|Home | About | Journals | Submit | Contact Us | Français|
4-1BB (CD137) is expressed on dendritic cells (DCs) and its biological function has remained largely unresolved. By comparing 4-1BB-intact (4-1BB+/+) and 4-1BB-deficient (4-1BB-/-) DCs, we found that 4-1BB was strongly induced on DCs during the maturation, and DC maturation was normal in the absence of 4-1BB. However, DC survival rate was low in the absence of 4-1BB, which was due to the decreased Bcl-2 and Bcl-XL in 4-1BB-/- DCs compared with 4-1BB+/+ DCs after DC maturation. Consistent with these results, 4-1BB-/- DCs showed increased turnover rate in steady state and more severely decreased in spleen by injecting LPS compared with 4-1BB+/+ DCs. When OVA-pulsed DCs were adoptively transferred to recipient mice along with OVA-specific CD4+ T cells, 4-1BB-/- DCs did not properly migrate to the T cell zone in lymph node and poorly induced proliferation of CD4+ T cells although both DCs comparably expressed functional CCR7. Eventually 4-1BB-/- DCs generated reduced number of OVA-specific memory CD4+ T cells compared to 4-1BB+/+ DCs. To further assess the role of 4-1BB on DC longevity in vivo, 4-1BB+/+ and 4-1BB-/- C57BL/6 were administrated with Propionibacterium acnes that develops liver granuloma by recruiting DCs. Number and size of granuloma were reduced in the absence of 4-1BB, but inflammatory cytokine level was comparable between the mice, which implied that the granuloma might be reduced due to the decreased longevity of DCs. These results demonstrate that 4-1BB on DCs controls the duration, DC-T interaction and therefore, immunogenicity.
4-1BB (CD137), an inducible T cell surface receptor, presents on a variety of cells including CD4+ and CD8+ T cells, NK cells, CD4+CD25+ regulatory T cells, neutrophils, and dendritic cells (DCs) (1-4), and provides CD28-independent costimulation of T cell activation (5). 4-1BB-mediated signaling plays a critical role in preventing activation-induced cell death (6) and increasing T cell cytolytic potential and IFN-γ production (7). The immunological role of 4-1BB on T cells is well characterized, but its role on DCs is less well understood. A few reports have demonstrated that 4-1BB triggering increases IL-6 and IL-12 production from DCs (4, 8).
DCs serve as sentinels against foreign antigens and activate naïve T cells by homing to adjacent lymph nodes (LN) (9). When DCs undergo the maturation process following antigen uptake and migration into LN, they down-modulate anti-apoptotic molecules to regulate an antigen-specific immune response (10). Therefore, mature DCs could not survive without survival signals, such as CD40 ligand, TRANCE (tumor necrosis factor-related activation-induced cytokine), inflammatory cytokines, TLRs, and costimulation from T cells (11-14). These signals activate NF-kB transcription factor proteins and increases anti-apoptotic factor Bcl-2 and Bcl-XL (15, 16).
To evaluate the role of 4-1BB in DC function, we compared DCs from 4-1BB-intact (4-1BB+/+) and 4-1BB-deficient (4-1BB-/-) mice in the survival rate, capability for CD4+ T proliferation and memory formation, localization of DCs in LN, and granuloma formation. Current studies demonstrated that 4-1BB provided DCs with survival signal which increased the life span of DCs. 4-1BB signal on DCs, therefore, promoted antigen-specific primary and memory CD4+ T responses.
Homozygous 4-1BB-/- mice (17) were backcrossed for at least 12 generations to wild-type (4-1BB+/+) BALB/c or C57BL/6 mice from The Jackson Laboratory (Bar Harbor, ME). RAG2-/- C57BL/6 mice were purchased from Taconic Farms and crossed with 4-1BB-/- C57BL/6 mice to generate 4-1BB-/-RAG2-/- mice. All mice were maintained under specific-pathogen-free conditions in the animal facility of the Immunomodulation Research Center, University of Ulsan, Ulsan, Korea, and used at 6 to 8 weeks of age. Lipopolysaccharides (LPS; Escherichia coli O55:B5) was purchased from Sigma-Aldrich (St. Louis, MO), and CD4- and CD11c-MicroBeads from Miltenyi Biotec (Auburn, CA). Cytokines including recombinant mouse TNF-α, IFN-γ, IL-12, and GM-CSF were purchased from PeproTech (Rocky Hill, NJ). Anti-mouse 4-1BB mAb-producing hybridoma cells (3E1) were a gift from Dr. Robert Mittler (Emory University, Atlanta, GA). The following antibodies were purchased from eBioscience (San Diego, CA): FITC-conjugated anti-mouse MHC class I (34-1-2S), anti-mouse MHC class II (M5/114.15.2), anti-CD80 (16-10A1), anti-CD86 (GL1), anti-CD3e (145-2C11), anti-CD4 (GK1.5) mAb, and annexin V: phycoerythrin (PE)-conjugated anti-CD11c (N418), anti-B220 (RA3-6B2), KJ1.26 mAb: and phycoerythrin-cychrome5 (PE-Cy5)-conjugated anti-CD8 (53-6.7) mAb.
Spleens of 4-1BB+/+ or 4-1BB-/- mice in BALB/c or C57BL/6 background were dissected, cut into pieces, and digested with collagenase/DNase I for 30 min at 37°C. Following the hypotonic lysis of red blood cells, splenocytes were incubated with CD11c-microbeads for 30 min on ice and CD11c+ cells were purified by loading the cells over a MACS column. To induce maturation, purified CD11c+ DCs were cultured with several reagents including 5 μg/ml LPS, 100 U/ml TNF-α, 150 ng/ml IFN-β, 20 ng/ml IL-12, or 20 ng/ml GM-CSF for 36 h.
To determine 4-1BB expression on CD11c+ DCs, splenocytes were isolated from normal BALB/c mice and stimulated with 5 μg/ml of LPS for 36 h. Freshly isolated or LPS-treated DCs were pre-incubated with Fc blocker 2.4G2 for 10 min at 4°C and stained with anti-4-1BB-FITC and anti-CD11c-PE for 30 min. In a separate experiment, mature DCs were stained with anti-CD11c-PE and anti-CD8a-PE-Cy5 as well as anti-4-1BB-, anti-MHC class I-, anti-MHC class II-, anti-CD80-, or anti-CD86-FITC. All samples were subsequently analyzed on FACScalibur (Becton Dickinson, San Jose, CA).
The purified DCs were cultured in 96-well plates at 1 × 106 cells/ml density and treated with 5 μg/ml LPS. The survival rate of the DCs was determined by measuring cell viability using propidium iodide staining.
CD11c+ DCs from 4-1BB+/+ and 4-1BB-/- mice were cultured in 24-well plates and treated with 5 μg/ml LPS as well as 5 μg/ml anti-4-1BB mAb or rat IgG for 36 h. Proteins were extracted with lysis buffer (10 mM Tris-HCl (pH 7.4), 50 mM NaCl, 5 mM EDTA, 30 mM NaF, 0.1 mM Na3VO4, 1% Triton X-100, 0.5% Nonidet P-40, 1 mM PMSF, and protease inhibitor mixture). Equal amounts of protein from each sample were diluted with 4× SDS sample buffer, applied to SDS-PAGE gels, separated, and transferred to nitrocellulose membranes (Millipore, Bedford, MA). Bcl-2 and Bcl-XL were detected with anti-Bcl-2 or anti- Bcl-XL Ab (Santa Cruz Biotechnology, Santa Cruz, CA) and secondary Ab-HRP. Bound Abs were detected by ECL (Amersham Pharmacia Biotech, Little Chalfont, UK).
4-1BB+/+ and 4-1BB-/- BALB/c mice received 10μg LPS via intravenous (i.v.) route and splenocytes were collected from the mice 0, 24, 48, and 96hrs after LPS injection. Splenocytes were first stained with PE-conjugated anti-CD11c and PE-cy5-conjugated anti-CD8 mAb following the incubation with 2.4G2 Fc blocker and then intracellularly stained with FITC-conjugated anti-Bcl-2 mAb (BD Bioscience; San Diego, CA) or anti-Bcl-XL Ab (Cell Signaling Technology Beverly, MA) and FITC-conjugated anti-rabbit Ab according to the manufacturer's instruction. All samples were analyzed on FACScalibur (BD Bioscience).
CD11c+ DCs were purified from 4-1BB+/+ and 4-1BB-/- mice 0, 24, 48hrs after LPS injection and total RNA was extracted from the DCs by using TRIzol (Invitrogen Life Technologies) and reverse-transcribed into cDNA using a SuperScriptII™ (Invitrogen Life Technologies). SYBR Green I-based real-time quantitative PCR was carried out on a continuous fluorescence detection system (Opticon DNA Engine, MJ Research Inc., Waltham, MA) with first-strand cDNA. Gene-specific primers used were as follows: Bcl-2, forward 5'-tgggatgcctttgtggaactat-3', reverse 5'-agagacagccaggagaaatcaaac-3'; Bcl-XL, forward 5'-ctgggacacttttgtggatctct-3', reverse 5'-gaagcgctcctggccttt-3'; GAPDH, forward 5'-gaacgggaagcttgtcatcaa-3', reverse 5'-ctaagcagttggtggtgcag-3'. The relative gene expression was determined using the comparative CT method. To simplify the representation of Bcl-2 and Bcl-XL expression, the gene expression data normalized for GAPDH are shown as fold increases compared to levels in 0hr sample of 4-1BB+/+ DCs and are means ± SD of triplicate experiments.
4-1BB+/+, 4-1BB-/-, RAG2-/-, RAG2-/-4-1BB-/- C57BL/6 mice were initially injected i.p. with 1mg of BrdU (Sigma, St. Louis, MO) and then continuously given BrdU (0.8mg/ml) in sterile drinking water. Three days after the BrdU injection, spleens were digested with collagenase-DNase I for 20 min at 37°C CO2 incubator and splenocytes were then surface stained with anti-CD11c-PE and anti-CD8-PE-Cy5 following Fc blocking with 2.4G2 for 5 min. Incorporation of BrdU was assessed by intracellularly staining the surface-stained splenocytes with FITC BrdU Flow kit (BD Pharmingen).
To follow the migration of CD11c+ DCs in draining lymph nodes (DLNs) after footpad injection, DCs were labeled with 10 μM carboxyfluoroscein succinimidyl ester (CFSE) for 8 min at 37°C according to the manufacturer's protocol (Molecular Probes). CFSE-labeled DCs were injected in both hind footpads (1 × 106 cells per footpad) and popliteal LNs (PLN) were collected at various time points. For flow cytometry, PLNs were digested with collagenase/DNase I for 30 min at 37°C and total LN cells were counted and stained with anti-CD11c-PE for 30 min at 4°C. By using FACScalibur, 5 × 105 events were acquired to quantify the absolute number of CFSE-labeled DCs per LN. For confocal microscopy, PLNs were embedded in optimum cutting temperature (OCT) compound (Sakura Finetek, Torrance, CA). Next, 6-μm-thick frozen sections were prepared and stained with antibodies. LN structure was visualized by staining the sections with anti-ERTR7 mAb (Acris Antibodies, Hiddenhausen, Germany) and secondary Ab-PE. To analyze the co-localization of DCs with T or B cells, the sections were stained with anti-CD3-FITC and anti-B220-PE. All samples were visualized with a confocal laser scanning microscope (FV500, Olympus, Center Valley, PA).
Spleens from 4-1BB+/+ and 4-1BB-/- mice were digested with collagenase-DNase and splenic DCs were isolated by incubating splenocytes with CD11c-microbeads (Miltenyi Biotec). CD11c+ DCs were stimulated with 5μg/ml of LPS for 24hrs to induce maturation of DCs and CCR7 expression. The LPS-stimulated DCs were stained with FITC-conjugated anti-CD11c and PE-conjugated anti-CCR7 mAb to assess the expression level of CCR7 on mature DCs. Mature DCs (5×105 cells/transwell; 3.0μm 24-well transwell; Costar Corning, Cambridge, MA) were allowed to migrate for 1.5hrs in response to 0, 10, or 100ng/ml of CCL19 or CCL21. The DCs that migrated to the lower chamber were counted, and CCL19- or CCL21-dependent migration was calculated as the ratio of migrating cells to starting cells.
Purified DCs from 4-1BB+/+ and 4-1BB-/- BALB/c mice were pulsed with OVA peptide (323-339, ISQAVHAAHAEINEAGR, 50 μg/ml) for 24 h in the presence of 5μg of LPS. Live DCs were enriched up to 80% by centrifuging the cultured DCs in low speed (x 170g) three times and then fixed with 0.5% paraformaldehyde. To prepare OVA-specific CD4+ T cells, CD4+ T cells were purified from DO11.10 transgenic (Tg) mice using CD4-microbeads. Naïve OVA-specific CD4+ T cells were first added at a density of 2 × 105 cells/well in 96-well round plates and were mixed with 1, 5, or 10% live or fixed DCs. The cells were cultured for 72 h and labeled with 1 mCi of [3H]-thymidine for the last 8 h. Cellular DNA was harvested and counted by liquid scintillation spectroscopy.
OVA-specific CD4+ T cells were purified from DO11.10 Tg mice using CD4-microbeads, labeled with 10 μM CFSE for 8 min at 37°C, washed, and i.v. transferred into 4-1BB+/+ mice at a density of 5 × 106 per mouse. After 24 h, the purified DCs from 4-1BB+/+ and 4-1BB-/- BALB/c mice were pulsed with 50 μg/ml of OVA peptide for 2 h, washed with PBS, and injected into the footpads of 4-1BB+/+ mice at a density of 1 × 106 cells per footpad that had received the OVA-specific CD4+ T cells. PLN cells were collected from each group of mice 4 days after the DC transfer and stained with KJ1.26-PE mAb. In a separate experiment, OVA-pulsed DCs from 4-1BB+/+ and 4-1BB-/-BALB/c mice were injected into the footpads of 4-1BB+/+ mice that received OVA-specific CD4+ T cells as described above. Lymphocytes were collected from PLNs and spleen on PI day 40 and stained with anti-CD4-FITC and KJ1.26-PE. The absolute number of CD4+KJ1.26+ cells was calculated by multiplying the percentage measured by flow cytometry by the total numbers of viable cells recovered from the same culture.
4-1BB+/+ and 4-1BB-/- C57BL/6 mice were injected i.p. with 0.5 mg of heat-killed P. acnes (Korean Type Culture Collection). Frozen sections (5 μm) were prepared from the livers, embedded in OCT compound (Sakura Finetek), fixed with formalin, stained with hematoxylin and eosin (H&E), and observed under a light microscope. To detect CD11c+ DCs in the granuloma, the frozen sections were preincubated with 2.4G2 Fc blocker for 10 min and further incubated with anti-CD11c-PE mAb (N418). Samples were mounted with Fluoromount-G (Southern Biotechnology, Birmingham, AL) and visualized with an confocal laser scanning microscope (Olympus FV500).
Six days after 4-1BB+/+ and 4-1BB-/- C57BL/6 mice received injections of 0.5 mg heat-killed P. acnes, 1 μg LPS was injected to induce endotoxin shock and serum was collected from each mouse 2 h later. The serum cytokines were quantified using a cytometric bead array kit (CBA; Mouse Inflammation Kit, BD Bioscience) on a FACSCaliber cytometer equipped with CellQuestPro and CBA software according to the manufacturer's instructions. Aspartate aminotransferase (GOT) and alanine aminotransferase (GPT) activity assays were performed using specific kits and reagents (Asan Pharm, Seoul, Korea), according to the manufacturer's instructions.
We first examined 4-1BB expression on immature and mature splenic DCs. 4-1BB was barely expressed on freshly isolated DCs, but was readily detected on LPS-stimulated DCs (Fig. 1A). Although this result is not consistent with previous reports that showed expression of 4-1BB on immature DCs (8), under our experimental conditions only CD4+CD25+ T cells expressed detectable levels of 4-1BB in naïve mice, but not CD4+ and CD8+ T cells, monocytes, B220+ B cells, or DCs (data not shown).
Because 4-1BB expression on DCs was induced by treating with LPS, we sought to determine whether the induction of 4-1BB was dependent on DC maturation itself or required specific factors such as cytokines. We examined the induction of 4-1BB on DCs by treating with several stimulators, including LPS, TNF-α, IFN-γ, IL-12, and GM-CSF, for 36 h. To evaluate 4-1BB expression on CD8- and CD8+ DCs, the cells were stained with anti-4-1BB-FITC, anti-CD11c-PE, and anti-CD8a-PE-Cy5. CD11c+ DCs were gated and analyzed for their expression of CD8 and 4-1BB. Both CD8+ and CD8- DCs expressed similar levels of 4-1BB at any condition (Fig. 1B).
We next tested the involvement of 4-1BB in antigen presentation by examining antigen presentation-related molecules such as MHC I, MHC II, CD80, and CD86. Purified splenic DCs were cultured with TNF-α or LPS in the presence of anti-4-1BB mAb or rat IgG for 36 h. CD11c+ DCs were gated and analyzed for their surface expression of CD8 as well as MHC I, MHC II, CD80, and CD86. 4-1BB triggering did not significantly alter the expression of the molecules that were involved in the functions of antigen presenting cells (APCs) (Fig. 1C). DCs were also treated with higher doses of anti-4-1BB (10 μg/ml or 50 μg/ml) in the presence of LPS, but the results were similar to those of 5 μg/ml anti-4-1BB mAb treatment (data not shown). Therefore, we concluded that DC maturation itself led to 4-1BB induction, but the 4-1BB signaling appeared not to be crucial for DC maturation or the induction of MHC I and II, CD80, and CD86.
Since 4-1BB triggering did not significantly alter the expression of DC surface molecules that are involved in antigen presentation, we next tested whether 4-1BB signaling was involved in survival of DCs. We compared the survival rate of DCs between 4-1BB+/+ and 4-1BB-/- BALB/c mice in the presence or absence of agonistic anti-4-1BB mAb. CD11c+ DCs were purified from 4-1BB+/+ and 4-1BB-/- mice and stimulated with LPS in the presence of anti-4-1BB mAb or rat IgG as a control. DC survival rates were determined by assessing the viability of DCs with propidium iodide staining on the indicated day of culture (Fig. 2A). The kinetics of DC survival rate indicated that the half life of 4-1BB+/+ DCs was ~36 h, which increased to > 48 h with anti-4-1BB mAb treatment, whereas 4-1BB-/- DCs had a half life of about 24 h (Fig. 2A). Because the reduced survival rate of 4-1BB-/- DCs could be due to the impaired maturation, 4-1BB+/+ and 4-1BB-/- DCs were analyzed for the expression of MHC class I/II, CD80, and CD86 following LPS treatment for 36hr. The two groups of DCs showed comparable levels of expression of those molecules (Fig. 2B).
Since Bcl-2 and Bcl-XL are crucial in maintaining the survival of DCs (10, 18), we first measured the levels of Bcl-2 and Bcl-XL proteins in 4-1BB+/+ and 4-1BB-/- DCs in vitro. CD11c+ DCs were stimulated with 5 μg/ml LPS in the presence or absence of anti-4-1BB mAb for 36 h, and equal amounts of cell lysates were applied to western blotting of Bcl-2 and Bcl-XL. Both anti-apoptotic proteins were detected in 4-1BB+/+ DCs, and their levels were slightly increased by treatment with anti-4-1BB mAb. However, 4-1BB-/- DCs barely expressed Bcl-2 and Bcl-XL (Fig. 2C). To further characterize Bcl-2 and Bcl-XL expression in DCs in vivo, mice were i.v. injected with 10 μg LPS to induce the maturation of DCs in vivo and the expression of Bcl-2 and Bcl-XL was assessed by intracellularly staining splenocytes with specific antibodies on the indicated days. First, CD11c+ DCs increased an expression of MHC II, CD80, and CD86 on 24hrs after the LPS injection, and then, the induced molecules were declined thereafter. DC maturation appeared to be comparable between both DCs because both 4-1BB+/+ and 4-1BB-/- DCs similarly expressed those molecules (Fig. 2D). 4-1BB+/+ DCs expressed relatively high level of intracellular Bcl-2 that was transiently decreased following LPS injection as previously reported (10). However, 4-1BB-/- DCs barely expressed Bcl-2 even in the immature status, which was lower than that of 4-1BB+/+ DCs at any maturation status. Bcl-XL was similarly expressed in both immature DCs, and which was transiently decreased by LPS stimulation, but Bcl-XL expression was lower in 4-1BB-/- DCs than 4-1BB+/+ DCs at 24hrs after LPS treatment (Fig. 2D). As an internal control, Bcl-2 and Bcl-XL expression in CD8+ T cells were assessed by flow cytometry because Bcl-2 expression in DCs was relatively weak. Unlike our expectation, Bcl-2 and Bcl-XL were higher in 4-1BB-/- CD8+ T cells compared to that of 4-1BB+/+ CD8+ T cells, which appeared to be due to the hyperresponsiveness of 4-1BB-/- CD8+ T cells in vivo (17, 19). Since the Bcl-2 and Bcl-XL expression in DCs and CD8+ T cells were determined in same samples, these results clearly demonstrated that 4-1BB-/- DCs expressed significantly lower Bcl-2 levels than 4-1BB+/+ DCs in vivo.
In a separate experiment, we assessed Bcl-2 and Bcl-XL transcripts in 4-1BB+/+ and 4-1BB-/- DCs to further confirm the results of flow cytometry. CD11c+ DCs were purified from 4-1BB+/+ and 4-1BB-/- mice 0, 24, and 48 hrs after LPS injection. Complementary DNA was synthesized with total RNA from the DCs and SYBR Green I-based real-time semi-quantitative PCR was carried out with the cDNA and gene-specific primers. Level of Bcl-2 and Bcl-XL was comparable between 4-1BB+/+ and 4-1BB-/- DCs before the LPS injection (Fig. 2E). Twenty-four hrs after LPS injection, both Bcl-2 and Bcl-XL were rapidly decreased in DCs, but the expression levels of both anti-apoptotic genes were lower in 4-1BB-/- DCs compared with that of 4-1BB+/+ DCs. The decreased anti-apoptotic genes were gradually recovered 48hrs after LPS injection. Consistent with the results of intracellular Bcl-2 and Bcl-XL staining, both anti-apoptotic gene transcripts were decreased in mature DCs following LPS injection, and the expression levels were lower in 4-1BB-/- DCs compared with that of 4-1BB+/+ DCs.
Taken together, these results indicate that 4-1BB-/- DCs underwent normal maturation, but survived poorly after maturation that might be due to the low expression of anti-apoptotic molecules. Therefore, we concluded that 4-1BB signaling was not essential for DC maturation, but was required to sustain the survival of mature DCs by increasing the expression of anti-apoptotic molecules.
Since DCs showed the shorter survival and the reduced anti-apoptotic molecules in the absence of 4-1BB signaling, it was reasonable to expect that DC turnover would be higher in 4-1BB-/- mice. To assess DC turnover rate, 4-1BB+/+ and 4-1BB-/- mice were continuously given BruU for 3 days and BrdU incorporation was examined by flow cytometry. CD11c-gated DCs were analyzed for CD8 expression and BrdU incorporation. DC turnover was higher in both CD8+ and CD8- DCs of 4-1BB-/- mice than 4-1BB+/+ mice (Fig. 3A, upper panel). We also examined the DC turnover by using 4-1BB+/+ and 4-1BB-/- mice in RAG2-/- background, which were lack of T/B lymphocytes and thus, the indirect effects of hyper-responsiveness of 4-1BB-/- mice could be excluded in DC turnover. Consistent with the results of WT background mice, BrdU incorporation rate was 49.6 ± 5.7% in 4-1BB+/+RAG2-/- mice and 65.6 ± 8.3% in 4-1BB-/-RAG2-/- mice (Fig. 3A, lower panel). These results indicate that DC turnover was higher in 4-1BB-/- mice compared with 4-1BB+/+ mice.
We have determined the percentages and absolute numbers of CD11c+ DCs in 4-1BB+/+ and 4-1BB-/- mice to assess the differences in the DC numbers. The results indicated that the frequency and numbers of DCs were higher in 4-1BB-/- mice (Fig. 3B). The results appear to reflect previous findings that 4-1BB-/- mice do not have a 4-1BBL-mediated suppression signal on DC development and consequently contain more DCs compared with 4-1BB+/+ mice (17, 19). 4-1BB is induced when the DC received the maturation signal (Fig. 1A). We, therefore, hypothesize that 4-1BB functions as a DC survival factor when the DCs undergo the maturation. We, therefore, compared the DC numbers between 4-1BB-/- and 4-1BB+/+ mice after i.v. injection of LPS which induced maturation and turnover of DCs (20). The number of DCs was decreased much more severely in 4-1BB-/- compared with 4-1BB+/+ mice; 41% in 4-1BB+/+ mice vs. 74% in 4-1BB-/- mice. The ratio of DC was severely decreased in 4-1BB-/- mice similar to the decrease in number; 38% in 4-1BB+/+ mice vs. 75% in 4-1BB-/- mice (Fig. 3B). Similar result was found in 4-1BB-/- mice in RAG2-/- background, 4-1BB-/-RAG2-/- mice retained more DCs in secondary lymphoid organs and more severely decreased the number of DCs by injecting LPS compared with 4-1BB+/+RAG2-/- mice. Moreover, it was notable that the recovery of DCs from LPS-induced apoptosis was delayed more than 2 days in 4-1BB-/-RAG2-/- mice compared with 4-1BB+/+RAG2-/- mice (Fig. 3C).
To further confirm whether DCs were more susceptible to the maturation-induced apoptosis in the absence of 4-1BB signaling, splenocytes were collected from 4-1BB+/+ and 4-1BB-/- mice 6hr or 12hrs after the LPS injection, and stained with anti-CD11c-PE and anti-CD8a-PE-Cy5 as well as FITC-conjugated annexin V. CD11c+ DCs were gated and analyzed for the expression of CD8 and the binding of annexin V. Apoptosis rates were 11.9 ± 3.21% and 16.9 ± 2.67% 6hr after LPS injection, 25.7 ± 4.13% and 34.8 ± 3.27% 6hr after LPS injection in 4-1BB+/+ mice and 4-1BB-/- mice, respectively (Fig. 3D). The apoptosis rate was also determined 24hrs after the LPS injection, but the result was not conclusive due to the severe decrease of DC frequency.
Taken together, these results indicated that 4-1BB signaling was protecting the DCs from the maturation-mediated apoptosis.
In that DCs showed a reduced survival rate in the absence of 4-1BB signaling in vitro, we sought to determine whether 4-1BB signaling is required to regulate the longevity of DCs in vivo. CD11c+ DCs were isolated from 4-1BB+/+ or 4-1BB-/- BALB/c mice, pulsed with OVA, labeled with CFSE, and injected into the footpads of 4-1BB+/+ or 4-1BB-/- mice at 1 × 106 DCs per footpad. PLNs were isolated from four groups of mice every day for 4 days and the absolute numbers of CFSE+CD11c+ DCs were determined by staining PLN cells with anti-CD11c-PE (Fig. 4A). 4-1BB+/+ DCs transferred into 4-1BB+/+ or 4-1BB-/-recipient were readily detected in the PLN 24 h after the transfer, peaked at 2 days post-injection (PI), and then declined steadily. The number of 4-1BB-/- DCs in PLN was half that of 4-1BB+/+ DCs, but the migration pattern was similar, with 4-1BB+/+ DCs also peaking on PI day 2 and declining steadily thereafter (Fig. 4B)
We also examined the localization of 4-1BB+/+ and 4-1BB-/- DCs in PLNs. CFSE-labeled DCs were injected into the footpads of 4-1BB+/+ BALB/c mice as described above and frozen sections were prepared from PLN of each group of mice 1 and 2 days later. On day 1, many DCs were located at the outer region of the cortex in PLN and there were no noticeable differences in distribution patterns between 4-1BB+/+ and 4-1BB-/-DCs (Fig. 4C). On day 2, 4-1BB+/+ DCs had moved into the inner region of the cortex, which might be the T cell zone, whereas markedly fewer 4-1BB-/- DCs had moved to the inner region of the cortex and many were still localized in the outer region.
It was also possible that the decreased migration of DCs to T zone could be due to a defect in DC trafficking to DLN. We examined whether 4-1BB-/- DCs had a defect in migration toward CCL19 and CCL21 that were the ligand of CCR7 and pivotal for DC migration in vivo (21). Purified CD11c+ DCs from 4-1BB+/+ and 4-1BB-/- mice were stimulated with LPS for 24hrs and their CCR7 expression was determined by flow cytometry. CCR7 expression was comparable between 4-1BB+/+ and 4-1BB-/- mature DCs (Fig. 4D). In vitro transwell assay indicated that mature 4-1BB-/- DCs have little defect in trafficking properties toward CCL19 and CCL21 (Fig. 4E). 4-1BB deficiency, however, could cause other migratory defects yet to be determined.
To further confirm the location of 4-1BB+/+ and 4-1BB-/- DCs in the lymph node substructure, on day 2 PLN sections were stained with ER-TR7 mAb, which visualized the reticular network of LN by detecting reticular fibroblasts (22). Consistent with Fig. 4C, most of the 4-1BB+/+ DCs were localized in the T cell zone rather than the cortical ridge (CR) or paracortical cord (PCC) (Fig. 5A) (23), while many of the 4-1BB-/- DCs were still localized around the CR or PCC (Fig. 5B). Dual staining of PLN with anti-CD3 and anti-B220 mAb again showed that 4-1BB+/+ DCs were localized inside the T cell zone, but many of the 4-1BB-/- DCs were localized around the interface of the T and B cell zones (Fig. 5C and D).
Therefore, it seems that 4-1BB signaling increases the longevity of mature DCs in DLN by increasing DC survival.
We tested whether presence or absence of 4-1BB influences the priming of DC. Since the fixed immature DCs did not express 4-1BB on their surface and would poorly induce T cell proliferation in vitro, 4-1BB+/+ and 4-1BB-/- DCs were stimulated with OVA peptide and LPS for 24 hrs to allow DCs to fully process and display the engulfed antigens during the maturation. After 24hr culture, live mature DCs were enriched by removing dead cells by low speed centrifugation such that both 4-1BB+/+ and 4-1BB-/- DC preparations contained approximately 80% of live DCs. Some of the peptide-pulsed DCs were fixed with paraformaldehyde. Live or fixed DCs were co-cultured with OVA-specific CD4+ T cells from DO11.10 mice at different ratios. Although 4-1BB-/- DCs showed a reduced survival rate, compared with 4-1BB+/+ DCs (Fig. 2A), there were no statistical differences in T cell proliferation with 4-1BB+/+ or 4-1BB-/- DCs whether they were fixed or not (Fig. 6A). However, there was a tendency of less T cell proliferation only when we used live 4-1BB-/- DCs in higher ratio.
To test CD4+ T cell priming by 4-1BB+/+ and 4-1BB-/- DCs in vivo, the DCs were pulsed with OVA peptide and injected into the footpads of BALB/c mice that had received CFSE-labeled CD4+ T cells from DO11.10 mice. PLN cells were collected from each group of mice 4 days after DC transfer and the dilution rate of CFSE in the transferred CD4+ T cells was analyzed. Both 4-1BB+/+ and 4-1BB-/- DCs clearly induced the division of OVA-specific CD4+ T cells (Fig. 6B). However, 4-1BB+/+ DCs were more efficient in inducing clonal expansion of CD4+ T cells compared with 4-1BB-/- DCs; 84.5% ± 4.8 % of CD4+ T cells were induced to divide more than 5 times by 4-1BB+/+ DCs whereas only 63% ± 3.1 % of CD4+ T cells underwent similar rates of division induced by 4-1BB-/- DCs. The ratio of each division also showed that 4-1BB+/+ DCs induced clonal expansion of CD4+ T cells more efficiently than 4-1BB-/- DCs (Fig. 6C).
Next, we examined whether 4-1BB-/- DCs were defective in memory formation in CD4+ T cells due to their reduced ability to expand antigen-specific CD4+ T cells. Following the adoptive transfer of OVA-specific CD4+ T cells as well as OVA-pulsed 4-1BB+/+ and 4-1BB-/- DCs as described above, the number of KJ1.26+CD4+ T cells was calculated in PLN and spleens on PI day 40. The number of memory CD4+ T cells was reduced by ~20 % in the PLN and by ~50% in the spleens of 4-1BB-/- DC-transferred mice compared to 4-1BB+/+ DC-transferred mice (Fig. 6D). Therefore, we concluded that the absence of 4-1BB on DCs decreased the clonal expansion of antigen-specific CD4+ T cells, and which led to the reduction in memory CD4+ T cells.
Here we found the significant reduction of T cell proliferation in vivo by 4-1BB-/-DCs compared with 4-1BB+/+ DCs, but not in vitro. We reasoned that the dead DCs, which were not removed by phagocyte system as efficiently as in vivo (24), seemed to induce the T cell proliferation in vitro as shown in Fig. 6A. This implied that in vitro T cell proliferation assay could not reliably reflect the difference of DC survival rate.
To further test the function of 4-1BB on DCs in vivo, 4-1BB+/+ and 4-1BB-/- C57BL/6 mice were injected with 0.5 mg of heat-killed P. acnes to induce granulomas that develop by recruiting circulating DC precursors into liver (25). Six days after the P. acnes injection, frozen sections were prepared from liver and stained with H&E. P. acnes induced granulomas in the livers of both 4-1BB+/+ and 4-1BB-/- mice, but the number of granulomas was reduced in 4-1BB-/- mice compared with 4-1BB+/+ mice (Fig. 7A and B) and the granulomas in the 4-1BB-/- mice were also smaller than those in the 4-1BB+/+ mice (Fig. 7C). As expected, the liver granulomas were composed of aggregates of CD11c+ DCs, which were readily detected in both 4-1BB+/+ and 4-1BB-/- mice under the confocal microscope (Fig. 7D). Similarly, when the P. acnes-primed mice received 1 μg of LPS to induce cytokine shock on PI day 6, the levels of serum GOT and GPT, an index for hepatocyte damage, was lower in the 4-1BB-/- mice than in the 4-1BB+/+ mice (GOP: 379.5±21.27 vs. 269.6±34.36, and GPT: 419.5 ± 18.32 vs. 321.7 ± 23.13) (Fig. 7E). Nevertheless, serum levels of proinflammatory factors such as IL-6, MCP-1, IFN-γ, and TNF-α were comparable in both kinds of mice as determined 2 h after the LPS injection (Fig. 7F).
Since the cytokine production was comparable between 4-1BB+/+ and 4-1BB-/-mice following LPS injection, it seemed likely that both DCs were appropriately primed by P. acnes. However, because many 4-1BB-/- DCs died on route to liver and were unable to survive in liver, the liver granuloma was reduced in 4-1BB-/- mice compared to 4-1BB+/+ mice. Consequently, we believe that the decreased survival rate of the DCs was the primary contribution to the reduction of granulomas in liver in the 4-1BB-/- mice.
In the present study, we examined the functional role of 4-1BB as a survival factor for DCs. 4-1BB has been intensively investigated as a costimulatory molecule for T cells and NK cells, but the function of 4-1BB on DCs remained to be elucidated. Activation of DCs with several stimuli up-regulated 4-1BB on the surface of DCs (Fig. 1A and B) and the expression level was comparable with that of CD4+ and CD8+ T cells (data not shown). 4-1BB induction was not dependent on the stimuli for DC activation and seemed to be included in the programmed maturation of DC. Signaling through the 4-1BB receptor did not influence the expression of antigen-presentation-related molecules such as MHC class I/II and CD80/86 (Fig. 1 and and2B).2B). However, 4-1BB-/- DCs showed reduced survival, which might be due to decreased levels of Bcl-2 and Bcl-XL (Fig. 2A and C-E). In particular, 4-1BB-/- mice showed shorter DC survival compared with 4-1BB+/+ mice in vivo and enhanced turnover in steady state, which was more evident in the absence of T/B lymphocytes (Fig. 3). Although the 4-1BB deficiency on DCs just led to a decrease in DC survival in vitro and in vivo (Fig. 2 and and3),3), the series of events in vivo was not as simple: the 4-1BB deficiency on DCs in vivo decreased the clonal expansion of antigen-specific CD4+ T cells (Fig. 6B and C), the formation of memory T cells (Fig. 6D), and granuloma formation in liver by P. acnes (Fig. 7A).
Consistent with these results, when CFSE-labeled DCs were transferred to 4-1BB+/+ recipient mice, the numbers of CFSE+ DCs in DLNs were lower in the mice that had received 4-1BB-/- DCs compared with those receiving 4-1BB+/+ DCs (Fig. 4B). Moreover, histological analysis of the DLN showed that 4-1BB-/- DCs did not move normally to the inner cortical region (T cell zone) (Fig. 5). Nevertheless, both 4-1BB+/+ and 4-1BB-/- DCs were properly matured by several stimuli (Fig. 2B) and induced comparable levels of CD4+ T cell proliferation in vitro (Fig. 6A). Therefore, we interpreted these results in the following way: 4-1BB-/- DCs matured normally and migrated to the DLN although some of these DCs might have undergone apoptosis during the maturation and migration process. In the DLN, the DCs induced the proliferation of antigen-specific CD4+ T cells by encountering them in the outer cortical region. After contact with T cells, it appeared that many of 4-1BB-/- DCs were not able to survive until they reached the inner cortical region due to the absence of the survival signal from the 4-1BB receptor; additionally new DCs migrated to the outer cortical region of the DLN. Therefore, many 4-1BB-/- DCs were still found in the non-T cell zone. In contrast, 4-1BB+/+ DCs formed clusters with antigen-specific T cells, migrated to the inner cortical region, and persisted for a long time, inducing a clonal expansion of T cells (26, 27) (Fig. 6B). However, there was still possibility that 4-1BB-/- DCs had defects in homing or migrating abilities, which could result in a decreased DC migration and defective proliferation of T cells in DLNs. In vitro transwell assay indicated that mature 4-1BB-/-DCs have little defect in trafficking properties toward CCL19 and CCL21 (Fig. 4E). However, there are still possibilities that the 4-1BB deficiency cause the defect of DC migration through CCL19- or CCL21-independent mechanisms. Therefore, the decreased migration of 4-1BB-/- DCs appeared to be due to a defect in DC survival and/or other mechanisms yet to be determined. Since the inhibition of apoptosis in DCs led to spontaneous T cell proliferation (28), it is reasonable to think that defective DC survival due to 4-1BB deficiency contributed to the reduction in T cell responses. Because DCs were known to interact with T cells and DCs in DLN (29), APCs and T cells expressing 4-1BB ligand seemed to provide the signals to 4-1BB on DCs (4, 30).
Granuloma formation by P. acnes indicated similar results. The granulomas developed in response to P. acnes were reduced in number and size in the livers of 4-1BB-/- mice compared to 4-1BB+/+ mice (Fig. 7A-C), although cytokine production was comparable between the two groups (Fig. 7F). Therefore, it seemed that P. acnes similarly primed both 4-1BB+/+ and 4-1BB-/- DCs, but due to the reduced survival of 4-1BB-/- DCs the number and size of liver granulomas seemed to be decreased in 4-1BB-/-mice.
For a decade, 4-1BB has been investigated as a costimulatory molecule for T cells and used in the form of agonistic anti-4-1BB mAb or recombinant 4-1BB ligand protein to enhance immune responses against viruses and tumors, which eventually increased memory T cells (31, 32). Here we showed that 4-1BB signaling also functioned as DC survival signal, particularly for mature DCs. However, it was not clear which type of cells were the source of 4-1BBL for 4-1BB triggering on DCs. 4-1BBL expression has been detected on activated T cells, B cells, DCs, macrophages, lymphomas, and carcinoma lines of epithelial origin (1, 33, 34). Therefore, we expected that there would be three possible sources of 4-1BBL for 4-1BB triggering on DCs; activated T cells, residential DCs in LN, and epithelial cells in blood vessel. It might be most natural to expect that activated T cells provided 4-1BB signaling to DCs by forming T-DC clusters in LN. However, it is still controversial whether activated T cells express functional 4-1BBL. Residential DCs in LN might also provide 4-1BB signal on DCs by DC-DC interaction as previously described (35). The T-DC or DC-DC interaction should occur in the DLN. However, as shown in Fig. 4B, 4-1BB-/- DCs were poorly migrated to DLN, which implied that Ag-pulsed DCs were matured during the migration and some of mature DCs were died before they arrived at LN. Therefore, 4-1BB signaling on mature DCs might be also needed during the migration. Since activated endothelial cells (EC) in blood vessel express various costimulatory molecules (36), it might be possible that ECs provide the 4-1BB signaling on DCs during LN homing in an inflammatory condition. Taken together, 4-1BB signaling was needed for the survival of mature DCs that regulated the immunogenicity of antigen uptaked by DCs, and could be provided by the interaction of DCs with non-hematopoietic cells such as ECs.
GRANT SUPPORT: This work was supported in part by grants from the National Cancer Center, Korea (NCC-0810720-1); Korean Research Foundation (KRF-2005-201-E00008, KRF-2005-084-E00001 to BSK); Arthritis Foundation (Innovative Research Award to BSK); and NIH Grant R01EY013325 (BSK).