We previously reported the characterization of a MHC class IIlowCD4CD103+ (CD4) subset of dendritic cells (DC) in rat spleen that exhibit a Ca2+-, Fas ligand-, TRAIL- and TNF-α-independent cytotoxic activity against specific targets in vitro. In this study, we demonstrate that this DC subset was also found in lymph nodes. Freshly extracted and, therefore, immature CD4 DC exhibited a potent cytotoxic activity against a large panel of tumor cell lines as well as primary endothelial cells. The cytotoxic activity of immature CD4 DC required cell-to-cell contact and de novo protein expression. CD4 DC-mediated cell death resembled apoptosis, as evidenced by outer membrane phosphatidylserine exposure and nuclear fragmentation in target cells, but was caspase as well as Fas-associated death domain and receptor-interacting protein independent. Bcl-2 overexpression in target cells did not protect them against DC-mediated cell death. Immature CD4 DC phagocytosed efficiently apoptotic cells in vitro and, therefore, rapidly and specifically engulfed their victims following death induction. Maturation induced a dramatic down-regulation of the killing and phagocytic activities of CD4 DC. In contrast, CD4+ DC were both unable to kill target cells and to phagocytose apoptotic cells in vitro. Taken together, these data indicate that rat immature CD4CD103+ DC mediate an unusual cytotoxic activity and can use this function to efficiently acquire Ag from live cells.

Dendritic cells (DC)7 are key inducers of adaptive immune responses (1). In peripheral tissues, immature DC scan for potential pathogen intrusion, collecting all environmental Ag by phagocytosis, endocytosis, or macropinocytosis. Upon encountering a danger signal, DC mature and migrate to secondary lymphoid organs where they initiate adaptive immune responses by stimulating naive T cells. More recently, DC have been shown to play an effector role during innate immune responses. In fact, DC are thought to link innate and adaptive immune responses (2). DC express pattern recognition receptors such as TLR and, therefore, can recognize pathogen-related molecules. This recognition usually induces DC maturation and, consequently, their capacity to induce primary adaptive immune responses (3).

DC can also exert direct effector functions during the innate immune response. For instance, the immediate precursors of plasmacytoid DC, also known as IFN-producing cells, secrete enormous amounts of type I IFN upon viral infection (4). Mature DC also produce cytokines that can activate cells of the innate immune system such as IL-12, that is known to stimulate the NK cells function (5). We have previously described in the rat that splenic DC exhibited a direct cytotoxic activity toward tumor cells in vitro (6). The cytotoxic activity of splenic DC was restricted to a specific subset with a MHC IIlowCD4CD103+CD11b+CD172α phenotype (hereafter referred to as CD4 DC) and was Ca2+-independent (7). The nonsecretory pathway of cell-mediated cytotoxicity is known to involve molecules of the TNF superfamily that interact with specific death-induced receptors on target cells. However, we found that rat DC-mediated killing was not dependent on Fas ligand (FasL), TNF-α, or TRAIL molecules (7).

Several studies have now confirmed that human as well as murine DC can have cytolytic activities against tumoral target cells. For instance, human blood CD11c+ DC and monocyte-derived DC were able to kill various tumor cell lines either naturally (8, 9) or after IFN stimulation (10, 11) and also normal proliferating endothelial cells but not other normal cells tested (8). Other stimuli such as dsRNA and CD40L could also induce monocyte-derived DC cytotoxicity (12). In another study, unstimulated monocyte-derived DC exhibited a potent cytostatic activity toward several tumor cells (13). Interestingly, although several groups have demonstrated a role for TRAIL in the killing activity of human DC, other molecules are also thought to be implicated. However, the precise molecular mechanism of human DC-mediated killing also needs to be clarified.

In vivo, the role of cytotoxic DC has not yet been elucidated. Conceptually, killer DC could play a direct effector role in tumor immune responses as well as an indirect role by initiating adaptive immune responses to tumor Ags. In contrast, killer DC could play a role in T cell tolerance as DC were found to efficiently kill tumor T cells in vitro. However, an important step to elucidate the in vivo role of cytotoxic DC is to determine their mechanism of killing and the fate of target cells following killing. In this study, we found that rat splenic CD4 DC induced rapid cell death in a large number of both hemopoietic and nonhemopoietic tumor cell lines as well as normal endothelial cells. Target cell death exhibited apoptosis-like features but was caspase independent. Moreover, unlike for human monocyte-derived DC, Bcl-2 overexpression in target cells did not protect them from rat CD4 DC-induced cell death. Killed target cells were rapidly and specifically phagocytosed by CD4 DC.

Sprague Dawley and Lewis rats were obtained from the Centre d’Elevage Janvier and were used at 6–10 wk of age.

Cycloheximide, brefeldin A, polyinosinic-polycytidylic acid (Poly (I:C)), PKH-26, and 3,3′-dioctadecyloxacarbocyanine perchlorate (DiOC18) were purchased from Sigma-Aldrich. Z-vad.Fmk was from R&D Systems. Collagenase D was purchased from Boehringer Mannheim and annexin V-biotin was from Immunotech Beckman Coulter. TO-PRO-1 and CFSE were purchased from Molecular Probes. LPS was purchased from InvivoGen and CpG2006 from Sigma-Genosys. Soluble CD40L was kindly provided by Prof. Y. Choi (Department of Pathology and Laboratory Medicine, University of Pennsylvania School of Medicine, Philadelphia, PA).

The following mAbs were purchased from BD Biosciences: OX6-allophycocyanin-Cy7 and OX6-PE (anti-MHC class II), W3/25-FITC (anti-CD4), anti-24F-FITC (anti-CD86), and FITC-3H5 (anti-CD80). Purified OX6, OX62, and W3/25 mAb and biotinylated OX62 were prepared in our laboratory from hybridoma supernatants. AlexaFluor 633-conjugated anti-mouse Ab was purchased from Molecular Probes.

The As84, RIN, C6, C6.9, and A15A5 cell lines were kindly provided by Dr. M. Grégoire (Institut National de la Santé et de la Recherche Médicale Unité 601, Nantes, France). The EL-4, NS1, MT450, C58, and MCF7 cell lines were kindly provided by Dr J. Le Pendu (Institut National de la Santé et de la Recherche Médicale Unité 601). Jurkat, L929, and PEC cell lines were kindly provided by Dr. I. Anegon (Institut National de la Santé et de la Recherche Médicale Unité 643, Nantes, France). The YAC-1, A20, P815, K562, SP2/O, Fao, and BRL cell lines were obtained from the European Collection of Cell Cultures. The OSRGA, FLS, and ROS cell lines were provided by Dr. F. Rédini (Nantes University Medical School, Nantes, France). Stable L5178Y transfectant cells for FasL and control L5178Y cells (7) were kindly provided by Dr. H. Yagita (Juntendo University School of Medicine, Tokyo, Japan). Fas-associated death domain (FADD)-deficient (14) and caspase 8-deficient (15) Jurkat cell lines were provided by Dr. J. Blenis (Harvard Medical School, Boston, MA). Receptor-interacting protein (RIP)-deficient Jurkat cells (16) were provided by Dr. B. Seed (Massachusetts General Hospital, Boston, MA). Stable YAC-1 (17) and Jurkat transfectants for Bcl-2 were kindly provided by Dr. J. Trapani (MacCallum Cancer Institute, Victoria, Australia). RNK16 cells were provided by Dr. M. Nakamura (University of California, San Francisco, CA). Cell lines were cultured in complete RPMI.

Rat DC.

Lymphoid organs (spleens and lymph nodes (LN)) were chopped into small pieces and digested in 2 mg/ml collagenase (Boehringer Mannheim) in RPMI 1640 1% FCS for 30 min at 37°C. EDTA at 10 mM was added for the last 5 min, and the cell suspension was then pipetted up and down several times and filtered. Cells were washed once in PBS/2 mM EDTA/1% FCS, and low density cells were obtained after centrifugation on a 14.5% Nycodenz gradient as previously described (6). Cells were washed twice and then incubated with a saturating concentration of biotinylated OX62 mAb at 4°C for 20 min, washed twice, and then mixed with streptavidin microbeads following the manufacturer’s instructions (Miltenyi Biotec). Alternatively, DC were selected with directly OX62-conjugated microbeads (Miltenyi Biotec) with similar results. Positive selection was performed on a MiniMacs type column or Automacs (Miltenyi Biotec). The purity was routinely >90%. For CD4 DC subset purification, a CD4 depletion was performed before OX62 positive selections. Low density cells were stained with W3/25 mAb at 4°C for 10 min. Cells were washed twice and then incubated with anti-mouse IgG-conjugated Dynabeads (Dynal).

Human monocyte-derived DC.

PBMC (from healthy volunteers) were collected by separation of blood over a Ficoll gradient (Eurobio). Monocytes were purified with CD14-conjugated MACS microbeads (Miltenyi Biotec) and cultured for 7 days in complete RPMI in the presence of human GM-CSF at 1000 U/ml (Leukomax; Novartis) and human IL-4 at 1000 U/ml (R&D Sytems).

For phenotyping of immature and mature DC, 5 × 104 cells were labeled with mAbs in 96-well plates, washed, and analyzed on a FACSCalibur (BD Biosciences). For cell sorting of CD4+ and CD4MHC IIlow subsets of spleen and LN DC, OX62+ cells were incubated with OX6-PE (CMH II)- and W3/25-FITC (CD4)-conjugated mAbs and sorted using a FACSVantage (BD Biosciences).

Flow cytometry.

Target cells were stained with CFSE for 2 min at room temperature and centrifuged through a Ficoll gradient to remove dead cells. After three washes in PBS, target cells were cultured with CD4 DC for 4 h. Cells were then harvested and DC were stained with OX6-allophycocyanin-Cy7 mAb and analyzed on a FACS Aria (BD Biosciences). Data were analyzed with FlowJo version 6 software (Tree Star) To induce apoptosis in YAC-1 cells, cells were first labeled with CFSE and then cultured for 24 h in the absence of FCS. Control staining with annexin V indicated that >90% of the cells were positive but still excluded DAPI (data not shown).

Confocal microscopy.

Target cells were stained with DiOC18 (10 μg/ml) and cultured with CD4 DC for 4 h. Cells were harvested and incubated on alcyan blue-treated coverslips in serum-free RPMI 1640 medium at 37°C to enable their adherence to the coverslips. Cells were then fixed with 4% paraformaldehyde and stained with OX6 mAb followed by Alexa 633-conjugated anti-mouse Ab. Coverslips were mounted on slides with ProLong AntiFade kit (Molecular Probes) and dried at 4°C overnight. The fluorescence was observed by confocal microscopy (LEICA TCS-SP1). The images were imported into Adobe Photoshop 6.0, pseudo-colored, and, in some cases, overlapped to produce merged images.

Annexin V binding.

Target cells were incubated with DiOC18 (10 μg/ml) for 1 h at 37°C. After three washes, cells were cultured with CD4 DC for 4 h at 37°C and then stained with OX6-PE mAb, annexin V-biotin followed by Streptavidin-TRICOLOR (Caltag). DiOC18+ CMH II target cells were assessed for annexin V binding using a FACSCalibur.

Nuclear fragmentation.

Target cells were stained with the PKH-26 marker and cultured with CD4 DC for 4 h. All cells were incubated on alcyan blue-treated coverslips as previously described, fixed with 4% paraformaldehyde and incubated with TO-PRO-1 dye after permeablization with saponin. Slides were analyzed by confocal microscopy as described above.

The cytotoxic activity of DC populations was assessed in a standard 5-h 51Cr-release assay. Briefly, target cells were labeled with sodium 51Cr for 45 min at 37°C in complete medium. Serial dilutions of effector cells in complete medium were mixed with 2000 target cells in triplicate in V-bottom 96-well plates, centrifuged for 3 min at 1500 × g, and incubated for 6 h at 37°C, 5% CO2. The supernatants were harvested and 51Cr release was determined (Packard Instruments) using standard scintillation procedures. Specific release was calculated as 100 × (experimental release − spontaneous release)/(maximum release − spontaneous release). Results are expressed as mean ± SD of triplicate wells.

Purified CD4 and CD4+ DC were cultured overnight with 2000 target cells (ratio 25:1) in complete RPMI in U-bottom 96-well plates. Supernatants were collected and stored at −20°C until analysis. The levels of IL-10 and TNF-α in the supernatants were measured using ELISA kits (OptEIA set; BD Biosciences) according to the manufacturer’s instructions. Rat IL-12p40 was detected using ELISA kit from Biosource International according to the manufacturer’s instructions.

Statistical analysis was performed using the Student t test.

DC were isolated from spleens and LN by positive selection using the OX62 mAb that recognizes the rat αE2 integrin chain, also known as CD103 (18). Based on expression levels of MHC class II and CD4 molecules, we were able to define two subsets of CD103+MHC IIlow DC (hereby referred to as CD4+ and CD4 DC) in the spleen (Fig. 1,A) and three (CD4+MHC IIlow, CD4MHC IIlow, and CD4−/+MHC IIhigh) in the LN (Fig. 1,A) (7). The CD4+ DC also strongly expressed CD5, CD90, and CD172α, but not CD200, whereas CD4 DC were negative for CD5, CD90, and CD172α but expressed high levels of CD200 (19). Both subsets expressed high levels of CD11b. These opposing phenotypes were also found for CD4+MHC IIlow and CD4MHC IIlow DC from LN (data not shown), suggesting they represent the same subsets as spleen CD4+MHC IIlow and CD4MHC IIlow DC. We previously reported that spleen but not LN DC exhibit a strong cytotoxic activity against YAC-1 cells in vitro, suggesting that this unusual function of DC was restricted to a specialized subset (7). As we previously showed (7), the cytotoxicity of spleen DC is indeed restricted to the CD4 subset of DC (Fig. 1,A). To determine whether the CD4MHC IIlow DC found in LN also exhibited a cytotoxic activity, we sorted both CD4+ and CD4MHC IIlow DC subsets from total CD103+ LN DC. Similar to spleen DC, LN CD4MHC IIlow DC but not CD4+MHC IIlow DC showed a potent cytotoxic activity against YAC-1 cells (Fig. 1 B). A very rare subset of CD4MHC IIlowCD11b+CD103+ cells was also identified in the blood; however, these cells could not be purified in sufficient numbers to perform a cytotoxic assay.

FIGURE 1.

CD4CD103+ DC from the spleen and LN exhibit a cytotoxic activity. Spleen and LN OX62+ DC were purified using magnetic beads following selection of low-density cells. A, Cells were stained with FITC anti-CD4 and PE anti-MHC II mAbs and analyzed using a cytofluorometer. In the spleen and LN, two subsets of MHClow DC were identified, one CD4 and one CD4+. In the LN, a large subset of MHChigh DC was also identified. B, Cells in gates R1 (MHClowCD4) and R2 (MHClowCD4+) were sorted from both spleen and LN DC and tested for their capacity to induce cell death in YAC-1 cells in a 5-h 51Cr-release assay at a E:T cell ratio of 50:1. Similar results were obtained in four and three independent experiments for spleen and LN DC, respectively.

FIGURE 1.

CD4CD103+ DC from the spleen and LN exhibit a cytotoxic activity. Spleen and LN OX62+ DC were purified using magnetic beads following selection of low-density cells. A, Cells were stained with FITC anti-CD4 and PE anti-MHC II mAbs and analyzed using a cytofluorometer. In the spleen and LN, two subsets of MHClow DC were identified, one CD4 and one CD4+. In the LN, a large subset of MHChigh DC was also identified. B, Cells in gates R1 (MHClowCD4) and R2 (MHClowCD4+) were sorted from both spleen and LN DC and tested for their capacity to induce cell death in YAC-1 cells in a 5-h 51Cr-release assay at a E:T cell ratio of 50:1. Similar results were obtained in four and three independent experiments for spleen and LN DC, respectively.

Close modal

The cytotoxic activity of immature CD4 spleen DC was screened against a panel of tumoral and nontumoral target cells (Table I). Spleen CD4 DC killed a large panel of transformed or tumoral cell lines (Table I). Although most of the normal cells tested (T cells, hepatocytes, rat aorta endothelial cells) were not killed by DC, HUVECs appeared to be sensitive indicating that the killing activity of CD4 DC is not strictly restricted to tumor cells. This cytolytic activity was not species specific because spleen CD4 DC were able to kill rat (e.g., AS/84), murine (e.g., YAC-1), and human (e.g., Jurkat) cells and, moreover, DC-mediated killing was observed against both hemopoietic (e.g., YAC-1, Jurkat, SP2/0) and nonhemopoietic cell lines (e.g., AS/84, RIN) (Table I). Although numerous lymphoid-derived cell lines (e.g., YAC-1, Jurkat, SP2/0, NS1, LY5178) appeared to be sensitive to DC-mediated cell death, the murine thymoma EL-4 and the rat thymoma C58 were resistant. Moreover, both virus-induced tumor cells, such as YAC-1, and chemical-induced tumor cells, such as LY5178, were killed by CD4 DC.

Table I.

Tumor cell lines and normal cells tested and sensitivity to rat CD4MHC IIlow DC-mediated killing

CellsSpeciesHistotypeCD4 DC-Mediated Killinga
Tumor cells    
 NS1 Mouse Myeloma ++++ 
 AS84 Rat Pancreatic adenocarcinoma +++ 
 FLS Rat Fibroliposarcoma +++ 
 YAC-1 Mouse T cell lymphoma +++ 
 Jurkat Human Acute T cell leukemia ++ 
 L5178Y Mouse T cell lymphoma ++ 
 OSRGA Rat Osteosarcoma ++ 
 RIN Rat Insulinoma ++ 
 Sp2/0 Mouse Myeloma ++ 
 K562 Human Chronic myelogenous leukemia 
 MT 450 Rat Mammary adenocarcinoma 
 NMU Rat Mammary adenocarcinoma 
 A15A5 Rat Glioma − 
 A20 Mouse Reticulum cell sarcoma; B lymphocyte − 
 BRL Rat Hepatocarcinoma − 
 C58 Rat T cell lymphoma − 
 L1210 Mouse Lymphoid leukemia − 
 C6 Rat Glioma − 
 EL-4 Mouse T cell lymphoma − 
 Fao Rat Hepatocarcinoma − 
 ROS Rat Osteosarcoma − 
 MCF7 Human Mammary adenocarcinoma − 
 P815 Mouse Mastocytoma − 
Normal cells    
 HUVEC Human Primary umbilical vein endothelial cells ++ 
 T cells Rat Lymph node − 
 RAEC Rat Primary aortic endothelial cells − 
 Hepatocytes Rat Liver − 
CellsSpeciesHistotypeCD4 DC-Mediated Killinga
Tumor cells    
 NS1 Mouse Myeloma ++++ 
 AS84 Rat Pancreatic adenocarcinoma +++ 
 FLS Rat Fibroliposarcoma +++ 
 YAC-1 Mouse T cell lymphoma +++ 
 Jurkat Human Acute T cell leukemia ++ 
 L5178Y Mouse T cell lymphoma ++ 
 OSRGA Rat Osteosarcoma ++ 
 RIN Rat Insulinoma ++ 
 Sp2/0 Mouse Myeloma ++ 
 K562 Human Chronic myelogenous leukemia 
 MT 450 Rat Mammary adenocarcinoma 
 NMU Rat Mammary adenocarcinoma 
 A15A5 Rat Glioma − 
 A20 Mouse Reticulum cell sarcoma; B lymphocyte − 
 BRL Rat Hepatocarcinoma − 
 C58 Rat T cell lymphoma − 
 L1210 Mouse Lymphoid leukemia − 
 C6 Rat Glioma − 
 EL-4 Mouse T cell lymphoma − 
 Fao Rat Hepatocarcinoma − 
 ROS Rat Osteosarcoma − 
 MCF7 Human Mammary adenocarcinoma − 
 P815 Mouse Mastocytoma − 
Normal cells    
 HUVEC Human Primary umbilical vein endothelial cells ++ 
 T cells Rat Lymph node − 
 RAEC Rat Primary aortic endothelial cells − 
 Hepatocytes Rat Liver − 
a

As assessed in a 6-h 51Cr-release assay at an E:T cell ratio of 50:1. –, <5% of specific lysis; +, 5–20%; ++, 20–40%; +++, 40–60%; ++++, >60%. Each cell line was tested in at least three independent experiments.

To determine whether the killing activity of CD4 spleen DC was modified by their maturation state, we measured the cytotoxic activity of CD4 DC before and after overnight culture in the presence or in the absence of the TLR-ligands poly I:C, LPS, CpG2006 or soluble CD40L. As shown in Fig. 2,A, freshly extracted CD4 DC exhibited an immature phenotype whereas TLR or CD40L-stimulated cells exhibited a mature phenotype as evidenced by the up-regulation of CD80, CD86, and MHC class II molecules. DC cultured in the absence of TLR ligand up-regulated MHC class II but not CD80 and CD86 molecules and exhibited poor survival (19). Whatever the maturation stimulus used, the cytotoxic activity of mature CD4 DC was 5-fold lower as compared with the killing activity of fresh and immature CD4 DC (Fig. 2 B). Moreover, the low cytotoxicity observed with mature CD4 DC was inhibited by EGTA whereas fresh DC used a Ca2+-independent mechanism suggesting that fresh DC and mature DC use different mechanisms. Finally, the addition of a TLR ligand during the 5-h cytotoxic assay with fresh DC did not modify their killing activity (data not shown) indicating that its down-regulation during maturation is rather a late event.

FIGURE 2.

Effect of DC maturation state on killing activity. A, The expression of MHC class II, CD80, and CD86 (bold line) molecules was analyzed on CD4 DC before (fresh cells) or after overnight culture in the absence (none) or in the presence of Poly I:C (25 μg/ml), soluble CD40L, LPS (0.5 μg/ml), or CpG 2006 (10 μM). The dotted line represents the isotype control mAb staining. B, The same cells were tested for their cytotoxic activity against YAC-1 cells. The assay was performed in the absence or the presence of EGTA (1 mM) at a ratio of 25:1. The results of one experiment representative of three are shown. ∗, p < 0.05.

FIGURE 2.

Effect of DC maturation state on killing activity. A, The expression of MHC class II, CD80, and CD86 (bold line) molecules was analyzed on CD4 DC before (fresh cells) or after overnight culture in the absence (none) or in the presence of Poly I:C (25 μg/ml), soluble CD40L, LPS (0.5 μg/ml), or CpG 2006 (10 μM). The dotted line represents the isotype control mAb staining. B, The same cells were tested for their cytotoxic activity against YAC-1 cells. The assay was performed in the absence or the presence of EGTA (1 mM) at a ratio of 25:1. The results of one experiment representative of three are shown. ∗, p < 0.05.

Close modal

To analyze the mechanism of DC-mediated killing we first sought to determine whether the killing required cell-to-cell contact or involved a soluble molecule. In preliminary experiments we could not observe any cytotoxic activity in the supernatant of cultured CD4 DC or of DC-target cell coculture (7). We then performed cytotoxicity assays in which spleen CD4 DC, alone or cocultured with unlabeled YAC-1 cells were separated from chromium-labeled YAC-1 target cells by a semipermeable membrane (Fig. 3,A). In these conditions, chromium-labeled target cells in the upper compartment were not killed, indicating that DC-mediated killing required intimate cell-to-cell contact similar to what was observed with a rat NK cell line (RNK16) (Fig. 3 A).

FIGURE 3.

The cytotoxic activity of CD4 DC requires de novo synthesis of a membrane protein. A, A Transwell cytotoxicity assay was performed in 24-well plates using unlabeled (YAC-1) or 51Cr-labeled (YAC-1*) YAC-1 cells and spleen CD4 DC or RNK16 at a ratio of 25:1. CD4 DC-mediated cell death as well as RNK16-mediated cell death required cell contact. Data represent the mean ± SD of three independent experiments. B, Spleen CD4 DC were cultured for 3 h in the presence of cycloheximide (10 μg/ml), brefeldin A (10 μg/ml), or the diluent, extensively washed, and then used in a 2-h cytotoxic assay at a ratio of 12:1. Histograms represent the mean ± SD of three independent experiments.

FIGURE 3.

The cytotoxic activity of CD4 DC requires de novo synthesis of a membrane protein. A, A Transwell cytotoxicity assay was performed in 24-well plates using unlabeled (YAC-1) or 51Cr-labeled (YAC-1*) YAC-1 cells and spleen CD4 DC or RNK16 at a ratio of 25:1. CD4 DC-mediated cell death as well as RNK16-mediated cell death required cell contact. Data represent the mean ± SD of three independent experiments. B, Spleen CD4 DC were cultured for 3 h in the presence of cycloheximide (10 μg/ml), brefeldin A (10 μg/ml), or the diluent, extensively washed, and then used in a 2-h cytotoxic assay at a ratio of 12:1. Histograms represent the mean ± SD of three independent experiments.

Close modal

Although 51Cr release from target cells started at 2 h and reached a plateau at 6 h with freshly extracted CD4 DC (data not shown), substantial levels of 51Cr release could be observed at 2 h provided that DC were cultured for 3 h before the cytotoxic assay (Fig. 3,B), suggesting that DC need to synthesize a putative killer factor during the first 3 h of the assay. To test this hypothesis, DC were cultured for 3 h in the presence of cycloheximide or brefeldin A, extensively washed, and used in a 2-h cytotoxic assay. Cycloheximide and brefeldin A increased CD4 DC death by 40 and 30% (means of three independent experiments), respectively. After 3 h of culture, the expression of MHC class II molecule but not CD80 and CD86 was slightly up-regulated on DC and this up-regulation was inhibited at the same extent by cycloheximide and brefeldin A (data not shown). Despite these very similar effects of cycloheximide and brefeldin A on DC survival and maturation during this 3 h culture time, we observed that the cytotoxic activity of DC was strongly inhibited by cycloheximide but not brefeldin A (Fig. 3 B). Taken together, these results suggest that the cytotoxic activity required the de novo synthesis of a putative killer membrane protein and that the killing of target cells by DC is likely to be rapid once this molecule is expressed at their surface.

We then sought to characterize whether CD4 DC-mediated cell death involved the necrosis or apoptosis of target cells. First, we compared the 51Cr-release assay with the Just Another Method test that is thought to measure apoptosis rather than necrosis, and we observed similar kinetics and levels of specific cytotoxicity (data not shown). Second, we found that CD4 DC induced exposure of phosphatidyserine at the surface of >50% of target cells as evidenced by annexin V binding (Fig. 4,A). Third, we examined nuclear morphology in target cells following exposure to CD4 DC. Apoptosis is characterized by chromatin condensation and then fragmentation leading to the formation of apoptotic bodies. As shown in Fig. 4,B, CD4 DC-induced cell death in Jurkat cells was associated with rapid chromatin condensation and fragmentation similar to that observed during classical apoptosis such that induced in Jurkat cells by Fas triggering. Although only 10% of visible Jurkat cells exhibited a fragmented nucleus when cultured with CD4 DC, such cells were not observed in control slides with Jurkat cultured alone for 5 h (data not shown). As described later, this discrepancy is probably due to the rapid phagocytosis of killed target cells by CD4 DC (see Fig. 7). Therefore, we concluded that killer DC induced apoptosis-like cell death in sensitive target cells.

FIGURE 4.

CD4 DC induce apoptosis-like cell death in sensitive target cells. A, Jurkat cells were labeled with PKH-26 and then mixed with spleen CD4 DC at an E:T ratio of 25:1 for 5 h at 37°C (bold histogram) or 4°C (thin histogram). Cells were then stained with FITC MHC II mAb and biotin annexin V followed by Streptavidin-TRICOLOR and analyzed by flow cytometry. The binding of annexin V was analyzed on gated MHC II PKH26+ cells. B, PKH-26-labeled Jurkat cells alone or mixed with CD4MHC IIlow DC or L5178Y-FasL transfectants (ratio 25:1) were cultured for 5 h at 37°C. Cells were then attached to coverslips, fixed, and labeled with TO-PRO-1 and analyzed by confocal microscopy.

FIGURE 4.

CD4 DC induce apoptosis-like cell death in sensitive target cells. A, Jurkat cells were labeled with PKH-26 and then mixed with spleen CD4 DC at an E:T ratio of 25:1 for 5 h at 37°C (bold histogram) or 4°C (thin histogram). Cells were then stained with FITC MHC II mAb and biotin annexin V followed by Streptavidin-TRICOLOR and analyzed by flow cytometry. The binding of annexin V was analyzed on gated MHC II PKH26+ cells. B, PKH-26-labeled Jurkat cells alone or mixed with CD4MHC IIlow DC or L5178Y-FasL transfectants (ratio 25:1) were cultured for 5 h at 37°C. Cells were then attached to coverslips, fixed, and labeled with TO-PRO-1 and analyzed by confocal microscopy.

Close modal
FIGURE 7.

CD4 DC rapidly engulf their victims following killing. A, A cytotoxicity assay was performed using CD4 DC and YAC-1 or EL-4 cells (ratio 50:1). B, YAC-1 and EL-4 cells were labeled two different concentrations of CFSE (5 and 1 μm, respectively) and cultured in V-bottom 96-well plates with CD4 DC at a ratio of 25:1 for 4 h. Cells were harvested every hour, stained with allophycocyanin-Cy7-MHC II mAb, and analyzed by flow cytometry (upper panel). Alternatively, CD4 DC were cultured for 4 h with either YAC-1 or EL-4 CFSE-labeled cells (bottom right panel). As a control for CFSE staining, CFSE-labeled target cells were cultured alone for 4 h (bottom left panel). C, CD4 DC were cultured for 4 h with DiOC18-labeled YAC-1 or EL-4 cells at a ratio of 25:1 attached to coverslips, fixed, stained with MHC II mAb followed by anti-mouse Alexa 633 Ab, and analyzed by confocal microscopy. Similar results were obtained in three independent experiments.

FIGURE 7.

CD4 DC rapidly engulf their victims following killing. A, A cytotoxicity assay was performed using CD4 DC and YAC-1 or EL-4 cells (ratio 50:1). B, YAC-1 and EL-4 cells were labeled two different concentrations of CFSE (5 and 1 μm, respectively) and cultured in V-bottom 96-well plates with CD4 DC at a ratio of 25:1 for 4 h. Cells were harvested every hour, stained with allophycocyanin-Cy7-MHC II mAb, and analyzed by flow cytometry (upper panel). Alternatively, CD4 DC were cultured for 4 h with either YAC-1 or EL-4 CFSE-labeled cells (bottom right panel). As a control for CFSE staining, CFSE-labeled target cells were cultured alone for 4 h (bottom left panel). C, CD4 DC were cultured for 4 h with DiOC18-labeled YAC-1 or EL-4 cells at a ratio of 25:1 attached to coverslips, fixed, stained with MHC II mAb followed by anti-mouse Alexa 633 Ab, and analyzed by confocal microscopy. Similar results were obtained in three independent experiments.

Close modal

To determine the role of caspases in CD4 DC-mediated apoptosis, we performed cytotoxic assays in the presence of the pan-caspase inhibitor Z-vad.Fmk. As shown in Fig. 5,A, Z-vad.Fmk did not inhibit CD4 DC-induced YAC-1 apoptosis. We confirmed that CD4 DC-induced cell death in Jurkat, AS84, and Sp2O was also caspase independent (data not shown). In control experiments, we found that rat NK-mediated killing of YAC-1 cells and Fas-mediated killing of A20 cells was, respectively, partially and totally inhibited by Z-vad.Fmk at the same concentration (Fig. 5,A). This indicates that YAC-1 cells can undergo a caspase-dependent cell death. Previously, we showed that DC-induced cell death cannot be inhibited by FasL-Fas, TRAIL-TRAIL-R, or TNF-TNF-R interaction blocking reagents (7). As FasL, TRAIL, and TNF induce apoptosis mainly through the recruitment of FADD and caspase 8 in target cells (20), we analyzed whether these two molecules were required for CD4 DC-induced apoptosis. As shown in Fig. 5,B, caspase 8 deficient Jurkat cells were killed as efficiently as wild-type Jurkat cells. Similarly, we found that FADD was not required for CD4 DC-induced apoptosis (Fig. 6,A). Therefore, the mechanism of killing induced by these DC involved a FADD and caspase-independent apoptosis-like cell death. Recently, FasL, TRAIL, and TNF were shown to induce cell death independently of caspase 8, by a mechanism involving the recruitment of the RIP (21). However, we found that RIP-deficient Jurkat cells were killed by CD4 DC even in the presence of Z-vad.Fmk (Fig. 6 B), thus excluding a role for this pathway in DC-mediated cell death. This is consistent with the fact that CD4 DC appear to induce FADD-independent apoptosis-like rather than necrosis in target cells (21).

FIGURE 5.

CD4 DC-induced apoptosis is caspase independent. A, A cytotoxicity assay was performed with CD4 DC and YAC-1 cells, RNK16 and YAC-1 cells, or L5178Y-FasL and A20 cells (ratio 50:1) in the absence or in the presence of the pan-caspase inhibitor Z-vad.Fmk (10 μM). B, A cytotoxicity assay was performed using CD4 DC and wild-type (Jurkat) or caspase 8-deficient (Jurkat I9.2) Jurkat cells (ratio 50:1). Similar results were obtained in three independent experiments. ∗, p < 0.05.

FIGURE 5.

CD4 DC-induced apoptosis is caspase independent. A, A cytotoxicity assay was performed with CD4 DC and YAC-1 cells, RNK16 and YAC-1 cells, or L5178Y-FasL and A20 cells (ratio 50:1) in the absence or in the presence of the pan-caspase inhibitor Z-vad.Fmk (10 μM). B, A cytotoxicity assay was performed using CD4 DC and wild-type (Jurkat) or caspase 8-deficient (Jurkat I9.2) Jurkat cells (ratio 50:1). Similar results were obtained in three independent experiments. ∗, p < 0.05.

Close modal
FIGURE 6.

CD4 DC-induced apoptosis is FADD- and RIP-independent and is not inhibited by Bcl-2 overexpression. A, A cytotoxicity assay was performed with CD4 DC or L5178-FasL transfectants as effector cells and wild-type or FADD-deficient as target cells (ratio 50:1). B, A cytotoxicity assay was performed using CD4 DC as effector cells and wild-type or RIP-deficient Jurkat cells (ratio 50:1) as target cells in the absence or presence of Z-vad.Fmk (10 μM). Similar results were obtained in two independent experiments. C, A cytotoxicity assay was performed using rat spleen CD4 DC and wild-type or Bcl-2-overexpressing Jurkat cells (ratio 50:1) in the absence or presence of Z-vad.Fmk (10 μM). Bcl-2 protection was verified using human monocyte-derived immature DC as target cells (inset). Similar results were obtained in three independent experiments. ∗, p < 0.05.

FIGURE 6.

CD4 DC-induced apoptosis is FADD- and RIP-independent and is not inhibited by Bcl-2 overexpression. A, A cytotoxicity assay was performed with CD4 DC or L5178-FasL transfectants as effector cells and wild-type or FADD-deficient as target cells (ratio 50:1). B, A cytotoxicity assay was performed using CD4 DC as effector cells and wild-type or RIP-deficient Jurkat cells (ratio 50:1) as target cells in the absence or presence of Z-vad.Fmk (10 μM). Similar results were obtained in two independent experiments. C, A cytotoxicity assay was performed using rat spleen CD4 DC and wild-type or Bcl-2-overexpressing Jurkat cells (ratio 50:1) in the absence or presence of Z-vad.Fmk (10 μM). Bcl-2 protection was verified using human monocyte-derived immature DC as target cells (inset). Similar results were obtained in three independent experiments. ∗, p < 0.05.

Close modal

Caspase-independent apoptosis has been described previously in various cell lines as well as in T cells, and has been shown to be mediated through mitochondrial release of proapoptotic molecules such as apoptosis-inducing factor or Omi/Htra (22). As several of those pathways were shown to be inhibited by the anti-apoptotic molecule Bcl-2 (22), we assessed the role of Bcl-2 in DC-mediated apoptosis. As shown in Fig. 6,C, Jurkat cells overexpressing Bcl-2 were killed as efficiently as control cells by spleen CD4 DC even in the presence of the caspase inhibitor Z-vad.Fmk. Similar results were obtained with Bcl-2 overexpressing YAC-1 cells (data not shown). In contrast and as previously shown by others (9), Bcl-2 overexpression protected Jurkat cells against human monocyte-derived DC-mediated cell death (Fig. 6 C, inset).

We then sought to determine whether CD4 DC could acquire Ags from killed target cells. For this purpose, we used two mouse lymphoma cell lines, YAC-1 cells that are sensitive to CD4 DC-induced cell death and EL-4 that are resistant (Fig. 7,A). For flow cytometry analysis of phagocytosis, YAC-1 and EL-4 cells were labeled with two different concentrations of CFSE and cultured with fresh CD4 DC at a ratio of 25 DC to 1 target cell. Cells were then analyzed every hour until 4 h after staining with a rat MHC class II mAb. This allows correlation of the phagocytic activity of DC with the number of YAC-1 and EL-4 cells in the same well. As shown in Fig. 7,B, the phagocytosis of target cell by CD4 DC was already obvious at 1 h after the beginning of the coculture (9% of DC) (upper panel). The percentage of DC having ingested target cells increased rapidly until 2 h (22%) and then more slowly until 4 h (27%). It should be reminded that 51Cr release from target cells was not observed during the first 2 h of the cytotoxic assay suggesting that CD4 DC start to phagocytose their victims before they lose membrane integrity. A decrease in the intensity of CFSE staining was observed both for YAC-1 and EL-4 cells when cultured alone (Fig. 7,B, bottom left panel) or with CD4 DC (Fig. 7,B, upper panel) as compared with cells analyzed right after the CFSE staining. This is likely due to spontaneous leakage of CFSE rather than proliferation. However, CD4 DC cultured for 4 h with EL-4 alone did not exhibit any CFSE staining, indicating that double positive cells correspond to DC having phagocytosed YAC-1 cells rather than having acquired soluble CFSE. The kinetics of phagocytosis were negatively correlated with the numbers of YAC-1 cells in the well, whereas the numbers of EL-4 cells were not modified throughout the experiment. Similar results were obtained when CD4 DC were cultured with either YAC-1 cells or EL-4 cells (Fig. 7 B, bottom right panel).

The same cells were analyzed by confocal microscopy. For this experiment, target cells were labeled with DiOC18 instead of CFSE. As shown in Fig. 7,C, we confirmed that CD4 DC have acquired fragments of YAC-1 cells but not EL-4 cells. (Fig. 7,C). These results indicate that the killing activity of CD4 DC enables these cells to acquire Ags from live cells. As shown in Fig. 7,B, CD4 DC that have been cocultured with YAC-1 target cells exhibit an increase in MHC class II molecule expression. This is due to spontaneous maturation of DC that occurs in vitro after tissue extraction. This maturation was also observed with EL4 cells (Fig. 7 C, bottom right panel), indicating that it was not induced by the killing or phagocytosing activity of CD4 DC.

Because CD4 DC lost their killing activity upon maturation (see Fig. 2,B), we sought to determine whether it is also the case for the phagocytosis. As a control, we used in this experiment CD4+ spleen DC that do not have a cytotoxic activity (see Fig. 1). Thus, immature, i.e., freshly extracted, or mature, i.e., overnight cultured in the presence of LPS, CD4 and CD4+ DC were cultured for 4 h with either live or apoptotic CFSE-stained YAC-1 cells and then analyzed by FACS (Fig. 8). As expected, unlike CD4 DC, CD4+ DC were unable to acquire fragments of target cells when cultured with live YAC-1 cells. However, CD4+ DC were also very poor at phagocytosing apoptotic YAC-1 cells as compared with CD4 DC. The capacity of CD4 DC to phagocytose apoptotic cells was dramatically down-regulated following in vitro maturation with LPS (Fig. 8 B). Moreover, correlating with the down-regulation of their killing activity, mature CD4 DC were unable to acquire fragments when cultured with live YAC-1 cells.

FIGURE 8.

Immature CD4 but not CD4+ nor mature CD4 or CD4+ spleen DC phagocytose efficiently apoptotic cells in vitro. Immature (A) or LPS-matured (B) CD4 and CD4+ spleen DC were cultured for 4 h with CFSE-labeled live or apoptotic YAC-1 cells at a ratio of 25:1. Cells were then stained with allophycocyanin-Cy7-MHC II mAb and analyzed by FACS. As a control, DC were incubated with apoptotic YAC-1 cells at 4°C. Similar results were obtained in four independent experiments.

FIGURE 8.

Immature CD4 but not CD4+ nor mature CD4 or CD4+ spleen DC phagocytose efficiently apoptotic cells in vitro. Immature (A) or LPS-matured (B) CD4 and CD4+ spleen DC were cultured for 4 h with CFSE-labeled live or apoptotic YAC-1 cells at a ratio of 25:1. Cells were then stained with allophycocyanin-Cy7-MHC II mAb and analyzed by FACS. As a control, DC were incubated with apoptotic YAC-1 cells at 4°C. Similar results were obtained in four independent experiments.

Close modal

To analyze whether interaction between the killer DC and target cells could have an effect on their function, CD4 DC were cultured in the absence or in the presence of YAC-1 or EL-4 cells and cytokine production as assessed (Table II). The coculture with target cells increases the production of IL-10, TNF-α and IL-12p40 compared with the amount of cytokines produced by the DC alone. However no significant difference was observed between cocultures with sensitive YAC-1 cells and resistant EL-4 cells. Finally, no differences were observed in the APC function of killer DC in allogenic MLR following incubation with YAC-1 or EL-4 cells (data not shown).

Table II.

Cytokine production during coculture of CD4 DC with target cellsa

IL 10 (pg/ml)TNF-α (pg/ml)IL12p40 (pg/ml)
CD4− YAC-1 64 (58) 30 (43) 4810 (240) 
CD4 EL-4 158 (26) 20 (17) 5507 (1604) 
CD4 alone 0 (0) 0 (0) 1374 (440) 
IL 10 (pg/ml)TNF-α (pg/ml)IL12p40 (pg/ml)
CD4− YAC-1 64 (58) 30 (43) 4810 (240) 
CD4 EL-4 158 (26) 20 (17) 5507 (1604) 
CD4 alone 0 (0) 0 (0) 1374 (440) 
a

CD4 DC were cultured overnight in the absence or in the presence of YAC-1 or EL-4 cells at a ratio of 25:1. Cytokine production was assessed by ELISA test. Means (SD) of three experiments.

Besides their capacity to efficiently stimulate CD4+ T cells and to drive Th1 polarization (19), the CD4CD103+ subset of rat DC that is found in spleen and in lower numbers in LN also exhibited a potent and rapid cytolytic activity in vitro toward select tumor and normal cells (7). This activity that required cell-to-cell contact was restricted to the immature stage of CD4 DC and involved a caspase-independent apoptosis like cell death. We also showed that immature but not mature CD4 DC rapidly phagocytose large cellular fragments from their victims following killing.

Although the capacity of in vitro generated monocyte-derived DC to induce cytolysis or cytostasis in tumor cells has been demonstrated by numerous group, very few studies have examined the cytolytic activity of normal human blood DC (11). In one study, Fanger et al. (11) demonstrated that purified CD11c+ DC but not IL3-Rα+ plasmacytoid DC exhibited cytolytic activity against tumor cells. However, this function, that was specifically TRAIL-dependent, needed to be induced by IFN-α or IFN-γ (11). Recently, a subset of cytolytic DC has been described in mice (23). These cells exhibited a NK/DC phenotype as they expressed both CD11c and two NK markers namely CD49d (DX5 mAb) and NK1.1, as well as a NK/DC function in that they could efficiently kill YAC-1 cells and process and present Ag to T cells. Similar to the cytotoxic subset of DC we describe in this study, the aforementioned murine killer DC exhibited a CD11b+CD4 phenotype. This subset of DC was found in unmanipulated animals and strongly expanded following viral infection. Interestingly, these NK/DC cells were found to have a regulatory function following treatment with anti-CD40 mAb, suggesting a role for these cells in peripheral tolerance (23). In humans, plasmacytoid DC were shown to express granzyme (24); however, they have not been reported to exhibit cytolytic activity so far (11). We recently identified the rat counterpart of plasmacytoid DC (25) and found that they were unable to induce cell death in YAC-1 or Jurkat cells (F. X. Hubert and R. Josien, unpublished observation). Together with the present study, these results indicate that only the CD4 subset of DC exhibits a killer activity in rats. These cells are found in the spleen and LN, and a very similar subset of DC has been identified in the afferent lymph (26) suggesting that, at least some of the CD4 DC found in the LN correspond to migratory tissue DC. In the spleen, the CD4 DC were found in both the red pulp and T cell area (Ref. 27 and our unpublished observations) suggesting that at least some of these cells could be blood-derived DC.

The in vitro cytolytic activity of CD4 DC was found against numerous tumor cells. Among the normal cells we could tested, only HUVEC exhibited sensitivity to rat CD4 DC. It is interesting to note that in humans, monocyte-derived DC as well as blood CD11c+ DC were shown to kill tumor cells but not normal cells (8, 11, 28) except some endothelial cells (8). The mechanisms of killing by human DC have been investigated by several groups with different results, although the requirement for a cell contact between DC and their targets has been demonstrated in most of the studies. The involvement of TRAIL has been reproducibly demonstrated. However, the expression of TRAIL by DC required a stimulation by IFN-α (11, 29), IFN-β (10, 12), IFN-γ (11), or measles virus (30), and target cells needed to express functional TRAIL receptor. Others have shown that immature human DC derived from ascitic monocytes of patients with ovarian cancer could kill the autologous tumor cells via FasL-Fas interaction (31). In mice, Langerhans cells were shown to express functional FasL upon CD40 ligation (32). Human DC were also shown to induce cytotoxicity in human breast cancer cells (33) as well as growth inhibition of numerous tumor cell lines (13) via TNF-α, however the involvement of TNF in these studies was partial and was enhanced by LPS stimulation. Lu and colleagues (8, 34) demonstrated that human DC could express and use the four different cytotoxic molecules of the TNF superfamily TNF-α, lymphotoxin-α1β2, FasL, and TRAIL. Finally, another study has shown that human monocyte-derived DC efficiently killed various tumor cells using a TNF-α-, FasL-, and TRAIL-independent and so far unknown mechanism (9). Together, these studies indicate that human DC can use multiple pathways of cytotoxicity to induce killing of tumor cells in vitro.

We have shown that the killing activity of fresh rat CD4 DC was Ca2+ as well as FasL, TRAIL, and TNF-α independent (7). The lack of involvement of these apoptosis-inducing ligands is strongly supported by the fact that FADD and caspase 8 were not required in target cells as shown in this report. Our results indicate that CD4 DC induced caspase-independent apoptosis-like cell death in target cells, likely requiring the expression of a membrane killing protein on DC that appeared to be induced during the cytotoxic assay. Because CD4 DC spontaneously mature in vitro following tissue extraction, it is possible that this molecule is only expressed during the short maturation step preceding the cytotoxicity. Alternatively, this requirement for protein synthesis and short-term culture could be due to the stress of the tissue extraction procedure and sorting.

Caspase-independent apoptosis-like cell death has been described in tumor cells as well as in T cells (22). This type of cell death can involve mitochondrial outer membrane permeabilization and the release of proteins that could directly induce DNA cleavage such as apoptosis-inducing factor, EndoG, or Omi/Htra (22). However, the overexpression of the anti-apoptotic molecule Bcl-2, which is known to oppose mitochondrial outer membrane permeabilization, did not protect Jurkat cells against CD4 DC-induced cell death. The release of some cathepsins from lysosomes in the cytosol can also induce caspase-independent apoptosis-like cell death (22). However, most of these proteases are inhibited by Z-vad.Fmk that did not affect rat CD4 DC-mediated cell death in our study. Interestingly, Hubert et al. (28) have found that human monocyte-derived DC could induce cell death of papillomavirus-transformed keratinocytes but not of normal keratinocytes, suggesting that DC can specifically recognize transformed cells. Our results indicate that this recognition is not limited to virus-induced tumor as some chemically induced as well as radiation-induced tumor cell lines were found to be killed by DC. Although the cytotoxic activity of CD4OX62+ DC was strongly down-regulated upon in vitro maturation, the residual killing was totally Ca2+-dependent. This is in agreement with our previous study, in which we found that overnight-cultured and, therefore, mature spleen DC induced cell death in YAC-1 cells using a Ca2+-mechanism (6). However, in this study, the killing activity of culture spleen DC was much stronger than that of mature CD4 DC found in the present study. These discrepancies might be related to the different purification procedure used (negative selection vs positive selection) that might lead to different DC subset enrichments. Confirming our initial study, Alli et al. (35) have recently shown that mature spleen DC could induce a Ca2+-dependent cell death in select target cells using a NKG2D-dependent recognition mechanism. However, our preliminary results suggest that the cytotoxic activity of immature CD4 DC is not triggered by the expression of a NKG2D ligand on target cells.

The role of the cytotoxic activity of DC in vivo remains to be determined; however, several hypotheses can be drawn. First, because of the relative restricted cytolysis of tumor cells, one can suggest a role for CD4 DC in immune surveillance and killing of tumor as described for NK cells (36). Therefore, it might be interesting to determine whether these DC can be found in tumors and if so, whether they are surrounded by apoptotic tumor cells. However, the lack of specific markers for CD4 rat DC and the fact that CD4 can be down-regulated on CD4+ DC upon maturation make this analysis difficult in vivo. We are currently developing new mAbs specific for rat DC subsets that will be useful for in vivo studies. Because CD4 DC appear to rapidly engulf their victims following killing, it is possible that they use their cytolytic function to acquire tumor Ags and present them to naive CD4+ and CD8+ T cells in lymphoid organs (31). However, although CD4 DC can induce strong Th1 polarization in CD4+ T cells, they are very poor stimulators of CD8+ T cells in vitro as compared with CD4+ DC (19). Second, and as suggested by others (11, 37), cytolytic DC could induce apoptosis and deletion of T cells and, therefore, play a role in peripheral tolerance. However, our results do not support this hypothesis as we were unable to detect any cytolytic activity against resting or activated normal T cells in vitro (data not shown). Third, because CD4 DC can induce cell death in normal cells (i.e., endothelial cells in our study), they could use their killer activity to acquire self Ag in peripheral tissue. The current and dominant hypothesis concerning the role of DC in peripheral tolerance is that, in the steady state, some immature DC acquire self Ag in peripheral tissues and then constitutively migrate without fully maturing to secondary lymphoid organs where they maintain peripheral T cell tolerance (38). An important issue is to know whether or not this capacity of DC to acquire self Ag is restricted to a specialized subset of DC. Studies performed in mice indicated that a subset of CD8+ DC was specialized in cross-tolerizing CD8+ T cells to self Ag (39). In rats, it was shown that in the gut draining lymph, a subset of CD4 DC that is very likely equivalent to the cytotoxic subset of CD4 DC we described in this study, transported fragments of gut apoptotic epithelial cells (26). Importantly, CD4+ DC, that were also found in the lymph, did not contain cell fragments, suggesting that, at least in this tissue, the constitutive acquisition and transport of self Ag is restricted to CD4 DC (26). Correlating with this finding, we found that, in vitro, CD4 DC were by far, much more efficient at phagocytosing apoptotic cells than CD4+ DC. Although the mechanism of self Ag uptake by tissue DC has not been identified, it has been proposed that DC engulf apoptotic cells that appear during tissue self-renewal (40). However, because some tissues have very low self-renewal and because other phagocytes such as resident macrophages can rapidly engulf dead cells, this mechanism might not prove optimal for ensuring peripheral tolerance. Harshyne et al. have shown that immature DC can acquire Ag from live cells via a process called nibbling (41). We suggest that rat CD4 DC might use their killing activity to eliminate senescent or damaged cells before they undergo apoptosis in tissue and then acquire self Ag with high efficiency.

To conclude, the MHC class IIlowCD4CD103+ subset of rat DC is endowed with a powerful cytotoxic activity toward both hemopoietic and nonhemopoietic tumor cells and also normal endothelial cells. This activity is restricted to the immature form of DC. The molecular mechanism by which CD4MHC IIlow DC induced caspase-independent apoptosis like cell death remains to be determined but clearly differs from the classical pathways of cell-mediated cytotoxicity described so far. The identification of this mechanism will be very helpful to understand the in vivo roles of this unusual function of DC. We suggest that one of these roles is to allow immature CD4 DC to acquire large amounts of Ags from live cells.

We are grateful to Marc Grégoire, Jacques Le Pendu, Ignacio Anegon, Dominique Heymann, Hideo Yagita, John Blenis, Brian Seed, and Joseph Trapani for providing various cell lines. We thank Nelly Robillard for cell sorting, Pierre-François Cartron and Philippe Juin for help with the characterization of nuclear fragmentation, and Caroline Colombeix for help with confocal microscopy.

The authors have no financial conflict of interest.

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

1

This work was supported by the Institut National de la Santé et de la Recherche Médicale and by grants to R.J. from the Etablissement français des Greffe (AO 2000) and the Association pour la Recherche sur le Cancer (ARC 5901). B.T. was supported by the Fondation PROGREFFE, the Société Française de Transplantation, and the Association pour la Recherche sur le Cancer; C.C. was supported by the Ligue Régionale contre le Cancer; H.P. was supported by the Ministère de l’Education Supérieure et de la Recherche; and C.V. was supported by La Ligue Nationale contre le Cancer.

7

Abbreviations used in this paper: DC, dendritic cell; DiOC18, 3,3′-dioctadecyloxacarbocyanine perchlorate; FADD, Fas-associated death domain; LN, lymph node; RIP, receptor-interacting protein; Poly I:C, polyinosinic-polycytidylic acid; FasL, Fas ligand.

1
Banchereau, J., R. M. Steinman.
1998
. Dendritic cells and the control of immunity.
Nature
392
:
245
-252.
2
Medzhitov, R..
2001
. Toll-like receptors and innate immunity.
Nat. Rev. Immunol.
1
:
135
-145.
3
Schnare, M., G. M. Barton, A. C. Holt, K. Takeda, S. Akira, R. Medzhitov.
2001
. Toll-like receptors control activation of adaptive immune responses.
Nat. Immunol.
2
:
947
-950.
4
Shortman, K., Y. J. Liu.
2002
. Mouse and human dendritic cell subtypes.
Nat. Rev. Immunol.
2
:
151
-161.
5
Trinchieri, G..
1995
. Interleukin-12: a proinflammatory cytokine with immunoregulatory functions that bridge innate resistance and antigen-specific adaptive immunity.
Annu. Rev. Immunol.
13
:
251
-276.
6
Josien, R., M. Heslan, J. P. Soulillou, M. C. Cuturi.
1997
. Rat spleen dendritic cells express natural killer cell receptor protein 1 (NKR-P1) and have cytotoxic activity to select targets via a Ca2+-dependent mechanism.
J. Exp. Med.
186
:
467
-472.
7
Trinite, B., C. Voisine, H. Yagita, R. Josien.
2000
. A subset of cytolytic dendritic cells in rat.
J. Immunol.
165
:
4202
-4208.
8
Janjic, B. M., G. Lu, A. Pimenov, T. L. Whiteside, W. J. Storkus, N. L. Vujanovic.
2002
. Innate direct anticancer effector function of human immature dendritic cells. I. Involvement of an apoptosis-inducing pathway.
J. Immunol.
168
:
1823
-1830.
9
Vanderheyde, N., E. Aksoy, Z. Amraoui, P. Vandenabeele, M. Goldman, F. Willems.
2001
. Tumoricidal activity of monocyte-derived dendritic cells: evidence for a caspase-8-dependent, fas-associated death domain-independent mechanism.
J. Immunol.
167
:
3565
-3569.
10
Liu, S., Y. Yu, M. Zhang, W. Wang, X. Cao.
2001
. The involvement of TNF-α-related apoptosis-inducing ligand in the enhanced cytotoxicity of IFN-β-stimulated human dendritic cells to tumor cells.
J. Immunol.
166
:
5407
-5415.
11
Fanger, N. A., C. R. Maliszewski, K. Schooley, T. S. Griffith.
1999
. Human dendritic cells mediate cellular apoptosis via tumor necrosis factor-related apoptosis-inducing ligand (TRAIL).
J. Exp. Med.
190
:
1155
-1164.
12
Vidalain, P. O., O. Azocar, H. Yagita, C. Rabourdin-Combe, C. Servet-Delprat.
2001
. Cytotoxic activity of human dendritic cells is differentially regulated by double-stranded RNA and CD40 ligand.
J. Immunol.
167
:
3765
-3772.
13
Chapoval, A. I., K. Tamada, L. Chen.
2000
. In vitro growth inhibition of a broad spectrum of tumor cell lines by activated human dendritic cells.
Blood
95
:
2346
-2351.
14
Juo, P., M. S. Woo, C. J. Kuo, P. Signorelli, H. P. Biemann, Y. A. Hannun, J. Blenis.
1999
. FADD is required for multiple signaling events downstream of the receptor Fas.
Cell Growth Differ.
10
:
797
-804.
15
Juo, P., C. J. Kuo, J. Yuan, J. Blenis.
1998
. Essential requirement for caspase-8/FLICE in the initiation of the Fas-induced apoptotic cascade.
Curr Biol.
8
:
1001
-1008.
16
Ting, A. T., F. X. Pimentel-Muinos, B. Seed.
1996
. RIP mediates tumor necrosis factor receptor 1 activation of NF-κB but not Fas/APO-1-initiated apoptosis.
EMBO J.
15
:
6189
-6196.
17
Sutton, V. R., D. L. Vaux, J. A. Trapani.
1997
. Bcl-2 prevents apoptosis induced by perforin and granzyme B, but not that mediated by whole cytotoxic lymphocytes.
J. Immunol.
158
:
5783
-5790.
18
Brenan, M., D. J. Rees.
1997
. Sequence analysis of rat integrin αE1 and αE2 subunits: tissue expression reveals phenotypic similarities between intraepithelial lymphocytes and dendritic cells in lymph [Published erratum appears in 1998 Eur. J. Immunol. 28: 3891].
Eur. J. Immunol.
27
:
3070
-3079.
19
Voisine, C., F. X. Hubert, B. Trinite, M. Heslan, R. Josien.
2002
. Two phenotypically distinct subsets of spleen dendritic cells in rats exhibit different cytokine production and T cell stimulatory activity.
J. Immunol.
169
:
2284
-2291.
20
Baker, S. J., E. P. Reddy.
1998
. Modulation of life and death by the TNF receptor superfamily.
Oncogene
17
:
3261
-3270.
21
Holler, N., R. Zaru, O. Micheau, M. Thome, A. Attinger, S. Valitutti, J.-L. Bodmer, P. Schneider, B. Seed, J. Tschopp.
2000
. Fas triggers an alternative, caspase-8-independent cell death pathway using the kinase RIP as effector molecule.
Nat. Immunol.
1
:
489
-495.
22
Jaattela, M., J. Tschopp.
2003
. Caspase-independent cell death in T lymphocytes.
Nat. Immunol.
4
:
416
-423.
23
Homann, D., A. Jahreis, T. Wolfe, A. Hughes, B. Coon, M. J. van Stipdonk, K. R. Prilliman, S. P. Schoenberger, M. G. von Herrath.
2002
. CD40L blockade prevents autoimmune diabetes by induction of bitypic NK/DC regulatory cells.
Immunity
16
:
403
-415.
24
Rissoan, M. C., T. Duhen, J. M. Bridon, N. Bendriss-Vermare, C. Peronne, B. de Saint Vis, F. Briere, E. E. Bates.
2002
. Subtractive hybridization reveals the expression of immunoglobulin-like transcript 7, Eph-B1, granzyme B, and 3 novel transcripts in human plasmacytoid dendritic cells.
Blood
100
:
3295
-3303.
25
Hubert, F. X., C. Voisine, C. Louvet, M. Heslan, R. Josien.
2004
. Rat plasmacytoid dendritic cells are an abundant subset of MHC class II+ CD4+CD11bOX62 and type I IFN-producing cells that exhibit selective expression of Toll-like receptors 7 and 9 and strong responsiveness to CpG.
J. Immunol.
172
:
7485
-7494.
26
Huang, F. P., N. Platt, M. Wykes, J. R. Major, T. J. Powell, C. D. Jenkins, G. G. MacPherson.
2000
. A discrete subpopulation of dendritic cells transports apoptotic intestinal epithelial cells to T cell areas of mesenteric lymph nodes.
J. Exp. Med.
191
:
435
-444.
27
Turnbull, E., G. MacPherson.
2001
. Immunobiology of dendritic cells in the rat.
Immunol. Rev.
184
:
58
-68.
28
Hubert, P., S. L. Giannini, A. Vanderplasschen, E. Franzen-Detrooz, N. Jacobs, J. Boniver, P. Delvenne.
2001
. Dendritic cells induce the death of human papillomavirus-transformed keratinocytes.
FASEB J.
15
:
2521
-2523.
29
Santini, S. M., C. Lapenta, M. Logozzi, S. Parlato, M. Spada, T. Di Pucchio, F. Belardelli.
2000
. Type I interferon as a powerful adjuvant for monocyte-derived dendritic cell development and activity in vitro and in Hu-PBL-SCID mice.
J. Exp. Med.
191
:
1777
-1788.
30
Vidalain, P. O., O. Azocar, B. Lamouille, A. Astier, C. Rabourdin-Combe, C. Servet-Delprat.
2000
. Measles virus induces functional TRAIL production by human dendritic cells.
J. Virol.
74
:
556
-559.
31
Yang, R., D. Xu, A. Zhang, A. Gruber.
2001
. Immature dendritic cells kill ovarian carcinoma cells by a FAS/FASL pathway, enabling them to sensitize tumor-specific CTLs.
Int. J. Cancer
94
:
407
-413.
32
Shibaki, A., S. I. Katz.
2001
. Activation through CD40 ligation induces functional Fas ligand expression by Langerhans cells.
Eur. J. Immunol.
31
:
3006
-3015.
33
Manna, P. P., T. Mohanakumar.
2002
. Human dendritic cell mediated cytotoxicity against breast carcinoma cells in vitro.
J. Leukocyte Biol.
72
:
312
-320.
34
Lu, G., B. M. Janjic, J. Janjic, T. L. Whiteside, W. J. Storkus, N. L. Vujanovic.
2002
. Innate direct anticancer effector function of human immature dendritic cells. II. Role of TNF, lymphotoxin-α1β2, Fas ligand, and TNF-related apoptosis-inducing ligand.
J. Immunol.
168
:
1831
-1839.
35
Alli, R., B. Savithri, S. Das, C. Varalakshmi, N. Rangaraj, A. Khar.
2004
. Involvement of NKR-P2/NKG2D in DC-mediated killing of tumor targets: indicative of a common, innate, target-recognition paradigm?.
Eur. J. Immunol.
34
:
1119
-1126.
36
Smyth, M. J., Y. Hayakawa, K. Takeda, H. Yagita.
2002
. New aspects of natural-killer-cell surveillance and therapy of cancer.
Nat. Rev. Cancer
2
:
850
-861.
37
Suss, G., K. Shortman.
1996
. A subclass of dendritic cells kills CD4 T cells via Fas/Fas-ligand- induced apoptosis.
J. Exp. Med.
183
:
1789
-1796.
38
Steinman, R. M., M. C. Nussenzweig.
2002
. Avoiding horror autotoxicus: the importance of dendritic cells in peripheral T cell tolerance.
Proc. Natl. Acad. Sci. USA
99
:
351
-358.
39
Belz, G. T., G. M. Behrens, C. M. Smith, J. F. Miller, C. Jones, K. Lejon, C. G. Fathman, S. N. Mueller, K. Shortman, F. R. Carbone, W. R. Heath.
2002
. The CD8α+ dendritic cell is responsible for inducing peripheral self-tolerance to tissue-associated antigens.
J. Exp. Med.
196
:
1099
-1104.
40
Steinman, R. M., S. Turley, I. Mellman, K. Inaba.
2000
. The induction of tolerance by dendritic cells that have captured apoptotic cells.
J. Exp. Med.
191
:
411
-416.
41
Harshyne, L. A., S. C. Watkins, A. Gambotto, S. M. Barratt-Boyes.
2001
. Dendritic cells acquire antigens from live cells for cross-presentation to CTL.
J. Immunol.
166
:
3717
-3723.