Long-term antibody responses in mice following subcutaneous immunization with ovalbumin entrapped in biodegradable microparticles

Long-term antibody responses in mice following subcutaneous immunization with ovalbumin entrapped in biodegradable microparticles D.T. O'Hagan*, H. Jeffery and S.S. Davis Ovalbumin ( O V A ) was entrapped in microparticles prepared from three different poly(lactide-co-glycolide) (PLG) polymers and the microparticles were administered subcutaneously to mice as a single dose. Two weeks after immunization, the serum IgG antibody response to 0 VA entrapped in microparticles was significantly greater than the response to soluble 0 VA. The response to 0 VA entrapped in microparticles peaked at week 10 and remained high Jbr the full 1-year duration of the study. In a second study, the effect of particle size on the immunogenicity of PLG microparticles with entrapped OVA was assessed. Following booster immunizations in mice, microparticles of 1.5 pm were significantly more immunogenic than microparticles of 72.6 #m. Furthermore, although enhanced serum IgG responses were induced by immunization with 0 VA adsorbed to microparticles (1.0 #m), entrapment of the OVA in microparticles (1.5 #m) resulted in significantly better responses. Keywords: Microparticles; vaccines, controlled release; ovalbumin

The WHO Expanded Programme on Immunization (EPI), has resulted in the prevention of 2-3 million deaths annually and ensured that 70-80% of the world's children are reached by immunization servicesl. However, several constraints exist to restrict the likelihood of the achievement of the overall objective of the EPI, which is universal vaccine coverage. To be protected against diphtheria, pertussis and tetanus (DPT), three doses of the existing combined vaccines are needed. To be protected fully against neonatal tetanus, women of child-bearing age must receive five doses of vaccine. Unfortunately, the drop-out rate for those receiving the first dose of vaccine, but not the last required for full protection, is significantly high in many parts of the world. For example, the average drop-out rate for DTP vaccination was 46% in the Southeast Asia Region and similar figures have been recorded in Africa 2. It is a salutary lesson in the problems of vaccine policy implementation and the consequences of failure, that nearly 565000 newborns died in 1991 of tetanus worldwide 3, although effective vaccines have been available for many years. Because of the problems associated with the currently available vaccines, the WHO established the Programme for Vaccine Development (PVD) in 1984. The main Department of Pharmaceutical Sciences, University of Nottingham, University Park, Nottingham NG7 2RD, UK. *To whom correspondence should be addressed. (Received 3 September 1992; revised 2 November 1992; accepted 5 November 1992) 0264~410)(/93/09/0965-05 ~t- 1993 Butterworth-Heinemann Ltd

objective of the PVD is to develop a vaccine capable of being administered as a single dose soon after birth. More recently, the Children's Vaccine Initiative was established, to promote the development of new vaccines that could be used early in life, would require fewer doses, would not require refrigerated storage and would have simple immunization schedules through easier routes of administration. One means whereby vaccine implementation may be improved is the use of controlled-release antigen delivery systems, comprising polymeric microparticles with entrapped antigens. These controlled-release vaccines would be designed to release entrapped antigens at predetermined intervals following a single immunization and hence should obviate the need for booster doses of vaccines 3. The poly(lactide-co-glycolides) (PLG) are the primary candidates for use as polymers in the preparation of controlled-release vaccines. The PLG polymers have already been used for other biomedical applications, including absorbable suture material 4 and bone plates 5. In addition, PLG polymers have been used in the preparation of various controlled-release drug delivery systems6. The rate of release of macromolecules, e.g. peptides and proteins, from PLG microparticles is complex, but appears to be largely controlled by the degradation of the polymer 7. The rate of degradation for individual polymers is controlled by a number of variables including polymer composition, molecular weight, molecular weight distribution and morphologys. In the present study, a model antigen, ovalbumin (OVA), was entrapped in microparticles prepared from

Vaccine, Vol. 11, Issue 9, 1993

965

Controlled-release microparticles as vaccines: D.T. O'Hagan et al.

three different PLG polymers. The polymers had different molecular weights or copolymer compositions and so would degrade at different rates. The three different microparticle compositions with entrapped OVA were combined before administration and injected subcutaneously into mice as a single dose. For comparison, a group of mice were also immunized with OVA adsorbed to an aluminium hydroxide adjuvant (Alugel). The IgG antibody responses to OVA in microparticles and adsorbed to Alugel were followed for 1 year. Aluminium salts remain the only adjuvants currently approved for use in humans ° . A second study was designed to assess the fundamental mechanisms of the adjuvant effect achieved by the incorporation of antigens in microparticles. Eldridge et al. TM had previously reported that microparticles of 1 10 ~tm in diameter with entrapped staphylococcal enterotoxin B toxoid (SEB) induced more potent antibody responses than microparticles > 10/~m. This was thought to be due to the ability of microparticles < 10 ~m to be phagocytosed and transported to the draining lymph nodes in macrophages 1°. In the present study, the effect of microparticle size on the serum IgG response induced by entrapped OVA was assessed. Two batches of microparticles were prepared, with entrapped OVA, that were < 5 # m and >50/~m in size. Furthermore, the adjuvant effect achieved by the adsorption of OVA to preformed microparticles was assessed, to determine the need for the entrapment of antigen in microparticles. In the studies of Eldridge et al. TM, SEB mixed with 'empty' microparticles before immunization did not induce greater antibody responses than immunization with SEB alone. This was surprising, since the adjuvant effect achieved by the adsorption of antigens to microparticles had been previously reported on a number of occasions 11,12 MATERIALS AND METHODS

Animals Male Balb/c mice (Olac Ltd, Cirencester, UK) aged 8 weeks and weighing about 25g were used and maintained on a normal mouse diet throughout the studies.

larger volume of an aqueous solution of 10% w/v poly vinyl alcohol (PVA) (88% hydrolysed, Aldrich Chemical Company, Poole, Dorset, UK) and homogenized to produce a stable w/o/w double emulsion. The double emulsion was stirred overnight at ambient temperature and pressure to allow solvent evaporation to proceed, with resultant microparticle formation. Polymer 1 was also used to prepare the microparticles for the second study, whi,:h was designed to assess the effect of particle size on the immunogenicity of microparticles with entrapped OVA. The emulsification conditions were modified to allow the production of microparticles > 50/~m. Microparticles (1.0/tm) without entrapped OVA were also prepared with polymer 1 for this study and OVA was adsorbed to their surface before administration. After preparation, the microparticles were collected by centrifugation, washed three times to remove nonentrapped OVA and freeze-dried.

Determination of the degradation rate of microparticles To determine the rates of degradation of polymers 1 3 used for microparticle preparation, microparticles without entrapped OVA were prepared using each polymer. Each batch of microparticles was accurately weighed into a number of vials and incubated in phosphate-buffered saline (PBS) under constant stirring at 37°C. At various time intervals, vials were removed and the microparticles were recovered, freeze-dried and reweighed. The weight loss of the microparticles over time was recorded to determine the relative rates of degradation for each polymer.

Determination of OVA entrapment in microparticles The amount of antigen in the microparticles was determined in a bicinchoninic acid (BCA) protein assay (Sigma) following disruption of the microparticles and extraction of the entrapped protein using two established methods. Either an aliquot of the microparticles was dissolved in dichloromethane (HPLC Grade, Aldrich) and the protein was extracted into distilled water, as previously described 7'1°'14, or alternatively, the microparticles were dissolved in sodium hydroxide and sodium dodecyl sulfate, as previously described 15.

Mieroparticle preparation Controlled-release microparticles with entrapped OVA (Grade V, Sigma Chemical Company, Poole, Dorset, UK) were prepared using three poly(D,L-lactide-coglycolide) polymers (Resomers RG506, RG508 and RG755, Boehringer Ingelheim KG, Ingelheim, Germany). The three polymers had different ratios of lactide/ glycolide, or different molecular weights as follows; polymer 1, lactide/glycolide ratio 50/50 (molecular weight, 22 kDa ); polymer 2, lactide/glycolide ratio 50/50 (40 kDa); and polymer 3 lactide/glycolide ratio 75/25 (18kDa). Microparticles were prepared from each polymer using a water-in-oil-in-water (w/o/w) solvent evaporation technique as described by Jeffery et al. 13. Briefly, a 6% w/v solution of the polymer in dichloromethane (HPLC grade, May and Baker, Dagenham) was emulsified together with a 6% w/v solution of OVA in double-distilled water using a Silverson homogenizer (Silverson Machines Ltd, Chesham, Bucks, UK) to produce a w/o emulsion. This emulsion was added to a

966

Vaccine, Vol. 11, Issue 9, 1993

IMMUNIZATION PROTOCOLS

Controlled-release microparticles Two groups of ten mice were each immunized subcutaneously with a single 300/zg dose of OVA, either dissolved in PBS or entrapped in PLG microparticles. Immediately before administration, the required dose of freeze-dried microparticles was weighed and resuspended in PBS. One group of mice received 100/~g OVA entrapped in each of the three different PLG polymer compositions. Hence, the total dose of OVA administered in microparticles was 300/~g. A third group of six mice were immunized subcutaneously with a single 300 #g dose of OVA adsorbed to a 2% suspension of aluminium hydroxide (Alugel-S, Serva, Heidelberg, Germany). Blood samples were collected from the tail veins of the mice at 2-week intervals for 16 weeks after immunization. Thereafter, blood samples were collected from the tail veins at 4-week intervals until the termination of the

Controlled-release microparticles as vaccines: D.T. O'Hagan et al.

study. The blood samples were centrifuged and serum was collected and stored frozen at - 4 0 ° C until assayed in an ELISA. Sampling from the group of mice receiving soluble OVA was terminated at study week 16, because the serum IgG antibody levels had fallen to pre-immune levels.

Effect of particle size on the immunogenicity of OVA entrapped in microparticles Four groups of ten mice were each immunized subcutaneously with 100 #g OVA dissolved in PBS, entrapped in P L G microparticles ( 1.5 #m), entrapped in P L G microparticles (72.6/~m) or adsorbed overnight at room temperature to the surface of P L G microparticles (1.0 #m). The group of mice immunized with microparticles with adsorbed OVA received the same amount of microparticles (21.7 mg) as the group receiving microparticles ( 1.5 #m ) with entrapped OVA. Identical booster immunizations of OVA were administered to each study group 6 weeks after the primary immunizations. Blood samples were collected from the tail veins of the mice at 2-week intervals for 12 weeks after primary immunization. The blood samples were centrifuged and the sera were collected and stored frozen at - 4 0 ° C until assayed in an ELISA.

Measurement of IgG by ELISA The specific anti-OVA IgG antibody content of each serum sample was determined in an established ELISA as previously described 16 and standardized against a positive control antiserum obtained by hyperimmunization of a mouse with OVA in Freund's complete adjuvant (FCA) 16. Each serum sample from each mouse was assayed at four different dilutions. The results are expressed as mean antibody units for the groups of mice, calculated from the standard curve obtained from the hyperimmune mouse serum diluted between 1/500 and 1/64000. The value for each dilution falling in the standard curve and its value taken as the mean of the four separate dilutions.

Statistical analysis The results are expressed as mean -t- s.e. for each group of mice. An unpaired Student's t test was used to compare the means for each study group at the different sample times and to assess statistical significance. Results were considered statistically significant if p < 0.05.

OVA entrapment in microparticles The microparticles prepared with the three different P L G polymers and used in the controlled-release study contained the following amounts of OVA: polymer 1, 5.7% w/w; 2, 1.0% w/w; and 3 3.9% w/w. The volume mean diameters of the microparticles prepared from the three polymers as measured by laser diffractometry were: polymer 1, 2.5 #m; 2, 1.0 ~m; and 3 1.8 #m (Malvern laser sizer 2600D, Malvern Instruments, Malvern, UK). The microparticles used to assess the effects of particle size on immunogenicity contained the following amounts of OVA: 1.5 pm, 4.6% w/w and 72.6 pm, 1.3% w/w.

Controlled-release microparticles In comparison with the responses to soluble OVA, the mean serum IgG antibody response to OVA entrapped in microparticles was significantly enhanced from 2 weeks after immunization. The serum IgG response to OVA entrapped in microparticles peaked at week 10 and remained high for the full duration of the study. During the early stages of the study, up to about week 10, the response to OVA entrapped in microparticles was significantly greater than the response to OVA adsorbed to Alugel. Thereafter, there was no significant difference between the two groups (Figure 1 ).

Immunogenicity of OVA in microparticles: the effect of particle size From week 4 onwards, the responses to OVA entrapped in microparticles of both sizes (1.5 pm and 72.6 ktm) and also the response to OVA adsorbed to microparticles, were significantly greater than the response to soluble OVA. Following booster immunizations, the response to microparticles (1.5#m) was significantly greater than the response to larger particles (72.6 #m ). There was also a significant difference between these two groups 4 weeks after the primary immunization. Following booster immunizations, at weeks 8 and 10, the response to OVA entrapped in microparticles (1.5 #m) was significantly greater than the response to OVA adsorbed to microparticles (1.0 pm). At all other weeks, there was no significant difference between these two groups (Figure 2). 800

c

RESULTS

600

>. "0

Degradation rate of microparticles c

Microparticles prepared from polymer 1 showed a 50% weight loss in 40 days. Polymer 2 microparticles showed a 50% weight loss in 55 days, while polymer 3 microparticles showed only a 25% weight loss in 140 days. Polymer degradation was associated with a deterioration in particle-surface morphology which was visible using scanning electron microscopy. As the degradation study proceeded, the microparticles lost their spherical appearance and collapsed structures became visible (data not shown). Further details of the microparticle degradation profiles will be published elsewhere.

E

400

200

0

~

0

2

4

8 10 12 14 16 20 24 28 32 36 40 44 48 52 Week

Figure I Serum IgG antibody responses to single doses of 300/~g OVA administered subcutaneously as ( [ ] ) soluble OVA, ( • ) OVA entrapped in poly(lactide-co-glycolide) microparticles and ( [ ] ) OVA adsorbed to aluminium hydroxide adjuvant (Alugel)

Vaccine, Vol. 11, Issue 9, 1993

967

Controlled-release microparticles as vaccines: D.T. O'Hagan et al. 800

2

600

'~

400

% u~

2oo

0

2

4

6

8

10

12

Week Figure 2 Serum IgG antibody responses to primary and secondary immunization of 100/~g OVA administered subcutaneously as ([Z]) soluble OVA, (11) OVA entrapped in poly(lactide-co-glycolide) (PLG) microparticles (1.5/~m), (IN) OVA entrapped in PLG microparticles (73/~m) and ( [ ] ) OVA adsorbed to PLG microparticles (1.0/~m). Each group of mice received a booster immunization at study week 6

DISCUSSION The induction of a serum IgG antibody response of lasting duration following a single immunization with antigen entrapped in microparticles is encouraging. The results obtained in the present study serve to illustrate the potential of P L G microparticles as controlled-release antigen delivery systems. It is likely that in the near future several currently available vaccines, including tetanus toxoid, will be formulated into PLG microparticles and will be assessed in humans for their ability to induce long-term immunity following a single immunization 3. The use of controlled-release microparticles as vaccines may obviate the need for booster doses of vaccines. Clearly, this would be a considerable advantage, particularly in the developing world. Nevertheless, many issues remain to be addressed concerning the potential use of microparticles as vaccines, including the effects of microparticle entrapment on antigen integrity, vaccine safety, sterility of microparticles, manufacturing scale-up and the costs of manufacture. In relation to the first of these issues, analyses by polyacrylamide gel electrophoresis, isoelectric focusing and Western blotting have indicated that OVA is not significantly altered following entrapment in PLG microparticles 13. In selecting Alugel as the adjuvant of choice for these studies, we were aware that in previous studies in mice with a similar globular protein antigen, bovine serum albumin, aluminium hydroxide had been shown to possess similar adjuvant activity to FCA ~7. In the present study, aluminium hydroxide (Alugel) was able to induce long-lasting immunity following a single immunization. The ability of aluminium hydroxide to induce potent long-term antibody responses following a single immunization is consistent with findings from studies in humans TM It should be emphasized that the three polymers used in the present study, P L G copolymers with molecular weights up to 40 kDa, were polymers that degraded relatively rapidly. The serum IgG antibody response to

968

Vaccine, Vol. 11, Issue 9, 1993

OVA entrapped in microparticles displayed a single peak at week 10, which was due to the release of the majority of entrapped antigen by this time (Jeffery, H., Davis, S.S. and O'Hagan, D.T., unpublished observations). Nevertheless, subsequent studies will use polymers that degrade more slowly, and will be designed to provide pulsatile antigen release for several months after immunization. Pulsatile-release profiles may be designed to mimic booster doses of vaccines. Homopolymers of poly-Llactides (PLA) are commercially available with molecular weights of up to several hundred kilodaltons. These PLA polymers would degrade over more than 1 year after immunization and antigen release could be controlled for this period. Hence, PLG and PLA polymers offer considerable flexibility in release profiles and also a much longer duration of antigen release than aluminium adjuvants. In addition, entrapment of antigens in a range of PLG and PLA polymers allows the option of delivering booster doses of antigen at the same time as the primary dose. This is not possible with aluminium adjuvants. Furthermore, although aluminium adjuvants are generally regarded as safe 9'18, the range of antigens for which they are effective is restricted and, importantly, they do not induce cell-mediated immunity 19. Preliminary studies have indicated that the entrapment of antigens in microparticles does result in the induction of cellmediated immunity z°. Aluminium-adsorbed vaccines cannot be lyophilized and therefore require refrigeration. The use of microparticles as vaccines would eliminate the need for a cold chain, since the product would be freeze-dried. In addition, problems of variability between different batches of aluminium adjuvants have been noted 21. The ability of an antigen entrapped in PLG microparticles to induce greater antibody responses than antigen adsorbed to an aluminium adjuvant has recently been confirmed in a study in mice with tetanus toxoid 22. The ability of microparticles (1.5/~m ) to induce more potent antibody responses than microparticles (72.6/~m ) was expected. This result supports the findings of Eldridge et al. ~°, who suggested that 1-10/~m microparticles were more effective for the induction of potent antibody responses due to their ability to be phagocytosed by macrophages and carried to the local lymph nodes. Although the 72.6 #m microparticles were too large to be phagocytosed, they still induced significantly enhanced antibody responses in comparison with soluble OVA. This may be partly due to controlled release of antigen from the microparticles, but may also be partly due to the phagocytosis of fragments of particle with associated antigen, as microparticle degradation and breakdown occurred. The ability of OVA adsorbed to microparticles to induce significantly enhanced responses was also expected, since the adsorption of proteins to microparticles had previously been shown to be an effective means of promoting antibody responsesL1'12 The absorption conditions employed in the present study, 100/~g OVA adsorbed to 21.7 mg particles overnight at room temperature, would be expected to result in the complete adsorption of the OVA to the microparticles (T.I. Armstrong, personal communication). Thus, the OVA would be expected to be delivered into macrophages through phagocytosis of the microparticles. However, entrapment of OVA in 1.5 #m microparticles resulted in significantly enhanced responses in comparison with adsorption of OVA to 1.0/~m microparticles. This may be due to the ability of microparticles to protect

C o n t r o l l e d - r e l e a s e m i c r o p a r t i c l e s as vaccines: D.T. O ' H a g a n et al.

OVA against degradation in vivo, or it may result from desorption of OVA from microparticles in vivo. Both of these situations would result in a reduction in the delivery of intact OVA into macrophages, which may have resulted in the reduced responses to adsorbed OVA. However, the controlled release of OVA entrapped in 1.5 #m microparticles may also have contributed to the enhanced responses for this group of mice. It should be noted that although adsorption of antigens to microparticles appears to be an effective means of inducing potent antibody responses, the mechanism of controlled release from PLG microparticles cannot function to the same extent for material that is adsorbed rather than entrapped. Therefore, one of the main attractions and advantages of microparticles as potential vaccines is negated, if the antigens are adsorbed rather than entrapped. In the study of Eldridge et al. 1°, immunization with SEB (50/tg) mixed with 2.8 mg of microparticles (1 8/~m) did not result in significantly enhanced antibody responses. However, the adsorption conditions were not stated and if mixing was only undertaken immediately before administration, adsorption of the SEB to microparticles may not have had time to occur. ACKNOWLEDGEMENTS Vaccine development research at Nottingham is currently funded by the World Health Organization. H.J. is the recipient of a research studentship from the Science and Engineering Research Council (SERC).

6 7 8 9 10

11 12 13

14

15 16 17

REFERENCES 1 2 3 4 5

Editorial. WHO 1991, 69, 791 Bektimirov, T., Lambert, P.-H. and Torrigiani, G. Vaccine development perspectives of the World Health Organisation. J. Med. Virol. 1990, 31, 62 Aguado, M.T. and Lambert, P.-H. Controlled release vaccines Biodegradable polylactide/polyglycolide (PL/PG) microspheres as antigen vehicles. Immunobiology 1992, 184, 113 Reul, G.J. Use of Vicryl (polyglactin 910) sutures in general surgery and cardiothoracic procedures. Am. J. Surg. 1977, 134, 297 Christel, P., Chabot, F., Leray, L.F., Morin, C. and Vert, M. Biodegradable composites for internal fixation. In: Biomaterials ( Eds

18 19 20 21 22

Winter, G.D., Giobonnes, D.F. and Plenk, H.) John Wiley, New York, 1982, p. 271 Maulding, H.V. Prolonged delivery of peptides by microcapsules. J. Controlled Release 1987, 6, 167 Cohen, S., Yoshioka, T., Lucarelli, M. Hwang, L.H. and Langer, R. Controlled delivery systems for proteins based on poly(lactic/ glycolic) microspheres. Pharm. Res. 1991, 8, 713 Vert, M., Li, S. and Garreau, H. More about the degradation of LA/GA-derived matrices in aqueous media. J. Controlled Release 1991, 16, 15 Edelman, R. Vaccine adjuvants. Rev. Infect. Dis. 1980, 2, 372 Eldridge, J.H., Staas, J.K., Meulbroek, J.A., Tice, T.R. and Gilley, R.M. Biodegradable and biocompatible poly(oL-lactide-coglycolide) microspheres as an adjuvant for staphylococcal enterotoxin B toxoid which enhances the level of toxin-neutralising antibodies. Infect. Immun. 1991, 59, 2978 Kreuter, J., Liehl, E., Berg, U., Soliva, M. and Speiser, P.P. Influence of hydrophobicity on the adjuvant effect of particulate polymeric adjuvants. Vaccine 1988, 6, 253 O'Hagan, D.T., Palin, K.J. and Davis, S.S. Poly(butyl-2-cyanoacrylate) particles as adjuvants for oral immunization. Vaccine 1989, 7, 213 Jeffery, H., Davis, S.S. and O'Hagan, D.T. Preparation and characterisation of poly(lactide-co-glycolide) microparticles. I1. Water-in-oil-in-water emulsion solvent evaporation. Pharm. Res. 1993, 10, 362 Eldridge, J.H., Hammond, C.J., Meulbroek, J.A., Staas, J.K., Gilley, R.M. and Tice, T.R. Controlled vaccine release in the gut-associated lymphoid tissues. I. Orally administered biodegradable microspheres target the Peyer's patches. J. Controlled Release 1990, 11,205 Hora, M.S., Rana, R.K., Nunberg, J.H., Tice, T.R., Gilley, R.M. and Hudson, M.E. Release of human serum albumin from poly(lactideco-glycolide) microspheres. Pharm. Res. 1990, 7, 1190 O'Hagan, D.T., Jeffery, H., Roberts, MJ.J., McGee, J.P. and Davis, S.S. Controlled release microparticles for vaccine development. Vaccine 1991, 9, 768 Bomford, R. The comparative selectivity of adjuvants for humoral and cell-mediated immunity. Clin. Exp. Immunol. 1980, 39, 426 Aprile, M.A. and Wardlaw, A.C. Aluminium compounds as adjuvants for vaccines and toxoids in man: A review. Can. J. Publ. Health 1966, 57, 343 Warren, H.S. and Chedid, L.A. Future prospects for vaccine adjuvants. CRC Crit. Rev. Immunol. 1968, 8, 83 O'Hagan, D.T., McGee, J.P., Holmgren, J., Mowat, McI.A., Donachie, A.M., Mills, K.H.G. et al. Biodegradable microparticles as oral vaccines. Vaccine 1933, 11, 149 Hem, S.L. and White, J.L. Characterization of aluminium hydroxide for use as an adjuvant in parenteral vaccines. J. Parenter. Sci. Technol. 1984, 38, 2 Esparza, I. and Kissel, T. Parameters affecting the immunogenicity of microencapsulated tetanus toxoid. Vaccine 1992, 10, 714

Vaccine, Vol. 11, Issue 9, 1993

969