16 June 2004; accepted 24 June 2004 Available online 21 July 2004

Abstract This study was focused on the utilization of a recombinant expression system to produce a unique modied subunit vaccine possessing a self-contained delivery system which could potentially improve the uptake and delivery of vaccine products as well their immunogenic potential. For this purpose the A-layer protein (At-R) associated with the sh pathogen atypical Aeromonas salmonicida was cloned and modied by the genetic fusion of the protein transduction domain (MTS) derived from Kaposi broblast growth factor (At-MTS). The potential for these proteins to be employed as antigens for oral immunization of goldsh was examined by encapsulation of At-R, At-MTS and the control, BSA, into biodegradable alginate gel macrospheres which were fed to goldsh in place of standard pellet sh feed. The bead physical properties were modied only in the presence of At-R and the temporal release of proteins was signicantly less when At-MTS was employed. Western blot analysis of serum samples collected from sh following intubation with the recombinant proteins determined that the rate of protein uptake from the digestive tract into the blood system improved considerably when MTS was fused to At-R. Experimental sh were fed one of three proteinalginate formulae on a schedule of 3 days/week or 5 days/month for a period of 2 months. After 1 month, animals fed on the 5-day protocol demonstrated increased serum antibody titers while following an additional month of feeding this level decreased and titers were found to be higher in sh maintained on the 3-day regime. Fish fed At-MTS maintained the highest titer at the end of 2-month period. To determine whether the diminished antibody titers were a result of oral tolerance sh were injected intraperitoneally with the At-R antigen. Only experimental groups which had been fed At-R or At-MTS demonstrated increased antibody titers which paralleled a typical secondary humoral response. In spite of the presence of an increased titer to A-protein, vaccinated sh did not demonstrate resistance to infection with atypical A. salmonicida. 2004 Elsevier Ltd. All rights reserved.
Keywords: Alginate; Membrane translocation sequence (MTS); Oral immunization; Goldsh

1. Introduction When maintained in a physiologically stable condition, sh are capable of evoking signicant levels of immune response, both at the humoral and cellular levels when injected intraperitoneally (i.p.) with a suitable antigen [1]. Fish are highly susceptible to handling stress and the immunization process itself can lead to a compromised immune system and

Corresponding author. E-mail address: msarah@tx.technion.ac.il (S. Maurice).

secondary infections. In addition, the adjuvants employed concurrently with the vaccine preparations are often responsible for abdominal adhesions and failure to survive [2]. Studies into the structure of gut-associated lymphoid tissue (GALT) strongly suggests that the gut of the teleost sh is immunocompetent [35]. Introduction of antigens by means of anal intubation has been shown to induce high serum antibody levels and protection but the success of oral immunization of sh varies according to the methods employed [6,7]. It is generally considered that degradation of antigen in the stomach and anterior intestine prevents immunostimulation from

0264-410X/$ see front matter 2004 Elsevier Ltd. All rights reserved. doi:10.1016/j.vaccine.2004.06.022

S. Maurice et al. / Vaccine 23 (2004) 450459

451

occurring and that a small, but immunologically signicant amount of bioactive proteins are not totally hydrolyzed before their absorption into the lamina propria of the intestinal microvilli [810]. If antigens could be protected during passage through the foregut, an increased uptake should be expected and as a result, effective immunization could be achieved. One approach, which is directed at entrapment of the target antigen into biocompatible matrices, has been employed in order to circumvent the problem of digestive proteolysis. Encapsulation of antigens into microparticles with biodegradable polymers such as polylactide-co-glycolide (PLG) and poly(d/l-lactide-co-glycolide) (PLGA) [1113] and production of enteric coated antigen microspheres (ECAMs) [14,15] are the most common methods investigated. Recombinant proteins and synthetic peptides are considered to be safer potential vaccine sources than inactivated or live attenuated microorganism. Only one vaccine containing recombinant products is commercially available for aquaculture use [16] while effective synthetic peptide and protein vaccines have been prepared against animal viruses and human microbial diseases [17,18]. A major problem relating to the use of recombinant proteins in oral immunization is the limited amount of intact antigen that can be absorbed through the integumental layers and the gastrointestinal tract. Improved bioavailability has been accomplished by the use of protein transduction domains such as the HIV-1 Twin Arginine Translocation (TAT) protein [1922] and the Membrane Translocation Sequence (MTS) derived from Kaposi broblast growth factor [23]. These domains, once fused to recombinant proteins, can transport folded and oligomerized proteins across membranes in a concentration dependant manner thus allowing relatively large molecules to literally pass through cells and enter the blood system. Sodium alginate is a readily available, inexpensive water soluble polymer that can be used to encapsulate a wide variety of antigens under mild conditions. Mucosally administered antigens entrapped in alginate microspheres smaller than 10 m have been employed for the inducement of immunity in cattle [24,25], rodents [26,27] and sh [28] with varying degrees of success. The purpose of our study was to determine whether the addition of a MTS domain to recombinant A-layer protein would improve its rate of intestinal uptake and thus increase its immunogenic potential when administered orally to the agastric sh, Carassius auratus. For this purpose, the recombinant proteins were entrapped in large alginate gel beads and fed to the sh, as is normal pellet food.

ature was increased to 28 C for 1 month, then slowly cooled to 21 C and maintained at this temperature for the duration of the experiment. With the exception of the vaccination period, sh were fed twice daily (0.1% body weight) with oating pelleted food containing 30% animal protein. The sh were closely monitored over a 2-month period for signs of disease and only sh tanks which maintained a high level of health were used. The sh were anesthetized with Tricain (40 mg/L) prior to intubation and blood sample acquisition. 2.2. Construction of A-protein gene fused to MTS domain The atypical Aeromonas salmonicida A-layer protein coding sequence was amplied from total genomic DNA according to the methods described in Maurice et al. [29]. The pET-vapatyp plasmid was employed as a template and a core plasmid for building the pET-MTSvapatyp and pET-vapatyp MTS constructs. Primer sets for construction and identication of the fused genes are displayed in Table 1. PCR amplication of the MTS-A and A-MTS fragments was performed in a Rapid cycler (Idaho Tech, ID USA) using 50- L capillary tubes. The amplied PCR products were puried employing Qiquick Gel Extraction Kit (Qiagen Inc., CA, USA), digested with the appropriate restriction enzymes and ligated into pET-vapatyp which had been cut previously with the same enzymes. The resultant vectors, pET-MTSvapatyp and pET-vapatyp MTS, were transformed into the replication host, E. coli DH10 by heat-shock and selection of subcloned plasmids was performed on LB agar plates containing 200 g/mL ampicillin. Colonies containing the MTS modied plasmids were selected by analytical colony PCR employing primers S/MTS-A F and S/MTS-A R for the MTS-A plasmids and S/A-MTS F and S/A-MTS R primers for the A-MTS plasmids (Table 1). PCR positive plasmids were conrmed to contain the MTS domain and to be free of PCR errors by nucleotide sequencing with an ABI 377 prism DNA sequencer by the dye terminator cycle sequencing method at The Laboratory of DNA Sequencing, Hebrew University of Jerusalem. Veried positive vectors were used for the transformation of the expression vector, E. coli BL21(DE3) plysedS. 2.3. Expression, refolding and purication of recombinant proteins Three recombinant proteins were produced: A-layer protein (At-R), A-layer protein fused to a MTS domain at the N-terminal (MTS-At), and A-layer protein fused to a MTS domain at its C-terminal (At-MTS). The methods employed in the expression and purication of the MTS fused proteins were similar to those described for the production of recombinant A-protein. Briey, the E. coli expression vector was grown to optical density (OD600 nm) 0.9 and induction performed with isopropyl -D thiogalactopyranoside to a nal concentration of 0.5 mM. Inclusion bodies were isolated from

2. Materials and methods 2.1. Maintenance of sh Five-month-old Carassius auartus (weight 40 4.6 g) were acquired from Hazorea Aquatics, Israel in mid-summer. The goldsh were divided into groups of 30 sh and housed in 200 L aquaria containing two 5-L biolters. Water temper-

452

S. Maurice et al. / Vaccine 23 (2004) 450459

Table 1 Primers employed for modication of the vapatyp gene and for the identication of fused genes Primer name At F MTS-A F MTS-A R A-MTS F A-MTS R S/MTS-A F S/A-MTS R Sequence 5 to 3 aaa acc atg gat gtc gtg att agc ccg aaa acc atg gca gcc gtt ctt ctc cct gtt ctt ctt gcc gca ccc gat gtc gtg att agc ccc gga cta gta acc aga cca tcc aac gtg ac aac agt tac tag taa cgt tgg taa tcg cga aa cgc gga tcc gcg tta ggg tgc ggc aag aag aac agg gag aag aac ggc tgc cag agt gaa atc tac cag c gcg tta ggg tgc ggc aag aag aac agg g ca gcc gtt ctt ctc cct gtt ctt ctt gc Restriction site and/or modication NcoI NcoI MTS addition SpeI SpeI BamHI MTS addition

Unique restriction sites are underlined and MTS domains are indicated in bold script.

bacterial proteins, suspended by sonication in 20 mM Trizma Base containing 4.5 M urea, diluted to a nal protein concentration of 100 g/mL, refolded by dialysis against 10 mM Tris pH 8.0 and puried on a Q-sepharose ion exchange column using step-wise concentrations of NaCl (50200 mM) in 10 mM Tris pH 8.0. The eluted protein was further dialyzed against sodium bicarbonate to produce a nal Na bicarbonate:protein concentration ratio of 1:4, lyophilized and maintained at 20 C. Relative molecular weight values of the three recombinant proteins were determined by nonreducing SDSPAGE on a discontinuous 10% acrylamide [30] and by HPLC gel ltration. 2.4. Oral introduction of recombinant proteins At-R, MTS-At and At-MTS proteins dissolved in ddw were administered perorally, by intubation, employing an insulin syringe connected to teonTM coated tubing (bore size 0.5 mm). Each sh (n = 4) received a dosage of 0.5 mL (1 mg/mL) protein. 1 h after introduction, 200 L of blood were collected by caudal vessel puncture and allowed to coagulate. Serum was isolated by centrifugation (4000 g, 5 min) and maintained at 20 C until use. All studies were repeated twice. 2.5. Western blot analysis Serum samples were analyzed by 10% acrylamide discontinuous SDSPAGE and Western Blot [31] employing rabbit polyclonal antibodies (1:20,000) directed against recombinant A-protein [29]. 5 L of 5X sample buffer containing 2-mercaptoethanol (2%) and 5 L SDS (10% w/v) were added to 20 L serum samples prior to incubation at 50 C (10 min). Protein transfer from the SDSPAGE gel to a nitrocellulose membrane was carried out o/n (100 mA, 4 C) using Towbin transfer buffer in a miniblot system (Hoefer Scientic Instuments, CA, USA). Nonfat dry milk powder (5%) dissolved in 20 mM Tris-HCl, (pH 7.2) containing 150 mM NaCl and 0.1% (v/v) Tween 20 (TBS-T) was employed for blocking (2 h, 4 C) of the membrane and for dilution of the antibodies. Each membrane was washed (3 10 min) after every stage with TBS-T. Polyclonal goat anti-rabbit biotinylated HRP antiserum (1:16,000) (Sigma, St. Louis, USA) was used in conjunction with Amersham Super Signal (UK) for elucidation of the protein bands.

2.6. Alginate gel bead production The lyophilized proteins, At-R, MTS-At, At-MTS or BSA were added directly to a 0.5% (w/w) solution of CaCl2 2H2 O (J.T. Baker Chem., N.J. USA) in ddw to a nal concentration of 1 mg/mL. To eliminate the possibility of protein diffusion in or out of the formed bead, the same proteins were also incorporated into a 2% (w/w) aqueous solution of food-grade LV Na-alginate (G/M ration 39:61) (Sigma Chemicals Co., St. Louis MO, USA) This was performed by adding the powdered protein to the alginate solution in measured quantities while maintaining slow but constant mixing (60 rpm). The resultant alginateprotein solution was allowed to stand until air bubbles were no longer visible. A few drops of aqueous apple green food coloring (benzoic acid preservative C.I. No. 19140/44090, John Moir (PTY) Ltd.) were added to the water portion of both solutions. Using a Mini-pulse peristaltic pump (Gilson, Wisconsin USA), the alginateprotein mixture was drawn into polypropylene tubing (bore size 0.5 mm, ow rate 1.5 mL/min), pushed through a 26 gauge syringe needle and expelled in the form of small drops into the CaCl2 protein solution which was mixed constantly with the aid of a magnetic stirrer. The distance between the tip of the needle and the uid surface was approximately 5 cm. The process was allowed to proceed for a period of 2 min and the beads, which had been produced during this time period, were maintained in the CaCl2 solution for an additional 2 min. They were subsequently removed to a Buchner funnel, washed briey with ddw and the excess uid absorbed with Whatman No 1 lterpaper. The beads were then transferred to 50 mL plastic test tubes and maintained at 4 C until use. Two additional bead types were produced: (1) alginate beads containing the same concentration of protein but lacking food coloring were produced for the protein-release experiments; (2) beads containing food coloring but without protein were produced for the training period in which the sh were taught to eat the compound. 2.7. SDSPAGE analysis of protein incorporated beads The integrity of the encapsulated protein was determined prior to use by means of non-reducing SDSPAGE. A single bead was incubated (30 min, RT) in 20 L ddw containing 5 L 5X Laemmli sample buffer. The sample was then

S. Maurice et al. / Vaccine 23 (2004) 450459

453

heated to 100 C for 5 min, centrifuged (10,000 g, 5 min) and the supernatant loaded onto a 10% discontinuous polyacrylamide gel. Prestained protein standards (Biorad, USA) were included on each gel. Electrophoresis was accomplished employing a Laemmli buffer system. Gels were stained with 0.5% Coomassie Blue R-250 in 40% methanol/10% acetic acid for a minimum of 2 h and destained with 40% methanol/10% acetic acid. 2.8. Physical analysis of alginate beads Samples of the alginate beads containing one of the three proteins and control beads which did not contain protein were compressed between parallel lubricated plates to failure, at a constant deformation rate of 10 mm/min with an Instron Universal Testing Machine (UTM), Model 5544. The Instron was interfaced to an IBM-compatible computer with a card. Merlin software (Instron Co., Canton, MA) performed data acquisition and conversion of the Instrons continuous voltage vs. time output into digitized forcedeformation, forcetime, stressstrain or stresstime values with any desired denition of stress and strain. The force versus time data was converted to a pseudo-stress versus pseudo-engineering strain relationship using the following substitutions: = F and E = A0 D D0 .

2.10. Immunization protocols Prior to initiation of oral vaccination the sh underwent a training period in which they were taught to swallow alginate beads which did not contain protein. Following a 2-day starvation period the sh were fed with beads that had been coated with powdered shmeal (100 mg shmeal/10 beads). The amount of coating was decreased gradually over a period of 5 days until the sh were willing to ingest beads which had not come in contact with the sh meal. Feedings took place at the same hours each day and the beads were introduced into the water at two xed points and only when the sh were in close proximity to the water surface. Two immunization protocols were employed in which two aquaria (30 sh per aquarium) were used per treatment. All the sh were starved 2 days prior to receiving one of four different proteins, At-R, At-MTS, MTS-At or BSA. The bead dosage consisted of 5 beads/sh twice daily (125 beads per aquarium 2). Protocol 1 involved feeding the protein pellets for a period of 5 days followed by a rest period of 3 weeks and then a further 5 days of treatment. Protocol 2 consisted of 3 days of protein feedings, weekly, for a period of 2 months. Blood samples were taken at the termination of each month and examined for serum antibody titers by means of ELISA. The sh were closely monitored to determine if the alginateprotein feedings were tolerated and a group of 10 sh from each aquarium was weighed at the initiation and summation of the experiment. 2.11. Intraperitoneal vaccination of sh to determine oral tolerance Two weeks after summation of the oral vaccination protocol half of the sh (n = 15) from each aquarium were injected intraperitoneally (200 ug protein/sh) with a chimerical A-protein fused to a cellulose binding domain and bound to OrbicellTM cellulose beads (CBDcys -At/Orb) (Accurate Polymers Ltd., Ind, USA) [33]. Control sh were similarly injected with 200 L physiological saline. Blood serum samples were collected 2 weeks post injection and antibody titers determined by ELISA. 2.12. Analysis of anti-A-protein serum titers by means of ELISA Serum titers of vaccinated sh was determined by ELISA using the method described in Maurice et al. [33]. Briey, Nunc ImmunoTM plates were coated with 100 L At-R protein (50 g/mL) and blocked with 5% dried skim milk. Hundred microlitres of a four-fold serial dilution (1/40-2560) of each test serum were applied to the wells followed by rabbit anti-carp IgM (100 l/well, 1:30,000) and biotinylated HRP labeled goat anti-rabbit IgG (100 l/well, 1:35,000) (Sigma, USA). TMB peroxidase substrate (KPL, Gaithersberg, MD) and TMB stop solutions were used to develop the immune reaction. The color intensity was analyzed with a Sunrise

where is the pseudo-stress; E the dimensionless pseudoengineering strain; F the force needed to compress the bead at a given time; D0 the original diameter of the bead; D the total deformation and; A0 the cross-sectional area of the original bead. Beads with an average diameter of 3 mm were used for compression and tests were performed in six replicates employing beads taken from two separate batches. Statistical analysis was conducted with JMP software (SAS Institute, 1995) 2.9. Temporal release of protein from alginate gel beads Alginate beads containing protein, but lacking green food coloring, were used in order to determine the rate of protein release in an aqueous environment. Ten beads were place in a 1.5 mL Eppendorf tube containing 500 mL of 10 mM Tris buffer (pH 8.0) and incubated (RT, 30 min) with constant mixing on a rotamix (Elmi, Riga Latvia). The buffer was then transferred to a new tube and fresh buffer added. This process was repeated four more times at 0.5 h intervals (0.52.5 h). The experiment was executed with a total of six repeats for each protein and the buffer collected from two sample repeats were pooled. The protein concentration in each pooled sample was determined by the mini-Bradford method [32]. Protein concentrations were determined at OD595 nm and calculated from a titration curve created for each individual protein.

454

S. Maurice et al. / Vaccine 23 (2004) 450459

absorbance reader (Tecan, Switzerland) at OD450 nm and the mean optical density determined for each duplicate with the aid of Magellan reader control software. Each sample was tested in duplicate and in parallel with positive and negative controls on the same plate. The positive control consisted of pooled sera possessing a high anti-At-R titer that had been collected from large sh (600 gm) vaccinated with A-protein emulsied in Complete Freunds Adjuvant and boosted with the same protein in Incomplete Freunds. Pooled sera from nave sh were employed as negative controls. Intra and inter plate assay differences were determined prior to interpretation of the data. Serum antibody titers were examined using a one-way analysis of variance (ANOVA) with the aid of the program, GraphPad Prism, Version 2. The Dunnet post-test was employed for comparison of the experimental groups. In all cases, P < 0.05 with a Cl of 95% was considered signicant. 2.13. Production of rabbit anti-carp IgM Anti-IgM was prepared according to the method described in Fish Immunology Technical Communications 2 [34]. Briey, adult carp of the type Dor 70 (n = 10) were immunized (i.p.) with goat IgG at weeks 0, 2, and 5. One week after the nal injection the sh were bled and the serums pooled. Specic carp anti-goat IgM was afnity puried on Afgel-10 (Bio-rad, CA, USA) covalently bound with goat IgG. The resultant puried serum was employed in the production of rabbit polyclonal anti-carp IgM. 2.14. Challenge of sh with a typical A. salmonicida Each experimental group was challenged with a pathogenic load of atypical A. salmonicida (strain At3 -96) 6 weeks subsequent to termination of the feeding of protein incorporated alginate beads and 4 weeks after the i.p. injection. Five sh from each group were maintained as controls and were not challenged. The sh (n = 10) were moved to alternate aquaria (20 L), cooled to 10 C and exposed to a nal concentration of 0.2 104 /L unwashed, atypical A. salmonicida bacterial cells for a period of 1 h. They were washed briey, placed in clean aquaria, warmed to 21 C and observed for signs of ulceration. Two weeks after exposure each sh was examined and classied numerically according to external symptoms. The values allotted were as follows; 1 = no symptoms; 2 = isolated areas of clouding of the epidermis; 3 = damage or hemorrhagic areas in the ns; 4 = localized raised area of scales without necrosis; and 5 = dermal and muscular necrosis. 3. Results 3.1. Recombinant proteins All 3 recombinant proteins At-R, MTS-A and A-MTS were puried by ion exchange chromatography on a

Fig. 1. Analysis of recombinant proteins by gel ltration chromatography employing a SuperdexTM 75HR 10/30 column. Proteins were monitored by absorbance at OD 280 nm using a Merk-Hitachi D-200 integrator. The column was developed at a ow rate of 0.8 mL/min with 25 mM TrisHCl buffer (pH 8.0) containing 150 mM NaCl and calibrated with bovine serum albumin (66 kDa, RT = 10.81) and trypsin inhibitor (22.1 kDA, RT = 13.60). (A) A-protein (RT = 10.90); (B) MTS-At (RT = 10.69); (C) At-MTS (RT = 10.62). 2. SDSPAGE analysis of recombinant proteins. Electrophoresis was performed with a Laemmli discontinuous system on 10% acrylamide gel. Protein bands were elucidated by Coomassie staining and destaining. (Lane 1) molecular mass markers (top to bottom150, 100, 75, 50, 37 kDa); (lane 2) A-protein; (lane 3) MTS-A-protein; (lane 4) A-protein-MTS.

Q-Sepharose column. The proteins demonstrated a high afnity to the resin at pH 8.0 and were bound at 90% efciency with no target protein found in the breakthrough after application. The greatest part of the protein was eluted with 50 mM NaCl and an average nal yield of 500 mg of approximately 99% monomeric protein was produced from 5 L bacterial growth (results not shown). The retention times of the fused MTS-proteins, as measured by HPLC Superdex at a ow rate of 0.8 mL/min were 10.9 (At-R), 10.69 (MTS-A) and 10.62 (A-MTS) (Fig. 1AC). Compared to the protein standards, BSA (RT = 10.81; MW = 66 kDa) and trypsin inhibitor (RT = 13.6; MW = 22.3 kDa), the molecular weight of At-R was estimated at 51.1 kDa and that of the MTS proteins to be 52.6 kDa. This was similar to the molecular mass of 50 and 51.5 kDa determined by non-continuous 10% acrylamide SDSPAGE (Figs. 12AC). 3.2. A-protein serum levels in intubated sh Serum samples collected from goldsh (n = 4) 1 h after oral introduction of either At-R or At-MTS were analyzed by SDSPAGE and Western blot. Only one sh from the At-R treated group displayed a small amount of A-protein (Fig. 2A) while all four animals which had been intubated with At-MTS demonstrated large quantities of protein in their blood sera (Fig. 2B).

S. Maurice et al. / Vaccine 23 (2004) 450459

455

Fig. 2. Western blot of serum samples acquired from sh subsequent to oral intubation with recombinant proteins. Goldsh (n = 4) were orally intubated with 500 l recombinant protein (1 mg/mL). Blood serum samples were collected one hour after introduction and examined by Western blot. 20 l of serum in 5X reducing sample buffer were resolved by SDSPAGE on a 10% discontinuous polyacrylamide gel and trans blotted onto a nitrocellulose membrane. Prestained protein standards and 10 ng of pure recombinant protein were included on each gel. The blot was incubated with rabbit polyclonal anti A-protein IgG and detected by goat anti-rabbit peroxidase conjugate and ECL. (A) (Lanes 14) serum from A-protein treated sh; (lane 5) 10 ng pure A-protein. (B) (Lanes 14) serum from At-MTS treated sh; (lane 5) 10 ng pure A-MTS.

3.3. Alginate bead properties An average number of 425 beads were produced from 10 mL of alginate solution with an estimated protein concentration of 24 g per bead. The proteins maintained their integrity during the encapsulation procedure and could be maintained at 4 C for 8 days without any obvious change when examined by non-reducing SDSPAGE (results not shown). The mechanical properties of the various alginate bead formulae are summarized in Table 2. The addition of At-R to the alginate complex resulted in a modication of the stress and strain at failure measurements of the beads. The addition of BSA and MTS did not have a signicant effect on the strength (stress values 0.32 0.08 and 0.27 0.04 MPa respectively) when compared with the empty control beads (0.25 0.06 MPa), while the presence of At-R resulted in a weaker nal product (0.14 0.03 MPa). The average brittleness was altered in the same pattern. Strain at failure measurements of A-MTS and BSA beads (0.77 0.03 and 0.74 0.04 respectively) were not different from those of the control while At-R beads demonstrated signicantly lower readings (0.70 0.05). The calculated total micrograms of protein released per bead over the 2.5 hs was 6.7 (MTS-At), 15.6 (At-R) and 15.3
Table 2 Stress and strain at failure of the various alginate beads Bead type Alginate (control) Alginate + At Alginate + BSA Alginate + MTS Stress at failure (MPa) 0.25 a 0.06 0.14 b 0.03 0.32 a 0.08 0.27 a 0.04 Strain at failure () 0.78 a 0.04 0.70 b 0.05 0.77 a 0.03 0.74 a 0.04

Fig. 3. Temporal release of recombinant proteins from alginate beads. 10 beads were incubated in 500 l of 10 mM Tris buffer, pH 8.0 (RT, 30 min) with constant mixing on a rotamix. The buffer was then transferred to a new tube and fresh buffer added (ve times at 0.5 h intervals). Protein concentrations of pooled samples were determined from a titration curve created for each individual protein and using the mini-Bradford method. Each point on the graph represents the protein concentration in g/mL (mean S.E.M.) of three pooled samples at any specic time. The symbols on the graph represent the proteins: ( ) BSA; ( ) A-MTS; ( ) A-protein.

(BSA). All three proteins demonstrated similar patterns of release (Fig. 3). An approximate one-quarter of the proteins was released into the surrounding medium within 0.5 h; 26.4% for At-MTS, 20.7% for At-R and 24.6% for BSA. There was a slight decrease after one hour in all of the protein complexes followed by a pulse after 1.5 h. At all collection points the AtMTS beads released considerably less protein. The amount of protein discharged decreased after 1.5 h but could still be detected after 2.5 h. 3.4. Serum antibody titers in immunized sh Antibody titers of goldsh orally vaccinated with the various preparations and protocols (n = 30) were assessed after 1 (Fig. 4A) and 2 months (Fig. 4B) after commencement of the experiment. Fish treated with the 5 day protocol produced similar and signicant anti-At-R titers (P < 0.01) after 1 month. Those treated on the 3 day protocol demonstrated very low antibody levels. During the second month of the trial the antibody levels in the sh which had been vaccinated with At-R and At-MTS proteins, on the 3 day regime, increased signicantly (P < 0.05 for At-R and P < 0.01 for At-MTS) while those maintained on the 5 day schedule presented decreased titers (P < 0.001 and 0.05, respectively). Anti-At-R antibodies were not apparent in the sh treated with BSA alginate beads. Two weeks after the nal oral dosage, half of the sh (n = 15) were injected i.p. with At-R protein fused to a CBD domain and bound to OrbicellTM cellulose beads (Fig. 5). All of

Results are the average of six replicates taken from two separate batches. Different letters indicate signicant differences at (P < 0.05).

456

S. Maurice et al. / Vaccine 23 (2004) 450459

the goldsh that received i.p. injections of CBDcys -At/Orb produced elevated antibody levels when compared with the controls which were injected with saline (P = 0.001). No variance was observed between the antibody titers observed in sh orally vaccinated with At-R or At-MTS. The antibody levels demonstrated by i.p. injected sh which had been treated with A-protein on the 3-day (1.12 0.08) or 5-day (0.96 0.14) regime were the same as those fed At-MTS alginate beads (1.044 0.02 and 1.2 0.15, respectively). BSA treated sh presented signicantly lower anti-At-R titers (0.42 0.08 and 0.599 0.029) (P < 0.001). 3.5. Bacterial challenge of orally immunized sh Three and a half months after initiation of the oral immunization experiment the experimental sh, including those which had been injected i.p. with CBDcys -At/Orb, were challenged with At3 -96. No difference was observed in the incidence and intensity of ulceration in At-R, At-MTS and BSA orally immunized sh (Fig. 6A), while sh which had been injected i.p. with CBDcys -At/Orb demonstrated a slightly lower incidence of disease symptoms (Fig. 6B). Fish which had not been challenged remained symptom free.

Fig. 4. A-protein antibody titers induced by oral introduction of proteinincorporated alginate beads. Six groups of goldsh (n = 30) were fed twice daily with 125 alginate beads/feeding containing A-protein, A-protein-MTS or BSA. Two basic protocols for immunization were followed; 5 days of feeding, once a month for a period of 2 months or 3 days of feeding, once a week, for 2 months. Blood serum samples were taken (n = 10) at the termination of each month and examined for anti A-protein antibodies by means of ELISA. The results displayed on the graph are the mean S.E.M. of two experiments and are expressed as the ratio of the test sera OD to a positive reference serum. (A) Antibody titers evident 1 month after initiation of immunization trial. (B) Antibody titers seen after 2 months. Anti A-protein antibodies were not apparent in BSA vaccinated sh therefore do not appear on the graphs.

4. Discussion The vector pET 3d was employed in conjunction with the E. coli BL21(DE3) host for the over-expression of all three recombinant proteins in the form of intracellular inclusion bodies. Calculations of the molecular mass of each recombinant protein by means of SDSPAGE and gel ltration were consistent with projected estimations (Fig. 1). Oral introduction of antigens is the preferred form of vaccination for sh but absorption and proteolysis of protein antigens in the digestive system must be addressed when developing a vaccine delivery system. In the case of the agastric goldsh the problem of denaturation due to acidic pH was not a problem. Western blot analysis of blood serum

Fig. 5. Antibody stimulation of orally vaccination goldsh. Goldsh (n = 15), which had been immunized orally with A-protein, A-protein-MTS or BSA were injected intraperitoneally (200 ug protein per sh) with A-CBD bound to OrbicellTM cellulose beads. Control sh were injected with 200 L physiological saline. The results displayed on the graph are the mean S.E.M. of two experiments and are expressed as the ratio of the test sera OD to a positive reference serum.

S. Maurice et al. / Vaccine 23 (2004) 450459

457

Fig. 6. Bacterial challenge of orally immunized sh. Goldsh (n = 30/trial group) were immunized orally with one of three protein formulae incorporated in alginate beads while employing two different protocols. Two months subsequent to initiation of the study 12 sh from each group were injected i.p. with CBDcys -At. All of the experimental sh, excluding ve control sh/group, were exposed to 104 cfu of At3 -96. One week after exposure, sh were examined for symptoms of disease and graded in the following manner: (1) no symptoms; (2) areas of clouding of the epidermis; (3) hemorrhagic areas in the ns; (4) localized raised scales without necrosis; (5) ulceration. Each dot represents a single sh and the accompanying line, the mean of the 10 samples. (A) Fish, which were immunized perorally only. (B) Fish, which were injected i.p. following oral vaccination. At-R = recombinant A-protein; At-MTS = A-protein fused to MTS domain; BSA = bovine serum albumin. The number written in brackets following the protein code indicates whether the 3 or 5 day protocols were applied.

samples collected one hour after intubation with At-R revealed a small amount of the target protein in 1 out of 4 sh examined. In comparison, the MTS fused protein was taken up more efciently and all four experimental sh contained large concentrations of fused A-protein with only a slight degree of degradation (Fig. 2). Proteins introduced by the gastrointestinal tract are prone to enzymatic proteolysis and it can be assumed that the speed and efciency at which the proteins are absorbed is a primary factor for the transfer of intact molecules into the blood system. The results observed in the oral introduction of At-MTS indicate that the presence of a hydrophobic domain such as MTS can provide this capacity. Alginate microspheres containing antigen have been incorporated into feed and employed for the inducement of immunity in sh with inconsistent degrees of success [28]. In this study, BSA, At-R and MTS-fused At-R were entrapped in alginate gel beads (macrospheres) and administered to goldsh in place of pellet food. Proteins were mixed with the alginate solution and rapid and efcient polymerization of the complex was completed by cation exchange between the matrix (Na+ ) and a solution of CaCl2 . When examined by non denaturing SDSPAGE it was determined that the proteins were not affected by the encapsulation process and remained stable for up to 8 days when maintained at 4 C (results not shown). The mechanical properties of the beads are summarized in Table 2. Only the addition of At-R to the alginate matrix resulted in signicant changes in the strain and stress at failure, which resulted in the formation of a more brittle and weaker bead. The small increase in bead strength following the addition of both BSA and At-MTS can be an outcome of the increased solid percentage per bead resulting from water exchange by the protein. In the case of At, although the water was replaced by the protein as previously

described, the decreased strength may have been the result of repulsive effects resulting from the electro negative charge of the protein, or its interference with the cross-linking reaction of the negatively charged alginate. Preliminary experiments performed to determine the ideal parameters for the alginate pellet revealed the importance of this factor to the palatability and to the sensory evaluation of the product to the experimental animals. Starved sh captured, but regurgitate the beads which were weaker those observed with At-R. We determined that, in order to maximize sh feeding during these or other experiments, the strength of the bead should fall into a range of not less than 0.13 MPa, and not greater than 0.32 MPa. In vitro kinetic studies were carried out to determine the rate of protein release in an aquatic environment in order to determine to what extent of the proteins would undergo leaching after the beads enter the aquarium water and to estimate the possible rate of release in the sh digestive tract (Fig. 3). Proteins entrapped in the alginate matrix will diffuse out into the surrounding aqueous environment creating a continuous concentration gradient. Ionic interactions between the net charge of the protein, the matrix and the external buffer system affect the concentration and rate of proteins escaping from the beads over a given time period. BSA and At-R displayed the same level of protein discharge (P < 0.05) and in comparison, an average of 56% less protein was released from the At-MTS alginate beads; a reaction which could be related to the increased hydrophobicity of the protein. Although the total protein released differed signicantly the pattern of release was remarkably similar. An average of 23 1.5, 42 1.8, 70 3.0 and 88 1.5 (average S.E.M.) percent of the total protein released was determined after 0.5, 1.0, 1.5, and 2.0 h in all protein species. Approximately one quarter of the total protein associated with beads containing At-R and BSA was found in the surrounding buffer after 0.5 h;

458

S. Maurice et al. / Vaccine 23 (2004) 450459

a rate which would guarantee that enough protein would remain within the beads for intestinal uptake. In addition the time lapse between ingestion of the beads and excretion of their casts was approximately 22.5 h (personal observation), enough time for absorption of the major part of the protein. This in vitro study most likely does not mirror the true train of events, which occur in the sh digestive tract. It did not take into account the fact that the beads are crushed into a shapeless form by the pharyngeal dentiles before entering the digestive tube nor the action of the various digestive enzymes on the alginate and proteins. Results from the protein uptake experiments determined that the presence of a MTS sequence increased the uptake of A-protein in the intestinal tract (Fig. 4) but the in vitro release studies suggested that this protein was not released efciently from the alginate beads. In spite of this evidence, MTS-fused proteins were employed in the immunization experiment with the supposition that the increased rate of uptake would compensate for the quantity of protein released. Regardless of the protein species administered, serum antibody titers in sh fed protein beads for 5 consecutive days during the rst month were signicantly higher than the level of antibodies measured in animals which had been fed the same proteins on a 3-day schedule (Fig. 5A). During the second month, antibody titers in the sh that were treated only once during the month (5 days) decreased considerably while the titer of the sh fed four times during the month climbed to a level which might appear to be a secondary reaction (Fig. 5B). Following 8 weeks of treatment on the 3-day protocol, the At-MTS protein demonstrated a positive affect on the antibody titer, which was signicantly greater than observed in the At-R group (P < 0.01). This advantage did not appear to be long lasting and was not observed in the serum samples collected from the saline injected sh 12 weeks into the experiment (Fig. 6). In all the cases, the titers present in the orally immunized sh were noticeably less then the antibody titers generated during the i.p. immunization trials employing the At-R antigen [33]. In order to determine if the low antibody concentrations observed were due to the development of a state of immune tolerance, half of the orally immunized animals were vaccinated i.p. with At-CBDcys /Orb (Fig. 6). The post-i.p. antibody titers observed in serum samples collected from sh treated orally with both species of recombinant protein indicated recognition of the target antigen. BSA treated sh generated a weaker immune response which is comparable to a primary reaction. The results conrm that the oral feeding of alginate protein beads instigated a specic humoral immune response. Atypical A. salmonicida, the causative agent of goldsh ulcer disease, is considered to be a supercial organism which infects the host by colonization of the integument and gills [35]. In vitro studies have demonstrated that the A-layer is an adhesion molecule which promotes the association of the bacteria with the host extracellular matrix proteins [36]. The literature is divided as to the protective affect of serum antibodies to this protein but Gundmundsdottir et al. [37,38]

have reported a positive interdependence between A-protein antibody levels and survival during challenge with atypical A. salmonicida. To our knowledge, oral vaccination trials employing large concentrations of A-protein have not been carried out to date, and for this reason we chose to examine the potential of employing a recombinant form of the protein in order to reach immunological signicant serum levels. In spite of the presence of serum antibody titers, all of the sh became ulcerated at the same rate when exposed to atypical A. salmonicida (At3 -96). Similar to the results observed in the parenteral vaccination experiments described in Maurice et al. [33], a single i.p. injection of recombinant A-protein bound to cellulose beads had a slight effect on lowering the intensity of symptoms in challenged goldsh. Whether protection was not induced due to the method employed or to the fact that in both cases the monomeric form of the protein and not the native tetrameric form was used, remains to be studied. This study examined two methods, which can be employed in the enhancement of oral delivery systems of recombinant proteins to sh. A hydrophobic membrane translocation sequence fused to the core protein was shown to increase the rate of transfer of full-sized proteins from the intestinal tract to the blood system. The proteins were easily incorporated into alginate macro particles, which were ingested willingly and rapidly by the aquatic animals. This simple and inexpensive encapsulation method allowed for the introduction of relatively large dosages of proteins at maximum concentrations and could also be employed for numerous types of medication.

Acknowledgements This work was supported by grants received from the Israeli Ministry of Industry and Commerce and the Hebrew University of Jerusalem (agreement No. 0325878).The authors wish to thank Mr. David Shelach for his expert maintenance of the wet-lab and for his participation in the handling and feeding of the experimental sh.

References
[1] Muiswinkel WB, Weigertjes GF. Immune responses after injection vaccination of sh. Bull Eur Assoc Fish Pathol 1997;6(17):204. [2] Lillenhaug A, Lunder T, Poppe TT. Field testing of adjuvanted furunculosis vaccines in Atlantic salmon Salmo salar L. J Fish Dis 1992;6:15574. [3] Quentel C, Vigneulle M. Antigen uptake and immune responses after oral vaccination. Dev Biol Stand 1997;90:6978. [4] Roumbout JH, van den Berg AA, van den Berg CT, Witte P, Egberts E. Immunological importance of the second gut segment in carp. III. Systemic and/or mucosal immune responses after immunization with soluble or particulate antigen. J Fish Biol 1989;35:17989. [5] Palm Jr RC, Landolt ML, Busch RA. Route of vaccine administration: effects on the specic humoral response in rainbow trout Oncorhynchus mykiss. Dis Aquat Organ 1998;33(3):15766.

S. Maurice et al. / Vaccine 23 (2004) 450459 [6] Joosten PH, Kruijer WJ, Rombout J. Anal immunisation of carp and rainbow trout with different fractions of a Vibrio anguillarum bacterin. Fish Shellsh Immunol 1996;6:54151. [7] Ellis AE. Recent development in oral vaccine delivery systems. Fish Pathol 1995;30:293300. [8] Rombout JH, Lamers CH, Helfrich MH, Dekker A, TaverneThiele JJ. Uptake and transport of intact macromolecules in the intestinal epithelium of carp (Cyprinus carpio L.) and the possible immunological implications. Cell Tissue Res 1985;239(3):519 30. [9] Georgopoulou U, Vernier JM. Local immunological response in the posterior intestinal segment of the rainbow trout after oral administration of macromolecules. Dev Comp Immunol 1986;10(4):529 37. [10] Hertz Y, Tchelet A, Madar Z, Gertler A. Absorption of bioactive human growth hormone after oral administration in the common carp (Cyprinus carpio) and its enhancement by deoxycholate. J Comp Physiol 1991;B161:15963. [11] Lavelle EC, Jenkins PG, Harris JE. Oral immunization of rainbow trout with antigen microencapsulated in poly(d,l-lactide-coglycolide) microparticles. Vaccine 1997;15(10):10708. [12] Wong G, Kaattari SL, Christensen JM. Effectiveness of an oral enteric coated vibrio vaccine for use in salmonid sh. Immunol Invest 1992;21(4):35364. [13] ODonnell GB, Reilly P, Davidson GA, Ellis AE. The uptake of human gamma globulin incorporated into poly(d,l-lactide-co-glycolide) microparticles following oral intubation in Atlantice salmon Salmo salar L. Fish Shellsh Immunol 1996;6:50720. [14] Piganelli JD, Wiens GD, Zhang JA, Christensen JM, Kaattari SL. Evaluation of a whole cell, p57-vaccine against Renibacterium salmoninarum. Dis Aquat Organ 1999;36(1):3744. [15] Piganelli JD, Zhang JA, Christensen JM, Kaattari SL. Enteric coated microspheres as an oral method for antigen delivery to salmonids. Fish Shellsh Immunol 1994;4:17988. [16] Grudding RA, Lillenhaug A, Ebensen. Recent developments in sh vaccinology. Vet Immunol Immunopathol 1999;72:20312. [17] Partidos CD, Stanley CM, Steward MW. Immune responses in mice following immunization with chimeric synthetic peptides representing B and T cell epitopes of measles virus proteins. J Gen Virol 1991;72:12939. [18] Borras-Cuesta F, Petit-Camurdan A, Fedon Y. Engineering of immunogenic peptides by co-linear synthesis of determinants recognized by B and T cells. Eur J Immunol 1987;17:12135. [19] Becker-Hapak M, McAllister SS, Dowdy SF. TAT-mediated protein transduction into mammalian cells. Methods Enzymol 2001;24(3): 24756. [20] Ho A, Schwarze SR, Mermelstein SJ, Waksman G, Dowdy SF. Synthetic protein transduction domains: enhanced transduction potential in vitro and in vivo. Cancer Res 2001;61(2):4747. [21] Schwarze SR, Ho A, Vocero-Akbani A, Dowdy SF. In vivo protein transduction: delivery of a biologically active protein into the mouse. Science 1999;285(5433):156972. [22] Vocero-Akbani A, Lissy NA, Dowdy SF. Transduction of fulllength Tat fusion proteins directly into mammalian cells: analysis

459

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30] [31]

[32]

[33]

[34]

[35] [36]

[37]

[38]

of T cell receptor activation-induced cell death. Methods Enzymol 2000;322:50821. Rojas M, Donahue JP, Tan Z, Lin Y-Z. Genetic engineering of proteins with cell membrane permeability. Nat Biotechnol 1998;16:3705. Bowersock TL, HogenEsch H, Suckow M, Guimond P, Martin S, Borie D, et al. Oral vaccination of animals with antigens encapsulated in alginate microspheres. Vaccine 1999;17(13-14):180411. Rebelatto MC, Guimond P, Bowersock TL, HogenEsch H. Induction of systemic and mucosal immune response in cattle by intranasal administration of pig serum albumin in alginate microparticles. Vet Immunol Immunopathol 2001;83(1-2):93105. Kidane A, Guimond P, Tzdhu-chi JR, Sanchez M, Gibson J, Bowersock T. The efcacy of oral vaccination of mice with alginate encpasulated outer membrane proteins of Pasteurella haemolytica and One-Shot. Vaccine 2001;19:263746. Suckow MA, Bowersock TL, Park H, Park K. Oral immunization of rabbits against Pasteurella multocida with an alginate microsphere delivery system. J Biomater Sci Polym Ed 1996;8(2):1319. Joosten PH, Engelsma MY, van der Zee MD, Rombout JH. Induction of oral tolerance in carp (Cyprinus carpio L.) after feeding protein antigens. Vet Immunol Immunopathol 1997;60(1-2):18796. Maurice S, Hadge D, Dekel M, Friedman A, Gertler A, Shoseyov O. A-protein from achromogenic atypical Aeromonas salmonicida: molecular cloning, expression, purication, and characterization. Protein Exp Purif 1999;16(3):396404. Laemmli UK. Cleavage of structural proteins during the assembly of the head of the bacteriophage T4. Nature 1970;227:6805. Towbin H, Staehelin T, Gordon J. Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Bioechnology 1992;24:1459. Bradford M. A rapid and sensitive method for the quantitation of microgram quantities of protein utilizing the principle of proteindye binding. Anal Biochem 1976;72:24854. Maurice S, Dekel M, Shoseyov O, Gertler A. Cellulose beads bound to cellulose binding domain-fused recombinant proteins; an adjuvant system for parenteral vaccination of sh. Vaccine 2003;21(23):32007. Stolen JS, Fletcher TC, Anderson DP, Kaattari SL, Rowley AF. Techniques in Fish Immunology, vol. 12. NJ 07704-3303, USA: SOS Publications; 1992. p. 12630. Elliot D. PhD., University of Washington 1985. Daly JG, Kew AK, Moore AR, Olivier G. The cell surface of Aeromonas salmonicida determines in vitro survival in cultured brook trout (Salvelinus fontinalis) peritoneal macrophages. Microb Pathol 1996;21(6):44761. Gudmundsdottir BK, Magnadottir B. Protection of Atlantic salmon (Salmo salar L.) against an experimental infection of Aeromonsa salmonicida ssp achromogenes. Fish Shelsh Immunol 1997;7:5569. Gudmundsdottir BK, Jonsdottir H, Steinthorsdottir V, Magnadottir B, Dudmundsdottir S. Survival and jumoral antibody response of Atlantic salmon Sakmo salar L., vaccinated against Aeromonas salmonicida ssp achromogenes. J Fish Dis 1997;20:35160.

Methods 40 (2006) 118124 www.elsevier.com/locate/ymeth

Particulate delivery systems for animal vaccines

Jean-Pierre Y. Scheerlinck , Deanne L.V. Greenwood
The Centre for Animal Biotechnology, The University of Melbourne, Vic. 3010, Australia Accepted 5 May 2006

Abstract The requirements for veterinary vaccines are diVerent to those of human vaccines. Indeed, while more side eVects can be tolerated in animals than in humans; there are stricter requirements in terms of cost, ease of delivery (including to wildlife), and a need to develop vaccines in species for which relatively little is known in terms of molecular immunology. By their nature particulate vaccine delivery systems are well suited to address these challenges. Here, we review particulate vaccine delivery systems, ranging from cm-sized long-distance ballistic devices to nano-bead technology for veterinary species and wildlife. 2006 Elsevier Inc. All rights reserved.
Keywords: Particulate; Vaccines; ISCOMs; Inert micro-particles; Virus-like particles; Liposomes; Veterinary

1. Introduction The generation of veterinary vaccines has many common issues with vaccines developed for human use including: the need for protective (if not sterilising) immunity, minimisation of side eVects and ease of handling and administration. Despite these similarities, some challenges are speciWc to veterinary vaccines. Indeed, while side eVects might be less of an issue for animals compared to humans, there are more stringent requirements with respect to a number of other important parameters. These include the cost of production and delivery, particularly for species where a large number of animals with a relatively low commercial value are utilised (e.g. chickens). To some degree the requirements of vaccines for pets resemble more those of human vaccines in that cost is less important, while the absence of side eVects is more critical. This also implies that an ideal vaccine should be eVective with a single dose especially for wildlife and animals that are handled infrequently in remote, extensive farming systems.

Corresponding author. Fax: +61 3 9347 4083. E-mail address: j.scheerlinck@unimelb.edu.au (J.-P.Y. Scheerlinck).

Another aspect of veterinary vaccines is the genetic diversity of the species being considered and the requirement for generic systems that work across diVerent species. To a large degree this limits the use of molecular targeting techniques to cell surface markers and immune modulators such as cytokines, because for many species including wildlife, only minimal knowledge of these molecules is available. Thus adjuvants that rely on universal activation signals of the innate immune response (i.e. that are identical in diVerent species) are to be preferred. Taking these requirements into consideration, particulate vaccine delivery systems are well suited for veterinary and wildlife vaccine strategies (Table 1). Indeed, they are often applicable to a large range of species as they do not rely on speciWc ways of activating the immune system but rather on basic characteristics of the mostly innate immune responses. However, in some cases adaptations may be required due to anatomical diVerences. The use of particulate adjuvants also allows for additional routes of immunisation, which are better suited to veterinary and wildlife species including oral delivery and long-distance ballistic intramuscular delivery. These systems protect the antigen from degradation and impact during penetration through the skin and muscle. In some cases, they allow for targeting of several antigens by linking them together. Finally, and most

1046-2023/$ - see front matter 2006 Elsevier Inc. All rights reserved. doi:10.1016/j.ymeth.2006.05.023

J.-P.Y. Scheerlinck, D.L.V. Greenwood / Methods 40 (2006) 118124 Table 1 Particulate vaccine delivery systems in veterinary and wildlife species Species Horse Pathogen InXuenza Equine herpes virus 2 African horse sickness virus Pig Toxoplasma gondii InXuenza Mycoplasma hyopneumoniae Escherichia coli Atrophic rhinitis/pleuropenumonia Bovine herpes virus 1 Bovine virus diarrhoea virus Bovine respiratory syncytial virus Chlamydia Toxoplasma gondii Moraxella bovis Maedi-Visna virus Model antigen Staphylococcus aureus Blue tongue virus Louping ill virus Bovine rotavirus Bovine papilloma virus Bovine norovirus Dogs Parvovirus Rabies Echinococcus granulosus Atopic dermatitis Feline leukaemia virus Phocid herpes virus type 1/Feline herpes virus Mycoplasma gallisepticum InXuenza (H5N1) Newcastle disease virus Infectious bursal disease viruses Hematopoietic necrosis virus Aeromonas salmonicida Brucella abortus Pasteurella multocida Phocid Distemper Virus 1 Antigen HA, NA HA, NP (DNA) Envelope GnRF VP2 Rhoptry proteins HA (DNA) Whole inactivated Whole and Wmbriae Envelope Glycoprotein D (DNA) F and G surface E2 glycoprotein Whole virus Crude extract Cytotoxin Env. glycoprot. (DNA) Ovalbumin Lysate VP2, VP5, VP3, VP7 prME, NS1 SA11 L1 and L2 Capside protein Killed virus Glycoprotein surface antigens Allergen extract gp70/85, p15E Envelope p64 and p56 Killed virus Solubilised virus VP2/4/3, VP2 (DNA) Glycoprotein (DNA) At-R fusion protein Whole lysate Life organism Exctract Canine distemper virus Delivery ISCOM Gene gun ISCOM ISCOMATRIX VLP ISCOM Gene gun Encapsulated PLG microsphere Liquid nano-particles ISCOM Gene gun ISCOMATRIX Liposome ISCOM ISCOM PLG microsphere ISCOM ISCOMATRIX Gene gun Nano-beads PLG microsphere VLP VLP (rSFV) VLP VLP VLP + adjuvant ISCOMATRIX ISCOM ISCOM Liposomes DNA ISCOM ISCOM ISCOM ISCOM ISCOM Liposome Gene gun Macrospheres Liposome Biobullet PLG microsphere ISCOM References

119

[31,54], [55] [43,56], [57] [31,32] [31], PWzer, [35,36] [58] [59,60] [61,5] [13] [17] [31,62] [41,42] [31,63] [24] [31,64] [31] [10] [65] [66] [50] [3] [12] [35] [67] [38] [68] [39] [31] [31, 69] [70] [25] [31,71], [72] [31,73,74] [31,75] [31,33] [76] [77] [51] [78] [23] [53] [15] [31,79], [80]

Ruminants

Cats

Poultry

Fish

Bison Rabbit Seal

importantly, particulate vaccine delivery systems allow for the induction of qualitatively diVerent immune responses including CTL and other cell mediated immunity. These immune responses are often critical for immunity to key veterinary pathogens including most viral infections. In this review we will discuss recent developments of particulate vaccine delivery systems in veterinary and wildlife species including virus-like particles (VLP), DNA

delivery by gold coated particles, liposomes, inert macro-, micro- and nano-beads. We have excluded delivery through live viruses and the use of excipients, which are often delivered as powders but serve a diVerent function (i.e. facilitation of uptake of antigen across mucosal membranes). We will consider all species (including wildlife); excluding only human applications and species used as models for human diseases, predominantly rodents and non-human primates.

120

J.-P.Y. Scheerlinck, D.L.V. Greenwood / Methods 40 (2006) 118124

1.1. Inert particles The ability for vaccines to illicit a particular type of immune response is important in conferring protection. Inert particles coated or conjugated to antigen are uniquely designed to provide a targeted immune response with little or no side-eVects. Another important consideration is the ease of delivery since diVerent routes of inoculation can also inXuence local reactions and the type of immune response induced. Oral delivery of particulate vaccines has been shown to be successful and this approach is feasible in remote farming communities. Slow release systems have also been found to be eVective as they provide a continued supply of antigen over weeks or months, which is able to boost the immune response. Encapsulation of antigen has been widely used as it is easy to deliver, provides protection of the antigen from degradation and has been found to be eVective with a single dose. Micro- and nano-particles are solid particles made from inert materials or biodegradable polymers. Nano-particles are often in the range of 101000 nm, while micro-particles are larger at around 1100 m in diameter [1]. The use of these micro-particles is thought to improve immunogenicity with minimal reactogenicity due to the absence of additional inXammatory mediators [2]. The more simpliWed nano-particle system utilises antigen covalently conjugated to solid core nano-bead of a deWned size. Recent studies by Scheerlinck et al. [3] have demonstrated in sheep that antigen covalently linked to inert nano-beads with a size range of 4050 nm induced antibodies and cell mediated immune responses. The size of these nano-beads, which falls into the size-range of most viruses, is such that they appear to be taken-up preferentially by dendritic cells (DC) [4]. Interestingly, since both mouse and sheep DCs appear to take-up these particles it is possible that the mechanism of uptake of 40 nm nano-beads would be conserved within a wide range of species. From an evolutionary point of view it is conceivable that animals which have to be frequently combat viral invasion would have developed generic ways to manage these pathogens, based on their characteristic size. One way of protecting the antigen during oral delivery is to encapsulate it. A spray process has also been used to apply encapsulating material, Eudragit L30 D-55 to microspheres containing Mycoplasma hyopneumoniae antigen. These micro-spheres given orally to pigs have been shown to evoke a systemic IgG response when animals were subsequently challenged [5] (see Table 1). Biodegradable and biocompatible polyesters (polylactide-co-glycolides) have also been used extensively to encapsulate antigens and allow their slow release. Poly (D, L-lactide-co-glycolide) PLG is a polymeric ester which forms after the hydrolysis of -hydroxyacids: lactic and glycolic acids, yielding small spherical polymeric particles, in the size range of 1100 m [6,7]. These polymers are utilised in the manufacturing of medical prostheses and other medical-grade material such as sutures [8,9]. PLG micro-spheres have been used successfully in intranasal immunisation with Toxoplasma gondii in

sheep eliciting both humoral and cell mediated immunity [10]. In addition, the formulation of Staphylococcus aureus encapsulated biodegradable micro-spheres has also been shown to produce a humoral response, providing cows with some protection against mastitis [11,12]. In Wsh, the At-RMTS fusion protein against the Wsh pathogen Aeromonas salmonicida when encapsulated into alginate gel microspheres has also been shown to evoke a humoral response. Surprisingly, these vaccinated Wsh did not demonstrate resistance to atypical A. salmonicida when challenged. In contrast, Escherichia coli and detached Wmbriae incorporated into PLG micro-spheres given to new-born and weaned pigs did not evoke a signiWcant antibody response or protect animals against challenge with E. coli [13]. These results are at odds with other published data and may be a result of diVerences in the encapsulation process employed. The encapsulation process involves exposure of the antigen to organic solvents and high shear stress (low pH), during the degradation of the polymer [14]. There has been some suggestion that the integrity of the encapsulated antigen in the PLG may be comprised during this process [14]. PLG vaccines incorporating microencapsulated Pasteurella multocida antigen (in aliginate micro-spheres) with cholera toxin have been shown to be eVective in vaccinating rabbit populations [15]. To date, no adverse eVects of these materials have been reported [14]. The main feature of these adjuvants is the antigen is encapsulated into the polylactide-co-glycolide micro-particle and is released slowly over a number of weeks or months [16]. It is possible that combination of both fast-releasing and slow-releasing micro-particles could result in high levels of antibodies similar to those observed after multiple injections [1]. Water dispersed liquid nano-particles with a size which can vary from 10 to 500 nm have also been investigated. Trials in swine against atrophic rhinitis or pleuropenumonia with liquid nano-particles combined with immunostimulating compounds have demonstrated an enhanced immune response without inducing adverse reactions [17]. Vaccine trials in horses are currently ongoing in this area [17,18]. The process to manufacturer these liquid nanoparticle vaccines, often consists of a simple mixing of the antigen with the nano-micro-particles. Overall, the non-toxicity of the nano- and micro-particles and the ability to control release and presentation has wide beneWts for application in veterinary animals. 1.2. Liposomes Traditionally liposomes are composed of mainly phospholipids with an aqueous phase inside the particle madeup of a lipid bilayer. The antigen is either enclosed within the particle in the aqueous phase or intercalated into the lipid layer and therefore accessible from the outside. They have been shown to induce both humoral and CTL responses (reviewed in [1]). Several variations on this theme have been given diVerent names (reviewed in [19]). In many

J.-P.Y. Scheerlinck, D.L.V. Greenwood / Methods 40 (2006) 118124

121

cases products such as monophosphoryl lipid A have been included into liposomes hereby confounding the eVects of the liposome sensus strictu and immuno-stimulatory properties of its individual components. In addition, certain viral proteins incorporated into liposomes could have immunogenic, fusogenic, targeting and/or other intrinsic properties, which can make interpretation of results rather confusing [20]. To complicate matters the size of the lipid vesicles can also profoundly inXuence the traYcking of antigen, the uptake by antigen presenting cells and ultimately the type of immune response induced [21,22]. With the antigen entrapped into liposomes it is possible to deliver vaccines orally. In Wsh this has allowed the vaccination of carp against Aeromonas salmonicida infection [23]. Liposomes have also been used for the delivery of DNA vaccines. In one study cationic liposomes were used to deliver a plasmid DNA expressing BVDV type 1 major glycoprotein E2, to cattle as a DNA vaccine [24]. In these experiments liposomes were ineVective at enhancing immune responses to the DNA vaccine, with animals vaccinated with the liposome preparation less protected then those vaccinated with DNA alone. Liposomes have also been used to deliver allergen extracts as immunotherapy for refractory canine atopic dermatitis [25]. In this study noncoding plasmid DNA was used as a source of CpG motives to serve as adjuvant and redirect the immune response towards Th1. Archaeosome (i.e. liposomes formed with polar lipids of Archaeobacteria) have shown to be very eVective adjuvants inducing both Th1 and Th2 immune responses [1,20]. It is believed that these unique lipids present in the archaeosome from diVerent organisms are responsible for their adjuvant activity leading to diVerential antigen delivery and DC activation [26]. 1.3. ISCOM-based adjuvants Immunostimulatory complexes (ISCOMs) were initially described by Morein et al. more than 20 years ago [27], when viral antigens were incorporated into 40 nm particles composed of cholesterol, phospholipid and saponins derived from Quillaia saponaria Molina trees. The use of viral antigens in these initial experiments is not surprising considering hydrophobic membrane associated antigens are much easier to formulate with ISCOMs compared to soluble proteins. In addition, ISCOMs were reported to induce strong CTL responses which are often necessary to eVectively control viral infections. More recently, ISCOMATRIX has been developed which diVers from ISCOMs in that the antigen is not incorporated into the 40 nm particle but instead is mixed with generically prepared particulate adjuvant (reviewed in [28]). ISCOMATRIX induces both humoral and cell mediated immune responses including CTL. Since it is diYcult to envisage a mechanism by which soluble (i.e. non-associated) antigen could enter the Class I pathway; one can speculate that the CTL responses induced by ISCOMATRIX

occur when antigen is associated to the ISCOMATRIX particles, postmixing, through electrostatic or some other type of interaction. By virtue of being easier to manufacture and its ability to be used with a broad range of antigens including soluble antigens, ISCOMATRIX might have broad applicability in the veterinary Weld. ISCOMATRIX has been shown to induce profound eVects on the innate immune response at the level of the local lymph node, resulting in the induction of strong local cytokine responses which aVected cell traYcking within the draining lymph node [29] and resulted in strong immune responses in sheep [30]. The use of ISCOMs in veterinary species was reviewed recently by Morein et al. [31]. Considering the ability of ISCOM-based adjuvants to induce CTL responses; the majority of applications for ISCOM-based adjuvants in veterinary species are for vaccines against viral diseases (Table 1). Interestingly from a practical point of view, vaccination against one disease can have protective eYcacy against opportunistic diseases. For example, an ISCOM-based Equine herpes virus 2 (EHV-2) vaccine has been shown to protect not only against infection with this virus; but also against Rhodococcus equi pneumonia for which EHV-2 is a predisposing factor [32]. This type of cross-protection might increase in an unexpected way the economic viability of vaccinating animals. In contrast to other particulate adjuvants, ISCOMbased adjuvants have been tested in human clinical trials. It is therefore interesting to note that ISCOM-based H5N1 inXuenza vaccines were eVective in chickens while the unadjuvanted vaccine was not [33]. This suggests that ISCOM-based adjuvants might also be eVective in controlling this infection in humans, in case this strain would become highly transmissible between humans. Interestingly, ISCOMATRIX has also been used as an adjuvant with an analogue of gonadotropin releasing factor (GnRF) in a vaccine marketed as Equity (PWzer Animal Health). The vaccine induces an antibody response against endogenous GnRF and acts as an anti-conception vaccine for horses. The practical use of this vaccine is to suppress oestrus in mares and Wllies as a way of alleviating oestrusrelated behaviour also known as horsing or marey behaviour ([31] and PWzer Animal Health). 1.4. Virus-like particles Currently, many vaccines use live (attenuated) or inactivated (killed) virus. Incomplete activation/attenuation or reversion of virulence of these vaccines poses some risk to both humans and animals. In the last 10 years development of an alternate vaccine strategys has involved the identiWcation of viral proteins containing protective determinants. This has lead to the advance of recombinant systems such as Baculovirus, which enable the expression of large quantities of these proteins that mimic these viral proteins [34]. These VLPs morphologically are similar to the empty

122

J.-P.Y. Scheerlinck, D.L.V. Greenwood / Methods 40 (2006) 118124

capsids of viruses. As more information is ascertained with regard to the three-dimensional structure of these viruses it is now becoming possible to engineer vaccines with multiple viral epitopes (which lack genetic material and are therefore non-infectious) in order to illicit protective immunity [35]. VLPs have been used successfully in a number of diVerent species. Roy and colleagues [36] utilising Bluetongue virus (BTV) found that vaccination of sheep with VLPs consisting of all four major proteins (VP2, VP5, VP3 and VP7) gave long lasting protection for 15 months postimmunization. In addition, vaccination with African horse sickness virus (AHSV) VP2 protein in the presence of adjuvant provided protection against challenge in horses [35,36]. VLPs have also been used to boost antibodies in maternal mammary secretions. Boosting lactogenic immunity, due to passive transfer of antibodies to the neonate via colostrums and milk provided protection of calves from bovine rotavirus (BRV) [37,38]. In contrast, calves which were fed VLP containing colostrum were only partially protected against BRV. This observation is possibly due to the presence of colostral antibodies which inhibit vaccine virus replication [37]. VLPs have also been co-administered with oil, mutant E. coli heat-labile toxin (R192G) (mLT) or ISCOMs and were found to provide protection in gnotobiotic calves [39]. VLPs oVer a safe way to obtain highly puriWed, nonpathogenic targeted vaccines. The high immunogenicity of VLPs is most likely due to their interaction with DCs. Overall, VLPs provide a useful approach to eliciting protective immunity to a number of diVerent viruses regardless of the virus family or the complexity of the virus particle [40]. 1.5. Ballistic delivery of DNA DNA vaccines have traditionally been delivered either intramuscularly or intradermally. While intramuscular injection is regularly practiced in veterinary species, intradermal injection using a needle is too time consuming to be of practical use in most species. Ballistic delivery of DNA coated gold particles therefore oVers an attractive alternative, although current prices for the equipment hinders its widespread use in the Weld. While there have been some reports that gene gun delivery of DNA vaccines induces immune responses in some species [4144], it seems that this method of vaccine delivery will not be universally applicable. For example gene gun vaccination in sheep skin was ineVective at inducing immune responses [45]. Gene gun immunisation into the ear of dogs seems ineVective at inducing neutralising antibodies to rabies, while injection through the same route did induce neutralising antibodies [46]. Upon close inspection it appears that the anatomical features of skin of diVerent species are quite diVerent and that this can have an eVect on the level of expression of injected plasmid DNA [47]. Interestingly, a recent paper by

Watkins et al. [48], described the presence of DC expressing a reporter gene in aVerent lymph following gene gun delivery of a reporter gene. Unfortunately, in these experiments no immune responses were measured making it impossible to determine whether the presence of EGFP expressing DC leads to an immune response to this protein. Experiments in mice showed that removal of the site of gene gun delivery with a 24 h completely ablated the immune response while maximal immune response required the presence of gene gun targeted skin for at least three days [49]. This suggests that following gene gun delivery of DNA vaccines expression in the skin is essential for induction of immune responses. Hence, DC expressing plasmid DNA in aVerent lymph in the Wrst few days might not play a critical role in the induction of immunity. Thus even with higher plasmid DNA concentrations as used in the paper by Watkins et al [48], it is not clear that strong immune responses can be induced in sheep following delivery of plasmid DNA by gene gun to the skin. In addition in our experiment [47], the skin at the site of gene gun delivery was shedded shortly after three days which might not have been suYcient to allow the immune response to develop to its full extend. If one accepts that the cellular turnover is lower in the vulva, this might explain the apparent discrepancy between our results and the success in immunising sheep following gene gun vaccination against Maedi-Visna virus antigens in the vulva of sheep [50]. Taken together these results suggest that depending on the species the precise site of gene gun delivery of a DNA vaccine might be critical to the success of the strategy. DNA vaccines for Wsh appears at present to be the furthest developed, with some products currently on the market. In Wsh gene gun delivery of plasmid DNA resulted in the induction of protective immunity to hematopoietic necrosis virus, in the absence of neutralising antibodies [51]. This lead to the suggestion that cell mediated immunity might be responsible for the observed protection. 1.6. Long-distance ballistic vaccination of wildlife The vaccination of large wildlife species presents additional dilemmas of having to approach each animal individually and delivering the vaccine with minimal disturbance to the animal and maximal safety to the operator. To achieve this goal ballistic vaccine delivery systems have been developed which can be shot into the muscle of large wildlife species (e.g. bison) from distances of about 20 m using air-powered riXes. These biodegradable Biobullets made of photopolymerized PEG hydrogels can serve as biodegradable bullets that deliver volumes of about 90 l [52], while protecting the carried load from the impact upon penetration into the muscle to a depth of 10 cm. Interestingly these Biobullets can be used to deliver particulate vaccines as demonstrated by the fact that 1 m Xuorescent beads could be eVectively delivered intramuscularly using this system. Using this method Olsen et al. demonstrated that the bison can be

J.-P.Y. Scheerlinck, D.L.V. Greenwood / Methods 40 (2006) 118124

123

eVectively vaccinated against Brucella abortus from a safe distances of 20 m [53]. 2. Concluding remarks A large variety of particulate vaccine delivery systems have been used in veterinary species. These range from the Biobullets (12 cm-sized) shot at wildlife with an air-riXe, to the micro-particle (100.1 m-sized) used for DNA vaccination, to nano-beads/ISCOM (40 nm-sized) used as adjuvants in injected formulations. Several of these have been used commercially in veterinary species or in the wild. In all cases it seems that veterinary applications have used the intrinsic properties of these systems to solve practical issues associated with veterinary vaccines, including diYculty of access to the animal and need for particular type of immune response. One interesting aspect that warrants further attention is the issue of veterinary vaccine registration with respect to particulate vaccine delivery systems. Depending on the particular species used, it is likely the regulatory requirements will be very diVerent. Indeed, when animals are bred for human consumption the regulatory requirements in terms of biodegradability could be much stricter than in the case of animals bred as pets or as production animals where the products are not entering the food chain. When the vaccine is for wildlife it is likely that environmental issues will dominate the regulators thinking. Another trend which should be considered is the issue of side eVects. At present, side eVects that would be unacceptable in humans can to some degree be tolerated in animals. However, as public pressure on animal welfare increases it is likely that less reactogenicity vaccine preparations will be favoured. Thus adjuvants such as oil-in-water emulsions that have signiWcant side eVects, but are currently used for veterinary applications could well be phased-out when less reactogenic alternatives become available. Here also it is likely that the pet market will lead the way as higher prices for vaccines can be tolerated while side eVects cannot. References
[1] A.R. Spickler, J.A. Roth, J. Vet. Intern. Med. 17 (2003) 273281. [2] T. FiWs, A. Gamvrellis, B. Crimeen-Irwin, G.A. Pietersz, J. Li, P.L. Mottram, I.F. McKenzie, M. Plebanski, J. Immunol. 173 (2004) 31483154. [3] J.P. Scheerlinck, S. Gloster, A. Gamvrellis, P.L. Mottram, M. Plebanski, Vaccine (2005). [4] A. Gamvrellis, D. Leong, J.C. Hanley, S.D. Xiang, P. Mottram, M. Plebanski, Immunol. Cell Biol. 82 (2004) 506516. [5] J.H. Lin, C.N. Weng, C.W. Liao, K.S. Yeh, M.J. Pan, J. Vet. Med. Sci. 65 (2003) 6974. [6] T. Kissel, R. Koneberg, A.K. Hilbert, K.D. Hungerer, Behring Inst. Mitt. (1997) 172183. [7] Y. Waeckerle-Men, M. Groettrup, Adv. Drug Deliv. Rev. 57 (2005) 475482. [8] S.D. Putney, P.A. Burke, Nat. Biotechnol. 16 (1998) 153157. [9] H. Okada, H. Toguchi, Crit Rev Ther. Drug Carrier Syst. 12 (1995) 199. [10] A.C. Stanley, D. Buxton, E.A. Innes, J.F. Huntley, Vaccine 22 (2004) 39293941.

[11] C.N. OBrien, A.J. Guidry, A. Fattom, S. Shepherd, L.W. Douglass, D.C. WesthoV, J. Dairy Sci. 83 (2000) 17581766. [12] C.N. OBrien, A.J. Guidry, L.W. Douglass, D.C. WesthoV, J. Dairy Sci. 84 (2001) 17911799. [13] C.B. Felder, N. Vorlaender, B. Gander, H.P. Merkle, H.U. Bertschinger, Vaccine 19 (2000) 706715. [14] I. Jabbal-Gill, W. Lin, P. Jenkins, P. Watts, M. Jimenez, L. Illum, S.S. Davis, J.M. Wood, D. Major, P.D. Minor, X. Li, E.C. Lavelle, A.G. Coombes, Vaccine 18 (1999) 238250. [15] M.A. Suckow, T.L. Bowersock, H. Park, K. Park, J. Biomater. Sci. Polym. Ed. 8 (1996) 131139. [16] H. Tamber, P. Johansen, H.P. Merkle, B. Gander, Adv. Drug Deliv. Rev. 57 (2005) 357376. [17] J. Aucouturier, L. Dupuis, V. Ganne, Vaccine 19 (2001) 26662672. [18] J. Aucouturier, V. Ganne, A. Laval, Ann. N Y Acad. Sci. 916 (2000) 600604. [19] G. Kersten, D. Drane, M. Pearse, W. Jiskoot, A. Coulter, in: S.H.E. Kaufmann (Ed.), Novel Vaccination Strategies, Wiley-VCH Verlag Gmb H & Co. KGaA, Weinheim, 2004, pp. 173196. [20] M. Singh, D.T. OHagan, Int. J. Parasitol. 33 (2003) 469478. [21] J.M. Brewer, L. Tetley, J. Richmond, F.Y. Liew, J. Alexander, J. Immunol. 161 (1998) 40004007. [22] J.M. Brewer, K.G. Pollock, L. Tetley, D.G. Russell, J. Immunol. 173 (2004) 61436150. [23] T. Irie, S. Watarai, T. Iwasaki, H. Kodama, Fish ShellWsh Immunol. 18 (2005) 235242. [24] S. Harpin, D.J. Hurley, M. Mbikay, B. Talbot, Y. Elazhary, J. Gen. Virol. 80 (Pt 12) (1999) 31373144. [25] R.S. Mueller, J. Veir, K.V. Fieseler, S.W. Dow, Vet. Dermatol. 16 (2005) 6168. [26] G.D. Sprott, S. Sad, L.P. Fleming, C.J. Dicaire, G.B. Patel, L. Krishnan, Archaea 1 (2003) 151164. [27] B. Morein, B. Sundquist, S. Hoglund, K. Dalsgaard, A. Osterhaus, Nature 308 (1984) 457460. [28] M.J. Pearse, D. Drane, Adv. Drug Deliv. Rev. 57 (2005) 465474. [29] R.G. Windon, P.J. Chaplin, L. Beezum, A. Coulter, R. Cahil, W. Kimpton, D. Drane, M. Pearse, A. Sjlander, J.M. Tennent, J.-P.Y. Scheerlinck, Vaccine 19 (2000) 572578. [30] R.G. Windon, P.J. Chaplin, P. McWaters, M. Tavarnesi, M. Tzatzaris, W. Kimpton, R. Cahil, L. Beezum, A. Coulter, D. Drane, A. Sjolander, M. Pearse, J.-P.Y. Scheerlinck, J. Tennent, Vaccine 20 (2001) 490497. [31] B. Morein, K.F. Hu, I. Abusugra, Adv. Drug. Deliv. Rev. 56 (2004) 13671382. [32] A. Nordengrahn, M. Rusvai, M. Merza, J. Ekstrom, B. Morein, S. Belak, Vet. Microbiol. 51 (1996) 5568. [33] G.F. Rimmelzwaan, E.C. Claas, G. van Amerongen, J.C. de Jong, A.D. Osterhaus, Vaccine 17 (1999) 13551358. [34] J.I. Casal, Biotechnol. Appl. Biochem. 29 (Pt 2) (1999) 141150. [35] P. Roy, G. Sutton, Arch. Virol. Suppl. 14 (1998) 177202. [36] P. Roy, Dev. Biol. (Basel) 114 (2003) 169183. [37] L.J. Saif, F.M. Fernandez, J. Infect Dis. 174 (1996) S98S106. [38] M.R. Fernandez-Fernandez, J.L. Martinez-Torrecuadrada, J.I. Casal, J.A. Garcia, FEBS Lett. 427 (1998) 229235. [39] M.G. Han, S. Cheetham, M. Azevedo, C. Thomas, L.J. Saif, Vaccine (2005). [40] R. Noad, P. Roy, Trends Microbiol. 11 (2003) 438444. [41] R.P. Braun, L.A. Babiuk, B.I. Loehr, H. van Drunen Littel-van den, Virology 265 (1999) 4656. [42] S.C. Oliveira, J.S. Harms, G.M. Rosinha, R.S. Rodarte, E.L. Rech, G.A. Splitter, J. Immunol. Methods 245 (2000) 109118. [43] D.P. Lunn, G. Soboll, B.R. Schram, J. Quass, M.W. McGregor, R.J. Drape, M.D. Macklin, D.E. McCabe, W.F. Swain, C.W. Olsen, Vaccine 17 (1999) 22452258. [44] C.W. Olsen, Vet. Microbiol. 74 (2000) 149164. [45] R. De Rose, J. Tennent, P. McWaters, P.J. Chaplin, P.R. Wood, W. Kimpton, R. Cahill, J.P. Scheerlinck, Vet. Immunol. Immunopathol. 90 (2002) 5563.

124

J.-P.Y. Scheerlinck, D.L.V. Greenwood / Methods 40 (2006) 118124 [62] M. Trudel, G. Boulay, C. Seguin, F. Nadon, G. Lussier, Vaccine 6 (1988) 525529. [63] U. Carlsson, S. Alenius, B. Sundquist, Vaccine 9 (1991) 577580. [64] S. Hagglund, K.F. Hu, L.E. Larsen, M. Hakhverdyan, J.F. Valarcher, G. Taylor, B. Morein, S. Belak, S. Alenius, Vaccine 23 (2004) 646655. [65] A. Lunden, Vet. Parasitol. 56 (1995) 2335. [66] J.A. Angelos, J.F. Hess, L.W. George, Vaccine 23 (2004) 537545. [67] M.M. Morris-Downes, B.J. Sheahan, M.N. Fleeton, P. Liljestrom, H.W. Reid, G.J. Atkins, Vaccine 19 (2001) 38773884. [68] R. Kirnbauer, L.M. Chandrachud, B.W. ONeil, E.R. Wagner, G.J. Grindlay, A. Armstrong, G.M. McGarvie, J.T. Schiller, D.R. Lowy, M.S. Campo, Virology 219 (1996) 3744. [69] M. Fekadu, J.H. Shaddock, J. Ekstrom, A. Osterhaus, D.W. Sanderlin, B. Sundquist, B. Morein, Vaccine 10 (1992) 192197. [70] H. Carol, A. Nieto, Vet. Immunol. Immunopathol. 65 (1998) 2941. [71] A. Osterhaus, K. Weijer, F. UytdeHaag, P. Knell, O. Jarrett, L. Akerblom, B. Morein, Vaccine 7 (1989) 137141. [72] A. Osterhaus, K. Weijer, F. Uytdehaag, O. Jarrett, B. Sundquist, B. Morein, J. Immunol. 135 (1985) 591596. [73] B.E. Martina, M.I. Airikkala, T.C. Harder, G. van Amerongen, A.D. Osterhaus, Vaccine 20 (2001) 943948. [74] B.E. Martina, M.W. van de Bildt, T. Kuiken, G. van Amerongen, A.D. Osterhaus, Vaccine 21 (2003) 24332440. [75] B.G. Sundquist, G. Czifra, L. Stipkovits, Vaccine 14 (1996) 892897. [76] A. Homhuan, S. Prakongpan, P. Poomvises, R.A. Maas, D.J. Crommelin, G.F. Kersten, W. Jiskoot, Eur. J. Pharm. Sci. 22 (2004) 459468. [77] J. Li, Y. Huang, X. Liang, M. Lu, L. Li, L. Yu, R. Deng, Virus Res. 98 (2003) 6374. [78] S. Maurice, M. Dekel, O. Shoseyov, A. Gertler, Vaccine 21 (2003) 32003207. [79] I.K. Visser, M.W. vd Bildt, H.N. Brugge, P.J. Reijnders, E.J. Vedder, J. Kuiper, P. de Vries, J. Groen, H.C. Walvoort, F.G. UytdeHaag, et al., Vaccine 7 (1989) 521526. [80] I.K. Visser, E.J. Vedder, M.W. van de Bildt, C. Orvell, T. Barrett, A.D. Osterhaus, Vaccine 10 (1992) 435438.

[46] D.L. Lodmell, M.J. Parnell, J.T. Weyhrich, L.C. Ewalt, Vaccine 21 (2003) 39984002. [47] U.R. Hengge, P.S. Walker, J.C. Vogel, J. Clin. Invest. 97 (1996) 2911 2916. [48] C. Watkins, J. Hopkins, G. Harkiss, Vaccine 23 (2005) 42474256. [49] C.A. Torres, A. Iwasaki, B.H. Barber, H.L. Robinson, J. Immunol. 158 (1997) 45294532. [50] B. Gonzalez, R. Reina, I. Garcia, S. Andres, I. Glaria, M. Alzueta, M.I. Mora, B.M. Jugo, I. Arrieta-Aguirre, J.M. de la Lastra, D. Rodriguez, J.R. Rodriguez, M. Esteban, M.J. Grillo, B.A. Blacklaws, G.D. Harkiss, Y. Chebloune, L. Lujan, D. de Andres, B. Amorena, Vaccine 23 (2005) 43424352. [51] S. Corbeil, G. Kurath, S.E. LaPatra, Fish ShellWsh Immunol. 10 (2000) 711723. [52] R.J. Christie, D.J. Findley, M. Dunfee, R.D. Hansen, S.C. Olsen, D.W. Grainger, Vaccine (2005). [53] S.C. Olsen, R.J. Christie, D.W. Grainger, W.S. StoVregen, Vaccine (2005). [54] C.F. Crouch, J. Daly, D. Hannant, J. Wilkins, M.J. Francis, Vaccine 23 (2004) 418425. [55] C.F. Crouch, J. Daly, W. Henley, D. Hannant, J. Wilkins, M.J. Francis, Vet. Immunol. Immunopathol. 108 (2005) 345355. [56] C.C. Breathnach, H.J. Clark, R.C. Clark, C.W. Olsen, H.G. Townsend, D.P. Lunn, Vaccine (2005). [57] G. Soboll, D.W. Horohov, B.M. Aldridge, C.W. Olsen, M.W. McGregor, R.J. Drape, M.D. Macklin, W.F. Swain, D.P. Lunn, Vet. Immunol. Immunopathol. 94 (2003) 4762. [58] J.L. Garcia, S.M. Gennari, I.T. Navarro, R.Z. Machado, I.L. Sinhorini, R.L. Freire, E.R. Marana, V. Tsutsui, A.P. Contente, L.P. Begale, Vet. Parasitol. 129 (2005) 209217. [59] D.L. Larsen, A. Karasin, C.W. Olsen, Vaccine 19 (2001) 2842 2853. [60] D.L. Larsen, C.W. Olsen, Am. J. Vet. Res. 63 (2002) 653659. [61] C.N. Weng, Y.L. Tzan, S.D. Liu, S.Y. Lin, C.J. Lee, Res. Vet. Sci. 53 (1992) 4246.