Regulation of Mucosal Dendritic Cell Function by Receptor Activator of NF-κB (RANK)/RANK Ligand Interactions: Impact on Tolerance Induction

Mice

Female BALB/c mice (6–10 wk of age) were obtained from Taconic Farms (Germantown, NY) and maintained in a specific pathogen-free facility at Immunex in accordance with approved ethical guidelines. BALB/c DO11.10 OVA TCR transgenic (Tg) mice (20) were bred and maintained in the specific pathogen-free facility at Immunex.

Analysis of RANK expression on peripheral and gut-associated lymphoid tissues (GALT) DC using flow cytometry

SPL, PLN (two inguinal and popliteal lymph nodes per animal), MLN, and PP were removed from groups of five mice that had been injected i.p. daily for 10 days with 10 μg purified CHO-derived human Flt3 ligand (produced and purified at Immunex as previously described (21)), and single-cell suspensions were prepared. Previous studies have shown that Flt3 ligand (Flt3L) expands all DC populations without inducing their activation (2, 22). Cells were incubated for 40 min at 4°C with biotinylated anti-RANK Ab (clone M395; Immunex), together with FITC-labeled anti-CD11c (HL3, hamster IgG), both used at 10 μg/ml, in 50 μl blocking buffer (PBS containing 10 μg/ml anti-CD16 (BD PharMingen, San Diego, CA), 10% normal goat serum, 2% normal rabbit serum, and 2% normal rat serum). After being washed twice with PBS/2% FBS, cells were incubated with 10 μg/ml PE-labeled streptavidin (10 μg/ml; Molecular Probes, Eugene, OR) for an additional 20 min. Samples were washed three times, resuspended in PBS supplemented with 1% paraformaldehyde, and analyzed on a FACScan flow cytometer (BD Biosciences, San Jose, CA). At least 30,000 cells were analyzed per sample.

DC isolation and purification

Groups of 10 mice were injected daily with 10 μg Flt3L given i.p. for 10 days to expand DC in vivo. The animals were sacrificed; their SPL, PLN (two axillary, brachial, cervical, inguinal, and popliteal lymph nodes per animal), MLN, and PP were removed; and individual tissue types were pooled. Single-cell suspensions were prepared by pressing the excised lymphoid tissues through nylon mesh. The resultant suspension was then spun over a Nycodenz density gradient (Nycomed Pharma, Oslo, Norway) at 1700 × g. DC-enriched cell populations were harvested from the interface, washed, and incubated with anti-mouse CD11c-coated magnetic beads for 20 min. After washing, CD11c⁺ cells were purified by positive selection using MACS separation columns (Miltenyi Biotec, Auburn, CA) according to the manufacturer’s instructions. In each of three separate, independent experiments, the purified DC populations were between 96 and 99% CD11c positive (Table I⇓).

Stimulation of DC with RANKL in vitro

Purified DC (3 × 10⁵/well) were cultured for 18 h for RNA studies and for up to 120 h for DC survival studies in 96-well, flat-bottom, tissue culture plates in either complete RPMI (RPMI supplemented with 10% FBS, penicillin/streptomycin, l-glutamine, and 2-ME) or complete RPMI containing 10 μg/ml soluble murine RANKL-isoleucine zipper fusion protein, produced and purified at Immunex, in a total volume of 200 μl.

Isolation of RNA from DC cultures

RNA was prepared following homogenization of DC in guanidinium isothiocyanate buffer (4.5 M guanidinium isothiocyanate, 50 mM sodium citrate, 0.5% (w/v) sodium sarcosyl) containing 2% 2-ME (Life Technologies, Gaithersburg, MD). RNA was then extracted using acid phenol, as described previously (23). RNA samples were subsequently treated with DNase (DNA-free; Ambion, Austin, TX) to eliminate contaminating genomic DNA, and the resulting purified RNA samples were reverse transcribed to cDNA with random hexamers according to the manufacturer’s instructions (TaqMan Reverse Transcription Reagents; PerkinElmer, Boston, MA).

Primers and fluorogenic probes for real-time PCR

Oligonucleotide primers and TaqMan probes for murine IL-10, IL-12 p40, and RANK were purchased from PerkinElmer. Primers and probes for murine β-actin and porphobilinogen deaminase (PBGD) were designed using Primer Express software and were synthesized by PerkinElmer. Forward primer, reverse primer, and probe sequences (5′-3′) for β-actin (forward primer, TCCTTCGTTGCCGGTCCAC; reverse primer, ACCAGCGCAGCGATATCGTC; probe, CCGCCACCAGTTCGCCATG) and PBGD (forward primer, CTGGCCTACAGCGCATGG; reverse primer, CATTCCTCTGGGTGCAAAATCT; probe, TGGCCCACCCGGTTCTGCCA) are shown in parentheses.

Quantification of gene expression using real-time PCR

Levels of RANK, IL-10, and IL-12 p40 mRNA were determined by TaqMan PCR analysis. Sequence-specific amplification of the genes of interest was detected by an increased fluorescent signal of FAM (reporter dye) during amplification. Murine β-actin and PBGD were used as endogenous controls in a multiplex TaqMan PCR (TaqMan Universal PCR Master Mix; PE Biosystems, Foster City, CA) to allow for relative mRNA quantification. Cytokine mRNA levels are presented as the mean ± SD fold increase in gene expression observed in triplicate wells of RANKL-treated DC relative to untreated DC and are representative of data achieved in three separate, independent experiments.

Assessment of oral tolerance in adoptive transfer mice

For adoptive transfer of OVA TCR Tg T cells, BALB/c mice were injected i.v. with 2.5 × 10⁶ clonotypic TCR⁺ (CD4⁺KJ1.26⁺) Tg cells from DO11.10 mice, as previously described (2). Groups of three BALB/c mice were then treated with PBS or 10 μg/ml RANKL-isoleucine zipper fusion protein, given i.p. at −18, −3, and 0 h relative to OVA feeding. Mice were fed a single dose of 25 or 0.5 mg OVA in 0.2 ml saline, or saline only, 2 days after transfer of Tg T cells and were immunized s.c. in the footpad with 100 μg OVA in RIBI adjuvant (RIBI Immunochemicals, Hamilton, MT) 5 days later. After another 4 days, draining PLN were removed, and the Ag-specific proliferative capacity of the lymph node cells was examined by culturing the cells in complete RPMI medium at a density of 2 × 10⁵ cells/well for 48–96 h. Cultures were performed in triplicate in 96-well flat-bottom plates in a total volume of 200 μl, either alone or in the presence of 1 or 0.3 mg/ml OVA. Proliferation was assessed by addition of 1 μCi/well [³H]thymidine (Amersham, Little Chalfont, U.K.) 18 h before harvesting. The amount of radioactivity incorporated into DNA was measured using a Matrix-96 cell harvester (Inotech, Lansing, MI) and a direct beta counter (Packard, Meridan, CT). In addition, the proportion of Tg T cells in adoptively transferred mice was determined by FACS analysis. Cells were incubated with mAb KJ1.26 FITC, which detects the clonotypic Tg TCR, and anti-CD4 PE (BD PharMingen) in 50 μl blocking buffer containing 10 μg/ml anti-CD16 (BD PharMingen), 10% normal goat serum, and 1% normal mouse serum and were analyzed on a FACScan (BD Biosciences) as described above.

Induction and assessment of conventional oral tolerance

BALB/c mice (six mice per group) were injected i.p. with PBS or 10 μg/ml RANKL, given at −18, −3, and 0 h relative to OVA feeding (given as a single dose of 0, 0.5, or 25 mg OVA in 0.2 ml saline by gavage). Ten days after OVA feeding, mice were immunized s.c. into the footpad with 100 μg OVA in 50 μl adjuvant (RIBI adjuvant). Twenty-one days after immunization, mice were assayed for systemic delayed-type hypersensitivity (DTH) responses by measuring the increase in footpad thickness 24 h after challenge with 100 μg heat-aggregated OVA in 50 μl saline as previously described (24). The mice were then sacrificed, and the draining PLN were removed for assessment of in vitro Ag-specific proliferative capability, as described above.

RANK is expressed at similar levels on SPL, PLN, MLN, and PP DC

To determine expression levels of RANK on DC isolated from peripheral and mucosal lymphoid tissues, DC were purified from SPL, PLN, MLN, and PP of Flt3L-treated mice. RNA was then isolated, and levels of RANK mRNA were measured using TaqMan Real-Time PCR. As shown in Fig. 1⇓A, RANK mRNA was expressed at similar levels in DC isolated from each of these tissues. To confirm these findings at the protein level, we also conducted flow cytometric analysis to determine RANK expression on gated CD11c⁺ SPL, MLN, and PP DC. Again, RANK was detected at similar levels on each of these DC populations regardless of the anatomical site (Fig. 1⇓B). Although the level of staining was low, the specificity of the FACS staining was confirmed by cold competition with unlabeled anti-RANK mAb (data not shown).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 1.

Expression of RANK mRNA (A) and protein (B) in DC isolated from peripheral and mucosal lymphoid tissues. A, Levels of RANK mRNA were determined in CD11c⁺ DC purified from SPL, MLN, PP, and PLN of Flt3L-treated mice using TaqMan real-time PCR analysis. RANK mRNA levels are presented as Δ cycle threshold (Δ CT), i.e., the cycle number at which the amplification of the gene lies within the exponential phase of amplification, normalized to the internal housekeeping gene. Data shown are representative of results achieved in three separate, independent experiments. B, RANK protein expression was determined on freshly isolated cells from SPL, MLN, and PP of Flt3L-treated mice using flow cytometry. The histograms represent the levels of RANK on gated CD11c⁺ DC from each of the indicated tissues. Filled histograms show isotype control Ab staining. Data are representative of four independent experiments.

Effect of RANKL on purified peripheral and mucosal DC

Given the similar levels of RANK expression on DC from peripheral and mucosal lymphoid tissues, we next examined how DC isolated from each of these different anatomical compartments responded following RANK engagement. To examine the effect of RANK engagement on cytokine induction by each of these DC populations, DC were again purified from the SPL, PLN, MLN, and PP of Flt3L-treated mice and were incubated in the presence or the absence of soluble murine RANKL for 18 h. RNA was then isolated from the DC cultures, and levels of cytokine mRNA were measured using TaqMan real-time PCR.

Analysis of cytokine mRNA levels indicated that DC isolated from SPL, PLN, MLN, and PP exhibited notably disparate gene expression profiles for IL-10 and IL-12 p40 following RANKL stimulation (Fig. 2⇓). While PLN, MLN, and PP-derived DC did not show an appreciable difference in the levels of IL-12 p40 mRNA after exposure to RANKL in vitro, expression of IL-12 p40 mRNA was up-regulated several-fold in SPL DC. In contrast, whereas IL-10 mRNA expression was not significantly altered in SPL, MLN, or PLN DC following RANKL treatment, elevated IL-10 mRNA expression was observed in RANKL-treated PP DC. Interestingly, IL-18 mRNA expression was consistently down-regulated in SPL, MLN, and PP DC, but not in PLN DC, following RANKL treatment in vitro (data not shown). Taken together, these data suggest that DC isolated from different tissues behave differently in response to stimulation with RANKL.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 2.

Differential effect of RANKL on cytokine gene expression in purified peripheral and mucosal DC. Purified DC from SPL, MLN, PP, and PLN were cultured in the presence or the absence of soluble RANKL for 18 h, after which the RNA was isolated, and the levels of IL-10 and IL-12 p40 were determined by TaqMan PCR analysis. Murine PBGD and β-actin were used as endogenous controls to allow for relative mRNA quantification. Cytokine mRNA levels are shown as the mean (±1 SD) fold increase in gene expression observed in triplicate wells of RANKL-stimulated DC relative to unstimulated DC and are representative of three separate independent experiments.

RANKL treatment of bone marrow-derived DC in vitro has been reported to induce cell aggregation and enhance survival. We tested whether RANKL treatment would prolong the life span of the primary peripheral and mucosal DC cultured in vitro in this study. RANKL treatment induced DC aggregation (Fig. 3⇓A), but no increase in the survival of primary peripheral and mucosal DC was observed (Fig. 3⇓B).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 3.

RANKL treatment does not prolong the survival of mucosal DC. RANKL treatment in vitro induced DC aggregation, which could be visualized under a light microscope (original magnification, ×20; A), but did not alter the survival of DC in culture (B). Purified DC were cultured in the presence or the absence of soluble RANKL, and viability was determined for up to 120 h by differential cell counts. Survival of PP was assessed at 72, 96, and 120 h only, owing to limiting cell yields. Cell yields of individual cultures are shown from a single experiment and are representative of three separate independent experiments.

Treatment with RANKL in vivo enhances oral tolerance induction in an adoptive transfer model

We have recently highlighted a central role for DC as APC for oral tolerance induction (2, 10) and identified GALT as the inductive site for oral tolerance (25). Since we found that RANKL promoted increased expression of mRNA for the immunosuppressive cytokine IL-10 in PP DC, but not in DC from peripheral lymphoid tissues, we next wanted to determine whether RANK engagement in vivo could modulate the outcome of an immune response to orally administered soluble OVA.

We first used an adoptive transfer model in which OVA-specific Tg T cells from D011.10 mice are transferred into normal unirradiated BALB/c recipients and detected using the anti-clonotypic TCR mAb KJ1-26 (26). Previous studies have shown that the number of OVA-specific Tg T cells detectable in draining lymph node (DLN) of adoptively transferred mice is dramatically increased following immunization (26). In contrast, when Ag is administered in a tolerogenic manner such as by the oral route (27), before immunization, the number of Tg T cells detectable in the DLN is decreased, and these cells exhibit a reduced Ag-specific proliferative capacity in vitro. As we have previously observed (2), feeding OVA induced a dose-dependent reduction in the percentage (Fig. 4⇓A), absolute number (Fig. 4⇓B), and proliferative capacity (Fig. 4⇓C) of Tg T cells recovered from the DLN of adoptively transferred, immunized, control (saline-injected) mice compared with saline fed mice. A single low dose feeding of 0.5 mg OVA consistently promoted a small reduction in Tg T cell numbers in control mice (Fig. 4⇓, A and B), but was consistently ineffective at reducing the Ag-specific proliferative capacity of Tg T cells in vitro (Fig. 4⇓C). A single high dose feeding of 25 mg OVA, in contrast, promoted a much greater reduction in the percentage and absolute number of Tg T cells detected in DLN (Fig. 4⇓, A and B) and dramatically reduced the Ag-specific proliferative capacity of these Tg T cells (Fig. 4⇓C).

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 4.

Effect of RANKL treatment in vivo on oral tolerance induction in an adoptive transfer model. Relative proportions (A), absolute numbers (B), and proliferative capacity (C) of OVA Tg T cells in the DLN of RANKL- or saline-treated adoptive transfer mice fed OVA or saline, immunized 5 days later, and analyzed after an additional 4 days. CD4⁺ Tg T cell numbers in the DLN of adoptively transferred mice were quantitated by FACS staining using the clonotypic mAb KJ1-26 and anti-CD4. Saline- and RANKL-injected mice fed 25 mg OVA had a reduced frequency (A) and lower absolute numbers (B) of CD4⁺ Tg T cells than equivalent saline-fed mice (∗, p < 0.01). Proliferative capacity of the Tg T cells was assessed by [³H]thymidine incorporation into DNA in an 18-h culture period. DLN cells from control (saline-injected) mice fed 25 mg OVA had significantly reduced Ag-specific proliferation compared with control mice fed saline (∗, p < 0.001), whereas cells from control mice fed 0.5 mg OVA showed normal proliferative responses. Cells from RANKL-treated mice fed either low (0.5 mg) or high (25 mg) OVA showed significantly reduced proliferation compared with cells from RANKL-treated mice fed saline (#, p < 0.001). Notably, the proliferative response of cells from RANKL-treated mice fed low dose OVA was significantly reduced compared with the response of cells from saline-treated mice fed low dose OVA (¶, p < 0.001). The data are reported as the mean cpm ± 1 SD of triplicate wells and are representative of three separate experiments.

When we examined the number of Tg T cells recovered from the DLN of RANKL-treated mice fed OVA, we observed a similar dose-dependent reduction in the percentage (Fig. 4⇑A) and absolute number (Fig. 4⇑B) of Tg T cells compared with that seen in OVA-fed mice that did not receive RANKL. However, the functional capacity of DLN cells from RANKL-treated, OVA-fed mice was quite distinct from that of cells from saline-treated, OVA-fed mice. This was particularly evident at low doses of OVA, which were insufficient to induce tolerance in saline-injected mice. Accordingly, DLN cells from RANKL-treated mice fed low doses of OVA showed dramatically reduced OVA-specific proliferation in vitro compared with cells from saline-injected, OVA-fed mice (Fig. 4⇑C). The markedly reduced proliferation of DLN cells from RANKL-treated mice fed low dose OVA resembled the diminished functional capacity of cells from saline-injected mice fed high dose OVA (Fig. 4⇑C). RANKL treatment had little effect on the low levels of proliferation observed at feeding a high dose OVA (Fig. 4⇑C), most likely because these levels were already maximally suppressed.

Treatment with RANKL in vivo enhances tolerance in a conventional model of oral tolerance induction

The above findings led us to examine whether RANKL might also modulate mucosal tolerance using a well-characterized, conventional model of oral tolerance induction. Control (saline-injected) mice fed soluble OVA before immunization with OVA plus adjuvant showed a dose-dependent reduction in DTH and Ag-specific in vitro proliferative responses (Fig. 5⇓), as we have previously described (2). Thus, while the DTH and proliferative responses of control mice fed low dose OVA (0.5 mg) were essentially normal (Fig. 5⇓), control mice fed high dose OVA (25 mg), displayed significantly suppressed in vivo and in vitro T cell responses (Fig. 5⇓). The pattern of tolerance induced by feeding OVA to RANKL-treated mice, was however, quite distinct from that observed in control OVA-fed mice. RANKL-treated mice fed low dose OVA showed markedly suppressed DTH and in vitro proliferative responses compared with equivalent control mice fed this dose of Ag (Fig. 5⇓), such that RANKL-treated mice fed low dose OVA resembled control mice fed high dose OVA. RANKL had less effect on the profound degree of tolerance exhibited in mice fed high dose OVA. As such, RANKL-treated mice fed high dose OVA exhibited the same markedly suppressed responses as saline-injected controls fed high dose OVA (Fig. 5⇓). Again, the fact that RANKL had little effect on the suppressed responses observed in mice fed high dose OVA may be attributed to the fact that these levels were already maximally suppressed.

• Download figure
• Open in new tab
• Download powerpoint

FIGURE 5.

RANKL treatment enhances the induction of tolerance of in vivo and in vitro T cell responses in mice fed soluble OVA. A, OVA-specific DTH responses. Control (saline-injected) mice fed 25 mg soluble OVA before immunization with OVA plus adjuvant exhibited significantly reduced DTH responses compared with control mice fed saline (∗, p < 0.05), whereas control mice fed 0.5 mg before immunization exhibited normal DTH responses. Mice treated with RANKL at the time of feeding either low (0.5 mg) or high (25 mg) dose OVA showed significantly reduced DTH responses compared with responses in RANKL-treated mice fed saline (#, p < 0.01). Moreover, RANKL-treated mice fed low dose OVA exhibited significantly reduced DTH responses compared with those exhibited by saline-treated mice fed low dose OVA (¶, p < 0.01). The data are reported as the mean footpad increment ± 1 SEM achieved in groups of six mice from a single experiment. B, OVA-specific proliferation in vitro. DLN cells isolated from control (saline-injected) mice fed 25 mg OVA showed significantly reduced Ag-specific proliferative responses compared with the responses of cells from control mice fed saline (∗, p < 0.001), whereas cells from control mice fed 0.5 mg OVA showed proliferation similar to that in control mice fed saline. DLN cells from mice treated with RANKL at the time of feeding either low or high dose OVA displayed significantly decreased proliferative responses compared with responses of cells from RANKL-treated mice fed saline (#, p < 0.05; ##, p < 0.001). Cells from RANKL-treated mice fed low dose OVA showed significantly reduced proliferation compared with that of cells from saline-treated mice fed low dose OVA (¶, p < 0.001). The data are reported as the mean cpm ± 1 SD of triplicate wells from a single experiment.