~a~ec~~r hmunoiogy,Vol. 28,No. 3, pp. 287-294,1991
0161-58~/91 $3.004 0.00 Pergamon Pressplc
Printed in Great Britain.
BIODEGRADABLE MICROSPHERES AS A VACCINE DELIVERY SYSTEM JOHNH. ELDRIDGE,* JAY K. STAAS,~ JONATHAN A. MEULBROEK,* JERRYR. MCGHEE,* THOMASR. TICE~ and RICHARDM. GILLEY~ *Department of Microbiology, University of Alabama at Birmingham, Birmingham, AL 35294, U.S.A. and tcontrolled Release Division, Southern Research Institute, Birmingham, AL 35255, U.S.A. (Accepted 1 May 1990)
Abstract-The utility of biodegradable and biocompatible microspheres as a vaccine delivery system for the induction of systemic and di~eminated mucosal antibody responses was investigated. Intra~~toneal (ip) injection into mice of I-IOfirn microspheres, constructed of the copolymer poly(DL-lactide-coglycolide) (DL-PLG) which contained approximately 1% by weight a formalinized toxoid vaccine of staphylococcal enterotoxin B (SEB), dramatically potentiated the circulating IgG anti-toxin antibody response as compared to the free toxoid. The initiation of vaccine release was delayed in larger microspheres, and a mixture of l-10 and 20-50pm microspheres stimulated both a primary and an anamnestic secondary anti-toxin response following a single injection. However, neither free nor microencapsulated SEB toxoid induced a detectable mucosal IgA anti-toxin response following systemic injection. In contrast, three peroral immunizations with toxoid-microspheres stimulated circulating IgM, IgG and IgA anti-toxin antibodies and a concurrent mucosal IgA response in saliva, gut washings and lung washings. Systemic imm~izatjon with mi~r~ncapsulat~ toxoid primed for the induction of disseminated mucosal IgA responses by subsequent oral or intratracheal (it) boosting in microspheres,
while soluble toxoid was ineffective at boosting. ‘lbese results indicate that biodegradable and biocompatible microspheres represent an adjuvant system with potentially widespread application in the induction of both circulating and mucosal immunity.
INTRODUCTION
A wide variety of infectious agents are currently of intense investigation directed toward the identification and isolation of effective and safe vaccine antigens for use in man (Chanock and Lerner, 1984). However, the overriding concern for safety dictates in most cases that subunit vaccines of defined and purified proteins or polysaccharides are used, and these are in many cases poorly immunogenic. As a case in point, the recent advances in DNA technology have provided powerful approaches to the study of viral and bacterial gene structure, and numerous genes encoding proteins which encode epitopes recognized by protective antibodies have heen isolated, expressed in vectors and the proteins isolated. However, due to the poor immunogenicity of the isolated proteins, few effective vaccines have heen produced to date. Among the approaches to enhancing immunogenicity are conjugation to carriers (Arnon et al., 1980), delivery systems such as liposomes (Allison and Gregoriadis, 1974), controlled release implants (Preis and Langer, 1979), and adjuvants (Warren et al., 1986). Among the adjuvants, alum is the single compound currently approved for human use, and its use for booster immunizations has heen called into question. In addition to the need for appropriate methods to enhance the response to injected immunogens, there is an acute need to identify effective vaccination the subject
procedures against organisms which gain entry through, or exert their pathophysiolo~c effects at, the gastrointestinal, pulmonary, nasopharangeal or genitourinary surfaces. These mucosal surfaces represent a tremendous surface area; over 400 m2 in man. The absorptive function of the wet mucosal surfaces such as the gut and respiratory tract dictate that they are semi-permeable. These imperfect barriers constitute the major paths through which antigenic materials enter the body, and the site of replication or the portal of entry for a large number of pathogenic organisms. Unlike the blood, the major antibody isotype associated with the mucus secretions is IgA (Hanson, 196 1; Tomasi and Zigelbaum, 1963; Tomasi et al., 1965) and it is predominantly present in a dimer form (Tomasi, 1970) in association with J chain (Halpern and Koshland, 1970) and secretory component (Tomasi, 1970). The daily synthesis of secretory IgA (sIgA) exceeds that of the other isotypes combined (Mestecky and McGhee, 1987); illustrating the importance of sIgA in protection, However, sIgA antibodies are efficiently induced only through immunization of mucosal tissues, and not by the commonly employed parenteral routes of vaccination. Numerous approaches to mucosal immunization have been investigated (Mesteclcy et al., 1987; Bienenstock and Befus, 1980; Mestecky et al., 1980), with most centering on oral ingestion or intranasal instillation to achieve antigen absorption through the microfold cells which overlay the 287
288
JOHN
H. ELDRIDGE~? al.
gut-associated and bronchial-associated lymphoid tissues. Although varying degrees of success have been achieved, mucosal immunization with purified vaccine antigens would benefit from improvements in delivery to these IgA inductive sites, retention of vaccine integrity, and an appropriate adjuvant. Our own work in the area of vaccine delivery has centered on the use of monolithic microspheres constructed of poly(DL-lactide-co-glycolide) (DL-PLG). This biocompatible and biodegradable copolymer is of the same class of materials which is currently used in resorbable sutures, and has a history of safe application in man. The vaccine is homogeneously dispersed within the copolymeric matrix in a dry state, thus providing exceptional stability. In addition, the biodegradation rate, and thus the rate of vaccine release, can be controlled through a variety of means. Among these are the ratio of lactide to glycolide, which determines the rate at which the ester bonding is hydrolized to yield lactic and glycolic acids (Miller et al., 1977), and the size of the microspheres. Following introduction into the body microspheres < 10 pm in diameter are rapidly phagocytized by the cells of the reticuloendothelial system, which accelerate their degradation. More importantly, the delivery of antigens within microspheres of this small diameter results in a profound potentiation of the specific humoral immune response. In the case of mucosal immunization via oral or intertracheal application, the copolymer protects the antigen from degradation and targets delivery to the mucosally-associated lymphoid tissues. MATERIALS
AND METHODS
Mice
Specific pathogen-free BALB/c mice of mixed sexes were used throughout these experiments. They were bred and maintained in our barrier facilities at the University of Alabama at Birmingham. Except as specified in the text they were allowed food and water ad libitum, and were entered into experimental protocols at 8-12 weeks of age. Staphylococcal enterotoxin B vaccine
A formalinized vaccine of SEB was prepared as described by Warren et al. (1973). In brief, 1 g of enterotoxin was dissolved in 0.1 M sodium phosphate buffer, pH 7.5, to 2 mg/ml. Formaldehyde was added to the enterotoxin solution to achieve a formaldehyde:enterotoxin mole ratio of 4300: 1. The solution was placed in a slowly shaking 37°C controlled environment incubator-shaker and the pH was monitored and maintained at 7.5 f 0.1 daily. After 30 days the toxoid was concentrated and washed into borate buffered saline (BBS) using a pressure filtration cell (Amicon), and sterilized by filtration. Conversion of the enterotoxin to enterotoxoid was confirmed by the absence of weight loss in
3-3.5 kg rabbits injected intramuscularly toxoid material.
with 1 mg of
Microsphere production, size distribution and vaccine release
The microspheres were prepared by a solvent-evaporation process which has been previously described in detail (Cowsar et al., 1985). Size distributions of the preparations were obtained by measuring the diameters of 500 microspheres shown on scanning electron micrographs. The diameters were then used to calculate the size distribution as a function of internal volume. The vaccine content (core loading) was determined by dissolving a sample of the microspheres in methylene chloride, extracting the enterotoxoid, determining the amount of protein by the BCA assay (Pierce) and calculating the percent antigen by weight. In vitro vaccine release kinetics were determined by placing a sample of the microspheres in a receiving fluid consisting a 0.1 M HCI. After incubation at 37°C for 2 hr the HCl was removed and replaced with pH 6.8 phosphate-buffered saline (PBS). The PBS was exchanged at 6 and 24 hr and every 24 hr thereafter until termination of the study. The amount of protein in the aliquots of receiving fluid was quantified and related to the total protein in the sample of microspheres to determine the cumulative percent SEB release as a function of time. Immunizations
Groups of mice were orally administered free or microencapsulated enterotoxoid in 0.5 ml of 8 parts tap water:2 parts 7.5% sodium bicarbonate via a blunt-tipped feeding needle inserted into the stomach. Systemic immunizations were carried out by intraperitoneal (ip) injection in 0.2 ml PBS. Bronchopulmonary immunization was performed via intratracheal (it) instillation of 40 pg of toxoid vaccine in 40 ~1 of PBS to mice under sodium methohexital (Brevital, Lilly) anesthesia. After the onset of anesthesia the mice were suspended by their lower incisors from an inverted u of wire, protruding from a dissecting board maintained at a 45” angle, such that the head could be pulled over the edge of the board. The pharynx was transilluminated with the aid of a fiber optic lamp, and the fluid instilled through the shaft of a blunt tipped feeding needle, inserted through the glottis, which was attached to a Hamilton syringe with a stepped dispenser through a length of teflon tubing. Tissue-penetration studies
Mice were administered a single dose of 20mg of microspheres suspended in 0.5 ml of tap water using a blunt-tipped feeding needle inserted into the stomach. The mice were sacrificed at the times indicated in the text and three representative Peyer’s patches from the small intestine, the first mesenteric
Biodegradable microspheres as a vaccine delivery system lymph node proximal to the appendix, and the spleen were excised, mounted in OCT freezing compound (Miles), and snap frozen in liquid nitrogen. The tissues were cut into 4-6pm serial sections and all sections were viewed on a fluorescence microscope (Leitz) to quantify microspheres in each tissue. The size of individual microspheres was determined using a calibrated eyepiece micrometer, and their location within the tissue was recorded. Collection of biologic fluids
Blood was collected from a puncture of the retroorbital plexus in calibrated, heparinized capillary pipettes and the plasma harvested following centrifugation. Saliva and gut wash samples were collected as described by Elson et al. (1984). In brief, the mice were administered four 0.5 ml doses of lavage solution, isoosmotic with mouse intestinal secretions, at 15 min intervals. Following the last dose the mice were anesthetized and administered 0.1 mg of pilocarpine by ip injection. The intestinal discharge was collected into 5 ml of a solution of 0.1 mg/ml soybean trypsin inhibitor (Sigma) in 50 mM EDTA. At the same time as the intestinal discharge a large volume of saliva was secreted and 200 ~1 was collected per mouse. Lung wash fluids were obtained by total lavage with 1 ml of PBS. The samples were clarified by centrifugation and sodium azide, phenylmethylsulfonyl fluoride and fetal calf serum were added as preservative, protease inhibitor, and alternate substrate for protease activity, respectively. All samples were stored at -70°C until assayed for antibody activity. Radioimmunometric assays of toxin-speczjic antibodies
Radioimmunometric assays were performed in Immulon strips (Dynatech) coated with toxin at 1 pg/ml in BBS, pH 8.4, overnight at 4°C. Control strips were left uncoated, but all were blocked for 2 hr at 25°C with 1% bovine serum albumin (Sigma) in BBS, which was also used as the diluent for all samples and ‘251-labeled reagents. Various two-fold dilutions of test samples were added to washed triplicate wells and incubated for 6 hr at 25°C. After washing, 100,000 cpm of ‘251-labeled affinity-purified IgG goat anti-mouse IgM, IgG or IgA heavy chain specific antibody (Southern Biotechnology Associates) was added per well and incubated overnight at 4°C. Following the removal of unbound antibodies by washing, the bound 12sI-antibodies were detected in a gamma spectrometer (Beckman). The results are presented as the reciprocal of the sample dilution producing a signal significantly different from that of the group-matched prebleed at the same dilution (end-point titration). Assays of total IgM, IgG and IgA were carried out in the same manner except that the wells were coated with appropriate antiisotype reagents. Calibration curves were prepared for IgM and IgA using various dilutions of purified myeloma proteins, while the IgG standard was
a reference serum (Miles). interpolations of unknowns puter using a “Log&log” the Biomedical Computing Center (Vanderbilt Medical
289 Calibration curves and were obtained by comBASIC program from Technology Information Center).
RESULTS
Several studies have shown that parenteral immunization with antigens trapped within, or incorporated into, the lipid bilayer of liposomes results in a potentiated humoral immune response (Allison and Gregoriadis, 1974). Two mechanisms appear to play a role in this immune enhancement. First, the liposomes serve as a depot for extended release of the antigen (Kramp et al., 1982). Second, an increased targeting of the antigen to the accessory cells of the reticuloendothelial system has been cited as a mechanism resulting in a more efficient presentation to T cells (Dailey et al., 1977). We investigated the immunopotentiating activity of an analogous vaccine delivery system without the instability inherent to liposomes: biocompatible and biodegradable microspheres constructed with DL-lactide and glycolide (DL-PLG). Potentiation of the circulating through microencapsulation
antibody response
Groups of mice were systemically immunized by the injection of 25 pg of a formalinized toxoid vaccine of staphylococcal enterotoxin B (SEB), either free or within 1-10pm microspheres in which the wall material consisted of equal mole parts of DLPLG (50: 50 DL-PLG). Plasma samples were collected at lo-day intervals through day 90, when the mice were boosted with the same dose and form of vaccine. Samples of plasma were collected for an additional 50 days after the booster immunization, and then all the samples were assayed for their levels of IgG anti-toxin antibodies by end-point titration in an isotype-specific RIA using solid-phase adsorbed SEB (Fig. 1). Twenty-five micrograms per mouse is the optimal dose of free toxoid, and it elicited a primary circulating IgG response characteristic of a protein antigen which peaked at a titer of 3200 at 40 days after immunization and thereafter fell. In contrast, the same dose of toxoid delivery in microspheres elicited a more vigorous IgG anti-toxin response which reached a’ titer of 204,800 and which was maintained near this level through day 90. Following secondary immunization on day 90, the immunopotentiation provided by the microspheres was again evident (Fig. 1). Although both the free and microencapsulated forms of SEB toxoid stimulated anamnestic circulating anti-toxin responses, the microencapsulated form elicited a secondary response which maintained its level at 32-64 times that of the free vaccine. Additionally, titration experiments demonstrated that higher doses of toxoid could be delivered in
JOHNH. ELDRIDGE et al.
290
-, 10'
--D-
Free
b
Microencapsulated
106 106 104 103 102 10' 20
8
40
60
60
A Experimental
100
120
140
Day
Fig. 1. Potentiation of the circulating IgG anti-SEB toxin response by immunization with microencapsulated SEB toxoid. BALB/c mice were immunized by IP injection of 25pg of free or microencapsulated (I-1Opm 50:50 DLPLG) toxoid on day 0 and boosted with the same dose and form on day 90. Anti-SEB toxin antibodies of the IgG isotype were determined by end-point titration in a RIA.
microspheres without inducing high dose suppression of the immune response, as is seen with the free toxoid. In fact, systemic immunization with
increasing doses of encapsulated toxoid, through 200lgg/mouse, led to progressively higher peak primary titers of circulating anti-toxin antibodies (data not shown). Controlled vaccine release microspheres and a single injection multiple release formulation
An advantage of the copolymer microcapsule delivery system is the ability to control the time and/or rate at which the incorporated material is released. In the case of vaccines this allows for scheduling of the antigen release in such a manner as to maximize the antibody response following a single administration. Among the possible release profiles which would be expected to improve the antibody response to a vaccine is a pulsed release analogous to conventional primary and booster immunizations. This release
profile could be approached either through the blending of vaccine-microspheres with two different copolymer ratios, or by blending vaccine microspheres of two sizes. This latter approach is based on the observations that microspheres < 10 pm in diameter are phagocytized and release antigen at a substantially accelerated rate relative to microspheres of the same copolymer ratio which are too large to be phagocytized. The possibility of using size to achieve pulsed vaccine release was investigated by systemically injecting SEB toxoid to groups of mice either in l-10 pm, 20-125 pm, or as a mixture of l-10 pm and 20-125 pm microspheres. Mice receiving SEB toxoid in I-1Opm microspheres produced a plasma IgG anti-toxin response which was detected on day 10, rose to a maximal titer of 102,400 on days 30 and 40, and then decreased through day 60 to 25,600 (Fig. 2). In contrast, the response to the toxoid administered in 20-125 pm microspheres was not detected until day 30, and thereafter increased to a titer of 51,200 on days 50 and 60. The co-administration of equal parts of the toxoid in l-10 and 20-125pm microspheres produced an IgG response which was for the first 30 days essentially the same as that stimulated by the l-10 pm microcapsules administered alone. However, beginning on day 40 the response measured in the mice receiving the vaccine in concurrently administered l-10 plus 20-125 pm microspheres steadily increased to a titer of 819,200, a level far more than the additive amount of the responses induced by the two size ranges administered singly. Adsorption of microspheres from the gut lumen into the Peyer’s patches and their draining lymphatics
Secretory immunity is most often achieved by direct immunization of the mucosally-associated lymphoid tissues, of which the largest mass consists of the Peyer’s patches (PP) located along the gastrointestinal tract. The PP are separated from the gut lumen by a unique epithelium which is interspersed with specialized microfold cells which internalize
c 600000 'ii
0 l.’
E
H H
600000
q
a
IO
20
30
40
50
l-10urn 20-50um l-10+20-50 urn
60
Day Post SC Immunization Fig. 2. Controlled vaccine release through alteration in microsphere size. BALB/c mice were immunized with SEB toxoid in 50: 50 DL-PLG microspheres of l-10 pm in diameter, 2&125pm in diameter, or in a mixture of I-10 and 20-125 pm diameter. Anti-SEB toxin antibodies of the IgG isotype in the plasma were determined by end-point titration in a RIA.
Biodegra~ble
microsphere
antigenic materials from the gut lumen and transport them to the underlying lymphoreticular cells. The PP lymphoid population contains a large proportion of B cells that are committed to IgA synthesis, and following antigen sensitization they extravasate the PP and home to diverse mucosal tissues. Large numbers of these IgA lymphobl~ts infiltrate the glandular tissues and lamina propria regions underlaying the mucosal membranes and undergo terminal differentiation to IgA synthesis. Although it is clear that this common mucosal immune system can be induced through experimental oral immunization (Mestecky et al., 1987), the large antigen doses required to achieve this have made it impractical for purified vaccine antigens. However, in addition to adjuvant activity, micr~ncapsulation provides two advantages in oral vaccine delivery. First, the wall material protects the vaccine from degradation by gastric acidity and the proteolytic enzymes of the gut. Second, relatively hydrophobic particles of appropriate size are readily absorbed by the M cells (LeFevre and Joel, 1984; LeFevre et al., 1980). The possibility of using microspheres as an oral vaccine delivery system was investigated by intragastrically administering single 20 mg doses of l-10 pm 85: 15 DL-PLG microspheres loaded with the fluorochrome coumarin-6 to mice (Eldridge et al., 1989a). Fluorescence microscopic observation of frozen sectioned tissues revealed that internalized microspheres were present in the macrophages of the Peyer’s patch dome region at 24 hr post-administration, and at all times tested through 35 days (Fig. 3a). The total number of microspheres observed in three representative Peyer’s patches increased through day 4 and then decreased over the following 31 days to 15% of the peak number. At 1, 2 and 4 days the proportion of < 5 pm (7682%) and > 5 pm (18-23%) particles remained constant and reflected the size distribution of the input preparation. Concurrent with the decrease in total microsphere numbers in the PP, which began at day 7, a progressive shift in the size distribution was observed such that the ~5 pm sizes ceased to predominate, and the > 5 pm microspheres became predominant. In addition, the microspheres which were observed to penetrate deep into the Peyer’s patches were almost exclusively 6 5 pm in diameter. The lymphatic drainage from the intestine passes through the mesenteric lymph nodes, and microspheres were detected in this tissue as soon as 24 hr after ingestion (Fig. 3b). The peak number of microspheres in the mesenteric lymph nodes was observed on day 7, when the total number in the Peyer’s patches had fallen significantly. All of the microspheres observed in the mesenteric lymph nodes were , 10 pm in diameter were not adsorbed. &&cosffl immunization with microe~c~psuiff ted an t&en The data on microsphere adsorption into and migration out of the Peyer’s patches suggested that
JOHNH. ELDRIDGEet al.
292
the type of immune response elicited through oral immunization could be controlled through varying the microsphere size. Microspheres ~5 pm extravasate from the Peyer’s patches within macrophages which disseminate into systemic lymphoid tissues, where vaccine release would be expected to induce a circulating antibody response. In contrast, microspheres > 5 pm remain in the IgA inductive environment of the Peyer’s patches, where vaccine recognition would stimulate a disseminated mucosal IgA immune response. Perhaps the most attractive possibility, concurrent circulating IgG and secretory IgA immunity, should be achievable through oral immunization with vaccine-microspheres spanning the size range of l-10pm. Three peroral immunizations at 30-day intervals with 100 pg of SEB toxoid in microspheres effectively induced the appearance of concurrent circulating and disseminated mucosal anti-toxin antibodies 20 days following the tertiary administration (Fig. 4). The plasma anti-toxin response included IgM and IgA components at titers of 2560 and a specific IgG titer in excess of 40,000. Each of the tested mucosal secretions, saliva, gut wash fluid and lung wash fluid, contained significant levels of toxin-specific IgA, and the washings from the gut and lungs additionally contained specific IgG. Parallel oral immunizations with doses of non-encapsulated toxoid ranging from 1 to 100 pg were ineffective at inducing a detectable
Plasma
Salka
Gut Wash
immune response in any of the tested fluids. In addition, doses of non-encapsulated toxoid above 50 pg per immunization induced oral tolerance as demonstrated by suppressed circulating antibody levels to systemic challenge (data not shown). In contrast to oral immunization, two IP immunizations with 100 pg doses of microencapsulated toxoid, separated by 30 days, induced a plasma IgG anti-toxin response at a titer of 6,533,600 on day 20 after boosting (Fig. 4). Despite the fact that this circulating IgG response was 64-fold higher than that induced through oral immunization with microencapsulated toxoid, no significant IgA anti-toxin antibodies were detected in the plasma or mucosal secretions. This hyperimmunization was, however, capable of resulting in the appearance of IgG antibodies in the samples of each of the mucosal secretions, suggesting that they were the result of transudation from the serum. Interrelationship between the systemic and mucosal immune systems The relationship between the systemic and mucosal immune systems was investigated by determining the isotype and distribution of antibodies induced through oral or intratracheal (it) boosting of systemically primed mice. As in the preceeding groups of mice, 100 pg of encapsulated toxoid was used at each administration and the plasma and secretion samples
Plasma
Lur!gWash
n
Saliva
Gut Wash LugWash
Saliva
Gut Wash LqWaah
I!#
r-lHlecl
Plasma
Saliva
Qti Wash
l.ungWprh
Plasma
-
Fig. 4. Levels and isotypes of anti-SEB toxin antibodies induced in the plasma and mucosal secretions of mice immunized by various protocols with SEB
toxoid
in I-10pm
5O:SO DL-PLG microspheres.
Biodegradable microspheres as a vaccine delivery system were collected 20 days after secondary immunization for evaluation in toxin-specific RIAs (Fig. 4). Oral immunization following systemic priming boosted the serum IgG anti-toxin titer to 819,200 and, unlike systemic boosting, resulted in the appearance of circulating IgA antibodies. In addition, the IP-Oral protocol was nearly as effective in the induction of mucosal IgA anti-toxin antibodies as was tertiary oral immunization. Similar, but more dramatic, results were obtained when the boosting was performed by it instillation of the microencapsulated toxoid (Fig. 4). The level of circulating IgG antibodies was boosted, and plasma IgA anti-toxin was detected to a titer of 3200. In addition, each of the mucosal secretions contained significant levels of specific IgG and IgA. This booster method was particularly effective at inducing both IgG and IgA antibodies in the lung fluids, with anti-toxin titers reaching 20,480 and 2560, respectively. DISCUSSION
Microspheres constructed of DL-PLG have been under investigation in recent years as injectable reservoirs for the long-term controlled release of pharmaceuticals (Tice and Cowsar, 1984; Redding et al., 1984). Several characteristics of these microspheres make them ideal as a delivery system for vaccines. The lactide-glycolide copolymer system is biodegradable, non-inflammatory and yields the biocompatible molecules lactic and glycolic acids upon degradation. The injection of antigen in microspheres < 10 pm in diameter provides a strong potentiation of the circulating antibody response which appears to be a function of both a depot effect and the rapid phagocytosis of the microspheres by antigen-presenting accessory cells. Although the exact contribution by each of these mechanisms to the immune enhancement is difficult to determine, substantially less adjuvancy has been observed with microspheres which are of a size that precludes their uptake by phagocytosis. Thus, it may be that the major mechanism of immune enhancement is through the efficient loading of a relatively high concn of antigen into accessory cells for presentation to T cells. Consistent with this possibility is the observation that the immune response to T cell-independent antigens, such as bacterial carbohydrates, is generally not enhanced through delivery in small diameter microspheres. The ability to provide distinct “pulses” of antigen release by the injection of a mixture of vaccine-containing microspheres with various degradation times appears to be a promising vaccination stategy. In addition to taking advantage of the immunopotentiation provided by this vaccine delivery system, timed booster immunizations are released without the need for additional contact with a health care professional. At present, using a combination of variables, including copolymer ratio and size, four discrete releases of antigen over a 120-day period have been achieved.
293
Currently available techniques should allow these releases to be spread over a period in excess of 1 year. Potentially this approach may prove useful in human immunization, particularly in third world countries, and in veterinary applications. When used as an oral delivery system, the polymer wall of the microspheres protects the vaccine antigen from degradation in stomach acid and from proteolysis in the gut. Further, there is adsorption of appropriately sized microspheres into the immune inductive environment of the PP, and no tissue penetration in other regions of the gut where T cells with suppressor phenotype predominate. This targeted delivery may be as important as the adjuvant activity of the microspheres in the conversion of the orally tolerogenic SEB toxoid into a form which is immunogenic after oral administration. Detailed examination of the behavior of the microspheres following adsorption from the gut demonstrated that size influenced their retention in the PP versus redistribution to other tissues. All the microspheres within the PP appeared to be within macrophages as judged by PAS staining for intracellular glycogen. Microspheres which were seen to traffic to other tissues did so through the known pattern of lymphoid drainage, were seen to remain within macrophages, and were essentially all < 5 pm in diameter. Although we have not directly tested the possibility of differentially inducing the circulating as opposed to the mucosal immune system through the use of very finely sized particles, three spaced oral immunizations with microencapsulated SEB toxoid did induce high levels of IgG anti-toxin antibodies in the plasma concurrent with the production of IgA antibodies in diverse mucosal tissues. Systemic immunization was found to effectively prime for a mucosal IgA response when a booster was administered to either the gut-associated or bronchial-associated lymphoid tissues. Thus, under these experimental conditions it does not appear that systemic immunization has induced a tolerance or down regulation of a subsequent mucosal antibody response, as has been described in other systems. This finding has potentially important implications for the development of single immunization strategies for the induction of concurrent systemic and mucosal antibody responses. CONCLUSIONS
Microspheres l-10 pm in diameter composed of a vaccine antigen impregnated in the copolymer poly(oL-lactide-co-glycolide) represent a pharmaceutically acceptable delivery system with immunopotentiating activity for humoral antibody responses. The product is a free flowing powder with excellent stability, an extended shelf life and which is easily suspended in physiologic diluents for administration. Vaccine release is controllable through alterations in the ratio of copolymers and/or the size of the micro-
JOHNH. ELDRIDGE et al.
294
spheres, allowing the construction of single injection formulations with a priming and one or more timed booster releases of vaccine. Orally administered microspheres are adsorbed by the M cells overlaying the Peyer’s patches and passed to the immune inductive environment of both the Peyer’s patches and systemic lymphoid organs and induce concurrent systemic and mucosal antibody responses.
Halpern M. S. and Koshland M. E. (1970) Novel subunit of secretory IgA. Nature (London) ZUI, 1276. Hanson L. A. (1961) Comparative immunological studies of the immune globulins of human milk and of blood serum.
Acknowledgements-The
LeFevre M. E. and Joel D. D. (1984) Peyer’s patch epithelium: an imperfect barrier. In Intestinal Toxicology (Edited by Schiller C. M.), p. 45. Raven Press, New York. Mestecky J. (1987) The common mucosal immune system and current strategies for induction of immune responses _ in external secretions. J. Clin. Immun. 7, 265. Mesteckv J. and McGhee J. R. (1987) Molecular and cellula; interactions involved in ‘IgA ‘biosynthesis and immune responses. Ado. Immun. 40, 153. Mestecky J., McGhee J. R., Crago S. S., Jackson S., Kilian M., Kiyono H., Babb J. L. and Michalek S. M. (1980) Molecular-cellular interactions in the secretory IgA response. J. Reticuloendothelial Sot. 28, 455. Miller R. A., Brady J. M. and Cutwright D. E. (1977) Degradation rates of resorbable inplants (polylactates and polyglycolates): Rate modificatibn with-changes in PLAIPGA copolvmer ratios. J. Biomed. Mater. Res. 11, ill. - Preis I. P. and Langer R. S. (1979) A single-step immunization by sustained antigen release. J. Immun. Meth. 28,
Int. Arch. Allergy Appl. Immun. 18, 241.
Kramp W. J., Six H. R. and Kasel J. A. (1982) Post-immunization clearance of liposome-entrapped adenovirus type __ 5 hexon. Proc. Sot. eip. Biol. Med: 369, 135. LeFevre M. E.. Hancock D. C. and Joel D. D. (1980) Intestinal barrier to large particles in mice. J. Tdxicoi Envir. Health 6, 691.
authors wish to thank Mr J.
Douglas Morgan for his excellent technical assistance and to MS Amie Stoppelbein for her help in the preparation of this manuscript. This work was supported by USPHS grants AI 24772, AI 21774 and contract DAMD17-86-6160 from the U.S. Army Medical Research Acquisition Activity. REFERENCES Allison A. C. and Gregoriadis G. (1974) Liposomes as immunologic adjuvants. Nature 252, 252. Amon R., Sela M., Parant M. and Chedid L. (1980) Antiviral response elicited by a complete synthetic antigen with built-in adjuvance. Proc. natn. Acad. Sci. U.S.A. 71, 6769.
Bienenstock J. and Befus A. D. (1980) Mucosal immunology. Immunology 41, 249. Chanock R. M. and Lemer R. A. (1984) Proceedings of the Cold Spring Harbor Conference: Modern Approaches to Vaccines. Cold Spring Harbor Laboratory.
Cowsar D. R., Tice T. R., Gilley R. M. and English J. P. (1985) Poly (lactide-co-glycolide) microcapsules for controlled release of steroids. In Methods in Enzymology; Drug and Enzymes Targeting (Edited by K. J. Widder and R. Green), p: 101. A&den& Press, New York. Dailev M. 0.. Post W. and Hunter R. L. (19771 Induction of c&mediated immunity to chemically modified antigens in guinea pigs. II. The interaction between lipid conjugated antigens, macrophages and T lymphocytes. J. Immun. 118, 963.
Eldridge J. H., Gilley R. M., Staas J. K., Moldoveanu A., Meulbroek J. A. and Tice T. R. (1989a) Biodegradable microspheres: A vaccine delivery system for oral immunization. Curr. Top. Microbial. Immun. 146, 59. Eldridge J. H., Hammond C. J., Meulbroek J. A., Staas J. K., Gilley R. M. and Tice T. R. (1989b) Contro!led vaccine release in the gut-associated lymphoid tissues. I. Orally administered biodegradable microspheres target the Peyer’s patches. J. Controlled Release 11, 205. Elson C. O., Ealding W. and Lefkowitz J. (1984) A lavage technique allowing repeated measurement of IgA antibody in mouse intestinal secretions. J. Immun. Meth. 67, 101.
193.
Redding T. W., Schally A. V., Tice T. R. and Meyers W. E. (1984) Long-acting delivery systems for peptides: Inhibition of rat prostate tumors by controlled release of D-Trp6-LH-RH from injectable microcapsules. Proc. natn Acad. Sci. U.S.A. 81, 5845.
Tice T. R. and Cowsar D. R. (1984) Biodegradable controlled release parenteral systems. Pharm. Technol. J. 8, 26.
Tomasi T. B. and Zigelbaum S. (1963) The selective occurrence of y,A globulin in certain body fluids. J. clin. Invest. 42, 1552.
Tomasi T. B., Tam E. M., Solomon A. and Pendergast R. A. (1965) Characteristics of an immune system common to certain external secretions. J. exp. Med. 121, 101. Tomasi T. B. (1970) Structure and function of mucosal antibodies. A. Rev. Med. 21, 281. Warren F. R., Spero L. and Mezger 3. R. (1973) Antigenicity of formaldehyde-inactivated staphylococcal enterotoxoid B. J. Immun. 111, 885. Warren H. S., Vogel F. R. and Chedid L. A. (1986) Current status of immunological adjuvant. A. Rev. Immun. 4, 369.
0161-58~/91 $3.004 0.00 Pergamon Pressplc
Printed in Great Britain.
BIODEGRADABLE MICROSPHERES AS A VACCINE DELIVERY SYSTEM JOHNH. ELDRIDGE,* JAY K. STAAS,~ JONATHAN A. MEULBROEK,* JERRYR. MCGHEE,* THOMASR. TICE~ and RICHARDM. GILLEY~ *Department of Microbiology, University of Alabama at Birmingham, Birmingham, AL 35294, U.S.A. and tcontrolled Release Division, Southern Research Institute, Birmingham, AL 35255, U.S.A. (Accepted 1 May 1990)
Abstract-The utility of biodegradable and biocompatible microspheres as a vaccine delivery system for the induction of systemic and di~eminated mucosal antibody responses was investigated. Intra~~toneal (ip) injection into mice of I-IOfirn microspheres, constructed of the copolymer poly(DL-lactide-coglycolide) (DL-PLG) which contained approximately 1% by weight a formalinized toxoid vaccine of staphylococcal enterotoxin B (SEB), dramatically potentiated the circulating IgG anti-toxin antibody response as compared to the free toxoid. The initiation of vaccine release was delayed in larger microspheres, and a mixture of l-10 and 20-50pm microspheres stimulated both a primary and an anamnestic secondary anti-toxin response following a single injection. However, neither free nor microencapsulated SEB toxoid induced a detectable mucosal IgA anti-toxin response following systemic injection. In contrast, three peroral immunizations with toxoid-microspheres stimulated circulating IgM, IgG and IgA anti-toxin antibodies and a concurrent mucosal IgA response in saliva, gut washings and lung washings. Systemic imm~izatjon with mi~r~ncapsulat~ toxoid primed for the induction of disseminated mucosal IgA responses by subsequent oral or intratracheal (it) boosting in microspheres,
while soluble toxoid was ineffective at boosting. ‘lbese results indicate that biodegradable and biocompatible microspheres represent an adjuvant system with potentially widespread application in the induction of both circulating and mucosal immunity.
INTRODUCTION
A wide variety of infectious agents are currently of intense investigation directed toward the identification and isolation of effective and safe vaccine antigens for use in man (Chanock and Lerner, 1984). However, the overriding concern for safety dictates in most cases that subunit vaccines of defined and purified proteins or polysaccharides are used, and these are in many cases poorly immunogenic. As a case in point, the recent advances in DNA technology have provided powerful approaches to the study of viral and bacterial gene structure, and numerous genes encoding proteins which encode epitopes recognized by protective antibodies have heen isolated, expressed in vectors and the proteins isolated. However, due to the poor immunogenicity of the isolated proteins, few effective vaccines have heen produced to date. Among the approaches to enhancing immunogenicity are conjugation to carriers (Arnon et al., 1980), delivery systems such as liposomes (Allison and Gregoriadis, 1974), controlled release implants (Preis and Langer, 1979), and adjuvants (Warren et al., 1986). Among the adjuvants, alum is the single compound currently approved for human use, and its use for booster immunizations has heen called into question. In addition to the need for appropriate methods to enhance the response to injected immunogens, there is an acute need to identify effective vaccination the subject
procedures against organisms which gain entry through, or exert their pathophysiolo~c effects at, the gastrointestinal, pulmonary, nasopharangeal or genitourinary surfaces. These mucosal surfaces represent a tremendous surface area; over 400 m2 in man. The absorptive function of the wet mucosal surfaces such as the gut and respiratory tract dictate that they are semi-permeable. These imperfect barriers constitute the major paths through which antigenic materials enter the body, and the site of replication or the portal of entry for a large number of pathogenic organisms. Unlike the blood, the major antibody isotype associated with the mucus secretions is IgA (Hanson, 196 1; Tomasi and Zigelbaum, 1963; Tomasi et al., 1965) and it is predominantly present in a dimer form (Tomasi, 1970) in association with J chain (Halpern and Koshland, 1970) and secretory component (Tomasi, 1970). The daily synthesis of secretory IgA (sIgA) exceeds that of the other isotypes combined (Mestecky and McGhee, 1987); illustrating the importance of sIgA in protection, However, sIgA antibodies are efficiently induced only through immunization of mucosal tissues, and not by the commonly employed parenteral routes of vaccination. Numerous approaches to mucosal immunization have been investigated (Mesteclcy et al., 1987; Bienenstock and Befus, 1980; Mestecky et al., 1980), with most centering on oral ingestion or intranasal instillation to achieve antigen absorption through the microfold cells which overlay the 287
288
JOHN
H. ELDRIDGE~? al.
gut-associated and bronchial-associated lymphoid tissues. Although varying degrees of success have been achieved, mucosal immunization with purified vaccine antigens would benefit from improvements in delivery to these IgA inductive sites, retention of vaccine integrity, and an appropriate adjuvant. Our own work in the area of vaccine delivery has centered on the use of monolithic microspheres constructed of poly(DL-lactide-co-glycolide) (DL-PLG). This biocompatible and biodegradable copolymer is of the same class of materials which is currently used in resorbable sutures, and has a history of safe application in man. The vaccine is homogeneously dispersed within the copolymeric matrix in a dry state, thus providing exceptional stability. In addition, the biodegradation rate, and thus the rate of vaccine release, can be controlled through a variety of means. Among these are the ratio of lactide to glycolide, which determines the rate at which the ester bonding is hydrolized to yield lactic and glycolic acids (Miller et al., 1977), and the size of the microspheres. Following introduction into the body microspheres < 10 pm in diameter are rapidly phagocytized by the cells of the reticuloendothelial system, which accelerate their degradation. More importantly, the delivery of antigens within microspheres of this small diameter results in a profound potentiation of the specific humoral immune response. In the case of mucosal immunization via oral or intertracheal application, the copolymer protects the antigen from degradation and targets delivery to the mucosally-associated lymphoid tissues. MATERIALS
AND METHODS
Mice
Specific pathogen-free BALB/c mice of mixed sexes were used throughout these experiments. They were bred and maintained in our barrier facilities at the University of Alabama at Birmingham. Except as specified in the text they were allowed food and water ad libitum, and were entered into experimental protocols at 8-12 weeks of age. Staphylococcal enterotoxin B vaccine
A formalinized vaccine of SEB was prepared as described by Warren et al. (1973). In brief, 1 g of enterotoxin was dissolved in 0.1 M sodium phosphate buffer, pH 7.5, to 2 mg/ml. Formaldehyde was added to the enterotoxin solution to achieve a formaldehyde:enterotoxin mole ratio of 4300: 1. The solution was placed in a slowly shaking 37°C controlled environment incubator-shaker and the pH was monitored and maintained at 7.5 f 0.1 daily. After 30 days the toxoid was concentrated and washed into borate buffered saline (BBS) using a pressure filtration cell (Amicon), and sterilized by filtration. Conversion of the enterotoxin to enterotoxoid was confirmed by the absence of weight loss in
3-3.5 kg rabbits injected intramuscularly toxoid material.
with 1 mg of
Microsphere production, size distribution and vaccine release
The microspheres were prepared by a solvent-evaporation process which has been previously described in detail (Cowsar et al., 1985). Size distributions of the preparations were obtained by measuring the diameters of 500 microspheres shown on scanning electron micrographs. The diameters were then used to calculate the size distribution as a function of internal volume. The vaccine content (core loading) was determined by dissolving a sample of the microspheres in methylene chloride, extracting the enterotoxoid, determining the amount of protein by the BCA assay (Pierce) and calculating the percent antigen by weight. In vitro vaccine release kinetics were determined by placing a sample of the microspheres in a receiving fluid consisting a 0.1 M HCI. After incubation at 37°C for 2 hr the HCl was removed and replaced with pH 6.8 phosphate-buffered saline (PBS). The PBS was exchanged at 6 and 24 hr and every 24 hr thereafter until termination of the study. The amount of protein in the aliquots of receiving fluid was quantified and related to the total protein in the sample of microspheres to determine the cumulative percent SEB release as a function of time. Immunizations
Groups of mice were orally administered free or microencapsulated enterotoxoid in 0.5 ml of 8 parts tap water:2 parts 7.5% sodium bicarbonate via a blunt-tipped feeding needle inserted into the stomach. Systemic immunizations were carried out by intraperitoneal (ip) injection in 0.2 ml PBS. Bronchopulmonary immunization was performed via intratracheal (it) instillation of 40 pg of toxoid vaccine in 40 ~1 of PBS to mice under sodium methohexital (Brevital, Lilly) anesthesia. After the onset of anesthesia the mice were suspended by their lower incisors from an inverted u of wire, protruding from a dissecting board maintained at a 45” angle, such that the head could be pulled over the edge of the board. The pharynx was transilluminated with the aid of a fiber optic lamp, and the fluid instilled through the shaft of a blunt tipped feeding needle, inserted through the glottis, which was attached to a Hamilton syringe with a stepped dispenser through a length of teflon tubing. Tissue-penetration studies
Mice were administered a single dose of 20mg of microspheres suspended in 0.5 ml of tap water using a blunt-tipped feeding needle inserted into the stomach. The mice were sacrificed at the times indicated in the text and three representative Peyer’s patches from the small intestine, the first mesenteric
Biodegradable microspheres as a vaccine delivery system lymph node proximal to the appendix, and the spleen were excised, mounted in OCT freezing compound (Miles), and snap frozen in liquid nitrogen. The tissues were cut into 4-6pm serial sections and all sections were viewed on a fluorescence microscope (Leitz) to quantify microspheres in each tissue. The size of individual microspheres was determined using a calibrated eyepiece micrometer, and their location within the tissue was recorded. Collection of biologic fluids
Blood was collected from a puncture of the retroorbital plexus in calibrated, heparinized capillary pipettes and the plasma harvested following centrifugation. Saliva and gut wash samples were collected as described by Elson et al. (1984). In brief, the mice were administered four 0.5 ml doses of lavage solution, isoosmotic with mouse intestinal secretions, at 15 min intervals. Following the last dose the mice were anesthetized and administered 0.1 mg of pilocarpine by ip injection. The intestinal discharge was collected into 5 ml of a solution of 0.1 mg/ml soybean trypsin inhibitor (Sigma) in 50 mM EDTA. At the same time as the intestinal discharge a large volume of saliva was secreted and 200 ~1 was collected per mouse. Lung wash fluids were obtained by total lavage with 1 ml of PBS. The samples were clarified by centrifugation and sodium azide, phenylmethylsulfonyl fluoride and fetal calf serum were added as preservative, protease inhibitor, and alternate substrate for protease activity, respectively. All samples were stored at -70°C until assayed for antibody activity. Radioimmunometric assays of toxin-speczjic antibodies
Radioimmunometric assays were performed in Immulon strips (Dynatech) coated with toxin at 1 pg/ml in BBS, pH 8.4, overnight at 4°C. Control strips were left uncoated, but all were blocked for 2 hr at 25°C with 1% bovine serum albumin (Sigma) in BBS, which was also used as the diluent for all samples and ‘251-labeled reagents. Various two-fold dilutions of test samples were added to washed triplicate wells and incubated for 6 hr at 25°C. After washing, 100,000 cpm of ‘251-labeled affinity-purified IgG goat anti-mouse IgM, IgG or IgA heavy chain specific antibody (Southern Biotechnology Associates) was added per well and incubated overnight at 4°C. Following the removal of unbound antibodies by washing, the bound 12sI-antibodies were detected in a gamma spectrometer (Beckman). The results are presented as the reciprocal of the sample dilution producing a signal significantly different from that of the group-matched prebleed at the same dilution (end-point titration). Assays of total IgM, IgG and IgA were carried out in the same manner except that the wells were coated with appropriate antiisotype reagents. Calibration curves were prepared for IgM and IgA using various dilutions of purified myeloma proteins, while the IgG standard was
a reference serum (Miles). interpolations of unknowns puter using a “Log&log” the Biomedical Computing Center (Vanderbilt Medical
289 Calibration curves and were obtained by comBASIC program from Technology Information Center).
RESULTS
Several studies have shown that parenteral immunization with antigens trapped within, or incorporated into, the lipid bilayer of liposomes results in a potentiated humoral immune response (Allison and Gregoriadis, 1974). Two mechanisms appear to play a role in this immune enhancement. First, the liposomes serve as a depot for extended release of the antigen (Kramp et al., 1982). Second, an increased targeting of the antigen to the accessory cells of the reticuloendothelial system has been cited as a mechanism resulting in a more efficient presentation to T cells (Dailey et al., 1977). We investigated the immunopotentiating activity of an analogous vaccine delivery system without the instability inherent to liposomes: biocompatible and biodegradable microspheres constructed with DL-lactide and glycolide (DL-PLG). Potentiation of the circulating through microencapsulation
antibody response
Groups of mice were systemically immunized by the injection of 25 pg of a formalinized toxoid vaccine of staphylococcal enterotoxin B (SEB), either free or within 1-10pm microspheres in which the wall material consisted of equal mole parts of DLPLG (50: 50 DL-PLG). Plasma samples were collected at lo-day intervals through day 90, when the mice were boosted with the same dose and form of vaccine. Samples of plasma were collected for an additional 50 days after the booster immunization, and then all the samples were assayed for their levels of IgG anti-toxin antibodies by end-point titration in an isotype-specific RIA using solid-phase adsorbed SEB (Fig. 1). Twenty-five micrograms per mouse is the optimal dose of free toxoid, and it elicited a primary circulating IgG response characteristic of a protein antigen which peaked at a titer of 3200 at 40 days after immunization and thereafter fell. In contrast, the same dose of toxoid delivery in microspheres elicited a more vigorous IgG anti-toxin response which reached a’ titer of 204,800 and which was maintained near this level through day 90. Following secondary immunization on day 90, the immunopotentiation provided by the microspheres was again evident (Fig. 1). Although both the free and microencapsulated forms of SEB toxoid stimulated anamnestic circulating anti-toxin responses, the microencapsulated form elicited a secondary response which maintained its level at 32-64 times that of the free vaccine. Additionally, titration experiments demonstrated that higher doses of toxoid could be delivered in
JOHNH. ELDRIDGE et al.
290
-, 10'
--D-
Free
b
Microencapsulated
106 106 104 103 102 10' 20
8
40
60
60
A Experimental
100
120
140
Day
Fig. 1. Potentiation of the circulating IgG anti-SEB toxin response by immunization with microencapsulated SEB toxoid. BALB/c mice were immunized by IP injection of 25pg of free or microencapsulated (I-1Opm 50:50 DLPLG) toxoid on day 0 and boosted with the same dose and form on day 90. Anti-SEB toxin antibodies of the IgG isotype were determined by end-point titration in a RIA.
microspheres without inducing high dose suppression of the immune response, as is seen with the free toxoid. In fact, systemic immunization with
increasing doses of encapsulated toxoid, through 200lgg/mouse, led to progressively higher peak primary titers of circulating anti-toxin antibodies (data not shown). Controlled vaccine release microspheres and a single injection multiple release formulation
An advantage of the copolymer microcapsule delivery system is the ability to control the time and/or rate at which the incorporated material is released. In the case of vaccines this allows for scheduling of the antigen release in such a manner as to maximize the antibody response following a single administration. Among the possible release profiles which would be expected to improve the antibody response to a vaccine is a pulsed release analogous to conventional primary and booster immunizations. This release
profile could be approached either through the blending of vaccine-microspheres with two different copolymer ratios, or by blending vaccine microspheres of two sizes. This latter approach is based on the observations that microspheres < 10 pm in diameter are phagocytized and release antigen at a substantially accelerated rate relative to microspheres of the same copolymer ratio which are too large to be phagocytized. The possibility of using size to achieve pulsed vaccine release was investigated by systemically injecting SEB toxoid to groups of mice either in l-10 pm, 20-125 pm, or as a mixture of l-10 pm and 20-125 pm microspheres. Mice receiving SEB toxoid in I-1Opm microspheres produced a plasma IgG anti-toxin response which was detected on day 10, rose to a maximal titer of 102,400 on days 30 and 40, and then decreased through day 60 to 25,600 (Fig. 2). In contrast, the response to the toxoid administered in 20-125 pm microspheres was not detected until day 30, and thereafter increased to a titer of 51,200 on days 50 and 60. The co-administration of equal parts of the toxoid in l-10 and 20-125pm microspheres produced an IgG response which was for the first 30 days essentially the same as that stimulated by the l-10 pm microcapsules administered alone. However, beginning on day 40 the response measured in the mice receiving the vaccine in concurrently administered l-10 plus 20-125 pm microspheres steadily increased to a titer of 819,200, a level far more than the additive amount of the responses induced by the two size ranges administered singly. Adsorption of microspheres from the gut lumen into the Peyer’s patches and their draining lymphatics
Secretory immunity is most often achieved by direct immunization of the mucosally-associated lymphoid tissues, of which the largest mass consists of the Peyer’s patches (PP) located along the gastrointestinal tract. The PP are separated from the gut lumen by a unique epithelium which is interspersed with specialized microfold cells which internalize
c 600000 'ii
0 l.’
E
H H
600000
q
a
IO
20
30
40
50
l-10urn 20-50um l-10+20-50 urn
60
Day Post SC Immunization Fig. 2. Controlled vaccine release through alteration in microsphere size. BALB/c mice were immunized with SEB toxoid in 50: 50 DL-PLG microspheres of l-10 pm in diameter, 2&125pm in diameter, or in a mixture of I-10 and 20-125 pm diameter. Anti-SEB toxin antibodies of the IgG isotype in the plasma were determined by end-point titration in a RIA.
Biodegra~ble
microsphere
antigenic materials from the gut lumen and transport them to the underlying lymphoreticular cells. The PP lymphoid population contains a large proportion of B cells that are committed to IgA synthesis, and following antigen sensitization they extravasate the PP and home to diverse mucosal tissues. Large numbers of these IgA lymphobl~ts infiltrate the glandular tissues and lamina propria regions underlaying the mucosal membranes and undergo terminal differentiation to IgA synthesis. Although it is clear that this common mucosal immune system can be induced through experimental oral immunization (Mestecky et al., 1987), the large antigen doses required to achieve this have made it impractical for purified vaccine antigens. However, in addition to adjuvant activity, micr~ncapsulation provides two advantages in oral vaccine delivery. First, the wall material protects the vaccine from degradation by gastric acidity and the proteolytic enzymes of the gut. Second, relatively hydrophobic particles of appropriate size are readily absorbed by the M cells (LeFevre and Joel, 1984; LeFevre et al., 1980). The possibility of using microspheres as an oral vaccine delivery system was investigated by intragastrically administering single 20 mg doses of l-10 pm 85: 15 DL-PLG microspheres loaded with the fluorochrome coumarin-6 to mice (Eldridge et al., 1989a). Fluorescence microscopic observation of frozen sectioned tissues revealed that internalized microspheres were present in the macrophages of the Peyer’s patch dome region at 24 hr post-administration, and at all times tested through 35 days (Fig. 3a). The total number of microspheres observed in three representative Peyer’s patches increased through day 4 and then decreased over the following 31 days to 15% of the peak number. At 1, 2 and 4 days the proportion of < 5 pm (7682%) and > 5 pm (18-23%) particles remained constant and reflected the size distribution of the input preparation. Concurrent with the decrease in total microsphere numbers in the PP, which began at day 7, a progressive shift in the size distribution was observed such that the ~5 pm sizes ceased to predominate, and the > 5 pm microspheres became predominant. In addition, the microspheres which were observed to penetrate deep into the Peyer’s patches were almost exclusively 6 5 pm in diameter. The lymphatic drainage from the intestine passes through the mesenteric lymph nodes, and microspheres were detected in this tissue as soon as 24 hr after ingestion (Fig. 3b). The peak number of microspheres in the mesenteric lymph nodes was observed on day 7, when the total number in the Peyer’s patches had fallen significantly. All of the microspheres observed in the mesenteric lymph nodes were , 10 pm in diameter were not adsorbed. &&cosffl immunization with microe~c~psuiff ted an t&en The data on microsphere adsorption into and migration out of the Peyer’s patches suggested that
JOHNH. ELDRIDGEet al.
292
the type of immune response elicited through oral immunization could be controlled through varying the microsphere size. Microspheres ~5 pm extravasate from the Peyer’s patches within macrophages which disseminate into systemic lymphoid tissues, where vaccine release would be expected to induce a circulating antibody response. In contrast, microspheres > 5 pm remain in the IgA inductive environment of the Peyer’s patches, where vaccine recognition would stimulate a disseminated mucosal IgA immune response. Perhaps the most attractive possibility, concurrent circulating IgG and secretory IgA immunity, should be achievable through oral immunization with vaccine-microspheres spanning the size range of l-10pm. Three peroral immunizations at 30-day intervals with 100 pg of SEB toxoid in microspheres effectively induced the appearance of concurrent circulating and disseminated mucosal anti-toxin antibodies 20 days following the tertiary administration (Fig. 4). The plasma anti-toxin response included IgM and IgA components at titers of 2560 and a specific IgG titer in excess of 40,000. Each of the tested mucosal secretions, saliva, gut wash fluid and lung wash fluid, contained significant levels of toxin-specific IgA, and the washings from the gut and lungs additionally contained specific IgG. Parallel oral immunizations with doses of non-encapsulated toxoid ranging from 1 to 100 pg were ineffective at inducing a detectable
Plasma
Salka
Gut Wash
immune response in any of the tested fluids. In addition, doses of non-encapsulated toxoid above 50 pg per immunization induced oral tolerance as demonstrated by suppressed circulating antibody levels to systemic challenge (data not shown). In contrast to oral immunization, two IP immunizations with 100 pg doses of microencapsulated toxoid, separated by 30 days, induced a plasma IgG anti-toxin response at a titer of 6,533,600 on day 20 after boosting (Fig. 4). Despite the fact that this circulating IgG response was 64-fold higher than that induced through oral immunization with microencapsulated toxoid, no significant IgA anti-toxin antibodies were detected in the plasma or mucosal secretions. This hyperimmunization was, however, capable of resulting in the appearance of IgG antibodies in the samples of each of the mucosal secretions, suggesting that they were the result of transudation from the serum. Interrelationship between the systemic and mucosal immune systems The relationship between the systemic and mucosal immune systems was investigated by determining the isotype and distribution of antibodies induced through oral or intratracheal (it) boosting of systemically primed mice. As in the preceeding groups of mice, 100 pg of encapsulated toxoid was used at each administration and the plasma and secretion samples
Plasma
Lur!gWash
n
Saliva
Gut Wash LugWash
Saliva
Gut Wash LqWaah
I!#
r-lHlecl
Plasma
Saliva
Qti Wash
l.ungWprh
Plasma
-
Fig. 4. Levels and isotypes of anti-SEB toxin antibodies induced in the plasma and mucosal secretions of mice immunized by various protocols with SEB
toxoid
in I-10pm
5O:SO DL-PLG microspheres.
Biodegradable microspheres as a vaccine delivery system were collected 20 days after secondary immunization for evaluation in toxin-specific RIAs (Fig. 4). Oral immunization following systemic priming boosted the serum IgG anti-toxin titer to 819,200 and, unlike systemic boosting, resulted in the appearance of circulating IgA antibodies. In addition, the IP-Oral protocol was nearly as effective in the induction of mucosal IgA anti-toxin antibodies as was tertiary oral immunization. Similar, but more dramatic, results were obtained when the boosting was performed by it instillation of the microencapsulated toxoid (Fig. 4). The level of circulating IgG antibodies was boosted, and plasma IgA anti-toxin was detected to a titer of 3200. In addition, each of the mucosal secretions contained significant levels of specific IgG and IgA. This booster method was particularly effective at inducing both IgG and IgA antibodies in the lung fluids, with anti-toxin titers reaching 20,480 and 2560, respectively. DISCUSSION
Microspheres constructed of DL-PLG have been under investigation in recent years as injectable reservoirs for the long-term controlled release of pharmaceuticals (Tice and Cowsar, 1984; Redding et al., 1984). Several characteristics of these microspheres make them ideal as a delivery system for vaccines. The lactide-glycolide copolymer system is biodegradable, non-inflammatory and yields the biocompatible molecules lactic and glycolic acids upon degradation. The injection of antigen in microspheres < 10 pm in diameter provides a strong potentiation of the circulating antibody response which appears to be a function of both a depot effect and the rapid phagocytosis of the microspheres by antigen-presenting accessory cells. Although the exact contribution by each of these mechanisms to the immune enhancement is difficult to determine, substantially less adjuvancy has been observed with microspheres which are of a size that precludes their uptake by phagocytosis. Thus, it may be that the major mechanism of immune enhancement is through the efficient loading of a relatively high concn of antigen into accessory cells for presentation to T cells. Consistent with this possibility is the observation that the immune response to T cell-independent antigens, such as bacterial carbohydrates, is generally not enhanced through delivery in small diameter microspheres. The ability to provide distinct “pulses” of antigen release by the injection of a mixture of vaccine-containing microspheres with various degradation times appears to be a promising vaccination stategy. In addition to taking advantage of the immunopotentiation provided by this vaccine delivery system, timed booster immunizations are released without the need for additional contact with a health care professional. At present, using a combination of variables, including copolymer ratio and size, four discrete releases of antigen over a 120-day period have been achieved.
293
Currently available techniques should allow these releases to be spread over a period in excess of 1 year. Potentially this approach may prove useful in human immunization, particularly in third world countries, and in veterinary applications. When used as an oral delivery system, the polymer wall of the microspheres protects the vaccine antigen from degradation in stomach acid and from proteolysis in the gut. Further, there is adsorption of appropriately sized microspheres into the immune inductive environment of the PP, and no tissue penetration in other regions of the gut where T cells with suppressor phenotype predominate. This targeted delivery may be as important as the adjuvant activity of the microspheres in the conversion of the orally tolerogenic SEB toxoid into a form which is immunogenic after oral administration. Detailed examination of the behavior of the microspheres following adsorption from the gut demonstrated that size influenced their retention in the PP versus redistribution to other tissues. All the microspheres within the PP appeared to be within macrophages as judged by PAS staining for intracellular glycogen. Microspheres which were seen to traffic to other tissues did so through the known pattern of lymphoid drainage, were seen to remain within macrophages, and were essentially all < 5 pm in diameter. Although we have not directly tested the possibility of differentially inducing the circulating as opposed to the mucosal immune system through the use of very finely sized particles, three spaced oral immunizations with microencapsulated SEB toxoid did induce high levels of IgG anti-toxin antibodies in the plasma concurrent with the production of IgA antibodies in diverse mucosal tissues. Systemic immunization was found to effectively prime for a mucosal IgA response when a booster was administered to either the gut-associated or bronchial-associated lymphoid tissues. Thus, under these experimental conditions it does not appear that systemic immunization has induced a tolerance or down regulation of a subsequent mucosal antibody response, as has been described in other systems. This finding has potentially important implications for the development of single immunization strategies for the induction of concurrent systemic and mucosal antibody responses. CONCLUSIONS
Microspheres l-10 pm in diameter composed of a vaccine antigen impregnated in the copolymer poly(oL-lactide-co-glycolide) represent a pharmaceutically acceptable delivery system with immunopotentiating activity for humoral antibody responses. The product is a free flowing powder with excellent stability, an extended shelf life and which is easily suspended in physiologic diluents for administration. Vaccine release is controllable through alterations in the ratio of copolymers and/or the size of the micro-
JOHNH. ELDRIDGE et al.
294
spheres, allowing the construction of single injection formulations with a priming and one or more timed booster releases of vaccine. Orally administered microspheres are adsorbed by the M cells overlaying the Peyer’s patches and passed to the immune inductive environment of both the Peyer’s patches and systemic lymphoid organs and induce concurrent systemic and mucosal antibody responses.
Halpern M. S. and Koshland M. E. (1970) Novel subunit of secretory IgA. Nature (London) ZUI, 1276. Hanson L. A. (1961) Comparative immunological studies of the immune globulins of human milk and of blood serum.
Acknowledgements-The
LeFevre M. E. and Joel D. D. (1984) Peyer’s patch epithelium: an imperfect barrier. In Intestinal Toxicology (Edited by Schiller C. M.), p. 45. Raven Press, New York. Mestecky J. (1987) The common mucosal immune system and current strategies for induction of immune responses _ in external secretions. J. Clin. Immun. 7, 265. Mesteckv J. and McGhee J. R. (1987) Molecular and cellula; interactions involved in ‘IgA ‘biosynthesis and immune responses. Ado. Immun. 40, 153. Mestecky J., McGhee J. R., Crago S. S., Jackson S., Kilian M., Kiyono H., Babb J. L. and Michalek S. M. (1980) Molecular-cellular interactions in the secretory IgA response. J. Reticuloendothelial Sot. 28, 455. Miller R. A., Brady J. M. and Cutwright D. E. (1977) Degradation rates of resorbable inplants (polylactates and polyglycolates): Rate modificatibn with-changes in PLAIPGA copolvmer ratios. J. Biomed. Mater. Res. 11, ill. - Preis I. P. and Langer R. S. (1979) A single-step immunization by sustained antigen release. J. Immun. Meth. 28,
Int. Arch. Allergy Appl. Immun. 18, 241.
Kramp W. J., Six H. R. and Kasel J. A. (1982) Post-immunization clearance of liposome-entrapped adenovirus type __ 5 hexon. Proc. Sot. eip. Biol. Med: 369, 135. LeFevre M. E.. Hancock D. C. and Joel D. D. (1980) Intestinal barrier to large particles in mice. J. Tdxicoi Envir. Health 6, 691.
authors wish to thank Mr J.
Douglas Morgan for his excellent technical assistance and to MS Amie Stoppelbein for her help in the preparation of this manuscript. This work was supported by USPHS grants AI 24772, AI 21774 and contract DAMD17-86-6160 from the U.S. Army Medical Research Acquisition Activity. REFERENCES Allison A. C. and Gregoriadis G. (1974) Liposomes as immunologic adjuvants. Nature 252, 252. Amon R., Sela M., Parant M. and Chedid L. (1980) Antiviral response elicited by a complete synthetic antigen with built-in adjuvance. Proc. natn. Acad. Sci. U.S.A. 71, 6769.
Bienenstock J. and Befus A. D. (1980) Mucosal immunology. Immunology 41, 249. Chanock R. M. and Lemer R. A. (1984) Proceedings of the Cold Spring Harbor Conference: Modern Approaches to Vaccines. Cold Spring Harbor Laboratory.
Cowsar D. R., Tice T. R., Gilley R. M. and English J. P. (1985) Poly (lactide-co-glycolide) microcapsules for controlled release of steroids. In Methods in Enzymology; Drug and Enzymes Targeting (Edited by K. J. Widder and R. Green), p: 101. A&den& Press, New York. Dailev M. 0.. Post W. and Hunter R. L. (19771 Induction of c&mediated immunity to chemically modified antigens in guinea pigs. II. The interaction between lipid conjugated antigens, macrophages and T lymphocytes. J. Immun. 118, 963.
Eldridge J. H., Gilley R. M., Staas J. K., Moldoveanu A., Meulbroek J. A. and Tice T. R. (1989a) Biodegradable microspheres: A vaccine delivery system for oral immunization. Curr. Top. Microbial. Immun. 146, 59. Eldridge J. H., Hammond C. J., Meulbroek J. A., Staas J. K., Gilley R. M. and Tice T. R. (1989b) Contro!led vaccine release in the gut-associated lymphoid tissues. I. Orally administered biodegradable microspheres target the Peyer’s patches. J. Controlled Release 11, 205. Elson C. O., Ealding W. and Lefkowitz J. (1984) A lavage technique allowing repeated measurement of IgA antibody in mouse intestinal secretions. J. Immun. Meth. 67, 101.
193.
Redding T. W., Schally A. V., Tice T. R. and Meyers W. E. (1984) Long-acting delivery systems for peptides: Inhibition of rat prostate tumors by controlled release of D-Trp6-LH-RH from injectable microcapsules. Proc. natn Acad. Sci. U.S.A. 81, 5845.
Tice T. R. and Cowsar D. R. (1984) Biodegradable controlled release parenteral systems. Pharm. Technol. J. 8, 26.
Tomasi T. B. and Zigelbaum S. (1963) The selective occurrence of y,A globulin in certain body fluids. J. clin. Invest. 42, 1552.
Tomasi T. B., Tam E. M., Solomon A. and Pendergast R. A. (1965) Characteristics of an immune system common to certain external secretions. J. exp. Med. 121, 101. Tomasi T. B. (1970) Structure and function of mucosal antibodies. A. Rev. Med. 21, 281. Warren F. R., Spero L. and Mezger 3. R. (1973) Antigenicity of formaldehyde-inactivated staphylococcal enterotoxoid B. J. Immun. 111, 885. Warren H. S., Vogel F. R. and Chedid L. A. (1986) Current status of immunological adjuvant. A. Rev. Immun. 4, 369.
