Humoral immune responses to model antigen co-delivered with biomaterials used in tissue engineering

ARTICLE IN PRESS

Biomaterials 25 (2004) 295–304

Humoral immune responses to model antigen co-delivered with biomaterials used in tissue engineering Melissa M. Matzelle, Julia E. Babensee* Wallace H. Coulter Department of Biomedical Engineering, Georgia Tech/Emory Center for the Engineering of Living Tissues, Georgia Institute of Technology and Emory University, 315 Ferst Drive, Atlanta, GA 30332, USA Received 25 November 2002; accepted 19 June 2003

Abstract A model shed antigen, ovalbumin (OVA), was co-delivered with polymeric biomaterial carrier vehicles in C57BL6 mice to test whether the presence of the biomaterial acted as an adjuvant in the immune response towards the associated antigen. The biomaterials tested were non-biodegradable polystyrene microparticles and biodegradable 50:50 or 75:25 poly(lactic-co-glycolic acid) (PLGA) microparticles or scaffolds. For each biomaterial carrier vehicle, to assess the resulting time-dependent systemic humoral immune response towards the co-delivered OVA, the OVA-specific IgG concentration and isotypes (IgG2a or IgG1, indicating a predominant Th1 or Th2 response, respectively) were determined using ELISA. OVA co-delivered with biomaterial carrier vehicles supported a moderate humoral immune response that was maintained for the 18-week duration of the experiment. This humoral immune response was primarily Th2 helper T cell-dependent as indicated by the predominant IgG1 isotype. Furthermore, this humoral immune response was not material chemistry-dependent within the material set tested here. With the presence of the biomaterial resulting in an enhancement of the humoral immune response to co-delivered antigen, it appears that the biomaterial acts as an adjuvant in the development of an adaptive immune response to co-delivered antigen. r 2003 Elsevier Ltd. All rights reserved. Keywords: Poly(lactic-co-glycolic acid) microparticles and scaffolds; Adjuvant; Humoral immune response; Tissue engineering

1. Introduction Tissue engineering has the potential to revolutionize reconstructive surgery through the provision of engineered tissues and organs where no suitable alternative exists or to reduce morbidity associated with current procedures. Tissue-engineered constructs must evade the host defense system to avoid rejection by the immune system and/or the consequences of inflammation. Otherwise, the cells of the construct will die, lose function, and/or be replaced by non-functional fibrotic or granulation tissue. In a tissue-engineered device, implanted cells combined with a biomaterial can serve as a source of foreign shed antigens (allo- or xenoantigens) depending on the cell source. An immune response to *Corresponding author. Tel.: +1-404-385-0130; fax: +1-404-8944243. E-mail address: [email protected] (J.E. Babensee). 0142-9612/$ - see front matter r 2003 Elsevier Ltd. All rights reserved. doi:10.1016/S0142-9612(03)00531-3

such shed antigens would sensitize the host, leading to an antigen-specific immune response that could directly affect the successful integration of the tissue-engineered device into living systems [1]. Additionally, biomaterial implantation triggers an inflammatory reaction including important aspects of protein adsorption, complement activation, coagulation and neutrophil and macrophage adhesion and activation [2]. We hypothesize that the biomaterial component, by promoting an inflammatory response, recruiting antigen-presenting cells (APCs, e.g. macrophages and dendritic cells) and inducing their activation, acts as an adjuvant to the immune response towards foreign antigens originating from the cellular component of the device. The adjuvant effect of the biomaterial component of the tissueengineered device may intensify an immune response to shed antigens or shorten the time until its onset and ultimately lead to reduced viability and destruction of the device. The transplantation of allogeneic or

ARTICLE IN PRESS 296

M.M. Matzelle, J.E. Babensee / Biomaterials 25 (2004) 295–304

xenogeneic cells (without immunosuppression of the host or immunomodulation of the graft) typically results in a graft rejection process [3,4]. Combining these concepts, the host response to a tissue-engineered device may be viewed as an inflammatory response to the biomaterial and an immune response to the transplanted cells. The interconnections between the interactions of these two device components with the host are fundamental to understanding the in vivo response to tissueengineered devices. Using microencapsulated cells as an example of a tissue-engineered construct, antigens shed from encapsulated cells have been shown to stimulate the immune system leading to the destruction of encapsulated cells. Microencapsulation of living mammalian cells has been proposed as a means for cell delivery to provide a source of therapeutic biomolecules. Originally, the premise of immunoisolation was to physically separate allogeneic and even xenogeneic tissue from the host immune system using a polymer membrane to prevent contact with immunoglobulins, complement components, and immune and inflammatory cells, obviating the need for immunosuppression. In this way, the presence of the polymer membrane would prevent an immune response originating from the outside of the capsule from acting on cells inside the capsule. However, an immune response is potentially also possible towards antigens shed by encapsulated cells, such as foreign proteins secreted by the cells, including the therapeutic agent, cell surface molecules, or cell components (e.g., proteins, phospholipids, DNA) released upon cell death. These materials diffuse through the polymer membrane to be recognized as foreign by the host’s immune system. Degradation of biological capsule components, such as an extracellular matrix for cell attachment, could also generate immunologic or proinflammatory products. Shed antigens are internalized, processed and presented in association with the host Class II Major Histocompatibility Complex (MHC) molecules, most effectively by macrophages or dendritic cells, to host CD4+ helper T cells in the indirect pathway of antigen recognition [5]. In fact, survival of encapsulated xenogeneic tissue has been shown to depend on the presence of CD4+ helper T cells [6–8]. For example, depletion of CD4+, but not CD8+ T cells using monoclonal antibody treatment, in normal mice resulted in the survival of encapsulated monkey kidney xenogeneic tissue [7]. Furthermore, in athymic mice, reconstituted with the CD8+ cell-depleted preparation, death of encapsulated monkey cells occurred whereas the reconstitution with the CD4+ cell-depleted preparation did not. The study highlighted the critical role of CD4+ T cells, in the absence of CD8+ cells and B cells, in the processes leading to the ultimate destruction of encapsulated xenografts. Because of the cell-impermeable

membranes used, it was hypothesized that the most likely involvement of CD4+ T cells is in the indirect antigen recognition by these cells and subsequent stimulation of inflammatory cells. Other studies further demonstrated the importance of antigen presentation [9,10]. In vivo blockage of co-stimulatory molecule interaction (T cell surface antigen, CD28, with its APC ligands B7-1 and B7-2) with CTLA4-Ig inhibited indirect presentation of xenoantigens and the host T cell response and prolonged islet xenografts survival synergistically with microencapsulation in NOD mice [9,10]. The basis for the action of adjuvants is the role of innate immunity in stimulating adaptive immune responses. Dendritic cells recognize pathogens using conserved structures, uniquely characteristic of microbial pathogens, through their cognate binding receptors on the phagocytes resulting in their maturation such that they become efficient APCs [11]. For example, constituents of mycobacteria, the active component of the strong adjuvant complete Freund’s adjuvant (CFA), induce the maturation of dendritic cells as exemplified by induction of proinflammatory cytokine secretion (e.g., IL-12), upregulation of MHC molecules, and costimulatory molecules [12]. Concomitant with these characteristics of mature dendritic cells is the prolonged expression of antigen-MHC complexes on the surface of dendritic cells, again supporting the stimulation of T cells [13]. Adjuvants may also function through the prolonged retention of the antigen in the vicinity of the inflammatory stimulus—the depot effect—typical of controlled release devices for antigen delivery for vaccines [14]. Hence, given the immunological responses of potential consequence with tissue-engineered devices, the aim of this study was to test the hypothesis that the biomaterial component of a tissue-engineered construct acts as an adjuvant in the immune response towards associated antigens. For this purpose, an in vivo murine model was used to assess the extent to which the biomaterial component of the devices supports an immune response to a co-delivered model antigen. Using ovalbumin (OVA) as a model antigen, the systemic concentration of OVA-specific antibodies was determined for OVA delivered with polymeric carrier vehicles as compared to the positive and negative controls of OVA delivered with CFA and with phosphate-buffered saline (PBS), respectively. The level of enhancement of the antibody level as compared to the negative control was indicative of the ability of the polymer to act as an adjuvant in the enhancement of an antigen-specific immune response. We show that the presence of the biomaterial enhances the humoral immune response to co-delivered antigen, suggesting that the biomaterial acts as an adjuvant in the immune response to co-delivered antigen.

ARTICLE IN PRESS M.M. Matzelle, J.E. Babensee / Biomaterials 25 (2004) 295–304

2. Materials and methods 2.1. Animals Male C57BL/6 mice (6–8 weeks old) (Charles Rivers Laboratory, Wilmington, MA) were used as recipients for the antigen-vehicle combinations and allowed to acclimate to their new environment for 1 week prior to experimentation. Mice were housed in microisolator cages (3 per cage, 2 cages per treatment type) in the Winship Cancer Center animal facility at Emory University. Animal care and treatment were in compliance with the Institution Animal Care and Use Committee (IACUC) of Emory University according to protocol #040-2000. 2.2. Polymeric carrier vehicles preparation and characterization 2.2.1. Preparation of polymeric biomaterial microparticles Poly(dl-lactide-co-glycolide) (PLGA) microparticles (PLGA MP) were prepared using a single-emulsion solvent-extraction technique adapted from Wake et al. [15]. Briefly, 500 mg of 50:50 or 75:25 PLGA (Birmingham Polymers, Birmingham, AL) was dissolved in 20 ml dichloromethane (DCM) (Sigma Chemical Company, St. Louis, MO) overnight. The PLGA-DCM solution was then added to 200 ml of a 0.3% (w/v) poly(vinyl alcohol) (PVA) (Sigma) aqueous solution that had been sterile-filtered with a 0.22 mm cellulose acetate filter (Corning, Corning, NY). The PLGA-DCM-PVA solution was emulsified for 2 min at 9000 rpm using a PowerGen 700 homogenizer (115 V) with an external speed control and a 7
95 mm flat bottom generator (Fisher Scientific, Pittsburgh, PA). Precipitation of the PLGA MP was initiated with the addition of 200 ml of 2% (v/v) aqueous isopropanol (Sigma). The solvent was evaporated from the mixture while stirring for 16 h at room temperature. The PLGA MP were collected by centrifugation at 210g for 10 min, resuspended in the isopropanol solution, and washed with 2% isopropanol (twice) and sterile filtered distilled/deionized water (3 times). The resultant PLGA MP were finally resuspended in sterile filtered distilled/deionized water and placed under ultraviolet light (UV) in a laminar flow hood for 1 h for sterilization. The mean size and standard deviation of the PLGA MP were obtained using a Coulter Multisizer II (Coulter Corporation, Miami, FL). For 50:50 PLGA MPs, sizes ranged from 2 to 26 mm with a mean diameter of 8 mm and for 75:25 PLGA MPs, sizes ranged from 2 to 39 mm with a mean diameter of 7 mm. PLGA MP were stored at 4 C for no more than 12 h prior to injection. Prior to usage, endotoxin was removed from 6 mm diameter polystyrene microparticles (PS MP) (Poly-

297

sciences, Warrington, PA) using a series of basic, acidic, and ethanol washes with endotoxin-free reagents (Sigma): 0.1 n NaOH (2 times), 0.1 n HCl (2 times), 95% ethanol (one time) for 20 min each in a sonicator [16]. In between each wash, PS MP were washed with endotoxin-free water for 10 min. The PS MP were subsequently washed two additional times with EF water and resuspended in EF water at a concentration of 2.22
108 MP/ml prior to injection. 2.2.2. Preparation of PLGA scaffolds 50:50 and 75:25 PLGA scaffolds (SC) (50:50 PLGA SC and 75:25 PLGA SC, respectively) were prepared by a salt-polymer casting particulate-leaching technique with NaCl as the leachable component [17,18]. After milling with an analytical mill (Tekman, Cincinnati, OH), NaCl was sieved into particle size range using a sieve shaker (W.S. Tyler, Mentor, OH) and 355–425 mm NaCl particles were used as the leachable component. Briefly, 500 mg of 50:50 or 75:25 PLGA (Birmingham Polymers) was added to 5 ml DCM and the polymer was allowed to dissolve completely. The PLGA-DCM was poured over 4.5 g NaCl (355–425 m) in a 50 mmdiameter Teflon petri dish. The mixture was stirred completely, covered, and placed in a fume hood. Following overnight incubation, the cover was removed and the mixture remained overnight in the fume hood to allow the DCM to evaporate. The disks were then removed and placed in distilled/deionized water on a shaker to dissolve out the leachable component. The water was changed frequently for several hours and the disks were allowed to remain in the water overnight on a shaker. The porous disks were air dried on sterile pads in a laminar flow hood. Once dry, 0.7 cm diameter and 0.2 cm thick SC were punched out of the disk. Intact SC were stored in a dessicator at 4 C until sterilization. Twenty-four hours prior to implantation, the SC were washed with 70% ethanol (3 times) and sterile-filtered distilled/deionized water (3 times) for 10 min each on a shaker. The SC were dried completely on sterile pads in a sterile laminar flow hood and stored in a dessicator at 4 C. Several hours prior to implantation, the scaffolds were placed under UV light in a laminar flow hood for 25 min per side and returned to the dessicator at 4 C until preparation for implantation. 2.2.3. Polymer quantification The dry weight of PS MP, 50:50 PLGA MP, or 75:25 PLGA MP at a concentration of 1.6875
108 MP/ml or a single 50:50 PLGA SC or 75:25 PLGA SC, administered during immunizations, was determined using preweighed cryovials. The vials containing the polymers were flash frozen with liquid nitrogen and lyophilized overnight in a Freeze Dry System/Freezone 4.5 (Labconco, Kansas City, Missouri). Polymer weights were

ARTICLE IN PRESS M.M. Matzelle, J.E. Babensee / Biomaterials 25 (2004) 295–304

298

determined in triplicate and the mean weight and standard deviation was determined. 2.2.4. Quantification of protein associated with polymer carrier vehicles Bicinchoninic acid (BCA) protein assay (Sigma) was used to determine the amount of chicken egg OVA (Grade VI, Sigma) associated with pre-soaked polymeric carrier vehicles as well as the amount that remained in the incubating solution. Polymer microparticles (1.6875
108 MP/ml) or scaffolds were pre-soaked for 1 h in PBS containing OVA at concentrations of 0, 0.5, 1, or 3 mg/ml or 0, 1, 2, or 6 mg/ml, respectively (similar to the method used prior to implantation of these OVApolymer carrier vehicles). The suspension was centrifuged at 260g for 5 min and the supernatant removed and stored at 4 C for BCA analysis. The remaining MP pellet was washed with PBS (3 times, 5 min), resuspended in 2% (w/v) sodium dodecyl sulfate (SDS) in tris-buffered saline (25 mm TRIS base, 8 g/l NaCl, 0.2 g/l KCl, pH 7.4) and rotated overnight on a Labquake rotator (Barnstead/Thermolyne, Dubuque, IA) as previously described [19]. The eluted protein supernatant was recovered and stored at 4C until analysis by BCA. To determine the amount of protein eluted from the polymeric carrier, vehicles or remaining in solution, the BCA assay was preformed according to the manufactures instructions. Briefly, BCA solution (Reagent A) and copper (II) sulfate pentahydrate 4% solution (Reagent B) were combined in a 50:1 ratio to generate BCA working reagent. Protein standards ranging from 1 mg/ml to 200 mg/ml were made by making serial dilutions from the given 1 mg/ml bovine serum albumin (BSA) standard solution. The unknown samples and standards were added to a 96-well plate in a 1:8 ratio with the BCA working reagent. The plates were sealed and incubated at 37 C for 30 min. The absorbance values were read at 562 nm by a Powerwavex340 ELISA plate reader (Bio-Tek Instruments, Inc., Winooski, VT) and corresponding OVA concentration was determined using the standard curve. The mean concentration and standard deviation for each set of triplicate samples was determined.

2.3. Humoral immune response assay 2.3.1. Co-delivery of model antigen with polymeric carrier vehicles in mice Prior to murine injection, PS MP, 50:50 PLGA MP, or 75:25 PLGA MP were centrifuged (260g, 5 min) and resuspended (final concentration 1.6875
108 MP/ml) in PBS containing OVA at concentrations of 0.5, 1, or 3 mg/ ml for 1 h. As a negative control, MP were also suspended PBS alone. For the murine implantation, 50:50 or 75:25 PLGA SC were pre-soaked in PBS containing OVA at concentrations of 1, 2, or 6 mg/ml for 1 h. As a negative control, SC were also pre-soaked in PBS alone. C57BL/6 mice were given a 100 ml dorsal subcutaneous injection of the MPs in their associated OVA solution or a dorsal implantation of a PLGA SC with their associated OVA solution under sterile surgical conditions (Table 1). Immunization with 0, 0.5, 1, or 3 mg/ml OVA in PBS served as the negative controls and immunization with 0, 1, 2, or 6 mg/ml OVA in PBS in a 1:1 dilution of Complete Freud’s Adjuvant (CFA) (Sigma) served as the positive controls. Three weeks after primary immunization, mice received a booster of the same OVA concentration with the identical polymeric carrier vehicle. For the positive controls, Incomplete Freud’s Adjuvant (IFA) (Sigma) was used as the carrier instead of CFA. For the PLGA SC, the same OVA concentration was injected into the SC site. Six mice were placed in each treatment or control group. Blood samples were collected from the retro-orbital plexus of the mice at 2, 3 (prior to boosting), 4, 8, 12, and 18 weeks after primary immunization. Following clotting overnight at 4 C, the serum was removed after centrifugation (2300g, 10 min). The serum was stored at 20 C until analysis for anti-OVA IgG antibodies and isotypes by ELISA. 2.3.2. Quantification of anti-OVA IgG and isotypes by ELISA The production of anti-OVA total IgG, and the isotypes, IgG1 and IgG2a, in mouse serum samples was measured by ELISA as previously described [20]. Standard wells of Nunc Immunot MaxiSorp ELISA plates (Life Technologies, Paisley, UK) were coated with

Table 1 Polymeric carrier vehicles and OVA concentration (mg/ml) used for in vivo humoral immune analysis Carrier vehicle

[OVA] (mg/ml)

Polymeric MP

Polymeric SC

PBS

CFA

PS MP

50/50 PLGA MP

75/25 PLGA MP

50/50 PLGA SC

75/25 PLGA SC

0 0.5 1 3

0 0.5 1 3

0 0.5 1 3

0 0.5 1 3

0 0.5 1 3

0 1 2 6

0 1 2 6

ARTICLE IN PRESS M.M. Matzelle, J.E. Babensee / Biomaterials 25 (2004) 295–304

a 1 mg/ml solution of goat anti-mouse IgG, IgG1, or IgG2a (Southern Biotechnology Associates, Birmingham, AL) in 0.1 m sodium bicarbonate (NaHCO3) (pH 9.5). Sample wells of the plate were coated with a 40 mg/ ml solution of OVA in 0.1 m NaHCO3. After incubation overnight at 4 C, the plates were washed four times with 0.5% (v/v) Tween 20 in PBS (PBT) and blocked with 5% (w/v) condensed milk in PBT (PBT-CM) for 2 h at 37 C. After four washes with PBT, triplicates of each serum samples were diluted 1/100, 1/1000, or 1/10,000 in PBT-CM and added to the sample wells. To the standard wells, mouse IgG, IgG1, or IgG2a standard was diluted 1:1 in PBT-CM starting at 1 mg/ml to a final concentration of 0.488 ng/ml and added to the standard wells, each concentration in duplicate. Following incubation for 2 h at 37 C, the plates were washed four times with PBT and incubated for 2 h at 37 C with a 1:1000 dilution of alkaline phosphatase-conjugated goat anti-mouse IgG, IgG1, or IgG2a (Southern Biotechnology Associates) in PBT-CM. Following four washes with PBT, the reaction was developed using a p-nitrophenyl phosphate solution prepared from tablets dissolved in 1
diethanolamine buffer prepared from an alkaline phosphate substrate kit (BIO-RAD, Hercules, CA). The reaction was stopped after 2 min by the addition of 0.4 m sodium hydroxide (NaOH) and the absorbance values were read at 405 nm using a Powerwavex340 ELISA plate reader. The mean absorbance was calculated for each set of triplicate serum samples. The mean concentration and standard deviation for each set of samples was determined using a standard curve.

299

2.3.3. Data analysis Statistical analysis was performed using an ANOVA’s general linear model with Minitab software (Version 13, Minitab Inc., State College, PA). P-values of p0.05 were denoted as significant.

3. Results 3.1. Humoral immune response to antigen co-delivered with polymeric vehicles OVA co-delivered with the strong adjuvant, CFA (positive control) into C57BL6 mice elicited an intense humoral immune response characterized by the production of high levels of anti-OVA antibodies (Fig. 1). Codelivery of 3, 1, or 0.5 mg/ml of OVA with CFA elicited similar levels of anti-OVA IgG. In addition, CFA without added OVA (0 mg/ml) resulted in negligible production of anti-OVA IgG. Since the levels of antiOVA IgG levels were OVA concentration-independent, only the results using the 1 mg/ml OVA concentration with the polymeric carrier vehicles are shown. For all polymeric vehicles, a similar OVA concentration-independent trend was observed. OVA administered at varying concentrations in PBS (negative control) resulted in a low baseline level of antibody production for all OVA concentrations (data not shown). Co-delivery of OVA (1 mg/ml) with the polymeric microparticle or scaffold carrier vehicles into C57BL6 mice elicited moderate levels of anti-OVA IgG antibodies that were maintained for the 18-week duration of the experiment (Fig. 2). For all these polymeric carrier

Fig. 1. Production of anti-OVA IgG as a function of time by C57BL/6 mice as a result of 0, 0.5, 1, or 3 mg/ml OVA co-delivered with CFA. Mean7SD, n ¼ 426 mice. ‘‘’’ different from 0 mg/ml. ‘‘+’’ different from 0.5 mg/ml.

ARTICLE IN PRESS 300

M.M. Matzelle, J.E. Babensee / Biomaterials 25 (2004) 295–304

Fig. 2. Production of anti-OVA IgG as a function of time by C57BL/6 mice as a result of 1 mg/ml OVA co-delivered with various carrier vehicles: PBS, CFA, PS MP, 50:50 PLGA MP, 75:25 PLGA MP, 50:50 PLGA SC, and 75:25 PLGA SC. Mean7SD, n ¼ 426 mice, note that the ordinate is in log scale. ‘‘+’’ different from CFA. ‘‘’’ different from PBS.

vehicles, the anti-OVA IgG antibody levels were statistically different from the negative control of OVA delivered in PBS alone (except for PS MP at 2 weeks) and also statistically different from the positive control of OVA delivered in CFA. The co-delivery of OVA with PLGA SCs resulted in anti-OVA IgG levels that were slightly, but statistically greater than the levels observed with PLGA MPs at 2 and 3 weeks. However, by 4 weeks and after, the levels of antibody production were similar to that seen with the MP (except for the polystyrene beads which remained different from the 75/ 25 PLGA SC until 8 weeks). The humoral immune responses for all three types of MPs were similar, suggesting that the production of antigen specific antibodies is not material chemistry-dependent within the material set tested here. To assess the predominance of a Th1 versus a Th2 helper T cell response in the humoral immune response to OVA delivered with polymeric carrier vehicles, the IgG isotypes were determined. The IgG2a isotype is characteristic of a Th1 response, whereas the IgG1 isotype is characteristic of a Th2 response. OVA (1 mg/ ml) co-delivered with polymeric carrier vehicles elicited baseline levels of OVA-specific IgG2a, generally not statistically different from the levels observed for OVA delivered in PBS, both of which were statistically significantly less than that observed for OVA delivered in CFA (Fig. 3). Exceptions were for OVA co-delivered with 50:50 PLGA MP which elicited slightly higher IgG2a isotope levels at 2 and 3 weeks, but at 4 weeks and after, the level was similar to that seen with PBS. In addition, OVA co-delivered with 50:50 PLGA SC elicited slightly higher IgG2a isotype levels at 4, 8 and 18 weeks, and similarly PS MP elicited slightly higher

IgG2a isotype levels at 12 weeks, as compared to the levels observed with PBS. The generation of anti-OVA IgG1 antibodies in response to OVA co-delivered with each polymeric carrier vehicle (Fig. 4) showed similar trends to those observed for total IgG (Fig. 2). The IgG1 isotype was the predominant IgG isotype with similar moderate levels of OVA-specific IgG1 observed for all polymeric carrier vehicles being maintained for the 18-week duration of the experiment. Again, these responses were generally statistically different from the negative control of OVA delivered in PBS alone and also statistically different from the positive control of OVA delivered in CFA (Fig. 4). Furthermore, the polymer SC carrier vehicles with 1 mg/ml co-delivered OVA, generated slighter higher IgG1 antibody levels at 2 and 3 weeks than the MP carrier vehicles, but at 4 weeks and after, the antibody levels are similar to those associated with the MP.

4. Discussion Antigen shedding from tissue-engineered constructs is a potential means for immunorecognition of tissueengineered grafts even without direct contact of host immune cells with the donor cells. In this study, codelivery of OVA with polymeric biomaterial carrier vehicles in the form of microparticles or scaffolds resulted in antigen-specific antibody titers that indicated a moderate immune response to the co-delivered antigen as compared to control treatments. Specifically, the immune response as a result of antigen delivery with microparticles or scaffolds was less than that of antigen

ARTICLE IN PRESS M.M. Matzelle, J.E. Babensee / Biomaterials 25 (2004) 295–304

301

Fig. 3. Production of anti-OVA IgG2a as a function of time by C57BL/6 mice as a result of 1 mg/ml OVA co-delivered with various carrier vehicles: PBS, CFA, PS MP, 50:50 PLGA MP, 75:25 PLGA MP, 50:50 PLGA SC, and 75:25 PLGA SC. Mean7SD, n ¼ 426 mice, note that the ordinate is in log scale. ‘‘+’’ different from CFA. ‘‘’’ different from PBS.

Fig. 4. Production of anti-OVA IgG1 as a function of time by C57BL/6 mice as a result of 1 mg/ml OVA co-delivered with various carrier vehicles: PBS, CFA, PS MP, 50:50 PLGA MP, 75:25 PLGA MP, 50:50 PLGA SC, and 75:25 PLGA SC. Mean7SD, n ¼ 426 mice, note that the ordinate is in log scale. ‘‘+’’ different from CFA. ‘‘’’ different from PBS.

delivered in CFA, but greater than that of antigen delivered in PBS alone. These results indicated that the biomaterial acted as an adjuvant in the enhancement of the immune response to the co-delivered antigen. The presence of the polymeric carrier vehicle in the absence of antigen did not induce an immune response, but was similar in antibody titer to PBS alone, indicating that the biomaterial alone did not induce an immune response. Future assessment of the biomaterial adjuvant effect are focused on determining the level of in vivo proliferation of OVA-specific T cells derived from

transgenic mice [21] in response to OVA delivered with polymers as compared to appropriate controls. Preliminary results have indicated that the level of OVAspecific CD4+ T cell proliferation in response to OVA co-delivered with PLGA microparticles was greater than that observed with antigen delivered in PBS and similar to the level observed with the strong adjuvant, CFA [M.M. Matzelle and J.E. Babensee, unpublished data]. Future development of this assay will allow for quantification of the extent of in vivo T cell proliferation.

ARTICLE IN PRESS 302

M.M. Matzelle, J.E. Babensee / Biomaterials 25 (2004) 295–304

There is support for the idea of an adjuvant effect associated with biomaterials primarily in the area of vaccine delivery, in which an enhancement of an immune response is desirable. Current vaccine research has shown that in vivo administration of PLA or PLGA biodegradable microspheres with incorporated antigenic proteins are potential single dose vaccines, in which the polymeric biomaterial serves as an adjuvant [20,22,23]. The particulate form of the antigen may facilitate the uptake of antigen by APCs, resulting in enhanced activation of macrophages and dendritic cells [22]. Another study showed that the cell type phagocytosing injected PLGA microparticles depended on the site of injection, where these microparticles were taken up by macrophages following intraperitoneal injection and by dendritic cells following intradermal injection [24]. In vitro it was shown that dendritic cells were almost as effective as macrophages at phagocytosing PLGA microparticles [25,26] and that such phagocytosis resulted in dendritic cell maturation [26]. The presence of the biomaterial may augment the phagocytosis of antigen, expression of MHC class II and co-stimulatory molecules on APCs, and the subsequent presentation of antigen to CD4+ T cells [27]. The adjuvanticity mechanism observed with vaccines may also be similar to the adjuvant effect observed in this research. For example, in separate studies we have seen that the biomaterials in microparticulate or film form may act as maturation stimuli for dendritic cells, potent APCs, resulting in the moderate upregulation of CD40, CD80, CD83, CD86, HLA-DR, and HLA-DQ [28]. However, in contrast to vaccine delivery, immune responses are to be minimized, not enhanced, in the area of tissue engineering. An adjuvant effect in tissue-engineered devices is not desirable. Scaffolds with co-delivered antigen displayed the same general trends as those observed with the microparticles, except for slightly higher antibody levels at 2 and 3 weeks. This early response may be due to a combination of factors. Previous studies have shown that dendritic cells can become activated in the absence of foreign substances, through endogenous ‘‘danger signals’’ from necrotic or stressed cells [29]. According to this ‘‘danger signal’’ model, immunity may be mediated by ‘‘danger signals’’ released by damaged or necrotic cells. The implantation of the scaffolds as a model tissue-engineered device resulted in necrotic damage to adipose tissue, epithelial cells, capillaries, and other cellular components. These components may have acted as ‘‘danger signals’’ for the initiation of an immune response through the recruitment and activation of dendritic cells. Enhanced dendritic cell activity in the initial weeks of implantation may lead to increased antigen uptake, presentation to T cells, and resultant antibody production by B cells. The surgical implantation of the scaffolds would have mediated a greater

acute inflammatory response as compared to the subcutaneous injection of the microparticles, which may explain the greater antibody levels seen during the initial weeks. In addition to the inflammation associated with the implantation, the greater amount of polymer delivered with the SC compared to the MP may have mediated the greater antibody titers seen for the SC at weeks 2 and 3. The amount of polymer delivered with 1.6875
108 MP/ ml was 1.3770.30 mg for PS MP, 4.2770.23 mg for 50:50 PLGA MP and 370.28 mg for 75:25 PLGA MP. The amount of polymer delivered was 9.9671.73 mg for 50:50 PLGA SC and 8.9872.10 mg for 75:25 PLGA SC. The presence of the large non-phagocytosible polymer SC may have resulted in ‘‘frustrated phagocytosis’’ and induced enhanced APC recruitment to the implant site during the first 3 weeks following implantation. The recruitment and activation of these cells are responsible for the production of cytokines, growth factors, proteolytic enzymes, and reactive oxygen intermediates in an attempt to degrade, ingest, or remodel the foreign material. The generation of these molecular products may also be responsible for subsequent activations of more immunocompetent cells, such as B cells, that would mediate the enhanced antibody levels seen in the initial weeks. The actual infiltration of these APCs, such as DCs, into the site of SC implantation or MP injection as well as the subsequent activation of these immune cells should be examined with and without associated antigen in order to test this hypothesis. Preliminary results have indicated more infiltration of DCs into PLGA scaffolds with associated OVA than without [A. Paranjpe, and J.E. Babensee, unpublished data]. In addition, tighter control of the amount of polymer delivered with each of the carrier vehicles would further the elucidation of the role of polymer quantity in the recruitment and activation of APCs by the biomaterial. OVA was adsorbed to these polymeric carrier vehicles from a bulk solution and the combination delivered to C57BL6 mice. Determination of the protein associated with polymeric carrier vehicles and that remaining in the incubating solution indicated that the majority of the OVA remained in solution with little to no OVA associating with the polymer (data not shown). While the humoral immune response elicited by the SC appears to be comparable to that generated by the MP at 4 weeks and after, there are experimental factors that may have mediated this similarity. Since most of the OVA remains in solution and not associated with the polymer carrier vehicles, the implantation of the SC with little associated with OVA solution, which resulted in the delivery of small amount of antigen. Even though the SC were incubated in a solution with two times the antigen concentration of the MP incubating solution (1, 2, or 6 mg/ml OVA for SC compared to 0.5, 1, or

ARTICLE IN PRESS M.M. Matzelle, J.E. Babensee / Biomaterials 25 (2004) 295–304

3 mg/ml for MP), little of the total antigen remaining in the SC incubating solution would have been delivered to the host upon implantation. In contrast, the injection of the MP suspended in an OVA solution would have resulted in the delivery of not only the polymer associated antigen but also the OVA which remained in the incubating solution. Therefore, the amount of antigen associated with the SC in addition to the antigen found in the implanted incubating solution may be minimal compared to the amount of antigen in the injected MP-OVA suspension. As a result, injection of the MP could have delivered more antigen as compared to the SC. The similarity in the humoral responses to the MP and SC may not only be a result of the biomaterial but also the differences in the antigen concentrations in the host. More accurately controlling the amount of antigen delivered by the biomaterial carrier vehicles may more precisely elucidate the differences in the humoral immune responses generated by the SC as compared to the MP. We are further controlling this effect by using PLGA MP or SC with incorporated OVA for controlled release for delivery of the same amount of OVA to the host. The immune response elicited by the MP and SC carrier vehicles with associated antigen were dominated by a Th2-type immune response as indicated by the production of IgG1 antigen-specific antibodies. This indicated that the adjuvanticity of the biomaterial resulted in the activation of CD4+ T cells and the proliferation and isotype switching of B cells. A Th1 immune response, which is characterized by the activation of macrophages, does not appear to play a major role, at least based on these experiments. The role of Th1 and Th2 immune responses in the acceptance or rejection of an implanted tissue-engineered construct is currently unknown. Research with allogeneic and xenogeneic tissue transplantations has implicated Th2 cytokines with prolonged graft survival [30–32] and Th1 cytokines with graft rejection [31–33]. The basis for this theory is the reduced activation and proliferation of Th1 response as a result of the production of Th2 cytokines, but several studies indicate the contrary [31]. Whether or not the Th2-mediated immune response observed in this research is responsible for the acceptance or rejection of a tissue-engineered device is merely speculative. It is possible that the production of cytokines such as, IL-4, IL-5, IL-6, and IL-10, which inhibit macrophage activation, might be detrimental to the construct integrity. In addition, the production of the non-complement fixing antibody, IgG1, would not result in the activation of the classical pathway of complement activation [33]. In contrast, the production of IgG1 may still bind to the cellular component of the tissue-engineered implant and initiate neutrophil and macrophage activation. To elucidate these mechanisms, it would be necessary to understand

303

the cytokine production profiles produced as a result of the implantation of a tissue-engineered construct.

5. Conclusions Trends elucidated by this research are central to the understanding of host immune responses to tissueengineered devices. Antigen shedding from tissueengineered constructs, either in the form of cells seeded on polymer scaffolds or cells encapsulated within polymer matrices, is a potential means for immunorecognition of tissue-engineered grafts even without direct contact of host immune cell with the donor cells. The experiments described herein indicate that OVA codelivered with biomaterial carrier vehicles result in the production of a moderate, predominately Th2 helper T cell-dependent, humoral immune response that was maintained for the 18-week duration of the experiment. This result supports the hypothesis that the biomaterial acts as an adjuvant in the immune response to associated antigens. The response for the three types of microparticle materials was similar, suggesting that the production of antigen-specific antibodies is not material chemistry-dependent within the material set tested here. These simplified systems served as the first phase in evaluating the adjuvanticity of biomaterials that are commonly used in tissue-engineered constructs. Understanding the roles of both the biomaterial and the cellular component in the immune response is critical to developing constructs with long-term structural and functional integrity in vivo.

Acknowledgements This work was supported by funding from the Georgia Tech/Emory Center (GTEC) for the Engineering of Living Tissues, an Engineering Research Council (ERC) program of the National Science Foundation under award number EEC-9731643. The authors acknowledge helpful discussions with Professor Thomas Prihoda, Department of Pathology, University of Texas Health Science Center in San Antonio, Texas.

References [1] Babensee JE, Anderson JM, McIntire LV, Mikos AG. Host responses to tissue engineered devices. Adv Drug Del Rev 1998;33:111–39. [2] Anderson JM. Mechanisms of inflammation and infection with implanted devices. Cardiovasc Pathol 1993;2:33S–41S. [3] Platt JL. Mechanisms of tissue injury in hyperacute xenograft rejection. ASAIO J 1992;38:8–16. [4] Chandler C, Passaro E. Transplant rejection mechanisms and treatment. Arch Surg 1993;128:279–83.

ARTICLE IN PRESS 304

M.M. Matzelle, J.E. Babensee / Biomaterials 25 (2004) 295–304

[5] Sayegh MH, Watschinger B, Carpenter CB. Mechanisms of T cell recognition of alloantigen. The role of peptides. Transplantation 1994;57:1295–302. [6] Weber CJ, Zabinski S, Koschitzky T, Wicker L, Rajotte R, D’Agati V, Peterson L, Norton J, Reemtsma K. The role of CD4+ helper T cells in the destruction of microencapsulated islet xenografts in NOD mice. Transplantation 1990;49:396–404. [7] Loudovaris T, Mandel TE, Charlton B. CD4+ T cell mediated destruction of xenografts within cell-impermeable membranes in the absence of CD8+ T cells and B cells. Transplantation 1996; 12:1678–84. [8] McKenzie AW, Geogiou HM, Zhan Y, Brandy JL, Lew AM. Protection of xenografts by a combination of immunoisolation and a single dose of anti-CD4 antibody. Cell Transplant 2001;10:183–93. [9] Weber CJ, Hagler MK, Chryssochoos JT, Larsen CP, Pearson TC, Jensen P, Kapp JA, Lindsey PS. CTLA4-Ig prolongs survival of microencapsulated rabbit islet xenografts in spontaneously diabetic NOD mice. Transplant Proc 1996;28:821–3. [10] Weber CJ, Hagler MK, Chryssochoos JT, Kapp JA, Korbutt GS, Rajotte RV, Linsley PS. CTLA4-Ig prolongs survival of microencapsulated neonatal porcine islet xenografts in diabetic NOD mice. Cell Transplant 1997;6:505–8. [11] Janeway CA, Medzhitov R. Introduction: the role of innate immunity in the adaptive immune response. Semin Immunol 1998;10:349–50. [12] Tsuji S, Matsumoto M, Takeuchi O, Akira S, Azuma I, Hayashi A, Toyoshima K, Seya T. Maturation of human dendritic cells by cell wall skeleton of Mycobacterium bovis bacillus CalmetteGuerin: involvement of toll-like receptors. Infect Immunol 2000;68:6883–90. [13] Cella M, Engering A, Pinet V, Peieters J, Lanzavecchia A. Inflammatory stimuli induce accumulation of MHC class II complexes on dendritic cells. Nature 1997;388:782–7. [14] Zhao Z, Leong KW. Controlled delivery of antigens and adjuvants in vaccine development. J Pharm Sci 1996;85:1261–70. [15] Wake MC, Gerecht PD, Lu L, Mikos AG. Effects of biodegradable polymer particles on rat marrow-derived stromal osteoblasts in vitro. Biomaterials 1998;19:1255–68. [16] Gorbet MB, Sefton MV. Leukocyte activation and leukocyte procoagulant activities after blood contact with polystyrene and polyethylene glycol-immoblization polystyrene beads. J Lab Clin Med 2001;137:345–55. [17] Mikos AG, Thorsen AJ, Czerwonka LA, Bao Y, Langer R, Winslow DN, Vacanti JP. Preparation and characterization of poly(l-lactic acid) foams. Polymer 1994;35:1068–77. [18] Ishaug SL, Crane GM, Miller MJ, Yasko AW, Yaszemski MJ, Mikos AG. Bone formation by three-dimensional stromal osteoblast culture in biodegradable polymer scaffolds. J Biomed Mater Res 1997;36:17–28. [19] Babensee JE, Cornelius RM, Brash JL, Sefton MV. Immunoblot analysis of proteins associated with HEMA-MMA microcapsules:

[20]

[21]

[22]

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30]

[31]

[32]

[33]

human serum proteins in vitro and rat proteins following implantation. Biomaterials 1998;19:839–49. Coombes AGA, Lavelle EC, Davis SS. Biodegradable lamellar particles of poly(lactide) induce sustained immune response to a single dose of adsorbed protein. Vaccine 1999;17:2410–22. Barnden MJ, Allison J, Heath WR, Carbone FR. Defective TCR expression in transgenic mice contrasted using cDNA-based [a]and [b]-chain genes under the control of heterologous regulatory elements. Immunol Cell Biol 1998;76:34–40. Igartua M, Hernandez RM, Esquisabel A, Gascon AR, Calvo MB, Pedraz JL. Enhanced immune response after subcutaneous and oral immunization with biodegradable PLGA microspheres. J Control Rel 1998;56:63–73. Nakaoka R, Inoue Y, Tabata Y, Ikada Y. Size effect on the antibody production induced by biodegradable microspheres containing antigen. Vaccine 1996;14:1251–6. Newman KD, Elamanchili P, Kwon GS, Samuel J. Uptake of poly(d,l-lactic-co-glycolic acid) microspheres by antigen-presenting cells in vivo. J Biomed Mater Res 2002;60:480–6. Walter E, Dreher D, Kok M, Thiele L, Kiama SG, Gehr P, Merkle HP. Hydrophilic poly(dl-lactide-co-glycolide) microspheres for the delivery of DNA to human-derived macrophages and dendritic cells. J Control Rel 2001;76:149–68. Thiel L, Rothen-Rutishauser B, Jilke S, Wunderli-Allenspach H, Merkle HP, Waler E. Evaluation of particle uptake in human blood monocyte-derived cells in vitro. Does phagocytosis activity of dendritic cells measure up with macrophages. J Control Rel 2001;76:59–71. Ke Y, McGraw CL, Hunter RL, Kapp JA. Nonionic triblock copolymers facilitate delivery of exogenous proteins into the MHC Class I and Class II processing pathways. Cell Immunol 1997;176:113–21. Yoshida M, Babensee JE, Poly(lactic-co-glycolic acid) moderately enhances maturation and function of human monocyte-derived dendritic cells. J Lab Clin Med, in preparation. Gallucci S, Lolkema M, Matzinger P. Natural adjuvants: endogenous activator of dendritic cells. Nature Med 1999; 5:1233–45. Allman AJ, McPerson TB, Badylack SF, Merrill LC, Kallakury B, Sheehan C, Raeder RH, Metzger W. Xenograft extracellular matrix grafts elicit a TH2-restricted immune response. Transplantation 2001;71:1631–40. Piccotti JR, Chan SY, vanBuskirk AM, Eichwald EJ, Bishop DK. Are Th2 helper T lymphocytes beneficial, deleterious, or irrelevant in promoting allograft survival? Transplantation 1997;15:619–24. Bach FH, Ferran C, Hechenleitner P, Mark W, Koyamada N, Miyatake T, Winkler H, Badrichani A, Canadinas D, Hancock WW. Accommodation of vascularized xenografts: expression of ‘‘protective genes’’ by donor endothelial cells in a host th2 cytokine environment. Nature Med 1997;15:196. Chen N, Gao Q, Field EH. Prevention of Th1 responses is critical for tolerance. Transplantation 1996;61:1076.