Male DBA/1 LacJ mice and BALB/c mice (Jackson Laboratories, Bar Harbor, ME, USA) were kept in filter-top cages at the Animal Facility, University of Western Ontario according to National Canadian Council for Animal Guidelines. Mice were allowed to settle for 2 weeks before the initiation of experimentation, which had ethical approval from the university board.

DBA/1 mice, 7 weeks of age, were intradermally immunized at several sites into the base of the tail with 200 μg bovine type 2 collagen (CII) dissolved in 100 μl of 0.05 M acetic acid and mixed with an equal volume of complete Freund's adjuvant (Sigma, Oakville, ON, Canada). CII was dissolved at a concentration of 2 mg/ml by stirring overnight at 4°C. On day 21, the mice received an intraperitoneal booster injection with 200 μg CII in an equal volume (100 μl) of PBS. The booster injection was necessary to induce reproducible CIA, which normally developed at about day 28.

Each mouse was examined visually three times per week for the appearance of arthritis in limb joints, and the arthritis score was given as follows: 0, no detectable arthritis; 1, erythema and mild swelling confined to the mid-foot or ankle joint; 2, significant swelling and redness; 3, severe swelling and redness from the ankle to digits; and 4, maximal swelling and redness or obvious joint destruction associated with visible joint deformity or ankylosis. Each limb was graded and expressed as the average score per affected paw, resulting in a maximum score of 4 per animal. Scoring was performed by two independent observers, without knowledge of experimental protocols.

At day 0, bone marrow cells were flushed from the femurs and tibias of DBA/1 mice, and were washed and cultured in six-well plates (Corning, Acton, MA, USA) at 4 × 10⁶ cells/well in 4 ml complete medium (RPMI 1640 supplemented with 2 mM L-glutamine, 100 U/ml penicillin, 100 μg streptomycin, 50 μM 2-ME, and 10% FCS (all from Invitrogen, Grand Island, NY, USA) supplemented with recombinant granulocyte-macrophage colony-stimulating factor (GM-CSF) (10 ng/ml) and recombinant mouse IL-4 (10 ng/ml) (both from PeproTech, Rocky Hill, NJ, USA). Cultures were incubated at 37°C in 5% humidified CO₂.

Nonadherent cells were then removed (day 2) and fresh medium was added. At day 4 the DC were treated either with LF (5–10 μg/ml) or with PBS, and fresh medium was added every 24 hours. At day 7 we pulsed LF-treated DC or PBS-treated DC with CII (10 μg/ml) for 24 hours. DC were then activated with lipopolysaccharide (LPS) (10 ng/ml; Sigma) and tumor necrosis factor alpha (TNFα) (10 ng/ml; PeproTech) for an additional 24 hours, were washed extensively, and were used for subsequent transfer experiments. On day 12 after the CII priming, different groups of mice with four to six animals per group were injected intraperitoneally with these LF-treated DC or untreated DC (5 × 10⁶ cells/mouse).

In some experiments, day 4 bone marrow DC from BALB/c mice, cultured in GM-CSF/IL-4, were treated with LF (0.1, 1 or 10 μg/ml) or PBS, and fresh medium without LF was added every 24 hours. At day 7, we pulsed LF-treated or PBS-treated DC with keyhole limpet hemocyanin (KLH) (10 μg/ml) (Sigma) for 24 hours. DC were then activated with LPS/TNF-α for an additional 24 hours and injected subcutaneously (5 × 10⁵ cells/mouse) into syngeneic mice. The mice were sacrificed after 10 days, and T lymphocytes from draining lymph nodes and spleens were isolated. Finally, a KLH-specific recall response was performed as described later.

At day 4 of culture, bone marrow DC from DBA/1 LacJ mice were treated with LF (10 μg/ml) or PBS, followed by addition of LPS/TNFα at day 8 for 24 hours. Activated DC were irradiated (3,000 rad) and seeded in triplicate into a flat-bottom 96-well plate (Corning) as stimulators. Spleen T cells from BALB/c mice were isolated by gradient centrifugation over Ficoll-Paque (Amersham, Canada) and added as responders (5 × 10⁵ cells/well). The mixed lymphocytes were cultured at 37°C for 72 hours in 200 μl RPMI 1640 supplemented with 10% FCS, 100 U/ml penicillin, and 100 μg/ml streptomycin, and were pulsed with 1 μCi/well [³H-labeled] thymidine (Amersham) for the last 16 hours of culture. Finally, cells were harvested onto glass fiber filters, and the radioactivity incorporated was quantitated using a Wallac Betaplate liquid scintillation counter (Beckman, Fullerton, CA, USA). Results were expressed as the mean counts per minute of triplicate cultures ± SEM.

Proliferative responses to KLH and CII in subsequent groups of mice were measured with a standard microtiter assay using either draining lymph node cells or splenocytes, using KLH or CII, and using ³H-labeled thymidine. T cells at 5 × 10⁵/well were seeded into a 96-well flat-bottom microtiter plate in triplicate and mixed with serial dilutions of KLH or CII (5–50 μg/well). Following a 72-hour incubation, 1 μCi [³H] thymidine was added to each well for 16 hours. Using a cell harvester, the cells were collected onto a glass microfiber filter, and the radioactivity incorporated was measured by a Wallac Betaplate liquid scintillation counter.

CII-specific antibodies were evaluated using a standard indirect ELISA in which 500 ng CII was absorbed to each well of a 96-well microtiter plate. Following blocking and washing steps, serial dilutions of immune mouse serum (1:100-1:100,000) were added to the appropriate wells in duplicate and were incubated overnight at 4°C. To develop the ELISA, horseradish peroxidase-conjugated goat anti-mouse IgG Fc and orthophenylenediamine dihydrochloride substrate buffer (Sigma) were used. Finally, the optical density into each well was measured at 490 nm wavelength in an ELISA plate reader.

LF-treated DC of DBA/1 origin were cultured alone or with the allogeneic (BALB/c) T cells for 48 hours. Supernatants were collected and assessed for DC (IL-10, IL-12) and for T-cell cytokines (interferon gamma, IL-4). An ELISA (Endogen, Rockford, IL, USA) was used for detecting cytokine concentrations in the supernatants according to the manufacturer's instructions using a Benchmark Microplate Reader (Bio-Rad, Hercules, CA, USA).

Paws of freshly dissected mice were removed and joint tissues were immersion-fixed for 4 days in 10% (wt/vol) neutral buffered formalin in 0.15 M PBS (pH 7.4). After decalcification in Decalcifier I solution (Surgipath, Richmond, IL, USA) overnight and subsequent dehydration in a gradient of alcohols, tissues were rinsed in running water. The specimens were processed for paraffin embedding in paraplast (BDH, Dorset, UK) as routine procedure. Serial paraffin sections throughout the joint were cut at 5 μm thickness on a microtome, heated at 60°C for 30 minutes, and were deparaffinized. Hydration was achieved by transferring the sections through the following solutions: three times through xylene for 6 minutes, and then for 2 minutes through 100% ethanol twice, 95% ethanol, and 70% ethanol, respectively. The sections were stained with H & E and were mounted on glass slides.

Phenotypic analysis of cells was performed using flow cytometry on a FACScan (Becton Dickinson, San Jose, CA, USA). DC were pretreated with LF (5-10 μg/ml) beginning at day 4. Activation of DC maturation was performed by addition of TNFα/LPS for 24 hours. The cells were stained with FITC-conjugated mAbs against surface markers associated with DC maturation (anti-mouse CD11c, I-A, CD40, and CD86; Cedarlane, Hornby, ON, Canada). Immunoglobulins of the same isotype were used as controls.

Data are expressed as the mean ± SEM. Differences in the arthritis score between different populations of mice were compared using the Mann-Whitney U test for nonparametric data. P < 0.05 was considered significant.

Our previous studies have demonstrated that LF together with anti-CD45RB mAb can induce a population of tolerogenic DC in transplant recipients that are responsible for maintenance of tolerance 14. Furthermore, we have previously demonstrated that LF treatment of isolated DC in vitro is capable of inhibiting the maturation-inducing kinase IKK, as well as the downstream transcription factor NF-κB 14. We therefore investigated the potential of LF to generate immature tolerogenic DC that could be used for antigen-specific immunotherapy in vivo. Bone-marrow-derived DC were generated using a standard 7-day culture in GM-CSF/IL-4. LF was added at day 4 of culture, whereas control DC were treated with PBS alone. Activation of control DC and LF-treated DC was performed by addition of TNFα/LPS for 24 hours. Assessment of MHC class II, CD40, and CD86 expression by flow cytometry revealed that control DC underwent marked maturation, whereas LF-treated DC did not upregulate maturation markers (Figure 1a). Both nonactivated control DC and LF-treated DC expressed low levels of the maturation markers, similar to the TNFα/LPS-activated LF-treated DC (data not shown).

We next assessed whether LF is involved in regulation of DC cytokine expression. LF-treated DC following activation with LPS/TNFα were cultured alone for 48 hours. Supernatants were then used to measure levels of IL-12 and IL10 cytokines. As shown in Figure 1b, IL-12 production of LF-treated DC was reduced, whereas IL-10 production reciprocally upregulated.

Functional assessment of LF-treated DC was performed using these cells as allogeneic stimulators in a mixed lymphocyte reaction (MLR). In contrast to control-DC-expressed potent allostimulatory activity, LF-treated DC evoked a much weaker proliferative response (Figure 1c). Using LF-treated DC as stimulators of MLR resulted in preferential production by T cells of the Th2 cytokine IL-4 and reduction of the Th1 cytokine interferon gamma (Figure 1d), in contrast to stimulation with control DC. These data suggest that LF treatment can effectively endow DC with an immature phenotypic and functional state.

We next used LF-treated DC as a platform for the delivery of antigens in a tolerogenic context. It has previously been reported that antigen-pulsed DC with a blocked NF-κB pathway can induce specific hyporesponsiveness to that antigen 15. Since we have recently demonstrated that LF blocks NF-κB translocation 9, and we have shown here that LF treatment inhibits DC maturation, we sought to assess whether LF-treated DC could induce tolerance to a nominal antigen such as KLH.

Pulsing of DC with antigen requires active cellular phagocytosis and processing of the antigen. The in vivo administration of the antigen-pulsed DC is subjected to conditions that may induce maturation not normally present in vitro. Since this is the first use of LF for treatment of DC before antigen pulsing, we performed optimization experiments to determine the most effective concentration of LF. On day 4 of culture, bone marrow DC were treated with 0.1, 1, and 10 μg/ml LF, and control DC were treated with PBS. KLH was added to DC at day 7 for 24 hours, and subsequently cells were activated with TNFα/LPS. On day 9, 5 × 10⁵ DC were injected intraperitoneally into BALB/c mice.

To test the T-cell expansion and activation, the recall response to KLH was assessed in vitro 10 days after the administration of KLH-pulsed control DC and LF-treated DC. KLH-specific responses from lymph node T cells were suppressed at all KLH concentrations used, in an LF dose-dependent manner (Figure 2a). To determine whether bystander tolerization occurred in LF-treated DC-induced immune suppression, we used a 'double immunization' system, in which mice were immunized with CII-pulsed DC alone with an immunization with KLH. The immunization with LF-treated DC and CII antigen-pulsed DC only suppressed the immune response to CII specifically (Figure 2b), but not the immune response to the nonrelevant antigen KLH (Figure 2C).

The CIA model of arthritis is a well-established method of evaluating therapeutic interventions in autoimmune arthritis. Several induction protocols have been reported, all of which in essence induce a T-cell-dependent inflammatory infiltration of the synovial membrane, leading to cartilage destruction and bone erosion. Since we have been able to induce T-cell hyporesponsiveness to KLH using LF-treated DC (Figure 2), we sought to determine whether pulsing LF-treated DC with CII would inhibit CIA development and histopathology. On day 12 post CII priming, DBA/1 mice were administered 5 × 10⁶ intraperitoneal CII-pulsed LF-treated DC or control DC. A booster injection of CII was made at day 21. The clinical onset of CIA as determined by the average arthritis score per effected paw began approximately on day 28.

Initiation of arthritis was delayed by 7 days in the CII-pulsed LF-treated DC group as compared with the control group. Furthermore, the control group had an average score per affected paw twice as high as that of the LF-treated DC group (Figure 3), but a score that ranged from less than twofold to fivefold depending on the time point. These results imply that LF-treated DC are not only capable of inducing antigen-specific hyporesponsiveness, but are also capable of reducing clinical manifestations and delaying disease onset in a model of autoimmunity.

Given that T cells play a key role in the initiation of CIA 16, antigen-specific T-cell proliferative responses to CII were assessed. At the end of the monitoring of CIA development, mice were sacrificed and lymph node cells were collected for proliferative analysis in response to CII. In vitro ³H-labeled thymidine incorporation assays revealed that a decrease in CII-specific recall responses were observed of mice receiving LF-treated DC in comparison with those receiving control DC (Figure 4). The response was antigen specific since modulation of responses to other control antigens was not affected (data not shown). The hyporesponsiveness of CII-specific T cells confirms clinical observations that CII-pulsed LF-treated DC could be useful in therapeutic intervention for antigen-specific T-cell-associated diseases.

The importance of antibodies in development of CIA pathology is well known 17. Although it has been previously suggested that LF directly inhibits antibody production 18, the ability of the LF-treated DC to induce this effect has not been studied. Tolerogenic DC may directly block antibody production through inhibition of BlyS and APRIL, factors that DC use to directly induce immunoglobulin production and class switching in B cells 19. Alternatively, tolerogenic DC may indirectly prevent antibody production through the inhibition of T-cell helper function.

In order to assess whether LF-treated DC pulsed with CII actually inhibit CII-specific antibody responses, we evaluated the serum levels of anti-CII immunoglobulin in DBA/1 mice 37 days following the arthritis onset. Using the same protocol as for induction of CIA, we used mice receiving LF-treated DC pulsed with CII, mice receiving LF-treated DC pulsed with PBS, mice receiving PBS-treated DC pulsed with CII, and mice receiving PBS-treated DC pulsed with PBS. A high titer of anti-CII antibody was seen in control DC pulsed with CII (Figure 5). Administration of LF-treated DC pulsed with CII resulted in a marked decrease in antibody production, although there was no essential difference between the two concentrations of LF used on the DC (Figure 5).

The control for this experiment omitted DC immunization, in which there was no inhibition of antibody production as compared with animals that received CII-pulsed DC without LF treatment. This suggests that CIA is not augmented by CII-pulsed DC, but instead that the CII-pulsed LF-treated DC actually inhibit the initiated autoimmune process.

Although we have demonstrated a clear inhibition of arthritis manifestation using the average arthritis score per affected paw, we further sought to examine histological differences induced by treatment with the CII-pulsed LF-treated DC. Animals injected with LF-treated DC, or control animals, were therefore sacrificed 37 days after arthritis onset and their joints were examined in serial sections. We observed that control DC-treated mice exhibited severe synovitis, pannus formation, and bone erosion (Figure 6a). A marked mononuclear cell infiltration was also observed. In contrast, the joint histology of the mice injected with LF-treated DC revealed markedly attenuated morphological changes, cellular infiltration, and the preservation of normal-appearing cartilage (Figure 6b). The histological verification of the arthritis score (Table 1) strongly suggests that the CII-pulsed LF-treated DC are a potent tolerogenic agent that is useful for inhibition of T-cell-mediated autoimmune responses.