Biodegradable microspheres as controlled-release tetanus toxoid delivery systems

M a r i a J. A l o n s o *~;, R a j e s h K. G u p t a t, C a r o l i n e M i n * , G e o r g e R. S iber t and R o b e r t L a n g e r *§

Purified tetanus toxoid, a high-molecular-weight protein, was entrapped within poly(L-lactic acid) (PLA) and poly(o,L-lactic/glycolic acid) (PLGA) microspheres prepared by either a solvent extraction or a solvent evaporation method carried out in a multiple emulsion system (water-in-oil-in-water). The physical integrity and anti#enicity of the protein treated under different processing conditions were investigated. A reduction of antigenicity that was related to the percentage of aggregated protein was noticed under some experimental conditions. This partial loss of antigenicity was associated with the lyophilization process and affected by the nature of the organic solvent. All types of microspheres prepared with different molecular weight PLA and PLGA displayed a high protein-loading efficiency (>80%) but their size was strongly influenced by polymer molecular weight (3000 versus 100000). Protein release pattern was influenced by both polymer molecular weight and composition (PLA versus PLGA). A constant release pattern after an induction period of 10 days was observed for microspheres composed of hi#h-molecular-weight polymers ( PLA and PLGA ). The release rate was lower from PLA microspheres than from PLGA microspheres. In contrast, a continuously increasing release rate preceded by a burst was observed for low-molecular-weight (3000) PLGA microspheres. Microencapsulated tetanus toxoid was significantly more immunogenic in mice than fluid toxoid as determined by IgG anti-tetanus antibody levels and neutralizing antibodies. However, the magnitude and duration of the antibody response did not differ significantly from a similar dose of aluminium phosphate-adsorbed toxoid. We conclude that microencapsulated tetanus toxoid shows significant adjuvant activity. Further improvements in the formulation of microspheres which result in the release of higher concentrations of antiyenically active tetanus toxoid for more prolonged periods may result in higher and more sustained antibody levels.

Keywords: Vaccine delivery system; poly(lactic acid); poly(lactic/glycolic acid); neutralizing antibodies

Controlled delivery of bioactive macromolecules has become increasingly important 1. Vaccines are an important class of molecules that could potentially benefit from controlled release because many immun- ization schedules require administration of several spaced antigen doses. The use of polymers to control the release of an antigen to stimulate the immune response was first reported in 19792 . More recently, several studies have indicated the potential of biodegradable microspheres to enhance the immunogenicity of poorly immunogenic

*Department of Chemical Engineering, Massachusetts Institute of Technology, Cambridge, MA 02139, USA. +Massachusetts Public Health Biologic Laboratories, Boston, MA 02130, USA. tPresent address: Department of Pharma- ceutical Technology, School of Pharmacy, Santiago de Compostela 15706 Spain. §To whom correspondence should be addressed. (Received 4 February 1993; revised 7 June 1993; accepted 8 June 1993)

molecules 3'4, and specific vaccines s-7. In particular, microspheres based on poly(lactic acid) (PLA) and poly(lactic/glycolic acid ) (PLGA), polymers approved for human administration, seem to be a practical approach for antigen delivery. However, in spite of promising initial results, little information concerning the design of antigen-loaded microspheres with specific size and controlled-release properties is available. Among the most critical design parameters to be investigated are the effects of polymer molecular weight and composition on in vitro release behaviour and immunogenicity of encapsulated antigens. In addition, a major challenge in the formulation of microencapsulated vaccines is to maintain their integrity and antigenicity. However, this issue has been only indirectly explored in previous research.

The overall goal of this study was to encapsulate a high-molecular-weight protein antigen, tetanus toxoid, in order to control its release and consequently to prolong

0264-410X/94104/0299-08 © 1994 Butterworth-Heinemann Ltd Vaccine 1994 Volume 12 Number 4 299

Biodegradable tetanus toxoid microspheres: M.J. Alonso et al.

its immune response. The selection of this antigen was based on the practical consequences related to the requirement of repeated administration. According to the World Health Organization, prevention of neonatal tetanus by maternal immunization requires at least two doses of alum-adsorbed tetanus toxoid in previously non-immunized women. The logistics of delivering two doses of tetanus toxoid during pregnancy are difficult and compliance is frequently inadequate. As a conse- quence tetanus is endemic in 90 countries throughout the world and in some regions is responsible for 25% of infant mortality 8. Therefore the design of a vaccine delivery system that would provide an effective single-injection immunization is highly desirable.

In this paper we describe a study which defines experimental conditions to efficiently encapsulate and release tetanus toxoid as well as results of an in vivo study conducted to investigate the immunogenicity of the formulations developed. In a previous study we observed the importance of vaccine purity for the physical and controlled-release properties in vitro of PLA and PLGA microspheres 9. In the present study, to obtain accurate information about the efficiency of the microencapsulation process and the mechanism of protein release, highly purified monomeric tetanus toxoid protein was used.

M A T E R I A L S A N D M E T H O D S

Materials Poly(L-lactic acid ) (PLA) (M w 3000 and 100 000) and

poly(vinyl alcohol) (PVA) (Mw 25000, 88 mol% hydrolysed) were obtained from Polysciences Inc., Warrington, PA. Poly(D,L-lactic/glycolic) acid 50:50 (PLGA) (M w 100000) and poly(D,L-lactic/glycolic) acid 50:50 (PLGA) (M w 3000) were supplied respectively from Boehringer, Ingelheim, Germany and from Wako Chemical, Inc., Richmond, VA. Pluronic F 68 was supplied by BASF Co., Parsippany, NJ. An aqueous solution (sodium chloride 0.9%) of tetanus toxoid chromatographically purified was used in this study. This solution contains 3600 Lf ml-1 (Limes flocculation (Lf) is the International Unit for tetanus toxoid ) and a protein concentration of 9.0 mg ml- 1.

Chromatographically purified tetanus toxoid was prepared by modifying a previously described technique lO. Aluminium phosphate-adsorbed tetanus toxoid was prepared by addition of the toxoid to freshly prepared aluminium phosphate gel (pH 6.0) made by precipitation of aluminium chloride and trisodium phosphate. The aluminium phosphate-adsorbed toxoid contained 10Lfm1-1 tetanus toxoid and 4mgm1-1 aluminium phosphate.

Animals

Female CD-1 mice weighing about 20 g (six to nine mice per group) from Charles River Breeding Laboratory, Wilmington, MA, were used and maintained on a normal diet throughout the study.

Microsphere preparation

Microspheres were prepared by either the solvent evaporation or the solvent extraction methods carried out in a double emulsion system. The solvent evaporation technique has already been applied to the encapsulation

of classical drugs and peptides T M and was recently reported by our laboratory for the encapsulation of proteins 13. Briefly, 50#1 of a saline tetanus toxoid solution (3600Lfml -~) was emulsified in 1 ml of an organic solvent (methylene chloride or ethyl acetate) containing 200 mg of PLA or PLGA by sonication at output 4 (50 W) for 10 s (ultrasonic probe, Sonics & Materials Inc.) or homogenization at 15 000 rev min- for 10s (Omni 2000 homogenizer). The resulting emulsion was dispersed, by agitation for 10 s (vortex, maximum speed), in 1 ml of an aqueous solution of PVA ( 1% ) and then diluted in 100 ml of PVA aqueous solution (0.3%). The system was maintained under magnetic stirring for 3 h to allow solvent evaporation. When the solvent extraction technique was applied, 5 min after the formation of the double emulsion, 200 ml of an aqueous solution ofisopropanol (2% v/v) was added to accelerate extraction of the solvent to the external aqueous phase. In this case the system was maintained under magnetic stirring for 30 min. The microspheres, formed as a consequence of polymer precipitation, were collected, washed with double-distilled water and freeze-dried.

Protein stability study

The effects of different conditions typical of the microencapsulation process on the physical characteristics and antigenicity of tetanus toxoid were evaluated (Table 1). Briefly, 100#1 of tetanus toxoid aqueous solution were emulsified by sonication or homogenization (conditions specified above) into 1 ml of an organic solvent (methylene chloride or ethyl acetate) (samples 2, 3, 4, 5, 7, 9). In some of these emulsions a hydrophilic stabilizer (Pluronic F68, PEG 4600 or sodium glutamate) was added to the protein solution. Control samples lacking the organic solvent were also analysed (samples, 1, 6, 8, 10). All preparations (emulsions and solutions) were freeze-dried and then dissolved in 1 ml of phosphate-buffered saline (PBS) solution, filtered through a 0.22 #m filter (Millex-GV, Millipore, Mildford, MA) and diluted as required for analysis.

The aggregation and antigencity of the protein were determined by three different techniques:

1 High-performance liquid chromatography (HPLC) (Waters Millipore, Mildford, MA). Tetanus toxoid samples were eluted with PBS, pH 7.4, through a gel column (TSKgel G3000PWx,TosoHaas, Philadelphia,

Table 1 Aggregation extent of tetanus toxoid after different experimental conditions inherent to the microencapsulaUon process

Samp le Solvent" Emuls i f ica t ion ~ S tab i l i ze r ° % Aggregated

1 - - - 14.0 2 MC SC - 13.11/3.0 ~ 3 EA SC - 5.4 4 MC HG - 13.5/5.3" 5 MC SC PI. F 68 9.6/5.9" 6 - - PI. F 68 0.0 7 MC SC PEG 9.0/4.3 ~ 8 - PEG 3.9 9 MC SC SG 3.4/6.7 ~

10 - - SG 4.4

aMC, methylene chloride; EA, ethyl acetate ~SC, sonication; HG, homogenization °PI. F 68, Pluronic F 68 (1%); PEG, PEG 4600 (0.4%); SG, sodium glutamate (1%) ~Two peaks in the chromatogram: o l igomer ic /po lymer ic protein

300 Vaccine 1994 Volume 12 Number 4

PA). The percentage of aggregated protein was determined in each chromatogram by comparing the area under the peak corresponding to aggregated protein with that corresponding to monomeric plus aggregated protein.

2 Sodium dodecyl sulfate-polyacrylamide gel electro- phoresis (SDS-PAGE). Tetanus toxoid samples were analysed on a 4-12% Tris-glycine gel (Novel Experimental Technology, San Diego, CA) under reducing conditions using 2-mercaptoethanol and stained with Coomassie blue (Bio-Rad Laboratories, Richmond, CA )14.

3 Double immunodiffusion. Tetanus equine serum (Massachusetts Public Health Biologic Laboratories, Boston, MA) (15 #1) and the same volume of the tetanus toxoid samples were loaded in the central well and in the outer wells in a 1% agarose gel, respectively 15

Protein encapsulation efficiency

To determine this parameter, tetanus toxoid was radiolabelled with 1251 by using the IODOBEADS technique (Pierce, Rockford, IL). Specific activity was 0.05 mCi mg- 1. The loading efficiency was calculated by referring the percentage of encapsulated labelled protein to the total amount of labelled protein used to prepare the microspheres (actual loading x l00/theoretical loading).

Protein release studies

Microspheres (30mg) were placed in 5 ml tubes and then incubated in 3 ml phosphate buffer, pH 7.4, under agitation at 37°C. The samples were periodicall X collected and centrifuged for 20 min at 60000 (Sorvall ~ RC centrifuge, Du Pont Instruments), and the supernatants analysed for the released toxoid. The amount of toxoid was determined by a microBCA protein assay (Pierce, Rockford, IL) and the percentage released was calculated with respect to the theroetical loading. Release experiments were done independently in triplicate.

Polymer degradation studies

Molecular weights of PLA and PLGA before and after different incubation times were determined on a Perkin-Elmer GPC system with a refractive index detector. Samples of fresh microspheres and micro- spheres from the degradation experiments were freeze- dried, dissolved in chloroform, and filtered through a 0.22/zm filter (MilIex-FGS, Millipore, Mildford, MA). The samples were eluted with chloroform through a Phenogel column (linear 0-10000 K, mixed bed) (Phenomenex). The molecular weight was determined relative to polystyrene standard (Polysciences, Mw range: 1250-233 000).

Particle size and morphology

These microsphere properties were examined before and during degradation studies by scanning electron microscopy (s.e.m.) (Cambridge Instruments, 250 MK or Amray AMR 1000A). Samples for s.e.m, were freeze-dried, mounted on metal stubs and coated with

Biodegradable tetanus toxoid microspheres: M.J. Alonso et al.

gold to a thickness of 200-500 A,. Photographs were taken and the mean size and standard deviation were determined by measuring the diameter of 50-100 microspheres according to a reference scale.

Immunization protocol

Four microsphere formulations prepared with PLA and PLGA of molecular weight 3000 and 100000 were selected for this immunization study. Groups of female CD-1 mice (six to nine mice per group) were injected subcutaneously (22-gauge needle) on the left side of the abdomen with a single dose of tetanus toxoid- containing microspheres. The same single dose (5 Lf) of fluid toxoid in 0.9% sodium chloride and aluminium phosphate- adsorbed tetanus toxoid were studied as controls. Prior to injection, 7 mg of microspheres were suspended in 0.5 ml of an aqueous vehicle (0.5% sorbitol, 0.1% carboxymethylcellulose, 0.02% Tween 80). Blood was collected, by cardiac puncture, several times after administration of a single immunizing dose and the sera separated by centrifugation. The sera of mice from each vaccine group were pooled and assayed for tetanus antitoxin in terms of antitoxin units (AU) by the toxin neutralization test 16.17. The values of A U ml-1 of serum samples were determined against US standard tetanus antitoxin (lot no. E134, obtained from the Center for Biologic Evaluation and Research, FDA, Bethesda, MD). Individual mice sera were also evaluated for IgG antibodies to tetanus toxin by the enzyme-linked immunosorbent assay (ELISA). Briefly, high-binding easy-wash 96-well microtitre plates (Corning Glass Works, Corning, NY) were coated with 100 pl of purified tetanus toxin diluted to 5 #g m1-1 in PBS, pH 7.2 at room temperature (RT) overnight. The plates were washed three times with PBS containing 0.05% Tween 20 between each step. Hyperimmune anti-tetanus toxoid mouse serum containing 225 AU ml- 1 was provided by the Laboratory of Development of Molecular Immunity, NIH, Bethesda, MD, and used as a reference for the ELISA and included in each plate. The samples and the reference serum were serially diluted at twofold dilution steps in the plates using PBS with 0.1% Brij 35 and 0.5% bovine serum albumin as a diluent (PBB). The plates were kept at RT for 2 h and washed. Goat anti-mouse IgG alkaline phosphatase conjugate (Fisher Bio- technology, Springfield, N J) diluted 1:1000 in PBB was added to plates. The plates were again incubated at RT for 2 h and washed. Finally p-nitrophenyl phosphate (Sigma Chemical Co., St Louis, MO) diluted to 1 mg ml- ~ in 1 M diethanoleamine, 0.5 mM magnesium chloride buffer was added to plates. The plates were kept at RT for 30 min and read at 405 nm wavelength on an ELISA reader. The unitage of the serum samples in ELISA Antitoxin Units (EAU)m1-1 were calculated against hyperimmune mouse serum by extrapolation from a standard curve.

RESULTS

Polymer characterization

As determined by gel-permeation chromatography the number-average molecular weights of the polymers used in this study were 3000 and 100000 for both PLA and PLGA.

Vaccine 1994 Volume 12 Number 4 301

Biodegradable tetanus toxoid microspheres: M.J. Alonso et al.

Stability of tetanus toxoid

Results shown in Table 1 indicate that aggregation takes place during freeze-drying and is influenced by the type of solvent in which the aqueous protein solution is dispersed. Only a very small peak corresponding to dimeric protein appears after exposure to ethyl acetate but two peaks corresponding to dimeric and oligomeric protein are noticeable when methylene chloride was used. Aggregation extent was very low for all processing conditions, except for the treatment with methylene chloride without stabilizers. When the aqueous solution of the protein was exposed directly to methylene chloride and then freeze-dried, the resulting protein powder was not completely soluble in water. On the other hand, all three stabilizers reduced aggregation of the toxoid during freeze-drying. In particular, Pluronic F68 totally eliminates protein aggregation. However, when the toxoid was exposed to methylene chloride, the protective effect of Pluronic F 68 was not observed. Results obtained by SDS- PAGE are closely related to those obtained by size-exclusion chromatography. Figure la shows some of the results (samples 1 to 7) obtained by SDS-PAGE for the treated protein. The band of aggregated protein was more intense for those samples whose chromatographic peak representing aggregation was bigger. This suggests that the mechanism of aggregation involves a chemical interaction between monomeric protein units. Figure lb displays the precipitation lines obtained with antibody and treated tetanus toxoid (samples 1 to 4) in an agarose gel. The intensity of the lines for samples 1 to 10 (5 to 10 not shown) was inversely related to the aggregation extent of the protein. As shown in Figure lb, an intense line was observed for the toxoid that was not exposed to a solvent (sample 1) or exposed to ethyl acetate (sample 3 ). In contrast, only faint lines were detected for toxoid exposed to methylene chloride (samples 2 and 4). These differences in antigenicity of the treated toxoid were only detected when the toxoid solution was highly diluted (90 Lfml- 1 ).

Tetanus toxoid loading efficiency of microspheres

The amount of radiolabelled protein encapsulated (actual content) with respect to the total amount of protein used to prepare the microspheres (theoretical content) was largely independent of the polymer characteristics (composition and molecular weight) and solvent elimination procedure (solvent extraction or evaporation). In fact, the loading efficiency (actual content × 100/theoretical content) varied from 80 to 99% depending on the processing conditions.

Characteristics of tetanus toxoid-loaded microspheres

Scanning electron micrographs (not shown) indicate that polymer molecular weight plays an important role in controlling microsphere size. Indeed, the mean size of microspheres prepared with high-molecular-weight PLA or PLGA was approximately 60 + 20 #m. However, the mean sizes of microspheres prepared with PLA and PLGA of molecular weight 3000 were respectively 9 _ 5/~m and 6 ___ 3 #m. The emulsification technique and solvent elimination procedure did not influence micro- sphere size (results not shown). The yield of the microencapsulation process, which refers the quantity of microspheres recovered to the total amount of material (polymer plus protein) used to prepare the microspheres, was above 90%.

Tetanus toxoid release

The graphs in Figures 2a and b compare the release profiles of tetanus toxoid from microspheres composed of high-and low-molecular weight polymers, respectively. The release of tetanus toxoid from PLA and PLGA microspheres (Mw 100000) is characterized by a triphasic release profile (Figure 2a). A small burst (3% release over the first day) was followed by a time lag of 10 days with very slow release and then by a faster constant release. The last phase of release is affected by

a

M

I

t

~.J K

q l t4.4

I 2 ~ 4 5 ~ 7 (" ~t

Figure 1 (a) SDS-PAGE gel of tetanus toxoid before and after various treatments specified in Table 1 (lanes 1-7) (C: control, non-treated toxoid). Wide band represents monomeric protein and narrow band represents aggregate protein. Molecular weight standards are shown in the extreme right lane. (b) Double immunodiffusion of tetanus toxoid purified before and after various treatments specified in Table 1 (C: control). The intensity of the precipitation lines is related to the antigenicity of tetanus toxoid

302 Vaccine 1994 Volume 12 Number 4

Biodegradable tetanus toxoid microspheres: M.J. Alonso et al.

40

30 '

2 0

10 ¸

,-$ M

a

i i ,

10 20 30

80

6 0

40 '

2 0

(3 0 0 0

b

10 20 30

Time, days Time, days

Figure 2 Cumulative in vitro release of tetanus toxoid from different types of polymer microspheres, prepared under the same processing conditions, in PBS at 37°C. (a) (11) PLA (M w 100000); (17) PLGA (M w 100000); (b) (A) PLA (M. 3000); (/k) PLGA (M. 3000)

Figure 3 Scanning electron microscope photographs of tetanus toxoid-loaded PLGA (M w 100000) microspheres at different degradation stages in release medium at 37°C. (a) Immediately after preparation; (b) after 30 days

polymer composition (PLA versus PLGA ), being nearly constant at a rate of 0.61% protein per day for PLA microspheres and 1.17% protein per day for PLGA microspheres. A near constant release rate was observed from PLA (M w 3000) microspheres after a small initial burst (3% of protein released). In contrast, a continuously increasing release rate preceded by a large burst was observed for PLGA (Mw 3000) microspheres (Figure 2b ).

Degradation characteristics of microspheres

As shown in Figure 3, immediately after preparation PLGA microspheres show an overall intact outer surface (a); after one month incubation the microspheres appeared highly eroded and porous (b). In contrast, for PLA microspheres no increase in porosity was observed over a 50-day degradation study (results not shown).

As measured by the reduction in polymer molecular weight, the degradation rate of PLGA microspheres was much faster than that of PLA microspheres. In fact, after incubation for 1 month the number-average molecular weight of PLGA decreased from 100000 to 1000. Also, a very broad polymer molecular weight distribution was observed after degradation, suggesting the co-existence of low- and high-molecular-weight polymer chains. In contrast, the number-average molecular weight of PLA (Mw 100000) decreased only by 10% over the same period (results not shown).

Immunization study

The antibody levels induced in mice after subcutaneous immunization with aluminium phosphate-adsorbed, fluid and encapsulated tetanus toxoid are displayed in Figure 4 and Table 2. Titres of tetanus antitoxin and of IgG antibodies to tetanus toxin following the administration

Vaccine 1994 Vo lume 12 Number 4 303

Biodegradable tetanus toxoid microspheres: M.J. Alonso et al.

Table 2 IgG antibody response by ELISA of mice to various microsphere formulations containing tetanus toxoid, fluid toxoid and aluminium phosphate-adsorbed toxoid

Geometric mean IgG antibody (EAU m1-1) (95% c.i.) after:

Formulation (tool. wt) 4 weeks 8 weeks 13 weeks 26 weeks

PLGA (3000) PLA (3000) PLGA ( 100 000) PLA (100000) Fluid toxoid Aluminium phosphate-adsorbed toxoid

3.23 (1.37-7.60) a 1.19 (0.59-2.40)" 3.01 (2.03-4.45) a 2.39 ( 1.08-5.28)" 0.24 (0.11-0.53) 2.73 (1.70-4.38) =

3.53 ( 1.58-7.88)" 1.04 (0.49-2.21)'~ 2.77 ( 1.71-4.50)" 1.50 (0.72-3.14) "b 0.20 (0.07-0.57) 5.70 (2.76-11.8) =

2.51 ( 1.21-5.21 )" 0.86 (0.50-1.47) "b 2.17 (1.29-3.65)" 1.10 (0.48-2.54) "b 0.21 (0.07-0.66) 5.59 (2.67-11.7)"

2.71 ( 1.22-6.01 )" 0.56 (0.27-1.66) b 1.77 (0.90-5.43)" 1.13 (0.58-2.19) = 0.18 (0.04-0.71) 4.16 (2.20-7.85)"

Female CD-1 mice (six to nine per group) were inoculated subcutaneously with 5 Lf of tetanus toxoid present in microspheres, as a fluid or adsorbed onto aluminium phosphate =p < 0.05 by Dunnett's multiple range test comparing the microencapsulated and aluminium phosphate-adsorbed tetanus toxoids with the fluid toxoid ~p < 0.05 by Dunnett's multiple range test comparing the microencapsulated tetanus toxoids with the aluminium phosphate-adsorbed tetanus toxoid

1 0

Aluminum phosphate ~,~ PLGA, MW 100,000

PLGA, MW 3,000

i -~ PLA, MW 100,000 1

.~ PLA, MW 3,000 .~ Fluid

0 . 1 i , ; i i

0 4 8 13 26

Time (weeks)

Rgure 4 In vivo neutralizing antibody response (antitoxin units/ml) to a single dose (5 Lf) of tetanus toxoid in mice through immunization with four different formulations of microencapsulated tetanus toxoid, fluid toxoid, and aluminium phosphate-adsorbed toxoid. The neutralization titres were determined in the pooled sera of six to nine mice

of aluminium phosphate-adsorbed or encapsulated toxoid reached a maximum at 13 weeks and 4-8 weeks, respectively. The geometric mean IgG antibody levels elicited by microsphere preparations were up to 17 times higher (formulation PLGA Mw 3000) than those obtained after administration of fluid vaccine. The IgG antibody levels achieved after administration of the encapsulated or aluminium phosphate-adsorbed tetanus toxoid were significantly higher than those obtained after injection of fluid vaccine by Dunnett's multiple range test ~8 (p<0 .05) which compared all microsphere preparations and aluminium phosphate-adsorbed vaccine with the fluid toxoid as a control. The most immunogenic PLGA-encapsulated preparations produced similar IgG antibody concentrations to aluminium phosphate- adsorbed tetanus toxoid, but the PLA-encapsulated preparations had significantly lower IgG concentrations at 8 weeks and later (by Dunnett's test, comparing all microencapsulated vaccines with aluminium phosphate- adsorbed toxoid as a control). By the Kruskal-Wallis test 19 comparing all microencapsulated preparations with each other, the differences between the preparations did not reach statistical significance.

D I S C U S S I O N

In this study we investigated the suitability of PLA and PLGA microspheres to encapsulate and slowly release

tetanus toxoid vaccine. Several microencapsulation techniques have been described for the preparation of PLA and PLGA microspheres. Among the procedures to encapsulate water-soluble molecules we chose the multiple emulsion-solvent evaporation technique for its simplicity 11-~3. However, when the encapsulation of sensitive molecules is desired, one of the major limitations of this technique is the long exposure time of the protein to the organic solvent. In our study a way to accelerate solvent elimination (by extraction) and consequently minimize the contact of tetanus toxoid with the organic solvent was investigated. Our results indicate that the solvent extraction method is an efficient way to encapsulate tetanus toxoid. A critical task in developing delivery systems for macromolecules is preserving the activity of these sensitive molecules. In particular, aggregation is a common mechanism of protein instability 2°'2~. In our stability study we observed that ethyl acetate was superior to methylene chloride for microsphere preparation because it produced less protein aggregation. In addition we noticed that aggregation during freeze-drying was prevented by using the non-ionic surfactant Pluronic F 68.

To investigate the importance of the polymer in controlling the release of tetanus toxoid, microspheres based on two types of polymers (PLA and PLGA) with two very different molecular weights (3000 and 100 000) were prepared. The loading efficiency was very high for all formulations developed (>80%). It has been previously reported 22 that drug loss during a micro- encapsulation process using a double emulsion technique occurs because of the migration of the inner water droplets to the external aqueous phase. However, under the experimental conditions presented here, the polymer precipitates rapidly, immobilizing the inner aqueous droplets within the organic droplets.

Physical characteristics of the microspheres such as size and surface appearance were not influenced by the solvent elimination method (extraction versus evapor- ation). However, microsphere size was strongly affected by polymer molecular weight. Indeed, size reduction was observed when the polymer molecular weight decreases from 100000 to 3000. This fact may be explained by the high viscosity of the polymer solution for high-molecular- weight polymers and therefore their more difficult dispersion into the external aqueous phase during the emulsification process.

Release of macromolecules from PLA and PLGA

304 Vaccine 1994 Vo lume 12 Number 4

microspheres has been characterized by a rapid release phase (burst effect) followed by a slow release phase 9'13'23'24. The initial burst, due to the release of protein molecules close to the microsphere surface, has been extremely high (more than 50% in 24h) for previously described polymeric tetanus vaccine formu- lations 7. The second phase, associated with release of protein by diffusion through water-filled pores, has been related to polymer degradation and subsequent erosion of the polymer matrix. It is known that systems based on PLA and PLGA, in an aqueous environment, undergo a hydration process followed by bulk erosion 25. Consequently, during erosion, matrix porosity increases and protein release by diffusion is facilitated. On the other hand, in some cases an intermediate stage of release called the 'induction period' characterized by a very slow release has been observed. This intermediate phase is typical of high-molecular-weight polymers which require a particular degradation time before erosion of the polymer matrix becomes sufficient to release the macromolecules 26.

Our release data are consistent with the pattern described above although, with the exception of PLGA (M w 31300) microspheres, the burst effect was insignificant (3% released over the first day). The low initial release gives an indication of the efficient protein encapsulation. The induction period, comprising a very slow protein release, was noticeable for high-molecular-weight polymers (Figure 2a). After the induction period the release rate was clearly affected by polymer composition, being faster from PLGA than from PLA microspheres. An effect of polymer molecular weight on the release profile was also observed. The higher release rate from low-molecular- weight PLA and PLGA microspheres (Figure 2b) can be attributed to the increase in the hydrophilic regions in these microspheres and also to their smaller size (<10#m) . To understand the mechanism of protein release from the formulations developed we studied the relationship between microsphere degradation and protein release pattern. By comparing microsphere erosion processes (Figure 3) and protein release profiles (Figure 2), a major contribution of the PLGA degradation to microsphere erosion and subsequent tetanus toxoid release was noted. Moreover, the extended protein release from high-molecular-weight PLA is also understandable by the slow degradation rate of this polymer. In contrast, the faster release rate obtained from low-molecular- weight PLA microspheres is not explainable by polymer degradation (only 10% molecular weight reduction was observed) but by their small size and their high water uptake. This high hydration capacity is due to the high number of carboxylic groups in low-molecular-weight polymers.

Previous studies have shown that parenteral immun- ization with tetanus toxoid trapped in liposomes resulted in a potential immune response 27. Two mechanisms are postulated to play a role in this immune enhancement. First, liposomes may serve as a depot for extended release of the antigen and second, they may target the antigen to macrophages resulting in more efficient antigen presentation. Because of the potential long-term in vitro and in vivo instability of liposomes, the present strategy to achieve a long-term vaccine delivery system was focused on the design of biodegradable PLA and PLGA microspheres. We investigated the immunopotentiating activity of microspheres of various sizes and release

Biodegradable tetanus toxoid microspheres: M.J. Alonso et al.

profiles with the final aim of designing a formulation which induces a prolonged immune response against tetanus. Aluminium phosphate-adsorbed tetanus toxoid was also included as a control. We found that each of the microencapsulated preparations induced higher levels of IgG antibody and tetanus antitoxin activity than fluid tetanus toxoid. These levels are considerably higher than the estimated minimum protective levels in humans (0.01 AUml-1) 2s. This clearly demonstrates that micro- encapsulation of this antigen was associated with a significant adjuvant effect. However, none of the preparations induced higher levels of more prolonged tetanus toxoid antibody responses than the aluminium phosphate-adsorbed control. The reason for this result is not clear. One possibility is that the protein released by the microspheres after the initial period of release was denatured and thus no longer antigenically active. As described previously 29, the amount of antigenically active tetanus toxoid (expressed as a proportion of the encapsulated tetanus toxoid ) that was released from PLA microspheres was 0.5-1% and the proportion released from PLGA microspheres was 2.5-20%. This may account for the higher immunogenicity of the PLGA microspheres in our experiments. Further improvement in the immunogenicity of microencapsulated tetanus toxoid may be achievable by developing formulations which release a high proportion of the antigenically active form over a prolonged period of time.

With all tetanus toxoid formulations evaluated in this experiment, IgG antibodies reached a peak at 4 or 8 weeks. In contrast, neutralizing activity peaked at 13 weeks. This observation may be explained by progressive affinity maturation of tetanus antibodies induced by prolonged exposure to low concentrations of tetanus toxoid. However, on the basis of the data presented here no correlation between in vitro release profiles and in vivo antibody titres can yet be established. On the other hand, even though it has been established that the adjuvant capacity of microspheres is related to their particle size 6, results in this study did not reveal this relation. In fact, according to that hypothesis, PLGA (Mw 3000) microspheres, which have a smaller size ( < 10/~m) and release faster than PLGA (M w 100000) microspheres, should display a higher response. However, no significant differences in immunological activity were observed between these formulations. This leads to the conclusion that microsphere size is not the only parameter controlling immune response to encapsulated antigens. Reasonably, a combination of physicochemical properties of the polymer, microsphere size and antigen-release profile will influence the antigen presentation, and consequently the immune response. More extensive in vivo studies will be required to elucidate the ideal characteristics of microspheres for a single-dose formulation of tetanus vaccine. Nevertheless, the results presented here show significant long-term adjuvant activity of PLA and PLGA microspheres for tetanus vaccine.

A C K N O W L E D G E M E N T S

The authors are grateful to Joao Ferreira for his help in the protein iodination. This work was supported by grants from the World Health Organization and the National Institutes of Health (GM26698 and AI 33575). M.J.A. was a visiting scientist on leave from the

Vaccine 1994 Volume 12 Number 4 305

B i o d e g r a d a b l e tetanus toxo id m ic rospheres : M.J. A lonso et al.

University of Santiago de Compostcla under the support of 'Xunta de Galicia' (Spain).

REFERENCES

1 Langer, R. New methods of drug delivery. Science 1990, 249, 1527-1533

2 Preis, I. and Langer, R. A single-step immunization by sustained antigen release. J. Immunol. Methods 1979, 28, 193-197

3 Kohn, J., Niemi, S.M., Albert, E.C., Murphy, J.C., Langer, R. and Fox, J. Single-step immunization using controlled release biodegradable polymer with sustained adjuvant activity. J. Immunol. Methods 1986, 95, 31-38

4 O'Hagan, D.T., Rahman, D., Mcgee, J.P., Jeffery, H., Davies, Williams, P. et al. Biodegradable microparticles as controlled release antigen delivery systems. Immunology 1991, 73, 239-242

5 Singh, M., Singh, A. and Talwar, G.P. Controlled delivery of diphtheria toxoid using biodegradable poly(o,L-lacUde) micro- capsules. Pharmacol. Res. 1991, 8, 958-961

6 Eldridge, J.H., Staas, J.K., Meulbroek, J.A., Tice, T.R. and Gilley, R.M. Biodegradable and biecompatible poly(D,L-Lactide-co-Glyco- lide) microspheres as an adjuvant for Staphylococcal Enterotoxin B toxoid which enhances the level of toxin-neutralizing antibodies. Infect. Immun. 1991, 59, 2978-2986

7 Esparza, I. and Kissel, T. Parameters affecting the immunogenicity of microencapsulated tetanus toxoid. Vaccine 1992, 10, 714-720

8 Galazka, A., Gasse, F. and Henderson, R.H. Neonatal tetanus in the world and the global Expanded Programme of Immunization. In: Proceedings of the Eighth International Conference on Tetanus ( Eds Nistico, G., Bizzini, B., Bytchenko, B. and Triau, R.) Pythagora Press, Rome, 1989, pp. 470-487

9 Alonso, M.J., Cohen, S., Park, T.G., Gupta, R.K., Siber, G.R. and Langer, R. Determinants of release rate of tetanus vaccine from polyester microspheres. Pharmaceut. Res. 1993, 10, 945-953

10 Salenstedt, C.R. and Tirunarayanan, M.O. Purification of tetanus toxin with the aid of sephadex gels. Allerg. Kiln. Immunol. 1966, 130, 190-196

11 Vranken, M.N. and Claeys, D.A. Process for encapsulating water and compounds in aqueous phase by evaporation. US Patent 3523906, 1970

12 Ogawa, Y., Yamamoto, M., Okada, H., Yashiki, Y. and Shimamoto, T. A new technique to efficiently entrap leuprolide acetate into microcapsules of polylactic acid or copoly(lactic/glycolic) acid. Chem. Pharm. Bull. (Tokyo) 1988, 36, 1095-1103

13 Cohen, S., Yoshioka, T., Lucarelli, M., Hwang, L.H. and Langer, R. Controlled delivery systems for proteins based on poly(lactic/ glycolic acid) microspheres. Pharmacol. Res. 1991, 8, 713-720

14 Laemmli, U.K. Cleavage of structural proteins during the assembly

of the head of bacteriophage T4. Nature 1980, 2"/7, 680-685 15 Kabat, E.A. and Mayer, M.M. Kabat and Mayer's Experimental

Immunochemistry (Ed. Thomas, C.C.) Charles C. Thomas, Springfield, IL, 1961, pp. 22-96

16 Relyveld, E.H. Tritage in vivo des anticorps antidiphteriques et antitetaniques 8 plusiers niveaux. J. Biol. Stand. 1977, 5, 45-55

17 Gupta, R.K., Maheshwari, S.C. and Singh, H. The titration of tetanus antitoxin. Studies on the sensitivity and reproducibility of the toxin neutralization test. J. Biol. Stand. 1985, 13, 143-149

18 Glantz S.A. (Ed.). The special case of two groups: The t test. In: Primer of Biostatistics, Third Edition, McGraw-Hill, New York, 1992, pp. 67-109

19 Glantz, S.A. (Ed.). Alternatives to analysis of variance and t test based on ranks. In: Primer of Biostatistics, Third Edition, McGraw-Hill, New York, 1992, pp. 320-371

20 Arakawa, T., Kita, Y. and Carpenter, F. Protein-solvent interactions in pharmaceutical formulations. Pharmacol. Res. 1991, 8, 285-291

21 Liu, W.R., Langer, R. and Klibanov, A. Moisture-induced aggregation of lyophilized proteins in the solid state. Biotech. Bioeng. 1991, 37, 177-184

22 Alex, R. and Bodmeier, R. Encapsulation of water-soluble drugs by a modified solvent evaporation method. I. Effect of process and formulation variables on drug entrapment. J. Microencapsul. 1990, 7, 347-355

23 Wang, H.T., Schmitt, E., Flanagan, D.R. and Linhardt, R.J. Influence of formulation methods on the in vitro controlled release of protein from poly(ester) microspheres. J. Controlled Release 1991, 17, 23-32

24 Heya, T., Okada, H., Ogawa, Y. and Toguchi, H. Factors influencing the profiles of TRH release from copoly(d,l-lactic/glycolic acid) microspheres. Int. J. Pharm. 1991, 72, 199-205

25 Su Ming, L., Garreau, H. and Vert, M. Structure-property relationships in the case of the degradation of massive aliphatic poly-(~-hydroxy acids) in aqueous media. J. Mater. Sci. (Mat. Med.) 1990, 1, 123-130

26 Sanders, L.M., Kell, B.A., McRae, G.I. and Whitehead, G.W. Prolonged controlled-release of narfarelin from biodegradable polymeric implants: influence of composition and molecular weight of polymer. J. Pharm. Sci. 1986, 75, 356-360

27 Davis, D. and Gregoriadis, G. Liposomes as adjuvants with immunopurified tetanus toxoid: influence of liposomal character- istics. Immunology 1987, 61,229-234

28 Bizzini, B. Tetanus. In: Bacterial Vaccines (Ed. Germanier, R.) Academic Press, Orlando, FL, 1984, pp. 37-68

29 Gupta, R.K., Siber, G.R., Alonso, M.J. and I_anger, R. Development of a single-dose tetanus toxoid based on controlled release from biodegradable and biocompatible polyester microspheres. In: Vaccines 93 (Eds Brown, F., Chanock, R., Ginsberg, H. and Lamer, R.) Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, 1993, pp. 391-396

306 Vacc ine 1994 V o l u m e 12 N u m b e r 4