Abstract
We have recently demonstrated that dendritic cells (DC) prepared from nonobese diabetic (NOD) mice, a spontaneous model for insulin-dependent diabetes mellitus, exhibit elevated levels of NF-κB activation upon stimulation. In the current study, we investigated the influence of dysregulation of NF-κB activation on the APC function of bone marrow-derived DC prepared from NOD vs BALB/c and nonobese diabetes-resistant mice. NOD DC pulsed with either peptide or virus were found to be more efficient than BALB/c DC at stimulating in vitro naive Ag-specific CD8+ T cells. The T cell stimulatory capacity of NOD DC was suppressed by gene transfer of a modified form of IκBα, indicating a direct role for NF-κB in this process. Furthermore, neutralization of IL-12(p70) to block autocrine-mediated activation of DC also significantly reduced the capacity of NOD DC to stimulate T cells. Despite a reduction in low molecular mass polypeptide-2 expression relative to BALB/c DC, no effect on proteasome-dependent events associated with the NF-κB signaling pathway or Ag processing was detected in NOD DC. Finally, DC from nonobese diabetes-resistant mice, a strain genotypically similar to NOD yet disease resistant, resembled BALB/c and not NOD DC in terms of the level of NF-κB activation, secretion of IL-12(p70) and TNF-α, and the capacity to stimulate T cells. Therefore, elevated NF-κB activation and enhanced APC function are specific for the NOD genotype and correlate with the progression of insulin-dependent diabetes mellitus. These results also provide further evidence indicating a key role for NF-κB in regulating the APC function of DC.
Studies in the nonobese diabetic (NOD)3 mouse, a spontaneous model for insulin-dependent diabetes mellitus (IDDM), have demonstrated that autoimmune destruction of the insulin-producing β cells is primarily mediated by CD4+ and CD8+ T cells. Furthermore, pathogenic CD4+ Th cells specific for a variety of β cell autoantigens typically exhibit a Th1-like phenotype (1, 2). The critical events resulting in the breakdown of self-tolerance within the T cell compartment and the apparent skewing toward Th1 cell differentiation are poorly understood but are clearly influenced by both environmental and genetic factors (2). Notably, macrophages prepared from NOD mice exhibit defects in differentiation and function which may contribute to the development of pathogenic T cells (3, 4, 5, 6, 7, 8). For example, macrophages prepared from NOD vs nonautoimmune strains of mice secrete elevated levels of IL-12 which could aid in mediating preferential differentiation of Th1 and T cytotoxic 1 (Tc1) cells (3). Due to a limited capacity to stimulate a syngeneic lymphocyte reaction, NOD macrophages are also believed to be deficient in eliciting regulatory T effector cell function (7).
In contrast to macrophages, little attention has focused on the role of dendritic cells (DC) in the pathogenesis of IDDM. In general, DC are characterized by a potent capacity to stimulate naive CD4+ and CD8+ T cells. Depending on the state of maturation, DC exhibit distinct APC functions (9). For example, immature DC are highly efficient at Ag uptake and processing. Following Ag encounter, activated immature DC traffic from peripheral nonlymphoid tissues into T cell-rich regions of the lymph nodes (10, 11, 12). There, DC undergo maturation marked by a decreased capacity to phagocytose and process Ag. The cognate interaction with CD40 ligand-expressing T cells results in engagement of CD40 which delivers signals further promoting DC maturation. Subsequent up-regulation of surface MHC class I and II and costimulatory molecules such as CD40, CD80, and CD86 enhances the capacity of mature DC to activate naive T cells (9). In addition to providing activational signals, DC regulate commitment of Th and Tc subset differentiation through secretion of cytokines such as IL-12 (13). Because of the elevated capacity to stimulate T cells and drive T cell differentiation, DC would be expected to contribute to the development of β cell-specific T cell reactivity. Indeed, DC are among the first cells detected within the islets of Langerhans during the onset of IDDM in NOD mice (14). Furthermore, overt diabetes is induced in neonatal NOD mice by adoptive transfer of syngeneic DC pretreated with TNF-α (15). Recently, we demonstrated that DC derived from bone marrow or the spleen of NOD mice exhibit enhanced levels of activation of the transcriptional factor NF-κB following various types of stimulation compared with BALB/c and C57BL/6 DC (16). Importantly, the hyperactivation of NF-κB detected in NOD DC directly resulted in elevated levels of IL-12 secretion relative to DC prepared from control animals.
The NF-κB family of transcription factors has multiple roles in regulating events associated with an immune response (17, 18). NF-κB, which is composed of dimers of the Rel family of proteins, is found in the cytoplasm of resting cells complexed with the inhibitory proteins IκBα, IκBβ, and IκBε. Following stimulation, the IκB proteins are phosphorylated and then degraded in an ubiquitin- and proteasome-dependent manner, permitting NF-κB to enter the nucleus and activate gene transcription (17, 18). A role for NF-κB in various aspects of DC immunobiology has been reported by a number of studies. Mice lacking RelB-containing NF-κB complexes, for instance, fail to develop myeloid-derived DC (19, 20). In addition, LPS-induced maturation of a murine splenic DC line was shown to be dependent on activation of NF-κB (21). Furthermore, various stimuli including CD40 engagement, IL-1, IL-12, TNF-α, and LPS activate NF-κB in DC.
Our finding that IL-12 secretion is significantly increased due to dysregulation of NF-κB prompted us to investigate whether other aspects of APC function were also increased in NOD DC (16). Specifically, the ability of NOD DC to process and/or present Ag and in turn stimulate naive or previously activated CD8+ T cells was compared with DC prepared from BALB/c and nonobese diabetes-resistant (NOR) mice. We found that NOD DC pulsed with peptide or intact Ag exhibited an increased capacity to stimulate CD8+ T cells and promote Tc1 cell differentiation in vitro. Importantly, selective inhibition of NF-κB activation significantly impaired the capacity of NOD DC to stimulate T cells. These results demonstrate that elevated NF-κB activation results in an overall enhanced APC function of NOD DC, which may contribute to the pathogenesis of IDDM in NOD mice.
Materials and Methods
Mice
BALB/cJ and NOR/Lt mice were purchased from The Jackson Laboratory (Bar Harbor, ME) and maintained and bred in specific pathogen-free conditions at the University of North Carolina animal facilities (Chapel Hill, NC). NOD/LtJ mice were similarly housed and bred. Currently in our colony, IDDM develops in ∼80% of NOD/LtJ female mice by one year of age. The NOD.CL4 mouse line was established by back-crossing transgenes encoding the CL4 clonotypic TCR derived from BALB/c.CL4 mice (provided by Dr. R. Liblau, Institut National de la Sante et de la Recherche Medicale, Paris, France) onto the NOD/LtJ genotype for 12 generations. The CL4-clonotypic TCR is H2Kd-restricted and specific for an influenza hemagglutinin (HA) peptide spanning amino acid residues 512–520.
Preparation of primary DC
Bone marrow-derived DC were prepared from the femurs of male or female mice between 8 and 12 wk of age. Following lysis of RBCs, bone marrow was depleted of CD4 (mAb GR1.5)-, CD8 (mAb HO2.2)-, MHC class I (mAb M1/42.3.9)-, MHC class II (mAb B21.2, anti-I-Ab,d, and mAb 10.2.36, anti-I-Ag7)-, and B220 (mAb RA3–3A1/6.1)-expressing cells via complement-mediated lysis. DC precursors were plated on six-well low-cluster plates in RPMI 1640 medium containing 10% FBS and penicillin/streptomycin (base medium), 10 ng/ml murine IL-4 (PeproTech, Rocky Hill, NJ), and 10 ng/ml murine GM-CSF (PeproTech). On the second day of culture, nonadherent cells were harvested and cultured as above for 8 days. Culture medium was added on days 4–5 and 7–8. For all experiments DC were harvested on day 10 of the cultures. Flow cytometric analysis demonstrated that DC expressed DEC-205, MHC class I (H2Kd), MHC class II, CD11c, CD80, CD86, and CD40, but not CD8α.
Fluorescence staining
The following mAbs used for fluorescence staining were purchased from BD PharMingen (San Diego, CA): anti-CD40 (clone HM40-3), FITC-anti-CD86 (clone GL1), FITC-anti-CD80 (clone 16-10A1), FITC-anti-B220 (clone RA3-6B2), PE-anti-H-2Kd (clone SF1-1.1), FITC-anti-CD11c (clone HL3), PE-anti-CD4 (clone L3T4), PE-anti-CD8 (clone 53-6.7), FITC-anti-CD44 (clone IM7), biotinylated anti-CD69 (clone H1.2F3), and biotinylated anti-CD62 ligand (L) (clone MEL-14). Anti-I-Ad (clone MK-D6) and anti-I-Ag7 (clone 10.2.36) mAbs were provided by E. P. Reich (Immunologic, Palo Alto, CA). The anti-DEC 205 clone NLDC-145 was provided by Dr. R. Johnston (University of North Carolina). PE-anti-mouse and PE-streptavidin secondary reagents were purchased from BD PharMingen. Following staining, analysis was conducted on a FACScan (BD Biosciences, San Jose, CA) using Summit software (Cytomation, Ft. Collins, CO).
EMSA
DC cultured in base medium at 5 × 105 cells/well in a 24-well plate were stimulated with 10 ng/ml murine TNF-α, 25 μg/ml LPS, or 10 μg/ml anti-CD40 Ab for specified times. Nuclear and cytoplasmic extracts were prepared as described previously (22). EMSA were performed as described using the DNA probe 5′-CAGGCTGGGGATTCCCATCTCCACAGTTTCACTTC-3′, which contains the NF-κB binding site from the MHC class I H2Kb gene (23). For supershift experiments, anti-RelA (sc-109), anti-p50 (sc-114-G), anti-p52 (sc-298), anti-c-Rel (sc-071), and anti-RelB (sc-726) rabbit polyclonal Abs from Santa Cruz Biotechnology (Santa Cruz, CA) were added to each sample. Bands were visualized using a PhosphorImager (Molecular Dynamics, Sunnyvale, CA).
Western blotting
For Western blotting, 50 μg of cytoplasmic extract prepared as described (22) was analyzed by SDS-PAGE using a 12% separating gel and 5% stacking gel. Proteins were transferred to a nitrocellulose membrane (Osmonics, Minnetonka, Minnesota) using a semidry transfer system and blocked overnight at 4°C with 5% nonfat dried milk in PBS. Blots were probed with anti-murine low molecular mass polypeptide (LMP)-2, LMP-7 (Affiniti Research Products, Exeter, U.K.), or anti-IκBα (sc-371 and sc-847) (Santa Cruz Biotechnology) rabbit polyclonal Abs. Following incubation with HRP-labeled goat anti-rabbit secondary Ab (Promega, Madison, WI), blots were developed using ECL reagents (Amersham Pharmacia Biotech, Piscataway, NJ), and the intensity of the bands was determined by densitometric readings.
IL-12 and TNF-α secretion assays
DC were harvested on day 10 of culture and plated at 106 cells/ml in a 24-well plate with 1 ml of base medium. Cells were stimulated with LPS or anti-CD40 mAb for 72 h. Supernatants were collected and assayed for IL-12(p70) and TNF-α secretion using ELISA kits (BD PharMingen) for the respective cytokines.
DC:T cell assays
CD8+ T cells were purified from the spleens of NOD.CL4 or (NOD.CL4 × BALB/c)F1 (F1.CL4) mice using anti-CD8 magnetic beads (Miltenyi Biotech, Auburn, CA). Flow cytometry analysis confirmed that T cells were CD62highCD44lowCD69low and >90% CD8+.
HA peptide-pulsed DC.
Varying numbers of DC were incubated with specified concentrations of HA peptide in 1 ml of base medium supplemented with 50 μM 2-ME, 1 mM sodium pyruvate, 1× nonessential amino acids, and 1 mM glutamine (complete medium) in a 24-well plate. After 30 min, DC were washed four times with medium, and 1 × 105 T cells were added to each well in a total of 0.5 ml of complete medium. At 48 h, culture supernatants were harvested and 100-μl aliquots were analyzed for levels of IL-2, IL-4, IL-5, IL-10, and IFN-γ via capture ELISA as per the manufacturer’s instructions (BD PharMingen). Alternatively, T cells were harvested after 96 h, washed, and restimulated for 48 h with fresh DC, and cytokine secretion was measured as described above. HA peptide was synthesized using standard Fmoc chemistry on a Ranin Symphony at the Peptide Sequencing Facility at the University of North Carolina. The purity of the peptides was verified by reversed phase HPLC and mass spectroscopy. In certain experiments, DC were transfected with pCMV-FLAG or pCMV-IκB super-repressor (SR)-FLAG before incubation with peptide. Transfections were performed with Superfect (Qiagen, Valencia, CA) as recommended by the manufacturer for six-well plates. Briefly, a transfection mixture containing 1 μg of DNA and 8 μl of Superfect reagent in a volume of 0.4 ml of RPMI 1640 was added to 5 × 106 DC for 3 h at 37°C, after which DC were washed extensively in medium.
Influenza virus-pulsed DC.
DC were incubated with HA peptide, BSA protein, or influenza strain A/Puerto Rico/8/34 (PR8) (provided by Dr. J. Frelinger, University of North Carolina) previously heat-inactivated for 30 min at 56°C. After 4 h, DC were washed three times with PBS and cultured at varying numbers in 96-well microtiter plates. Naive CD8+ T cells (5 × 104) were added to the DC cultures in a final volume of 250 μl/well. At 48 h, culture supernatants were harvested and levels of cytokine secretion were determined as above.
We also performed experiments to confirm that proteasomal processing was occurring during whole virus experiments. BALB/c DC (1 × 105) were pretreated with 10 μM lactacystin prepared in 0.1% DMSO, a proteasome-specific inhibitor, or DMSO for 1 h and then incubated with 1 μM HA peptide or increasing hemagglutination activity units (hau) of heat-inactivated PR8 influenza virus for 4 h. The 10 μM lactacystin was found to be the lowest concentration which mediated maximum inhibition as determined by measuring the effect of a range of doses. DC were washed three times with PBS and cultured with 1 × 105 F1.CL4 T cells for 48 h. ELISA for IL-2 secretion was performed as described above.
Neutralization of IL-12 secretion from DC.
A total of 5 × 105 DC were pulsed with 1 μM HA peptide for 30 min, washed three times with PBS, and cultured with 2 × 105 NOD.CL4 CD8+
Infection of DC with adenovirus recombinants
Construction of the IκB-SR has been described in detail elsewhere (24). Recombinant adenovirus (either control or encoding the SR form of IκB) was prepared by the University of North Carolina Gene Therapy Center. DC prepared from 10-day old cultures were washed in PBS, resuspended in base medium at 106 cells/ml, and then infected with adenovirus encoding IκB-SR or β-galactosidase at multiplicity of infection (moi) of 50. Twenty-four hours after infection, DC were washed and then stimulated with 50 μg/ml LPS in base medium for 18 h and then analyzed by flow cytometry. β-Galactosidase assays indicated that >90% of DC were infected by adenovirus at 50 moi. Propidium iodide was used to exclude cellular debris.
Results
Peptide-pulsed NOD DC exhibit an enhanced capacity to stimulate CD8+ T cells in vitro
We have previously reported that activation of NF-κB by a variety of stimuli is significantly enhanced in bone marrow or splenic DC prepared from NOD vs BALB/c and C57BL/6 mice (16). Furthermore, dysregulation of NF-κB activation results in increased secretion of IL-12(p70) by NOD DC compared with DC prepared from the two control strains of mice. To test the possibility that other APC functions are enhanced due to increased NF-κB activation, NOD and BALB/c DC were examined for the capacity to stimulate naive T cells in a peptide-specific manner. Accordingly, NOD.CL4 mice transgenic for a TCR specific for influenza virus HA were used. Because the HA-specific TCR is H2Kd-restricted (25), a direct comparison could be made between H2Kd-expressing NOD and BALB/c DC. To ensure that possible differences in activation and/or cytokine secretion between NOD and BALB/c T cells did not influence the analysis, (NOD.CL4×BALB/c)F1 mice (F1.CL4) were used as a source of CD8+ T cells. Initially, F1.CL4 CD8+ T cells exhibiting a CD62LhighCD69lowCD44low phenotype were cultured with varying numbers of NOD or BALB/c DC pulsed with 1 μM HA peptide, and IL-2 secretion was measured via ELISA. As demonstrated in Fig. 1⇓A, IL-2 levels were significantly increased in cultures containing NOD vs BALB/c DC at all numbers of DC tested. Maximum IL-2 secretion (350 pg/ml) was detected with 1 × 104 NOD DC, a 10-fold increase compared with cultures containing a similar number of BALB/c DC (35 pg/ml).
Peptide-pulsed NOD DC exhibit an increased capacity to stimulate CD8+ T cells compared with BALB/c DC. NOD and BALB/c DC at various concentrations were pulsed with 1 μM HA peptide and cultured with 1 × 105 naive CD8+ F1.CL4 T cells for 48 h (A). Alternatively, F1.CL4 T cells were stimulated with NOD or BALB/c DC for 96 h, harvested, and then restimulated for 48 h with corresponding NOD or BALB/c DC pulsed with HA peptide for 48 h (B and C). Culture supernatants were analyzed for the production of IL-2 and IFN-γ by ELISA. Data represent mean values ± SD of triplicate wells and are representative of at least three independent experiments. ∗, Student’s t test confirmed that differences between NOD and BALB/c for a given number of DC were significant (p < 10−3).
NOD DC were also more effective than BALB/c DC at stimulating previously activated F1.CL4 T cells. At low numbers, NOD DC promoted significant IL-2 and IFN-γ secretion by primed F1.CL4 T cells, with maximum levels of cytokine secretion detected in cultures containing 5 × 104 DC (Fig. 1⇑, B and C). In contrast, maximum levels of IL-2 and IFN-γ secretion were observed with 5 × 105 BALB/c DC (Fig. 1⇑, B and C). Even at this relatively high number of DC, cytokine secretion in BALB/c cultures was less than that observed in cultures containing fewer NOD DC. Furthermore, increased numbers of NOD DC (>5 × 104) proved to be inhibitory for IL-2 and IFN-γ secretion by the F1.CL4 T cells. In the same experiment, 5 × 104 NOD or BALB/c DC only were stimulated with 50 μg/ml LPS, and 48 h afterward IFN-γ secretion was determined via ELISA. Whereas 280 pg/ml IFN-γ was detected in NOD DC cultures, no IFN-γ above background was observed in cultures of BALB/c DC. These results confirm that the IFN-γ detected in our DC:T cell cocultures was predominately derived from T cells. Secretion of IL-4, IL-5, and IL-10 was not detected above background in either NOD or BALB/c DC cultures.
Additionally, no significant difference was observed between NOD and BALB/c DC in the levels of H2Kd, CD40, CD80, and CD86 expression either before or following LPS stimulation (Table I⇓). This observation indicates that the heightened capacity of NOD DC to stimulate T cells was not due to increased surface expression of peptide:MHC complexes or costimulatory molecules compared with BALB/c DC.
Flow cytometry analysis of cell surface proteins expressed by bone marrow-derived DCa
NOD DC pulsed with influenza virus exhibit an enhanced capacity to stimulate CD8+ T cells in vitro
The above results demonstrated that peptide-pulsed NOD DC exhibited an elevated capacity to stimulate CD8+ T cells compared with BALB/c DC. Next, we determined whether a similar difference existed between the respective DC when the HA epitope was naturally processed and presented. Previous work reported that exogenous Ags are readily processed and presented by DC to CD8+ T cells (26, 27, 28). Accordingly, heat-inactivated influenza PR8 virus, which contains the p512–520 HA-specific epitope recognized by CL4 T cells, was used as a source of exogenous Ag. To ensure that the HA epitope was processed in a proteasome-dependent manner, 1 × 105 BALB/c DC were pulsed with HA peptide or virus in the presence or absence of 10 μM lactacystin and then examined for the capacity to stimulate F1.CL4 T cells. Lactacystin is a proteasome-specific inhibitor shown to block MHC class I processing and presentation of cytoplasmic Ags (29). DC treated with or without lactacystin and pulsed with HA peptide stimulated IL-2 secretion by F1.CL4 T cells to a similar extent (Fig. 2⇓A). In contrast, lactacystin significantly reduced the capacity of DC pulsed with virus to stimulate IL-2 secretion by F1.CL4 T cells relative to virus-pulsed DC not treated with the drug (Fig. 2⇓A). This result indicated that the HA epitope was being processed in a proteasome-dependent manner and presented by virus-pulsed DC.
NOD DC exhibit an increased capacity to stimulate naive CD8+ T cells via proteasome-dependent processing of HA compared with BALB/c DC. A, BALB/c DC were pretreated with 10 μM lactacystin prepared in 0.1% DMSO or DMSO only for 1 h and then incubated with 1 μM HA peptide or increasing hau of heat-inactivated PR8 influenza virus for 4 h. DC were washed three times and cultured with 1 × 105 F1.CL4 T cells for 48 h. B, A total of 1 × 104 NOD or BALB/c DC were pulsed with varying hau of heat-inactivated virus, as above, and cultured with 1 × 105 F1.CL4 T cells for 48 h. Supernatants were harvested and IL-2 was detected via ELISA. Data represent mean values ± SEM from duplicate wells and are representative of at least three independent experiments.
A comparison was then made between NOD and BALB/c DC (1 × 104) pulsed with varying doses of heat-inactivated virus or BSA protein to stimulate F1.CL4 T cells. NOD DC induced significantly higher levels of IL-2 secretion by F1.CL4 T cells compared with BALB/c DC, regardless of the dose of virus (Fig. 2⇑B). For example, at 6 hau of virus, 1755 and 38 pg/ml of IL-2 were detected in cultures containing NOD and BALB/c DC, respectively (Fig. 2⇑B). Furthermore, maximum levels of IL-2 secretion (2500 pg/ml) were initially detected in NOD DC cultures pulsed with 30 hau of virus (Fig. 2⇑B). In contrast, 973 pg/ml of IL-2 was detected in BALB/c DC cultures pulsed with 75 hau of virus, the highest dose of viral particles tested (Fig. 2⇑B). Neither NOD nor BALB/c DC pulsed with BSA induced levels of IL-2 secretion above background (data not shown).
Inhibition of NF-κB activation suppresses the capacity of NOD DC to stimulate CD8+ T cells
To determine whether the increased levels of NF-κB directly influenced APC function, NOD DC were transfected with plasmid DNA (pCMV-SR) encoding a specific inhibitor of NF-κB, the IκB-SR, then pulsed with HA peptide and examined for the capacity to stimulate naive NOD.CL4 T cells. Typically, 40–60% of DC were transfected under the conditions used. NOD DC transfected with the control plasmid pCMV stimulated IL-2 and IFN-γ secretion equivalent to that of mock-transfected NOD DC (Fig. 3⇓). In marked contrast, both IL-2 and IFN-γ secretion by NOD.CL4 T cells was significantly reduced in cultures containing DC transfected with pCMV-SR relative to mock-transfected DC (Fig. 3⇓). Suppression of IL-2 and IFN-γ secretion by F1.CL4 T cells was also observed when BALB/c DC were transfected with pCMV-SR but not pCMV (data not shown).
Specific inhibition of NF-κB activation suppresses the capacity of NOD DC to stimulate T cells. A total of 1 × 104 NOD DC were transfected with control pCMV or pCMV-SR plasmids or mock (no plasmid)-transfected 24 h before pulsing with 0.1 μM HA peptide, and then cultured with 2 × 105 NOD.CL4 T cells. Levels of IL-2 and IFN-γ were determined as described in Fig. 1⇑. ∗, Student’s t test confirmed that the differences between mock-transfected DC and DC transfected with pCMV-SR were significant (p ≤ 0.02). Data are representative of at least two independent experiments.
The inability of IκB-SR-expressing DC to stimulate T cells could be due to a block in up-regulation of cell surface MHC class I and/or costimulatory molecules. To test this possibility, NOD DC were infected with an adenovirus recombinant encoding IκB-SR (Adeno-SR). Adenovirus-based gene transfer was used because we consistently obtain >90% infection of DC using a moi of 50–100 with various recombinants. DC infected with Adeno-SR and stimulated with LPS exhibited levels of CD40, CD80, and CD86 expression comparable to noninfected NOD DC treated with LPS (Table I⇑). Importantly, DC infected under the same conditions with Adeno-SR and pulsed with HA peptide or PR8 virus failed to stimulate IL-2 secretion by NOD.CL4 T cells (Table II⇓). In contrast, DC infected with an adenoviral recombinant encoding β-galactosidase (Adeno-LacZ) effectively stimulated T cells (Table II⇓).
NOD DC infected with Adeno-SR no longer stimulate IL-2 secretion by CD8+ T cellsa
Neutralization of IL-12(p70) inhibits the capacity of NOD DC to stimulate CD8+ T cells in vitro
Recently, IL-12(p70) was shown in an autocrine manner to enhance the immunogenicity of peptide-pulsed DC initially activated by CD40 engagement (30). With this mind, we investigated the possibility that IL-12(p70) was contributing to the ability of NOD DC to stimulate CD8+ T cells. Cultures were prepared in which anti-IL-12(p70) Ab was added to HA-pulsed NOD DC, and cytokine secretion by naive NOD.CL4 T cells was measured. As shown in Fig. 4⇓, treatment with anti-IL-12(p70) Ab significantly reduced the levels of IL-2 and IFN-γ secretion by NOD.CL4 T cells. Addition of an isotype-matched control Ab to cultures had no suppressive effect on the capacity of DC to stimulate cytokine secretion by the T cells (data not shown).
Neutralization of autocrine IL-12(p70) inhibits the capacity of NOD DC to stimulate T cells. NOD DC were incubated with 1 μM HA peptide for 30 min, washed, and then cultured with varying amounts of anti-IL-12 Ab and 2 × 105 NOD.CL4 T cells for 96 h. Supernatants were analyzed for IL-2 and IFN-γ production via ELISA. Data represent mean values ± SEM of duplicate wells and are representative of at least two independent experiments. ∗, Student’s t test confirmed that the differences between control and anti-IL-12 Ab-treated DC were significant (p ≤ 0.01).
Reduced LMP-2 expression in NOD DC has no effect on p100 and p105 processing or IκBα degradation
A study recently reported that NOD splenocytes exhibit decreased NF-κB activity due to a proteasome defect resulting from the absence of the LMP-2 subunit (31). The decrease in NF-κB activity was attributed to a lack of the p50 and p52 subunits and IκBα degradation. p50 and p52 are products of the proteasome- and LMP-2-dependent processing of p105 and p100 precursor molecules, respectively (32, 33). Our data demonstrating enhanced NF-κB activation (16) and effective proteasome-dependent processing of HA suggested that similar defects may not reside in NOD DC. To examine this issue further, the levels of LMP-2 and LMP-7 subunit expression were compared between NOD and BALB/c DC stimulated with or without TNF-α. As demonstrated in Fig. 5⇓, NOD DC exhibited an approximate 2-fold reduction in LMP-2 expression compared with BALB/c DC at all times analyzed following TNF-α treatment. In contrast, no difference was observed in the expression of LMP-7 between NOD and BALB/c DC (Fig. 5⇓).
Decreased LMP-2 expression in NOD vs BALB/c DC. NOD and BALB/c DC prepared from bone marrow of 8- to 10-wk-old female mice were stimulated with 10 ng/ml TNF-α for various times. Cytoplasmic extracts were prepared and Western blot analysis was performed with 50 μg of protein. Blots were probed with Abs specific for LMP-2 or LMP-7. Data are representative of two independent experiments.
In addition, we assessed whether the decreased expression of LMP-2 in NOD DC affected p105 and p100 processing or IκBα degradation. Nuclear extracts from TNF-α-treated DC were analyzed by EMSA using Abs to supershift specific NF-κB complexes. Both p50 and p52 complexes were observed in NOD and BALB/c DC extracts (Fig. 6⇓A and data not shown, respectively). Moreover, the levels of these complexes were higher in nuclear extracts prepared from NOD vs BALB/c DC, consistent with our prior observations (16). Finally, we examined whether reduced levels of LMP-2 expression in NOD DC affected proteasome-dependent degradation of IκBα. As shown in Fig. 6⇓B, the level of degradation of IκBα was similar in cytoplasmic extracts prepared from NOD and BALB/c DC following either TNF-α or LPS stimulation.
NOD DCs efficiently process p100 and p105 and degrade IκBα. NOD and BALB/c DC were stimulated with 10 ng/ml TNF-α (A and B) or 25 μg/ml LPS (B). A, Nuclear extracts from NOD DC were prepared and EMSA was performed using a DNA probe containing the NF-κB consensus site. Supershifts were performed using various Abs specific to NF-κB subunits. B, Cytoplasmic extracts from NOD and BALB/c DC were probed on a Western blot with anti-IκBα Ab. Data are representative of at least three independent experiments.
Dysregulation of NF-κB activation and increased APC function are specific for the NOD mouse genotype
To determine whether enhanced NF-κB activation is indeed specific to the NOD genotype and a corollary for IDDM progression, DC prepared from the bone marrow of diabetes-resistant NOR mice were examined. The NOR mouse is a NOD-related MHC-syngeneic recombinant strain which contains ∼12% C57/KsJ-derived genes (34). These mice develop peri-insulitis; however, they develop only limited (if any) intrainsulitis and consequently do not develop overt diabetes. Treatment with anti-CD40 Ab or LPS and subsequent EMSA analysis of nuclear extracts demonstrated that NF-κB nuclear translocation in NOR DC was significantly reduced relative to NOD DC but comparable with that observed in BALB/c DC (Fig. 7⇓A). On repeated experiments both bone marrow and splenic DC prepared from NOR mice consistently exhibited nuclear levels of NF-κB below that observed in NOD DC after stimulation (data not shown).
NOD DC exhibit increased NF-κB nuclear translocation and NF-κB-dependent cytokine secretion compared with NOR and BALB/c DC. DC prepared from 8- to 10-wk old NOD, NOR, and BALB/c female mice were stimulated with 10 μg/ml anti-CD40 Ab or 25 μg/ml LPS for 30 min (A) or 48 h (B and C). A, Nuclear extracts were prepared from 5 × 105 DC and analyzed by EMSA using a DNA probe containing the NF-κB consensus site. Data are representative of two independent experiments. B and C, Supernatants from 1 × 104 NOD, NOR, or BALB/c DC treated with anti-CD40 Ab or LPS were analyzed for IL-12(p70) (B) and TNF-α (C) secretion by ELISA. ELISAs were performed from triplicate wells and data are representative of at least three independent experiments. Data shown are mean SD. ∗, Student’s t test-confirmed differences between NOD vs NOR or BALB/c DC were significant (p ≤ 0.001).
Next, DC prepared from NOD, NOR, and BALB/c mice were assessed for IL-12(p70) and TNF-α secretion following stimulation. NF-κB has been shown to play a role in the transcription of both the IL-12 p40 and TNF-α genes (35, 36, 37). Consistent with the EMSA data, NOD DC exhibited markedly increased levels of secretion of the two cytokines compared with NOR and BALB/c DC upon stimulation with either LPS (Fig. 7⇑B) or anti-CD40 Ab (Fig. 7⇑C).
Finally, a comparison was made between HA peptide-pulsed NOD and NOR DC to stimulate naive NOD.CL4 T cells. As demonstrated in Fig. 8⇓, a 5-fold greater concentration of IL-2 was detected in cultures containing NOD vs NOR DC. No significant differences were detected between NOD and NOR cell surface expression of H2Kd or costimulatory molecules during flow cytometric analysis (data not shown). These results indicate that NOD DC have enhanced APC function compared with NOR as well as BALB/c DC.
NOD DC exhibit an increased capacity to stimulate T cells compared with NOR and BALB/c DC. A total of 1 × 104 NOD, BALB/c, and NOR DCs were pulsed with 1 μM HA peptide and cultured with 1 × 105 naive CD8+ NOD.CL4 T cells for 48 h. Supernatants were analyzed for production of IL-2 by ELISA. Data shown represent mean values ± SD of triplicate wells and are representative of two experiments. ∗, Student’s t test-confirmed differences between NOD vs NOR or BALB/c DC were significant (p ≤ 0.001).
Discussion
Our previous work demonstrated that myeloid DC prepared from the bone marrow and spleen of NOD mice exhibit enhanced levels of NF-κB activation compared with BALB/c and C57BL/6 DC following stimulation with either IL-1, IL-12(p70), TNF-α, LPS, or anti-CD40 Ab (16). Furthermore, activated NOD DC were found to have an increased capacity to secrete IL-12(p70) relative to BALB/c and C57BL/6 DC, consistent with the fact that inducible expression of the IL-12(p40) gene is in part regulated by NF-κB (37). In the current study, we provide evidence that elevated NF-κB activation impacts on the ability of NOD DC to stimulate T cells. Namely, NOD DC pulsed with either peptide or virus exhibited a significantly enhanced capacity to stimulate HA-specific CD8+ T cells compared with BALB/c and NOR DC (Figs. 1⇑, 2⇑, and 8⇑). Interestingly, NOD DC pulsed with a mimotope peptide stimulated ∼5- to 10-fold greater levels of IL-2 secretion in vitro by CD4+ T cells prepared from NOD BDC2.5 TCR-transgenic mice vs NOR DC (B. Poligone and R. Tisch, unpublished results). Together, these observations demonstrate that NOD DC can stimulate both CD4+ and CD8+ T cells in an Ag-specific manner to a greater extent relative to BALB/c or NOR DC. A direct role for NF-κB in this function was demonstrated by gene transfer of IκB-SR and subsequent inhibition of the capacity of NOD DC to stimulate T cells (Fig. 3⇑ and Table II⇑). Finally, enhanced NF-κB activation and APC function by DC was found to be specifically associated with the NOD genotype.
The enhanced capacity to stimulate T cells by NOD vs BALB/c or NOR DC was independent of the level of expression of H2Kd and the costimulatory molecules CD40, CD80, and CD86. For instance, no significant difference between the respective strains of mice was observed in the cell surface phenotype of “resting” immature or “activated” mature DC (Table I⇑). Indeed, DC expressing the IκB-SR continued to up-regulate expression of costimulatory molecules (Table I⇑) despite an inability to efficiently stimulate T cells (Fig. 3⇑ and Table II⇑). Overexpression of molecules such as OX40L and 4-1BBL, which are associated with T cell activation and are directly or indirectly linked to the NF-κB signaling pathway, could partly explain the high T cell stimulatory capacity of NOD DC (38, 39). Furthermore, increased levels of activation may potentiate antiapoptotic effects associated with NF-κB (40, 41) and therefore increase the persistence and in turn the stimulatory capacity of NOD DC. However, a comparison of annexin V staining showed no significant differences in the rate or frequency of apoptosis in nonactivated or stimulated NOD and BALB/c DC (B. Poligone and R. Tisch, unpublished results), supporting earlier observations that NF-κB has minimal influence on DC survival (21).
Strikingly, neutralization of IL-12(p70) in cultures resulted in a significant reduction in T cell stimulation by NOD DC (Fig. 4⇑). This finding is consistent with recent work demonstrating that the adjuvant effect of anti-CD40 Ab treatment on DC is blocked by in vitro neutralization of IL-12(p70) secreted by the same DC (30). The study further demonstrated that the autocrine-mediated effect is IL-12 specific, because anti-CD40 Ab stimulated DC continued to elicit T cell responses despite neutralization of IL-1 or TNF-α secreted by the DC. Accordingly, the elevated levels of IL-12(p70) secreted (Fig. 7⇑B) and the increased sensitivity to IL-12(p70)-induced activation may result in enhanced CD40-mediated effects in NOD vs BALB/c and NOR DC. Signaling through CD40 and IL-12R may also contribute to the elevated levels and extended kinetics of NF-κB activation detected in stimulated NOD DC. However, the addition of various amounts of IL-12(p70) did not increase the capacity of BALB/c DC to stimulate F1.CL4 T cells in vitro (B. Poligone and R. Tisch, unpublished results), suggesting that autocrine IL-12(p70) may be necessary but not sufficient to mediate the enhanced APC function associated with NOD DC.
Our results demonstrating that expression of the IκB-SR by NOD DC inhibits IL-12(p70) secretion (16) and the ability to stimulate T cells (Fig. 3⇑ and Table II⇑) underscore the importance of NF-κB in regulating the APC function of DC. Studies have demonstrated that nonspecific inhibition of NF-κB activation using, for example, N-p-tosyl-l-lysine chloromethyl ketone or N-acetyl cysteine can prevent maturation of immature DC, which in turn have only a limited capacity to stimulate T cells (42, 43, 44). Interestingly, results obtained by specific inhibition of NF-κB via Adeno-SR suggest that NF-κB may regulate select events associated with maturation and APC function of DC. For example, NOD DC infected with Adeno-SR continued to up-regulate H2Kd, CD40, CD80, and CD86 following LPS stimulation (Table I⇑). Indeed, the process of infection with Adeno-SR alone was sufficient to partially activate NOD DC which exhibited increased H2Kd and costimulatory molecule expression (Table I⇑). Nevertheless, Adeno-SR-infected DC failed to secrete IL-12(p70) (16) and stimulate T cells, although LPS stimulation led to significant increases in H2Kd, CD40, CD80, and CD86 (Table II⇑). A recent study demonstrated that under the appropriate conditions adenovirus infection can induce complete maturation of bone marrow DC based on cell surface phenotype, detection of cytokine secretion, and an elevated allogeneic stimulatory capacity (44). This study also reported that an adenovirus recombinant encoding IκB-SR inhibited up-regulation of CD86 (and MHC class II) expression normally associated with the infection process, suggesting a role for NF-κB in this event. However, whether an increase in CD86 expression and other molecules occurred in the Adeno-SR-infected DC upon a subsequent activating stimulus, e.g., LPS, was not examined. In our system, a block in CD40-mediated signaling through NF-κB, coupled with inhibition of IL-12(p70) secretion and its subsequent autocrine effects, likely contributed to the inability of IκB-SR-expressing NOD DC to stimulate T cells.
A recent study reporting that some cell types from NOD mice, including splenocytes and Kupfer cells, have decreased NF-κB signaling due to the absence of LMP-2 expression prompted us to examine the latter in NOD DC (31). Interestingly, a 2-fold reduction in expression of LMP-2 but not LMP-7 was detected in NOD vs BALB/c DC (Fig. 5⇑). Despite the reduction in LMP-2 expression, no obvious defects in the NF-κB signaling pathway or Ag processing of HA were detected in NOD DC. For example, p50- and p52-containing NF-κB complexes were detected in nuclear extracts prepared from NOD DC (Fig. 6⇑A and data not shown, respectively), indicating that reduced expression of LMP-2 had no significant effect on processing of the p100 and p105 precursor molecules. In addition, no significant difference was observed between NOD and BALB/c DC in the proteasome-dependent degradation of IκBα (Fig. 6⇑A). Finally, virus-pulsed NOD DC were found to stimulate T cells to a greater extent compared with BALB/c DC (Fig. 2⇑B), demonstrating that LMP-2 expression was sufficient for proteasome-dependent processing of the p512–520 HA epitope (Fig. 2⇑A). However, it is possible that other proteasome subunits expressed by DC compensate for the reduced expression of LMP-2.
The marked differences between NOD and NOR DC regarding NF-κB activation (Fig. 7⇑A), cytokine secretion (Fig. 7⇑, B and C), and the capacity to stimulate T cells (Fig. 8⇑), coupled with the close similarity of the two genotypes (34), indicates that the enhanced APC function of DC is unique to the NOD genotype. Nevertheless, increased APC function by DC may be associated with other autoimmune diseases, especially in view of a growing consensus that susceptibility loci can be shared among different types of autoimmunity. The fact that NOR mice remain diabetes-free also demonstrates a correlation between the enhanced APC function of DC and the progression of IDDM in NOD mice. Interestingly, NOR mice exhibit extensive peri-insulitis, which surrounds but does not penetrate (intrainsulitis) the islets as in NOD mice. Therefore, one possibility is that NOD DC contribute to the progression of peri-insulitis to intrainsulitis. In this regard, the difference seen in the amounts of TNF-α secreted by activated NOD and NOR DC (Fig. 7⇑C) is noteworthy. Ectopic expression of TNF-α in β cells is associated with the recruitment and activation of APCs and subsequent presentation of autoantigens in the pancreas (15, 45). Furthermore, DC have been shown to be among the first cells infiltrating the pancreas (14), in addition to being an early source of TNF-α in NOD islet infiltrates (46).
Enhanced levels of IL-12 secretion by NOD vs NOR DC (Fig. 7⇑) may also contribute to the preferential development of β cell-specific Th1 (Tc1) effector cells. For example, significant β cell-specific Th1 cell reactivity is detected in NOD mice as IDDM progresses, whereas minimal if any is detected in NOR mice (47). Interestingly, we found that stimulation of F1.CL4 T cells was more readily inhibited in vitro with increasing numbers of NOD vs BALB/c DC (Fig. 1⇑B). The clonal anergy/deletion that was induced may in part account for the protection reported in NOD mice following multiple injections of relatively high numbers of syngeneic DC (48, 49). Efforts are ongoing to establish a direct link between the enhanced APC function of NOD DC and the progression of IDDM.
It is well established that IDDM is a polygenic disorder in which regulation of various aspects of immune function are disrupted (50). However, the precise nature of these defects remains largely ill defined. In this study we demonstrate that hyperactivation of NF-κB enhances the APC function of NOD DC. A number of naturally occurring defects in the NF-κB signaling pathway have been shown to be associated with cell transformation (51, 52, 53); however, our findings demonstrate that dysregulation of NF-κB activation can have significant effects on the function of nontransformed cell types such as DC. In view of their central role in initiating and regulating T cell responses, our results argue that DC are a contributing factor in the breakdown of T cell tolerance in NOD mice.
Acknowledgments
We thank Drs. Jeffrey A. Frelinger and Roland S. Liblau for supplying influenza strain PR8 and BALB/c.CL4 mice, respectively, and we thank Elaine S. Gilmore for thoughtful insight.
Footnotes
-
↵1 This work was supported in part by Grant 1-P60-DE 13079 from the National Institute of Dental and Craniofacial Research and by Grant 5RO1-DK-52365 from the National Institute of Diabetes and Digestive and Kidney Diseases. D.J.W. was supported in part by National Institute of Allergy and Infectious Diseases Training Grant 5-T32-AI07273. A.S.B. and B.P. were supported in part by Grant AI35098 from the National Institute of Allergy and Infectious Diseases.
-
↵2 Address correspondence and reprint requests to Dr. Roland Tisch, Department of Microbiology and Immunology, University of North Carolina, 635 Mary Ellen Jones Building, CB#7290, Chapel Hill, NC 27599-7290. E-mail address: rmtisch{at}med.unc.edu
-
↵3 Abbreviations used in this paper: NOD, nonobese diabetic; DC, dendritic cell; HA, hemagglutinin; hau, hemagglutination activity unit; IDDM, insulin-dependent diabetes mellitus; LMP, low molecular mass polypeptide; NOR, nonobese diabetes-resistant; Tc, T cytotoxic; moi, multiplicity of infection; SR, super-repressor; PR8, A/Puerto Rico/8/34; Adeno-SR, an adenovirus recombinant encoding IκB-SR; L,ligand.
- Received July 24, 2001.
- Accepted October 29, 2001.
- Copyright © 2002 by The American Association of Immunologists
References
In this issue
