Suppression of collagen-induced arthritis by single administration of poly(lactic-co-glycolic acid) nanoparticles entrapping type II collagen: A novel treatment strategy for induction of oral tolerance




Poly(lactic-co-glycolic acid) (PLGA), a biodegradable polymer, is a carrier for drug delivery systems. This study was undertaken to investigate the tolerogenic effect of single administration of PLGA entrapping type II collagen (CII) on the development of collagen-induced arthritis (CIA).


The biophysical properties of PLGA nanoparticles entrapping CII (PLGA-CII) were investigated by in vitro release testing of CII, immunohistochemistry analysis, and electron microscopy. PLGA-CII was fed singly to animals 14 days before immunization, and the effect on joint inflammation was assessed. Circulating IgG anti-CII antibodies and T cell responses to CII in draining lymph nodes were assayed by enzyme-linked immunosorbent assay and 3H-thymidine incorporation assay, respectively. The expression of messenger RNA (mRNA) for transforming growth factor β (TGFβ) and tumor necrosis factor α (TNFα) was determined by reverse transcriptase–polymerase chain reaction.


The in vitro release test showed that CII was slowly discharged from PLGA-CII over a period of a month. After single administration of PLGA-CII, numerous particles ∼300 nm in size were detectable in Peyer's patches, by electron microscopy and immunohistochemical staining for CII, 14 days after the original feeding. Mice fed a single dose of PLGA containing 40 μg of CII had significantly reduced values for incidence and severity of arthritis, serum IgG anti-CII antibodies, and CII-specific T cell proliferation as compared with mice fed solvent alone, those fed 6 doses of 20 μg CII alone, and those fed a single dose of PLGA alone. PLGA-CII was also able to suppress CIA after disease onset. Moreover, PLGA-CII–fed mice showed a higher level of TGFβ mRNA expression in Peyer's patches, but a lower level of TNFα mRNA expression in draining lymph nodes, compared with the other groups of mice.


Our data show that PLGA may serve as a powerful vehicle to promote the tolerance effect of oral CII and that single administration of PLGA-CII may hold promise as a new treatment strategy in rheumatoid arthritis.

Oral administration of autoantigen is a well-established procedure for inducing peripheral immune tolerance, which suppresses autoimmune responses in allergic encephalomyelitis, rheumatoid arthritis (RA), experimental uveitis, and diabetes in the NOD mouse (1–4). Several factors have been shown to influence the effectiveness of tolerance induction. These include the nature of the antigen (soluble or particulate), antigen dose, genetic background of recipient animal, and degree of antigen uptake (5). In particular, the magnitude of tolerance induction is largely dependent on the concentration of an immunologically relevant antigen in circulation. Unfortunately, the amount of antigen absorbed in intact form is extremely small, possibly due to intraluminal digestion (6). Thus, a strategy to enhance the rate of absorption of orally administered antigen would be important for improving the effect of oral tolerance.

Biodegradable nano- and microparticles have been widely studied for use in drug carrier systems, based on advantages such as enhanced absorption (7, 8), biocompatibility (9), and sustained drug release (10). Among these, poly(lactic-co-glycolic acid) (PLGA), a biodegradable polymer, has been utilized in the delivery of various active agents, including drugs (11), hormones such as leuprolide acetate and insulin (12, 13), and vaccines (14). In those trials, entrapped substances were slowly released into their surroundings for prolonged periods. Interestingly, it has been documented that several kinds of polymeric materials have a potential by themselves to suppress a pathologic immune response, especially after multiple injections (15–17). However, there has been no suggestion of use of singly administered biodegradable particles entrapping self antigens to suppress autoimmune diseases.

Although the etiologic agent(s) in RA remains unclear, type II collagen (CII) is a strong candidate because of its abundance in cartilage, an immunologically privileged tissue, and because of its ability to induce destructive immune-mediated polyarthritis (18, 19). Antigen-specific immunosuppression using CII has been tested both in an animal model of arthritis and in human RA. Even though orally administered CII has been shown to strongly suppress arthritis in the animal model, the results in human RA have not been entirely uniform (2, 20, 21). The efficacy of oral CII treatment in human RA may differ depending on the amount and type of CII. However, it might also be greatly influenced by the level of antigen delivery into Peyer's patches in the small intestine, the main machinery of oral tolerance induction.

We hypothesized that PLGA nanoparticles entrapping self antigen would be suitable for induction of oral tolerance because successive and extended exposure of the intestinal immune system to antigen is indispensable for successful oral tolerance. Application of polymer in the delivery of antigen would also be beneficial in that a single feeding might be enough to treat diseases, due to the slow and prolonged release of entrapped antigens. To investigate these possibilities, we conducted a series of oral tolerance induction studies using PLGA nanoparticles entrapping CII (PLGA-CII) in animals with collagen-induced arthritis (CIA), an experimental model of RA.



Male DBA/1 mice (Jackson Laboratories, Bar Harbor, ME) were maintained in groups of 3–5 in polycarbonate cages in a specific pathogen–free environment and fed standard mouse chow (Ralston Purina, St. Louis, MO) and water ad libitum. Neonatal mice were obtained by breeding mice in our facility.

Preparation of CII.

Bovine CII was extracted and purified from native-form calf articular cartilage, as described previously (22). Briefly, cartilage slices were solubilized with pepsin, and proteoglycan fractions were extracted by exposure to neutral 4M guanidine hydrochloride, followed by repeated precipitation from 0.5M acetic acid. After removal of pepsin by dialysis against 0.02M Na2HPO4, proteoglycans were isolated by DEAE chromatography and type I collagen was eliminated by differential salt precipitation at neutral pH. The purity of the resultant CII fraction was confirmed by sodium dodecyl sulfate–polyacrylamide gel electrophoresis. CII polypeptides were then lyophilized and dissolved in 0.1N acetic acid as a solvent at 1 mg/ml, dialyzed against 50 mM Tris, 0.2M NaCl, and then sterilized by filtering through a 0.2-μm micropore filter.

Preparation of PLGA entrapping CII.

Nanoparticles of PLGA-CII were prepared by the water-in-oil-in-water emulsion solvent evaporation technique, as described previously (10). A solution of bovine CII (10% [weight/volume]) in 0.1N acetic acid (internal aqueous phase) was initially emulsified for 1 minute with a solution of PLGA (3% [w/v]) in methylene chloride (oil phase), using an ultrasonic processor (VCX 600Watt Vibracell; Sonics & Materials, Danbury, CT) to form a water-in-oil emulsion. This water-in-oil emulsion was poured into a 5% (w/v) polyvinyl alcohol aqueous solution (external water phase), followed by a second ultrasonic emulsification for 1 minute, resulting in a water-in-oil-in-water emulsion. The water-in-oil-in-water emulsion was stirred overnight at room temperature to evaporate methylene chloride. To obtain PLGA nanoparticles, the resulting aqueous suspension was subjected to ultracentrifugation (Optima LE-80K; Beckman, Palo Alto, CA) at 30,000 revolutions per minute for 20 minutes. The pellets were washed twice with water and resuspended in water prior to lyophilization. Figure 1 shows the chemical formula of the PLGA used in this study.

Figure 1.

Structure of poly(lactic-co-glycolic acid).

Induction and evaluation of arthritis.

Mice were immunized with native CII at 8–12 weeks of age. For injection, CII was dissolved in 0.1N acetic acid at 2 mg/ml and emulsified (1:1 ratio) with Freund's complete adjuvant (CFA) at 4°C (23). Mice received 0.1 ml of the emulsion containing 100 μg of CII in the base of tail as a primary immunization. Booster injections into the footpad, with 50 μg of CII similarly dissolved and emulsified 1:1 with CFA, were administered 14 days after the primary immunization.

The incidence and severity of arthritis were determined by visual inspection. Mice were observed 2–3 times per week for onset, duration, and severity of joint inflammation, for a period of 10 weeks after primary immunization. Each limb was assessed on a 0–4 scale, by the following criteria, as described previously (24): 0 = normal; 1 = mild inflammation of a single area (i.e., midfoot, ankle, or toes); 2 = moderately severe arthritis involving toes and ankle or midfoot; 3 = severe arthritis involving entire paw; and 4 = severe arthritis resulting in ankylosis and loss of joint movement. The hind paw in which the booster immunization had been administered was excluded from the evaluation. Thus, the maximum possible arthritis score was 12. The mean arthritis index, incidence of arthritis (percent of animals), and percent arthritic limbs were used for data comparison among experimental groups.

Oral tolerance induction.

For tolerance induction, intragastric feeding was performed with a ball-tipped feeding needle. Mice were divided into 4 groups. Experimental mice were fed singly with various amounts of PLGA-CII on day −14. Mice fed CII 6 times were used as a positive control group. In this group, CII was fed on days −14, −12, −10, −7, −5, and −3 relative to the primary immunization with CII in CFA on day 0. The optimal dose of CII was determined in a preliminary study in which various concentrations of CII, ranging from 5 μg to 80 μg, were given to CIA mice. The maximal enhancement of tolerance by oral CII was achieved with a concentration of 20 μg (data not shown); thus, this concentration was used for the remainder of the study. Negative control animals were fed solvent (0.1N acetic acid) 6 times or fed singly with PLGA alone, in the same quantity as used in the PLGA-CII nanoparticles.

In vitro release test.

The amount of CII released from the PLGA-CII was monitored for 1 month, using PLGA-encapsulated fluorescein isothiocyanate (FITC)–labeled CII at 37°C, as described previously (10). PLGA containing FITC-labeled CII was generated, after labeling of CII, by the same method as was used for PLGA-CII. The particles were then dispersed in phosphate buffer (0.1M, pH 7.4) and incubated at 37°C with gentle stirring. At predetermined time intervals, aliquots of nanoparticle suspensions were collected and centrifuged (Optima LE-80K) at 30,000 rpm for 30 minutes. The amount of FITC-labeled CII in the supernatants was determined by measuring the fluorescence with a spectrofluorometer.

Characterization of PLGA-CII by electron microscopy

Scanning electron microscopy (SEM) was performed to determine the surface topography of the particles. A sample of particles was placed on double-sticky tape over aluminum stubs to achieve a uniform layer. Thereafter, gold coating of the sample was performed using sputter gold coater at a thickness of 20–30 nm, at 40 mA current and 50 millitorr pressure for 200 seconds. Gold-coated samples were cooled over dry ice prior to SEM observations with a JSM 840A microscope (JEOL, Tokyo, Japan).

To evaluate whether PLGA-CII survives in the intestine for a prolonged period and to characterize PLGA-CII particles in vivo, transmission electron microscopy (TEM) of Peyer's patches was performed. Peyer's patches were separated from the intestine 14 days after a single feeding of 3 mg PLGA-CII (120–150 μg/gm) containing 40 μg of CII. Control Peyer's patches were obtained from mice given 6 feedings of 20 μg CII or solvent (0.1N acetic acid) alone. Peyer's patches were then fixed with glutaraldehyde, cut into 1-mm3 blocks, and postfixed with 1% osmium tetroxide in phosphate buffered saline for 2 hours. The tissues were then dehydrated in a graded series of ethanol and embedded in Epon 812. One-micrometer semithin sections were stained with toluidine blue and examined by light microscopy. Ultrathin sections were stained with uranyl acetate and lead citrate and photographed with a transmission electron microscope (1200EX; JEOL).

Immunohistochemistry of Peyer's patches.

To determine the in vivo stability of CII entrapped in PLGA in the intestinal environment, immunohistochemical staining for CII was performed with sections of Peyer's patches. Briefly, Peyer's patches were obtained from mice 14 days after a single feeding of PLGA-CII; Peyer's patches from mice given 6 feedings of 20 μg CII or solvent alone were used as control tissues. After fixing with periodate-lysin paraformaldehyde solution overnight at 4°C, tissues were dehydrated with alcohol. Samples were then washed, embedded in wax, and sectioned into 5-μm slices. Sections were depleted of endogenous peroxidase activity by addition of methanolic H2O2 and blocked with normal sera for 1 hour. After overnight incubation at 4°C with goat polyclonal antibodies to bovine CII (Southern Biotechnology, Birmingham, AL), sections were hybridized for 2 hours with peroxidase-conjugated donkey anti-goat IgG F(ab′)2 (Jackson ImmunoResearch, West Grove, PA). Color development was achieved by placing sections into 0.1% 3',3′-diaminobenzidine in 0.05M Tris buffer for 5 minutes, after which H2O2was added to a final concentration of 0.01. Photography was performed with an Olympus (Tokyo, Japan) photomicroscope.

Assay for IgG antibodies to native CII.

Sera were collected from each group of mice on day 28 after primary immunization and stored at −20°C until assay. Levels of IgG anti-CII in sera were determined by commercial enzyme-linked immunosorbent assay (Chondrex, Redmond, WA). The optical densities of standard sera which were serially diluted 2-fold were expressed as 100, 50, 25, 12.5, 6.25, and 3.125 arbitrary units. The relationship of the optical density measured in standard serum diluted serially and arbitrary units showed good linear correlation in all determinations (r > 0.98) (data not shown). The concentrations of IgG anti-CII in the sera diluted 1:4,000 are presented as relative values (arbitrary units) compared with the optical density of the standard sera.

Assessment of T cell proliferation.

Inguinal and popliteal lymph nodes were removed from each group of mice 28 days after primary immunization and washed in RPMI 1640. Tissue from 4 or 5 mice was pooled, minced into single cell suspensions in RPMI 1640, and washed 3 times. Cells (5 × 105/well) were cultured in 96-well microtiter plates (Nunc, Roskilde, Denmark) with or without 40 μg/well of CII in 0. 3 ml of Click's medium supplemented with 0.5% mouse serum, at 37°C in 5% humidified CO2, for 4 days. Eighteen hours before the termination of the cultures, 1 μCi 3H-thymidine was added to each well. Cells were harvested onto glass fiber filters and counted on a Matrix 96 direct ionization beta counter (Packard, Meriden, CT). Data are presented as stimulation indices, calculated as the ratio of counts per minute in the presence of CII:cpm in the absence of CII.

Reverse transcriptase–polymerase chain reaction (RT-PCR) for semiquantitation of transforming growth factor β (TGFβ) and tumor necrosis factor α (TNFα) messenger RNA (mRNA).

Peyer's patches were removed from the intestines of the mice in each of the 4 groups 14 days after the first feeding. Draining lymph nodes were also obtained 4 weeks after the primary immunization (6 weeks after the first feeding). Tissue from 3–4 mice was pooled and made into single cell suspensions. Total RNA was extracted from Peyer's patches or lymph node cells using RNAzol B, according to the instructions of the manufacturer (Biotecx, Houston, TX). RT of 5 μg of total RNA was carried out using an RT system kit (Gibco BRL, Grand Island, NY) under conditions ensuring optimal complementary DNA (cDNA) synthesis. PCR was performed using aliquots of the resulting cDNA. To this was added 2.5 mM dNTP, 2.5 units Taq DNA polymerase (Takara Shuzo, Shiga, Japan), 0.25 μM of each sense and antisense primer, and PCR buffer (1.5 mM MgCl2, 50 mM KCl, 10 mM Tris HCl, pH 8.3) in a total volume of 25 μl.

The following sense and antisense primers (5′ to 3′) were used: TGFβ sense ACCGCAACAACGCCATCTAT, antisense GTAACGCCAGGAATTGTTGC; TNFα sense AGGCAATAGGTTTTGAGGGCCAT, antisense TCCTCCCTGCTCTGATTCCG; β2-microglobulin (β2m) sense TGACCGGCTTGTATGCTATC, antisense CAGTGTGAGCCAGGATATAG. Reaction mixtures were incubated in a DNA thermal cycler (Perkin-Elmer Cetus, Norwalk, CT). Cycling conditions for TGFβ, TNFα, and β2m were as follows: 20 seconds of denaturation at 94°C, 30 seconds of annealing at 56°C, and 30 seconds of elongation at 72°C, for 33 cycles. Amplifications were preceded by denaturation for 60 seconds at 94°C and followed by a final extension for 7 minutes at 72°C. PCR products were run on a 1.5% agarose gel and stained with ethidium bromide. To compare the expression of each cytokine mRNA, results are expressed as the ratio of the quantified cytokine product to the β2m product.

Statistical analysis.

Comparisons between groups were performed by Student's t-test or analysis of variance (ANOVA) for numeric data and by chi-square test for categorical data. P values less than 0. 05 were considered significant.


Sustained release of CII from PLGA-CII in vitro. As a first step, we generated PLGA nanoparticles entrapping CII with the desired size, by optimizing various formulation parameters and the energy source input. Figure 2A shows the discrete and spherical shape of PLGA-CII, with a mean ± SD diameter of 299.7 ± 4. 9 nm as seen on SEM. Based on the Lowry assay, 1.3% of the total weight was composed of CII; this was confirmed by hydroxyproline determination. In the in vitro release test, PLGA-CII particles showed an initial burst release, discharging 10–20% of encapsulated CII within the first few days (Figure 2B). The subsequent release of CII took place at a slower release rate for up to 15 days, during which a further 10–20% of CII was expelled. This lag phase was followed by a second burst release from day 15 to day 30. These data suggest that singly administered PLGA-CII could harbor CII for at least a month, due to the slow release characteristics of PLGA.

Figure 2.

Characteristics of poly(lactic-co-glycolic acid) nanoparticles entrapping type II collagen (PLGA-CII). A, Scanning electron microscopy image of PLGA-CII. A number of discrete, spherical-shaped PLGA-CII particles, with a mean ± SD diameter of 299.7 ± 4.9 nm, are seen. Arrows indicate 1 particle. B, Results of in vitro release testing using fluorescein isothiocyanate (FITC)–labeled CII. After generation of PLGA nanoparticles entrapping FITC-labeled CII, the particles were dispersed in phosphate buffer (0.1M, pH 7.4) and incubated at 37°C with mild stirring. Samples were maintained under these conditions, aliquots were obtained, and fluorescence was measured periodically with a spectrofluorometer. FITC-labeled CII was released from PLGA with a slow release profile over 4 weeks, showing a typical biphasic pattern. Values are the mean of 2 independent experiments.

Persistent in vivo stability of PLGA-CII. To determine whether PLGA-CII particles would be retained in the intestine, we administered 3.0 mg PLGA-CII containing 40 μg CII to mice and then examined for the presence of the particles in Peyer's patches, using immunohistochemical staining and TEM. The examination was performed 14 days after feeding, the time that soluble antigen cannot normally be detected in the intestine.

Immunohistochemistry analysis using CII-specific antibodies revealed dense staining for CII inside villi and in the dome area of the Peyer's patches from PLGA-CII–fed mice (Figure 3C), whereas this finding was not noted in the Peyer's patches from mice fed CII alone (Figure 3B) or solvent alone (Figure 3A). The staining was specific for CII, as indicated by the fact that isotype-control IgG antibodies (Sigma, St. Louis, MO) did not stain any Peyer's patches from PLGA-CII–fed mice (results not shown). TEM pictures of the same tissues demonstrated that numerous particles sized ∼300 nm were present in an aggregated pattern within the cytoplasm of Peyer's patch cells obtained from PLGA-CII–fed mice (Figures 4B and C), but not in those from mice fed solvent (Figure 4A) or 20 μg CII alone (results not shown). Combined with the results of the in vitro release test, these observations suggest that CII may be released slowly from PLGA-CII and may help protect against attacks by acids and enzymes in the gastrointestinal tract.

Figure 3.

Immunoreactivity to CII in Peyer's patches from PLGA-CII–fed mice. High reactivity to CII was found in Peyer's patches from mice fed PLGA-CII containing 40 μg CII (C; arrows), but not from those given 6 feedings of 20 μg CII (B) or solvent alone (A). See Figure 2 for definitions. (Original magnification × 200 in A;× 400 in B and C.)

Figure 4.

Transmission electron microscopy findings in Peyer's patches from mice fed PLGA-CII (B and C) or solvent alone (A). At 14 days after a single feeding, a number of particles were detectable in cytoplasm of Peyer's patch cells from PLGA-CII–fed mice (B), but not from solvent-fed mice (A). Area delineated by box in B is shown at higher magnification in C. Arrows indicate aggregated particles sized ∼300 nm. See Figure 2 for definitions. (Original magnification × 5,000 in A and B; × 30,000 in C.)

Protection against CIA by single administration of PLGA-CII. The next experiment was performed to determine the tolerizing effect of singly administered PLGA-CII in CIA. As shown in Figure 5A, mice fed a single dose of 3.0 mg PLGA-CII containing 40 μg CII (n = 20) had a significantly lower mean arthritis index, compared with mice fed solvent alone (n = 19), 3.0 mg PLGA alone (n = 19), or 20 μg CII (n = 20), as determined 5 weeks after immunization (mean ± SD arthritis index 4.7 ± 1.5 in solvent-fed mice, 3.5 ± 1.2 in PLGA-fed mice, 1.8 ± 1.2 in CII-fed mice, 0.8 ± 0.7 in PLGA-CII–fed mice) (P < 0.05, PLGA-CII–fed mice versus CII-fed mice; P <0.001, PLGA-CII–fed mice versus solvent-fed mice and versus PLGA-fed mice, by ANOVA). The incidence of arthritis and number of arthritic limbs were suppressed similarly to the mean arthritis index in PLGA-CII–fed mice compared with the other groups (Figures 5B and C) (for incidence of arthritis, P < 0.001 and P < 0.01, PLGA-CII–fed mice versus solvent-fed and PLGA-fed mice, respectively; for number of arthritic limbs, P < 0.001, PLGA-CII–fed mice versus solvent-fed and PLGA-fed mice, by ANOVA). The incidence of arthritis and number of arthritic limbs tended to be lower in the PLGA-CII–treated group than in the CII-treated group, but this trend did not reach statistical significance.

Figure 5.

Effect of singly administered PLGA-CII on the severity and incidence of arthritis. Seventy-eight mice were divided into 4 groups. Experimental mice (n = 20) were fed singly with 3 mg PLGA-CII containing 40 μg CII on day –14 relative to the primary immunization with CII in Freund's complete adjuvant (day 0). As controls, solvent (n = 19), 3.0 mg PLGA (n = 19), and 20 μg CII (n = 20) were administered; solvent and CII were fed 6 times (days –14, –12, –10, –7, –5, and –3), and PLGA singly on day –14. A, Mean arthritis index, B, arthritis incidence, and C, number of arthritic limbs were determined 5 weeks after immunization. D, Dose-dependent effect of PLGA-CII on arthritis suppression. Various amounts of PLGA-CII (0.375–12 mg), containing 5–160 μg of entrapped CII, were administered singly, to groups of 12–15 animals. Values are the mean and SD. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001 (in D, all versus no PLGA-CII). NS= not significant (see Figure 2 for definitions).

In another set of experiments, we administered various amounts of PLGA-CII (0.375–12 mg), containing 5–160 μg of entrapped CII, to groups of 12–15 animals. As shown in Figure 5D, a suppressive effect was evident with PLGA-CII at 0.375 mg (the smallest dose tested). This effect peaked with PLGA-CII at 3 mg and remained high with PLGA-CII at 6 mg and 12 mg, demonstrating the dose dependence of PLGA-CII–induced suppression of arthritis.

Suppression of autoimmune responses to CII in mice treated with PLGA-CII. To investigate the effect of PLGA-CII on CII-specific immune responses to CII, we measured IgG antibodies to CII in sera and T cell proliferative responses to CII in draining lymph nodes, 4 weeks after immunization. As expected, levels of circulating IgG antibodies to CII were significantly lower in mice fed PLGA-CII (mean ± SD 7.2 ± 6.6 arbitrary units; n = 20) than in mice fed solvent (66.3 ± 32.9 arbitrary units; n = 19), PLGA alone (41.6 ± 26.0 arbitrary units; n = 19), or CII alone (12.4 ± 11.0 arbitrary units; n = 20) (P < 0.001, PLGA-CII–fed mice versus solvent-fed and PLGA-fed mice, by ANOVA) (Figure 6A). In addition, the assay for T cell responses to CII in draining lymph nodes showed that mice fed CII alone or PLGA-CII had lower CII stimulation indexes compared with the other groups (mean ± SD background cpm 2,550 ± 562; mean ± SD stimulation index 3.6 ± 0.5, 3.0 ± 0.4, 1.8 ± 0.4, and 1.5 ± 0.2 in solvent-fed, PLGA-fed, CII-fed, and PLGA-CII–fed mice, respectively) (P < 0.01, CII-fed mice versus PLGA-fed mice; P < 0.001, CII-fed mice versus solvent-fed mice and PLGA-CII–fed mice versus PLGA-fed and solvent-fed mice, by ANOVA) (Figure 6B). These observations clearly show that singly administered PLGA-CII effectively suppresses systemic CII-specific immune responses in CIA.

Figure 6.

Suppression of the autoimmune response to CII by single administration of PLGA-CII. A, Levels of circulating IgG antibodies to CII in mice fed solvent alone (n = 19), 3.0 mg PLGA (n = 19), 20 μg CII (n = 20), or 3.0 mg PLGA-CII containing 40 μg CII (n = 20), as determined 4 weeks after immunization. Boxes show the 25th and 75th percentiles; vertical bars show the minimum and maximum values; boldface horizontal bars show the medians; plus signs show the means. B, T cell responses to CII in lymph node cells. Inguinal and popliteal lymph nodes were removed from mice 4 weeks after primary immunization. Pooled lymph node cells from 4 or 5 mice were cultured with or without CII (40 μg/well). The mean background cpm without CII was 2,550 ± 562. Values are the mean and SD stimulation index (arbitrary units) from 5 independent experiments. ∗∗ = P < 0.01; ∗∗∗ = P < 0.001. NS = not significant (see Figure 2 for definitions).

Inhibition of arthritis progression by PLGA-CII. To determine the effect of PLGA treatment on disease progression after immunization, we administered PLGA-CII to mice with recent-onset arthritis (n = 16). For these experiments, 3 mg of PLGA-CII was fed singly 3 weeks after primary immunization. Control mice (n = 18) were treated with solvent alone. Although the time course of disease did not differ considerably between the 2 groups, PLGA-CII–treated mice showed significantly lower mean arthritis index scores from 4 weeks to 8 weeks after immunization (Figure 7A). The peak incidence of arthritis during the experiment was also decreased in PLGA-CII–treated mice compared with control mice (43.8% versus 88.9%) (P < 0.01 by chi-square test). In addition, levels of IgG antibodies to CII, as determined 4 weeks after primary immunization, were significantly reduced in mice treated with PLGA-CII compared with those treated with solvent alone (mean ± SD antibody titer 28.6 ± 12.5 arbitrary units versus 78.5 ± 28.3 arbitrary units) (P < 0.01 by Student's t-test) (Figure 7B).

Figure 7.

Suppression of disease progression by PLGA-CII. PLGA-CII (n = 16) was fed singly 3 weeks (arrow) after primary immunization. Control mice (n = 18) were treated with solvent alone. A, Reduction of mean arthritis index in PLGA-CII–fed mice. B, Levels of circulating IgG antibodies to CII in PLGA-CII–fed versus solvent-fed mice, as determined 4 weeks after primary immunization. Values are the mean and SD. ∗ = P < 0.05; ∗∗ = P < 0.01; ∗∗∗ = P < 0.001. See Figure 2 for definitions.

Modulation of tissue cytokine mRNA expression by PLGA-CII. The primary mechanisms by which oral tolerance is mediated are deletion, anergy, and active suppression (25–28). In particular, relatively low doses of antigens administered repeatedly can induce expansion and/or activation of TGFβ-secreting cells that have been shown to antagonize proinflammatory responses (27, 28). We therefore analyzed whether PLGA-CII could induce TGFβ-secreting cells in Peyer's patches. As shown in Figure 8, TGFβ mRNA expression in Peyer's patches was up-regulated by treatment 6 times with 20 μg CII and more potently by a single treatment with 3 mg PLGA-CII, as determined on day 14 after feeding. PLGA alone slightly increased TGFβ mRNA expression, but solvent alone did not affect it. In contrast, the expression of TNFα mRNA in draining lymph nodes 4 weeks after primary immunization was strongly down-regulated by treatment with CII or PLGA-CII, but not by treatment with solvent or PLGA alone.

Figure 8.

Modulation of tissue cytokines by PLGA-CII. Transforming growth factor β (TGFβ) mRNA and tumor necrosis factor α (TNFα) mRNA levels were assessed by reverse transcriptase–polymerase chain reaction 14 days after first feeding and 4 weeks after immunization, respectively. Top, TGFβ mRNA expression in pooled Peyer's patch cells (left) and TNFα mRNA expression in draining lymph node cells (right) from mice treated 6 times with solvent alone (lanes 1 and 5), once with 3 mg PLGA (lanes 2 and 6), 6 times with 20 μg CII (lanes 3 and 7), or once with 3 mg PLGA-CII containing 40 μg CII (lanes 4 and 8). Bottom, Levels of mRNA expressed as the mean fold increase relative to mRNA from cells of solvent-fed mice, and corrected for the levels of β2-microglobulin mRNA signal. Values are the results from 1 of 3 similar experiments. See Figure 2 for other definitions.


Gut-associated lymphoid tissue (GALT) is unique in that it favors the induction of both Th2 cells and T cells secreting TGFβ in nonpathologic conditions (27, 28). It has been reported that particulate antigens administered orally, including PLGA nanoparticles, are absorbed efficiently into GALT, especially into Peyer's patches (29–31). Furthermore, soluble antigens entrapped by particles can be protected from acidic and enzymatically hostile environments found at mucosal surfaces, facilitating uptake of the antigens by M cells overlying Peyer's patches and the mucosal epithelium, thereby promoting antigen transport to local lymphoid follicles and the associated lymph nodes, and possibly enabling local accumulation of the antigen.

Torche et al, in a TEM analysis of Peyer's patches, demonstrated the traffic of PLGA particles throughout M cells, their transport into the basolateral invaginations of the M cells, and their subsequent migration into the dome area of Peyer's patches (32). In the present study, PLGA-CII slowly released CII for a month in vitro, and was retained up to 14 days in the dome area of Peyer's patches. These findings indicate that CII delivery using PLGA is an effective strategy to overcome gastrointestinal digestion of oral CII and also can eliminate the necessity to administer the oral CII daily. Moreover, singly administered PLGA-CII strongly suppressed CIA, and its efficacy was similarly potent or stronger compared with repetitive CII feeding. Thus, it is possible that PLGA-CII may release sufficient amounts of CII to stimulate intestinal immune cells for a prolonged period and then to show a strong tolerizing activity, comparable with repetitive CII administration. In this respect, the system using PLGA as a carrier would be applicable to the treatment of other autoimmune diseases using different kinds of self antigens.

Earlier studies have documented that several kinds of polymeric materials have a potential by themselves to suppress pathologic immune responses (15–17). Possible accumulation of particulate antigen in Peyer's patches and subsequent translocation to discrete anatomic compartments such as the liver and spleen may modulate both mucosal and systemic immune responses. For example, copolymer 1, a synthetic amino acid copolymer, has a beneficial effect on the development of experimental encephalitis that is associated with down-regulation of T cell immune responses to myelin basic protein and is mediated by Th2/3-type regulatory cells (15). In CIA, copolymer 1 also competitively inhibits the interaction of CII with RA-associated HLA–DR molecules and thereby suppresses CII-reactive T cell responses (16). Polyethylene glycol, another straight amphiphilic polymer, has also been shown to prevent arthritis (17). However, the effect of PLGA nanoparticles on the immune system has not been reported.

In this study, single administration of PLGA nanoparticles significantly suppressed the incidence and severity of arthritis, suggesting an immunomodulating activity of the PLGA nanoparticles themselves, although the effect was less potent than that of PLGA-CII. Therefore, the inhibition of arthritis by PLGA-CII may have been caused by a coordinated action via 2 different mechanisms, e.g., antigen-dependent immunosuppression by slowly released CII and antigen-independent bystander suppression by the PLGA nanoparticles themselves. It is unclear how the PLGA nanoparticles alone induce arthritis suppression. One potential explanation may be the activation of dendritic cells. Recent studies have demonstrated that hyperplasia or activation of dendritic cells in Peyer's patches plays a pivotal role in tolerance induction (33, 34). On TEM, we found that phagocytes taking up PLGA nanoparticles showed an increased amount of lysozyme granules (data not shown), indicating cellular activation, which might be associated with the suppression of arthritis. Further studies, especially with dendritic cells, are necessary to explore this possibility.

PLGA has been used in strategies to enhance the immunogenicity of entrapped foreign antigens, such as in vaccine delivery, which is opposite to the outcome of our intervention. The destination of oral antigens encapsulated in PLGA is likely to depend on multiple factors that are still not well understood. Recently, it has been demonstrated that the route of immunization and the size of the PLGA particles affect the type of immune response (35). For example, parenteral immunization with pertussis toxoid and hemagglutinin entrapped in PLGA microparticles elicited a potent Th1-type response; in contrast, nanoparticles entrapping the same antigens favored the induction of a Th2-type response (35). Another possible factor would be the dose of oral antigen. It has been documented that the oral tolerance effect is abrogated with increasing doses of soluble antigen within PLGA (36). In the present study, we used nanoparticles entrapping self antigen, and the amount of CII was very small (5–160 μg). A difference in properties of the antigen itself (e.g., self or foreign antigen), particle size, and antigen dose could partially explain this apparent discrepancy.

CIA is mediated by the synergistic actions of both CII-reactive T cells and antibodies to CII (37). We demonstrated previously that IgG anti-CII antibodies and T cell responses to CII were enhanced in RA patients, particularly in synovial fluid, suggesting a role of anti-CII reactivity in the pathogenesis of human RA (38–40). In this study, PLGA-CII strongly reduced the levels of circulating IgG anti-CII antibodies and T cell responses to CII in lymph node cells, which is consistent with the data obtained from visual inspection. Moreover, PLGA-CII prevented the progress of disease and the production of anti-CII antibodies in mice with recent-onset arthritis. Of note, in our studies with larger animals, such as dogs, no toxic effect of PLGA-CII, even at a concentration of 700 mg/kg, was found at autopsy (Kim W-U, et al: unpublished observations). Given the suppression of anti-CII reactivity and disease progression by PLGA-CII and the fact that PLGA-CII has no side effects, this form of treatment could be applicable to the subgroup of RA patients who have already demonstrated high CII reactivity. Studies evaluating the efficacy and safety of PLGA-CII in human RA are currently in progress.

Oral tolerance induction is mediated by 2 principal mechanisms, depending on the dose of antigen administered. Antigens fed repeatedly in higher doses can induce clonal deletion and/or anergy (25, 26). Conversely, relatively lower doses of antigens can induce active suppression by expanding regulatory cells that are capable of secreting interleukin-4, interleukin-10, and/or TGFβ (27, 28). In particular, TGFβ-secreting cells appear to represent a unique subset of regulatory cells, which have been termed “Th3 cells.” Endogenous TGFβ is associated with suppression of pathologic immune responses in experimental autoimmune diseases, including CIA (41). Administration of anti-TGFβ antibody reverses protection against disease by oral antigen (28). In contrast, cotreatment with TGFβ and CII potentiates the tolerizing effect of oral CII in CIA (42).

An important question from this study involves the mode of action of PLGA-CII. RT-PCR analysis revealed that TGFβ mRNA expression in Peyer's patch cells was much higher in mice fed PLGA-CII than in control mice, and also higher than in mice given 6 doses of CII alone. In contrast, TNFα mRNA expression in draining lymph node cells was dramatically reduced in mice fed PLGA-CII. These observations suggest that singly administered PLGA-CII may ameliorate the progression of arthritis by stimulating TGFβ-secreting cells in Peyer's patches, thereby inhibiting TNFα-secreting cells. In this context, active suppression by repetitive challenge with CII escaped from PLGA-CII in Peyer's patches may be the major mechanism responsible for arthritis suppression by PLGA-CII.

However, we also found that PLGA-CII–treated mice displayed higher TGFβ mRNA expression but lower TNFα mRNA expression in mesenteric lymph nodes than did other groups of mice (data not shown), suggesting that the above-described cytokine change is not limited to Peyer's patches. Moreover, oral tolerance can be induced in the absence of organized Peyer's patches (43). Therefore, although Peyer's patches are one of the primary areas in which tolerogenic responses are generated, the regulatory effect of PLGA-CII may involve additional mediators or mechanisms within other structures of GALT or mesenteric lymph nodes.

In conclusion, single oral administration of PLGA-CII suppressed the development of arthritis as well as autoimmune responses to CII and strongly induced TGFβ mRNA expression in Peyer's patches. These results suggest that single oral administration of PLGA-CII may be a new strategy for the treatment of RA.


The authors are grateful to Drs. J. W. Lee and H. S. Kim (Yuhan Research Center, Gunpo, Korea) for providing type II collagen. We also thank Dr. Andrew H. Kang for advice on this study.