C57BL/6 mice, 2 months old, were obtained from Sankyo Laboratory Service (Tokyo, Japan) and were maintained in the animal-care facility of Toin University of Yokohama. C57BL/6-TG(TcraTcrb)425-CbnJ (OT-II) mice [11] were obtained from the Jackson Laboratory (Sacrament, CA). This study was reviewed and approved by the research ethics committee of Toin University of Yokohama.

Polyethylene glycol conjugate of OVA (PEG-OVA) was prepared by the method described by Saito et al. [12]. Five milligrams of OVA (albumin from chicken egg white, grade VII; Sigma-Aldrich, St. Louis, MO) were dissolved in one ml of 0.1 M sodium tetraborate (pH 10.0), and then 1,200 mg of activated PEG₂ [2, 4-bis(Omethoxypolyethylene glycol)-6-chrolo-s-triazine] [13] (Seikagaku Kogyo, Ltd., Tokyo, Japan) was added to the solution. The reaction mixture was incubated under stirring at 37˚C for one hour. After the reaction, the mixture was diluted with 10 ml of 0.1 M borate buffer (pH 8.0). The unreacted polyethylene glycol was removed by extensive ultrafiltration using a PM-30 membrane (Merck Millipore, Darmstadt, Germany). Protein concentrations were measured by the Biuret method [14]. In the PEGOVA thus prepared, the average degree of modification of 22 free amino groups in an OVA molecule was 54%, as measured by the colorimetric titration of the free amino groups using trinitrobenzene sulfonate [15].

Female C57BL/6 mice, 3 months old, received intraperitoneal injections of PEG-OVA (100 μg in 200 μl of PBS) or PBS (200 μl) at weeks 0, 1 and 2. Subsequently, these mice were immunized with OVA (250 μg in 200 μl of PBS) at weeks 4, 5, 6 and 11. At week 13, blood samples were obtained from orbital venous plexus using a capillary tube, and were allowed to clot at 4˚C. Sera were stored at −20˚C until analysis. Levels of IgG class antiOVA antibodies were measured by ELISA.

Splenic lymphocytes from OT-II mice were labeled with carboxyfluorescein succinimidyl ester (CFSE; Molecular Probes Inc., Eugene, OR) by the method described by Quah et al. [16]. Briefly, splenic lymphocytes were isolated by density-gradient centrifugation using NycoPrep reagent (Axis-Shield PoC; AS, Oslo, Norway). Four μl of CFSE solution in dioxane (2.5 mM) was added to 2 ml of the lymphocyte suspension (10⁷/ml in PBS containing 2% FBS), and the solution was reacted at room temperature for 5 minutes. After washing with PBS containing 2% FCS, 10⁷ labeled cells were resuspended in 500 μl of PBS containing 2% FCS, and were transferred into the circulation of histocompatible C57BL/6 mice via the tail vein. The fluorescence at 535 nm of the CD4⁺ OT-II Th cells was measured by a FACS Vantage Immunocytometry System (Becton Dickinson, San Jose, CA).

Spleen cells from OT-II mice were suspended at 5 × 10⁶ cells/ml in RPMI 1640 medium containing 50 μM of 2-mercaptoethanol, 100 units/ml of penicillin G (potassium salt), 100 μg/ml of streptomycin sulfate, 250 ng/ml of amphotericin B, and 10% FCS. Aliquots (200 μl) of the cell suspension were distributed in wells of a polystyrene flat-bottomed plate, and then various concentrations of OVA_323-339 peptide were added [17]. The plate was incubated at 37˚C under 5% CO₂ for 24 hours. Incorporation of bromodeoxyuridine (BrdU) was tested during the last 8 hours of the incubation by adding BrdU to each well to obtain a final concentration of 100 μM. After incubation, the supernatant was carefully removed by aspiration. Subsequently, the cells were fixed to the plate by treatment with methanol and drying. Incorporated BrdU in the genomic DNA of the cells was measured by ELISA with anti-BrdU monoclonal antibody [18].

Levels of serum IgG class anti-OVA antibodies were measured by ELISA [19]. Each well of a polystyrene flat-bottomed plate (Nunc-Immuno Plate 442404; Nalge Nunc International, Roskilde, Denmark) was coated with 1 μg of OVA in 0.1 ml PBS at 4˚C overnight, blocked with 3% BSA in PBS for one hour, and reacted with diluted serum (50 μl) at 4˚C for one hour. After 6 washings with PBS containing 0.05% Tween 20, 50 μl of peroxidase-conjugated goat anti-mouse IgG antibodies was added to each well, and the plate was incubated at room temperature for 90 minutes. After extensive washing, 100 μl of 4.8 mM o-phenylenediamine (Wako Pure Chemical Industries, Ltd., Osaka, Japan) in 0.1 M citrate-phosphate buffer (pH 5.0) and hydrogen peroxide (2.6 mM) was added to each well. After incubation at room temperature for 5 minutes, the reaction was terminated by adding 50 μl of 4 N H₂SO₄. The absorbance at 490 nm was measured by a model 550 Plate Reader (Bio-Rad Laboratories, Hercules, CA).

The levels of transcription of the cytokine genes were measured by real-time PCR analysis. Spleen lymphocytes (4 × 10⁷) of the OT-II mice were stimulated in vitro with 2.2 μM of OVA_323-339 peptide in 8.0 ml of RPMI medium supplemented with 10% FCS and 50 μM of 2-mercaptoethanol. After 16 hours, the cells were harvested and lysed with Trisol Reagent [20]. Cytoplasmic RNA was extracted as described in the manufacturer’s instructions, and was further purified by using an RNAfree DNase in the presence of recombinant ribonuclease inhibitor (Takara Shuzo, Shiga, Japan). cDNA molecules were synthesized by using a reverse transcriptase and random hexamer, and were amplified by Taq thermostable DNA polymerase in the presence of SYBR Green [21]. The quantitative real-time PCR analysis was conducted by using a Roche Lightcycler 480 (Roche Diagnostics GmbH, Mannheim, Germany). The nucleotide sequences of the primers used for the analysis were as follows: for Mouse IL-2 [22], AACTGAAA CTCCCCAGGAT and AGGGCTTGTGAGATGATGC; for Mouse IL-4 [23], CCTCACAGCAACGAAGAACA and AAGTTAAAGCATGGTGGCTCA; for Mouse interferon γ (IFN-γ) [24], AAGCAATCAGGCCATCAGC and ATCAGCAGCGACTCCTTTTC; for Mouse IL-10 [25], CCAAGCCTTATCGGAAATGA and TGGCCT GTGACACCTTGG; and for Mouse IL-17 [26], TCTCTGATGCTGTTGCTGCT and GACAGATCTCTT GCTGGA.

Splenic lymphocytes were treated with a lysis buffer containing 0.15 M ammonium chloride, 0.01 M potassium hydrogen carbonate and 0.1 mM EDTA at room temperature for 5 minutes to remove erythrocytes. After washing with RPMI medium containing 10% FCS, splenic lymphocytes (2 × 10⁶) were stained with fluorescein-conjugated anti-TCR Vα2 mAb (clone B20.1; Becton Dickinson, San Jose, CA), and PE-Cy5-conjugated anti-mouse CD4 mAb (clone L3T4; Becton Dickinson) in the presence of 0.5% normal rat serum, and subsequently stained for dead cells with 7-amino-acitinomycin D (7-AAD, 2.5 μg/ml; Becton Dickinson). The stained lymphocytes were analyzed with a FACS Vantage Immunocytometry System (Becton Dickinson).

Blood kinetics of OVA and PEG-OVA were measured using OVA labeled with fluorescein-5-maleimide at sulfhydryl groups on the molecule. An aliquot of 960 μl of 1.1 mM fluorescein-5-maleimide (Wako Pure Chemical Industries, Ltd.) in dimethylformamide was added to 36 ml of 5 mg/ml OVA in 0.1 M borate buffer (pH 7.5). The mixture was incubated under stirring at 37˚C for one hour, and dialyzed extensively against 0.1 M borate buffer (pH 7.5) to obtain fluorescein-labeled OVA (FOVA). A portion of F-OVA solution (18 ml of 5 mg/ml) was dialyzed against 0.1M sodium tetraborate buffer (pH 10.0), and subsequently reacted with 5.4 g of activated PEG₂ to obtain PEG-conjugate of F-OVA (F-PEG-OVA) as described in Section 2.1. Both F-OVA and F-PEGOVA were purified extensively by ultrafiltration to remove residual fluorescein-maleimide. The protein concentrations of F-OVA and F-PEG-OVA were measured by the Biuret method [7], and both solutions were adjusted to 8 mg/ml of protein. F-OVA or F-PEG-OVA solution (200 μl) was injected into every 6 female C57BL/6 mice intraperitoneally. Blood samples were collected from each mouse at 15, 30, and 45 minutes, and 1, 2, 3, 8, 17.5, 45, 96 and 144 hours after injection. A 100 μl aliquot of each blood sample was immediately transferred to 900 μl of PBS containing 5 mM EDTA, and centrifuged to obtain the plasma sample. The concentrations of F-OVA and F-PEG-OVA in plasma samples were determined by fluorometric quantification using a Fuji FLA- 2000 imaging analyzer (Fujifilm Co., Ltd., Tokyo, Japan).

PEG-OVA was prepared by reacting the free amino groups of the OVA molecule with activated PEG₂ reagent having two methoxypolyethylene glycol chains with a molecular weight of 50 kilodaltons. The optimum degree of modification of the primary amino groups of the OVA molecule for tolerogenic conjugates was previously shown to be around 50% [12]. The degree of modification of the free amino groups of PEG-OVA used in the present study was 54%. To test the tolerogenicity of PEG-OVA conjugates, female C57BL/6 mice were pretreated with three weekly intraperitoneal injections of PEG-OVA, and subsequently immunized with OVA. The results are shown in Figure 1. Pre-treatment with PEGOVA induced a marked suppression of the IgG-class antiOVA immune response in mice, in keeping with the previous observations [7,8,12]. Similar profiles were observed for the anti-OVA responses of the IgM, IgG1, IgG2a, IgG2b and IgG3 subclasses (data not shown).

To characterize the response of OVA-specific naïve Th cells as elicited by tolerogenic PEG-OVA, splenic CD4⁺ T cells from OT-II mice with OVA-specific transgenic T-cell receptor genes [11] were labeled with carboxyfluorescein succinimidyl ester (CFSE), and were trans-

ferred into the histocompatible C57BL/6 mice. The recipient mice were subsequently injected intraperitoneally with either PBS, unconjugated OVA or tolerogenic PEGOVA. As shown in Figure 2, PEG-OVA, although tolerogenic with respect to the production of anti-OVA antibodies, induced a robust proliferation of naïve OVAspecific Th cells in vivo, as demonstrated by the appearance of CD4⁺ Th cells with gradually decreased fluorescence. The proliferative response induced by PEGOVA was comparable to or even higher than that induced by unconjugated OVA (Figures 2(d) and (e)).

To determine the effect of longer exposure to tolerogenic PEG-OVA, female OT-II mice were injected intraperitoneally three times either with PBS, OVA or PEG-OVA at weeks 0, 1 and 2. Mice were sacrificed at week 3, and the splenic CD4⁺ Th cells were tested in vitro for their response towards varying concentrations of OVA_323-339 peptide [17] that is recognized by OVAspecific Th cells. Figure 3 compares the dose-response curves observed for incorporation of BrdU in splenic Th cells from OT-II mice pretreated either with PBS, OVA or PEG-OVA. Half-maximal levels of the proliferation of naïve OT-II CD4⁺ T cells pretreated with PBS were observed at 1.6 × 10⁻⁹ M of OVA_323-339 peptide. The cells

from OT-II mice pretreated with OVA responded similarly, and the half-maximal level of response was observed at 2.0 × 10⁻⁹ M. In contrast, a higher concentration (1.1 × 10⁻⁸ M) of OVA_323-339 peptide was required to observe the half-maximal response. The plateau level of the response observed for the cells from mice pretreated with PEG-OVA was also lower than those observed for the cells from mice pretreated with OVA or PBS.

The lower response observed in OT-II mice pretreated with PEG-OVA may have been partly due to the paucity of the OVA-specific CD4⁺ T cells in the spleen. Therefore, the frequency of TCR Vα2⁺CD4⁺ cells in the spleen was compared in OT-II mice pretreated with either PBSOVA or PEG-OVA. Representative two-color profiles are shown in Figures 4(a), (b) and (c). There was a significant reduction in the frequencies of splenic TCR Vα2⁺ CD4⁺ T cells in mice treated with tolerogenic PEG-OVA as compared with those from OT-II mice pretreated with PBS.

To characterize the pattern of cytokine productions of CD4⁺ Th cells exposed to tolerogenic PEG-OVA, OT-II mice were pretreated either with PBS, OVA or tolerogenic PEG-OVA as in Figure 3. Subsequently, splenic CD4⁺ T cells from these mice were cultured in the presence of 2.2 μM of OVA_323-339 peptide in vitro. After 16 hours of incubation, the cells were harvested and analyzed for the expression of cytokine genes by real-time RT-PCR analysis. Figure 5 shows the relative abundance of transcripts as normalized on the basis of actin gene transcripts. Enhanced transcripts of IL-2, IL-4, IFN-γ, IL-10 and IL-17 were observed for the mice pretreated with OVA. In contrast, lower levels of cytokine gene transcripts were observed for the mice pretreated with PEG-OVA. The transcript of IL-2 was markedly suppressed by the pretreatment with PEG-OVA.

The functional features of OVA-specific Th cells exposed to tolerogenic PEG-OVA were analogous to those observed for anergic Th cells, and presumably caused by the persistent exposure to the antigenic peptide of OVA_323-339 derived from PEG-OVA. Therefore, in the present study, the blood kinetics of OVA and PEG-OVA were compared in mice. Figure 6 shows the blood kinetics of OVA and PEG-OVA. In mice injected with OVA, the concentration of OVA in the plasma reached a peak level (7 μg/ml) 30 minutes after the injection and then was cleared rapidly, becoming undetectable in 8 hours. In mice injected with PEG-OVA, the level reached a plateau (50 μg of OVA equivalent /ml) at 2 hours after the injection, and then declined very slowly. The estimated halflives for OVA and PEG-OVA were 2.0 and 59 hours, respectively. Therefore, PEG-OVA was found to be nearly 30 times as stable as OVA in the circulation, and was detectable in the plasma even 144 hours after the injecttion.