IFN-producing killer dendritic cells (IKDC) were initially described as B220+CD11c+CD3NK1.1+ tumor-infiltrating cells that mediated part of the antitumor effects of the combination therapy with imatinib mesylate and IL-2. In this study, we show their functional dependency on IL-15 during homeostasis and inflammatory processes. Trans-presentation of IL-15 by IL-15Rα allows dramatic expansion of IKDC in vitro and in vivo, licenses IKDC for TRAIL-dependent killing and endows IKDC with immunizing potential, all three biological attributes not shared by B220NK cells. However, IL-15 down-regulates the capacity of IKDC to induce MHC class I- or II-restricted T cell activation in vitro. Trans-presentation of IL-15 by IL-15Rα allows IKDC to respond to TLR3 and TLR4 ligands for the production of CCL2, a chemokine that is critical for IKDC trafficking into tumor beds (as described recently). We conclude that IKDC represent a unique subset of innate effectors functionally distinguishable from conventional NK cells in their ability to promptly respond to IL-15-driven inflammatory processes.

Interferon-producing killer dendritic cells (IKDC)4 represent a rare but unique entity sharing hybrid features in-between dendritic and NK cells (1, 2, 3, 4, 5, 6, 7). They could be found in trace amounts in all lymphoid organs at the steady-state but accumulated during certain inflammatory processes, such as tumor regression under the influence of imatinib mesylate (IM) + IL-2 (2) or Listeria infection (1). In such circumstances, IKDC expressed high levels of MHC class II molecules and maintained CD11c expression while lacking CD19 and CD40 molecules, thereby diverging from bona fide B cells. Chan et al. (1) could demonstrate that IKDC derived from lymph nodes of BALB/c mice behaved as conventional myeloid DC in their capacity to traffic to secondary lymphoid organs and to present Ag to MHC class II-restricted CD4+ TCR transgenic (Tg) T cells. Nevertheless, IKDC resembled NK cells in that they exhibited potent killing activity during tumor regression (2, 3, 8, 9, 10). Thus, NK cell experts may consider IKDC as a subset of immature NK cells (11, 12, 13, 14), that could display molecules such as B220, CD11c (15), or even MHC class II upon activation. In an attempt to delineate a functional hierarchy between IKDC and NK cells, we sought to determine the regulatory cytokine dictating homeostasis and activation of IKDC in vitro and in vivo. Although IL-15 is crucial for the homeostasis of both innate effectors at the steady state, trans-presentation of IL-15 endowed IKDC with unique biological functions that are not shared by conventional B220NK cells. Indeed, trans-presentation of IL-15 by IL-15Rα licensed IKDC for proliferation in vitro and in vivo, for killing in a TRAIL-dependent manner, and, finally, for induction of antitumor immunity. Therefore, these data underscore the fundamental role of IL-15 in IKDC biology, suggesting a potential involvement of IKDC not only in the control of tumor growth, but also in various inflammatory processes.

Female C57BL/6 wild-type (WT) mice were obtained from the Centre d’ Elevage Janvier (Le Genest St. Isle, France) and used at 6–10 wk of age. IFN type 1R−/−, CCL2/MCP-1−/−, and CD45.1+ mice backcrossed on a C57BL/6 background were provided by Centre d’Elevage d’Orléans (Centre de distribution, typage et archivage animal Orléans, France). IL-15Rα−/−, IL-15−/−, IL-2/IL-15−/−, IL-2−/−, and IL-15 Tg mice were backcrossed on a C57BL/6 six to eight times and maintained at the animal facility of S. Bulfone-Paus (Research Center Borstel, Borstel, Germany). IL-2Rγ−/− × Rag2−/− were kindly provided by E. Vivier, Centre d‘Immunologie de Marseille, France. Tg OTI and OTII mice were a kind gift by O. Lantz (Institut Curie, Paris, France). Animals were all maintained according to the Animal Experimental Ethics Committee Guidelines. B16F10 is a melanoma cell line syngeneic of C57BL/6 (provided by M. T. Lotze, University of Pittsburgh, PA) and was cultured in RPMI 1640 (Invitrogen) with 10% heat-inactivated FBS enriched with 5% l-glutamine, non-essential amino acids, sodium pyruvate, and antibiotics. MS-5-feeder cell lines (provided by W. Vainchenker, IGR, Villejuif, France) were cultured in IMDM (Sigma-Aldrich) containing 10% heat-inactivated FBS, 5% l-glutamine, sodium pyruvate, and antibiotics.

FACS analyses of IKDC, IKDC15, and NK cells were performed using allophycocyanin-conjugated anti-CD11c (HL3), PE-Cy7-conjugated anti-NK1.1 mAb (PK136), allophycocyanin-Cy7-conjugated anti-B220 mAb (RA3–6B2), and PerCP-conjugated anti-CD3 mAb (17A2). We further stained with PE- or FITC-conjugated mAb to examine the following molecules: MHC class II (AF6-120.1), CD40 (3/23), CD80 (16.10A1), CD86 (GL1), CD4 (GK1.5), CD8 (53-6.7), CD11b (M1/70), CD122 (TM-β), CD49b (Dx5), NKG2D (CX5), ckit (ACK2), CD69 (H1.2F3), CD19 (1D3), or CD27 (LG.3A10). Abs were purchased from BD Pharmingen or eBioscience. The anti-NKp46 mAb was kindly provided by E. Vivier, CIML, France. Cells were preincubated with Fc block for 20 min (CD16/CD32, 2.4G2; BD Pharmingen) in 2% FBS and 2% mouse serum and afterward stained for 20 min at 4°C with the different Abs at 1/200. Immediately before FACS analysis, 4′,6-diamidino-2-phenylindole dihydrochloride (Sigma-Aldrich) was added. FACS analysis was performed by LSRII (BD Biosciences) using FACSDiva software (BD Biosciences) and CellQuestPro software (BD Biosciences) or FlowJo (Treestar).

NK and IKDC cells from C57BL/6 mice were sorted on a Mo-Flo instrument (DAKO) in two steps. First, we enriched NK1.1+ cells. Second, we sorted CD3CD19CD11cintB220+NK1.1+ cells (defined as “IKDC”) and CD3CD19CD11c+/−B220NK1.1+ cells (defined as B220NK cells henceforth). Cells were stained with FITC-conjugated anti-CD3 and CD19 mAb, PE-conjugated anti-CD11c mAb, PE-Cy7-conjugated anti-NK1.1 mAb, and PB-conjugated anti-B220 mAb. The purity of cell separation exceeded 97%. Purified NK cells and IKDC were then used for functional experiments.

Freshly cell sorted IKDC were cultured in the presence of murine stromal cells MS-5 (16). One or 2 days before coculture with IKDC, MS-5 cells were plated in round-bottom 96-well plates (7500 cells per well). Cultures of IKDC were initiated by seeding 104 freshly sorted IKDC in MS-5 precoated 96-well plates in DMEM (Invitrogen) culture medium containing 4500 mg/l of glucose, 5% l-glutamin, pyruvate, and enriched with antibiotics, 10% Bovine Growth serum (Lot no. ANB 18298, HyClone), and 20 ng recombinant murine (rm) IL-15/ml (R&D Systems). Upon expansion of IKDC, stromal cells and culture medium were replaced twice a week. Inhibition of IL-15 trans-presentation was performed using anti-mIL-15Rα Ab (AF551, R&D Systems) at a concentration of 20–30 μg/ml in the presence of 20 ng/ml IL-15. Limiting dilution assays were also initiated in 96-well plates (1 cell per well) using the automated cell device unit. Although B220NK cells could not proliferate ex vivo in similar conditions as IKDC, we could maintain NK cells at high concentrations (5 × 105/ml) for 7 days on MS-5 feeders and rIL-15 to allow fair comparisons with IKDC.

A total of 105 freshly sorted NK cells and IKDC or IKDC15 (obtained at day 7 of ex vivo expansion) or NK and IKDC stimulated with rmIL-15 (20 ng/ml; R&D Systems) for 24 h were further incubated with LPS at 100 ng/ml (InvivoGen) or CpG oligodeoxynucleotide (ODN) 1668 (MWG Biotech) at 5 μg/ml. These in vitro cultures were performed in 200 μl RPMI (Invitrogen) 10% FBS (Invitrogen) in 96 round-bottom well plates. After 24–36 h, cell supernatants were collected and commercial LUMINEX kits were used to determine chemokine and cytokine profiles (used according to the manufacturer’s conditions, Linco Research/BioSource International).

FACS sorted 105 CD4+ resting OTII lymphocytes purified from naive OTII Tg mice were incubated at various effector/T cell ratios (1:1, 1:5, 1:20, and 1:100) with different effector cells (such as resting IKDC, B220NK, bone marrow-derived DC (BMDC), IKDC15, or NK15 cells) after a 24-h coculture of effector cells with B16 tumor cells in the presence of 1 mg/ml OVA protein followed by extensive washing (three times in PBS 1× to remove resting traces of OVA protein). After a 20-h incubation period, T cells were stained with anti-CD3, anti-CD4, anti-Vα2, and anti-CD69 Ab and analyzed by FACS.

51Cr release killing assay was performed according to standard protocols using 2 × 103 Na251CrO4-labeled B16F10 tumor cells (T) incubated with various E:T ratios (1:1, 5:1, 10:1, 15:1, and 30:1) of effector (E) cells (NK vs IKDC stimulated or not with rmIL-15 in trans-presentation) for 4, 8, or 12 h. Supernatants were harvested for the measurement of chromium release (E) using γ emission counting (Topcount NXT, Packard Instrument). Spontaneous 51Cr release (S) was counted in target cells alone, maximal 51Cr release (M) from target cells treated with 5% alkyltrimethylammonium bromide and specific lysis was calculated according to the following: % lysis = (E-S)/(M-S)×100. As an additional method, crystal violet assay was used. Effector and target cells were mixed at different ratios for 24 or 48 h. Live tumor cells were revealed using a crystal violet staining as previously reported (17). Cocultures of E:T were performed in the presence of neutralizing anti-TRAIL Ab (N2B2, provided by H. Yagita, Juntendo University School of Medicine, Tokyo, Japan), commercial anti-FasL mAb (CD95L; eBioscience) at 10 μg/ml, or EGTA (Sigma-Aldrich) at 1 mM or Concanamycin A (Sigma-Aldrich) at 20 nM.

WT C57BL/6, IL-15Rα−/− mice were treated with IL-2 alone (100,000 IU i.p. twice a day for 4 days), or rmIL-15 (R&D Systems at 0.5 μg i.p. daily for 4 days) or CpG ODN 1668 (MWG Biotec at 5 μg i.p. daily for 4 days). Mice received an i.p. injection of BrdU (100 μg/100 μl PBS) 1 day before sacrifice. Spleen cells were harvested and processed according to the manufacturer’s protocol (BD Biosciences BrdU Flow kit). Briefly, cells were stained for surface Ags, fixed, and permeabilized. DNase digestion followed by staining with anti-BrdU mAb were performed before flow cytometry analyses.

Cultured IKDC15 were resuspended in RPMI 1640, washed, and 5 × 104 cells were gently spread onto a slide coated with poly-l-lysine (Sigma-Aldrich). Slides were incubated for 45 min at 37°C. Cells were fixed in 4% paraformaldehyde and permeabilized with 0.1% SDS. After 20 min of blocking in 10% FBS and washing, cells were stained with the appropriate anti-MHC class II (NIMR-4; Southern Biotechnology Associates), anti-Perforin, and anti-Granzyme B mAbs (BD Pharmingen) in PBS containing 1% BSA for 1 h. Next, slides were extensively washed and incubated with the appropriate secondary Ab (Alexa Fluor 488 goat anti-rat IgG) for 1 h and, after an additional washing step, with DNA-labeling Topro3 (Invitrogen) for 10 min. Finally, 0.17-mm cover glasses were mounted on the slides. Stacks of confocal images were collected with a Zeiss LSM 510 laser scanning confocal microscope (Carl Zeiss) using a × 63 1.4 NA apochromat plan objective. Z-projection of slices and image analyses were performed using Zeiss LSM Image examiner software.

RNA was extracted using TRIzol reagent (Invitrogen) according to the manufacturer’s instructions. Low cell number samples were precipitated in the presence of 10 μg/sample GlycoBlue (Ambion). After RNA purification, samples were treated with DNase to remove contaminating genomic DNA (DNaseI Amplification grade, 18068; Invitrogen) and Superscript II Reverse transcriptase (Invitrogen). Gene specific primers were purchased from NBS BIOTECH Scrl; sequences and detailed amplification protocols are available upon request. The iQ SYBR Green Supermix (Bio-Rad) was used to run relative quantitative real-time PCR of the samples according to the manufacturer’s instructions. Reactions were run in triplicate on an iCycler (Bio-Rad) and generated products analyzed with the iCycler iQ Optical System software (Version 3.0a; Bio-Rad). Gene expression was routinely normalized both based on β-Actin mRNA and 18S rRNA contents with overlapping results. The amounts of target mRNAs are expressed in arbitrary units calculated as the relative change compared with spleen samples. Data are displayed as 2−ddCt values and are representative of at least three independent experiments.

Aberrant values were excluded using Dixon’s test. Normality of distributions was assessed using the Shapiro-Wilk’s test. Normal distributions were compared by the Student’s t test; non-normal samplings were compared using the Mann-Whitney test. Statistical analyses of survival curves were performed using Log-rank (Mantel Cox) test. Values of p inferior to 0.05 were considered significant. All tests were done using Prism 5 software (GraphPad).

IKDC were previously described as CD11cintB220+NK1.1+ cells mediating the antitumor effects of the combination therapy with IM and IL-2. IKDC represent ∼2% of bone marrow CD11c+ cells and 1–2% of spleen derived-CD11c+ in resting C57BL/6 mice and increased by 4-fold during the combination therapy with IM + IL-2 (2). Phenotype wise, IKDC are a specific cell population coexpressing CD11c, B220, NK1.1, and NKp46 (a recently described NK cell marker) (18) (Fig. 1 A). To avoid possible contamination with plasmacytoid DC, conventional DC. B lymphocytes, or NK cells, we sorted IKDC in two steps. First, we performed a preselection of CD3CD19NK1.1+ cells. Second, we gated on CD11cintB220+NK1.1+ cells (defined as “IKDC”) and CD11c+/−B220NK1.1+ cells (defined as B220NK cells henceforth).

FIGURE 1.

IL-15 drives homeostasis and inflammation-induced proliferation of IKDC. A, Phenotypic characterization of IKDC. Flow cytometry analyses were performed after staining splenocytes with FITC anti-CD3/CD19, PE anti-CD11c, PerCP anti-B220, PE-Cy7 anti-NK1.1, and Alexa 647 anti-NKp46 mAb. B, Adoptive transfer of IKDC in vivo. IKDC were purified from spleen and bone marrow of CD45.2+ mice and transferred into lethally irradiated CD45.1+ mice, which were rescued with 1 × 105 CD45.1+×CD45.2+ bone marrow cells. Flow cytometry was performed on day 12 on splenocytes using FITC anti-CD45.1, PE anti-CD45.2, PerCP anti-B220, allophycocyanin anti-CD11b, PE-Cy7 anti-NK1.1, and PB anti-CD11c mAb. C, Enumeration of IKDC and NK cells in mice deficient in γ common chain-dependent cytokines. Flow cytometry analyses were performed on splenocytes derived from different mouse genetic backgrounds after staining as shown in A. IKDC were defined as CD3CD11c+B220+NK1.1+, NK cells as CD3CD11c+/−B220NK1.1+ and B cells as the whole B220+ population. D, Enumeration of IKDC and NK cells in C57BL/6 mice after different stimuli in vivo. Flow cytometry was performed as described above. E, BrdU incorporation in IKDC and NK cells following stimulation in vivo. WT or IL-15Rα loss of function mice were treated with different stimuli and injected with BrdU before flow cytometry analysis of splenocytes according to protocols described in Materials and Methods. The mean percentages ± SEM of BrdU incorporation in 7–10 mice are presented; (*) indicates statistically significant differences at 95% confidence between IL-15Rα−/− compared with WT C57BL/6 mice. ND, not determinable.

FIGURE 1.

IL-15 drives homeostasis and inflammation-induced proliferation of IKDC. A, Phenotypic characterization of IKDC. Flow cytometry analyses were performed after staining splenocytes with FITC anti-CD3/CD19, PE anti-CD11c, PerCP anti-B220, PE-Cy7 anti-NK1.1, and Alexa 647 anti-NKp46 mAb. B, Adoptive transfer of IKDC in vivo. IKDC were purified from spleen and bone marrow of CD45.2+ mice and transferred into lethally irradiated CD45.1+ mice, which were rescued with 1 × 105 CD45.1+×CD45.2+ bone marrow cells. Flow cytometry was performed on day 12 on splenocytes using FITC anti-CD45.1, PE anti-CD45.2, PerCP anti-B220, allophycocyanin anti-CD11b, PE-Cy7 anti-NK1.1, and PB anti-CD11c mAb. C, Enumeration of IKDC and NK cells in mice deficient in γ common chain-dependent cytokines. Flow cytometry analyses were performed on splenocytes derived from different mouse genetic backgrounds after staining as shown in A. IKDC were defined as CD3CD11c+B220+NK1.1+, NK cells as CD3CD11c+/−B220NK1.1+ and B cells as the whole B220+ population. D, Enumeration of IKDC and NK cells in C57BL/6 mice after different stimuli in vivo. Flow cytometry was performed as described above. E, BrdU incorporation in IKDC and NK cells following stimulation in vivo. WT or IL-15Rα loss of function mice were treated with different stimuli and injected with BrdU before flow cytometry analysis of splenocytes according to protocols described in Materials and Methods. The mean percentages ± SEM of BrdU incorporation in 7–10 mice are presented; (*) indicates statistically significant differences at 95% confidence between IL-15Rα−/− compared with WT C57BL/6 mice. ND, not determinable.

Close modal

Importantly, the expression of B220 molecules on bona fide B220NK cells did not appear to represent a marker of activation (12, 13, 19) because a 24–48 h stimulation of NK cells with cytokines (such as IL-2, IL-12, IL-15, and IL-18) and/or maturing DC were not sufficient to convert B220NK cells into B220+IKDC counterparts (Table I). Conversely, IKDC did not convert or differentiate into B220NK cells neither in vitro (Table I) nor in vivo (Fig. 1,B). Indeed, we performed an adoptive transfer of FACS sorted IKDC from CD45.2 donor mice into irradiated congenic CD45.1 C57BL/6 recipient mice. To rescue recipient hosts, coinjection of bone marrow derived from CD45.1×CD45.2 chimera was performed in parallel. IKDC were cell sorted either from spleen or bone marrow. A similar number of CD45.2+IKDC could be recovered in the spleen at day 12, whether originating from spleen or bone marrow (Fig. 1,B). IKDC did not lose CD11c nor B220 cell surface markers and did not acquire LyC6/Gr1, CD4, or CD8α molecules (Fig. 1 B and our unpublished data). At later time points (3 wk), CD45.2+IKDC were almost undetectable (not shown). Therefore, these data support the notion that IKDC are terminally differentiated cells.

Table I.

Flow cytometric analyses of cytokine-stimulated B220NK and IKDC cellsa

Cytokine 1Cytokine 224-h Duration48-h Duration
B220 (%)Class II (%)B220 (%)Class II (%)
NKIKDCNKIKDCNKIKDCNKIKDC
IL-2  0.25 96, 39 0.43 1.88 0.72 94, 80 1.29 3, 63 
 IL-12 0.79 97, 38 1.56 1.24 0.82 96, 79 0.85 2, 99 
 IL-18 0.23 96, 46 0.43 2.35 1.67 92, 23 1.61 5, 22 
 IFN-α 0.51 97, 08 0.27 1.05 0.54 95, 65 0.63 3, 2 
IL-15  0.34 96, 32 1.33 3.08 0.24 97, 64 0.93 5, 12 
 IL-12 0.52 96, 83 1.34 2.06 0.46 95, 93 1.28 2, 59 
 IL-18 0.78 97, 41 1.15 2.27 1.35 94, 55 0.64 6, 09 
 IFN-α 1.11 97, 55 1.26 1.17 0.61 95, 52 0.98 3, 47 
BMDC IL-15 0.89 97, 18 1.02 11, 03 1.32 95, 84 0.87 7, 35 
Cytokine 1Cytokine 224-h Duration48-h Duration
B220 (%)Class II (%)B220 (%)Class II (%)
NKIKDCNKIKDCNKIKDCNKIKDC
IL-2  0.25 96, 39 0.43 1.88 0.72 94, 80 1.29 3, 63 
 IL-12 0.79 97, 38 1.56 1.24 0.82 96, 79 0.85 2, 99 
 IL-18 0.23 96, 46 0.43 2.35 1.67 92, 23 1.61 5, 22 
 IFN-α 0.51 97, 08 0.27 1.05 0.54 95, 65 0.63 3, 2 
IL-15  0.34 96, 32 1.33 3.08 0.24 97, 64 0.93 5, 12 
 IL-12 0.52 96, 83 1.34 2.06 0.46 95, 93 1.28 2, 59 
 IL-18 0.78 97, 41 1.15 2.27 1.35 94, 55 0.64 6, 09 
 IFN-α 1.11 97, 55 1.26 1.17 0.61 95, 52 0.98 3, 47 
BMDC IL-15 0.89 97, 18 1.02 11, 03 1.32 95, 84 0.87 7, 35 
a

B220NK (CD3CD19B220NK1.1+) and IKDC (CD3CD19B220+NK1.1+) were cell-sorted and incubated in vitro in the presence of different cytokines or with BMDC (1:1). After 24 or 48 h, cells were harvested and analyzed for B220 and MHC class II expression. This table indicates the percentage of B220+ or I-A/I-E+ cells amongst total NK1.1+ cells. IL-2 is used at 50,000 IU/ml, IL-15 at 20 ng/ml, IL-12 at 10 ng/ml, IL-18 at 2.5 ng/ml, and IFN-α at 2,500 IU/ml. Cells were plated at 100,000 cells per well in 96-well, round-bottomed plates in 100 μl of complete medium. These experiments have been performed three times with similar results.

IKDC might be considered as a subpopulation of the so-called “NKDC” (containing all CD11c+NK1.1+ cells) (20, 21). However, it is noteworthy that the B220 fraction of NKDC was not significantly different from bona fide NK cells for all the hallmark criteria that the manuscript will describe (our unpublished data).

IL-2Rγ-chain-dependent cytokines, such as IL-15, are critical to promote lymphoid homeostasis and, more specifically, to maintain survival and proliferation of NK cells. We have previously reported that CD11c+CD49b+B220+ cells (described as such by Chan et al. in BALB/c littermates) (1) could be found in old (>3 mo) Rag−/−×IL-2Rγ-chain−/− mice. However, those cells did not express NK1.1 molecules (not shown). An almost complete deprivation in IKDC (as defined in Fig. 1,A) was found in IL-15Rα−/− and IL-15−/− animals, supporting that IL-15 is a requirement for the differentiation of not only B220NK cells (which decreased by 10-fold in IL-15−/− or IL-15Rα−/− mice) but also of IKDC in vivo (Fig. 1,C). Moreover, IL-2 plays a redundant role in both the B220NK and IKDC developmental pathways (Fig. 1,C). Surprisingly, IL-15 Tg animals did not contain enhanced numbers of IKDC at the steady state (Fig. 1,C). In contrast, administration of rIL-15 was associated with the accumulation of IKDC in the spleen (Fig. 1,D) resulting from their proliferation, as assessed by incorporation of BrdU in IKDC more than in B220NK cells (Fig. 1,E). IL-15 but also IL-2 could promote a 2–4-fold accumulation of IKDC in the spleen (Fig. 1,D). About one third of IKDC underwent cell division during exogenous administration of IL-15 or IL-2 in WT mice while <20% did in the B220NK cell fraction (Fig. 1,E). Importantly, IKDC proliferation induced by TLR9 ligands (CpG 1668) was IL-15Rα-dependent in contrast to B220NK cells (as assessed by comparing BrdU incorporation in WT vs IL-15Rα−/− mice, Fig. 1,E). However, although TLR9L induced IKDC proliferation, IKDC did not appear to accumulate in the spleen following administration of CpG ODN (Fig. 1 D).

Hence, IKDC critically depended upon IL-15/IL-15Rα for their homeostasis and CpG driven-proliferation in vivo.

There are trace numbers of IKDC in lymphoid organs of naive animals (∼50,000/spleen). Based on the IL-15/IL-15Rα requirement for IKDC differentiation in vivo, we set up culture conditions allowing ex vivo IKDC proliferation and/or differentiation. Immunoblot analysis indicated that IKDC do not harbor IL-15Rα in contrast to B220NK cells or DC (Fig. 2,A). However, cell surface expression of IL-15Rα was detectable only on MS-5 stromal cells and DC using FACS analyses (Fig. 2 B). We used IL-15Rα expressing MS-5 to test the hypothesis of the role of trans-presentation of IL-15 by IL-15Rα in the biology of IKDC. “Trans-presentation” of IL-15 defines a phenomenon by which IL-15 is presented by IL-15Rα on a bystander cell to neighboring cells lacking IL-15Rα and responding through the IL-2/15Rβ and γ-chains (22).

FIGURE 2.

Trans-presentation of IL-15 allows ex vivo expansion of IKDC. IKDC fail to express IL-15Rα. A, Immunoblot analysis of IKDC, NK, conventional BMDC (grown in GM-CSF + IL-4), and MS-5 lysates for IL-15Rα expression (anti-IL15Rα at a dilution of 1/200, N-19; Santa Cruz Technology). B, Flow cytometry analyses of IKDC, NK, BMDC, and MS-5 using biotinylated anti-IL-15Rα Ab (plain line) and the isotype control Ab (dotted line). Ex vivo expansion of spleen IKDC. C, A total of 104 freshly sorted IKDC were cultured with or without MS-5 feeder cells in the presence of rmIL-15 (20 ng/ml) or rhIL-2 (50,000 U/ml) or both. D, IKDC and B220NK were cultured under the above mentioned conditions (rmIL-15 + MS-5). Proliferation of IKDC has further been tested in the presence or absence of transwells (//) physically separating MS-5 and by adding control mAb or neutralizing anti-IL-15Rα mAb. The graphs show cell numbers enumerated using trypan blue exclusion assays followed by FACS analysis to verify the surface expression of CD11c/B220/NK1.1 by IKDC. One representative experiment is shown as means + SEM. The experiments were reproduced >3 times with identical results. E, Histogramms of cloning efficiency of IKDC and B220NK cells after single cell sorting cultured on MS5 + rmIL-15 for 7–10 days. The percentage of positive wells is indicated for each culture condition. The experiment has been performed three times with similar results.

FIGURE 2.

Trans-presentation of IL-15 allows ex vivo expansion of IKDC. IKDC fail to express IL-15Rα. A, Immunoblot analysis of IKDC, NK, conventional BMDC (grown in GM-CSF + IL-4), and MS-5 lysates for IL-15Rα expression (anti-IL15Rα at a dilution of 1/200, N-19; Santa Cruz Technology). B, Flow cytometry analyses of IKDC, NK, BMDC, and MS-5 using biotinylated anti-IL-15Rα Ab (plain line) and the isotype control Ab (dotted line). Ex vivo expansion of spleen IKDC. C, A total of 104 freshly sorted IKDC were cultured with or without MS-5 feeder cells in the presence of rmIL-15 (20 ng/ml) or rhIL-2 (50,000 U/ml) or both. D, IKDC and B220NK were cultured under the above mentioned conditions (rmIL-15 + MS-5). Proliferation of IKDC has further been tested in the presence or absence of transwells (//) physically separating MS-5 and by adding control mAb or neutralizing anti-IL-15Rα mAb. The graphs show cell numbers enumerated using trypan blue exclusion assays followed by FACS analysis to verify the surface expression of CD11c/B220/NK1.1 by IKDC. One representative experiment is shown as means + SEM. The experiments were reproduced >3 times with identical results. E, Histogramms of cloning efficiency of IKDC and B220NK cells after single cell sorting cultured on MS5 + rmIL-15 for 7–10 days. The percentage of positive wells is indicated for each culture condition. The experiment has been performed three times with similar results.

Close modal

Accordingly, trans-presentation of rIL-15 by MS-5 to IKDC was successful and mandatory to promote ex vivo expansion (up to 10–30-fold) of spleen or bone marrow derived-IKDC within 7–12 days (Fig. 2, C and D). Indeed, IKDC proliferated only in the condition of IL-15 trans-presentation, neither on MS-5 alone, nor in rIL-15 alone (Fig. 2,C). IKDC cultured in rIL-15 in the absence of MS-5 feeder cells expanded by 5-fold by day 7 but lost their proliferative potential afterward (Fig. 2,C). IL-15 could not be substituted by rIL-2 and there was no additive effect with the combination of rIL-2 or rIL-18 + rIL-15 (Fig. 2 C and unpublished data).

IKDC proliferation on the MS-5 stromal cells in the presence of rIL-15 was dependent on cell to cell contact, as shown by transwell experiments (Fig. 2,D), and could be abrogated by anti-IL-15Rα neutralizing Abs (Fig. 2,D). Interestingly, this culture procedure did not allow the expansion of B220NK cells (Fig. 2,D), although NK cell proliferation also depended on trans-presentation of IL-15 in vivo (Fig. 1,E). The cloning efficiency of IKDC on MS-5 + rIL-15 as determined by limiting dilution analysis after single-cell sorting was ∼20% (Fig. 2,E). Clones derived from CD117+ or CD117 bone marrow or spleen IKDC expanded exponentially in culture to colonies of 3 × 103 cells by 7 days (not shown). Under these culture conditions, cloning efficiency of NK cells was not significant (Fig. 2 E).

Ex vivo expanded IKDC (referred to as “IKDC15” henceforth) acquired a large blastic cytoplasm and contained numerous granules and vacuoles (Fig. 3,A). At days 7–10 after expansion, IKDC15 still failed to harbor membrane expression of IL-15Rα (Fig. 3,B) but maintained the hallmark criteria of IKDC such as the expression of CD11c, B220, and NK1.1 (as well as CD11b, CCR7, and CD62L, Fig. 3,C). Interestingly, IKDC lost the expression of MHC class II molecules, both at the mRNA and protein levels (Figs. 4,A and 3,C). Accordingly, IKDC15 lost their capacity to activate MHC class II-restricted OTII cells in vitro. Following pulsing with soluble OVA protein, neither IKDC15 nor NK15 could induce CD69 expression on naive OVA-specific I-Ab-restricted Tg OTII cells (Fig. 4,B), while IKDC (but not NKB220 cells) and BMDC could do so (Fig. 4 C). Similarly, IKDC15 do not have the capacity to activate MHC class I-restricted OTI cells in vitro (not shown).

FIGURE 3.

Morphological and phenotypical analysis of expanded IKDC15. A, Morphological changes of cultured IKDC15. Cytospins of freshly sorted cells or expanding IKDC were stained with May-Grünwald-Giemsa (Sigma-Aldrich) and analyzed on a Leica Microsystems DMLB microscope with ×800 magnification, using a Camera SONY 3CCD and TRIBVN ICS Version 1.4 software. Representative photographs are depicted. B, IKDC15 do not express IL-15Rα. FACS analysis on IKDC15 as described in Fig. 2 B. Anti-IL-15Rα Ab (plain line) and isotype control Ab (dotted line). C, Phenotype of IKDC15 compared with freshly sorted IKDC. FACS analyses were performed on IKDC15 at day 7 of expansion compared with freshly sorted IKDC using the Abs listed in Materials and Methods. Mean fluorescence intensity fold increase and percentages of positive cells compared with the isotype matched control Ab are shown for each staining. A representative staining is shown. The experiments have been performed three times with identical results.

FIGURE 3.

Morphological and phenotypical analysis of expanded IKDC15. A, Morphological changes of cultured IKDC15. Cytospins of freshly sorted cells or expanding IKDC were stained with May-Grünwald-Giemsa (Sigma-Aldrich) and analyzed on a Leica Microsystems DMLB microscope with ×800 magnification, using a Camera SONY 3CCD and TRIBVN ICS Version 1.4 software. Representative photographs are depicted. B, IKDC15 do not express IL-15Rα. FACS analysis on IKDC15 as described in Fig. 2 B. Anti-IL-15Rα Ab (plain line) and isotype control Ab (dotted line). C, Phenotype of IKDC15 compared with freshly sorted IKDC. FACS analyses were performed on IKDC15 at day 7 of expansion compared with freshly sorted IKDC using the Abs listed in Materials and Methods. Mean fluorescence intensity fold increase and percentages of positive cells compared with the isotype matched control Ab are shown for each staining. A representative staining is shown. The experiments have been performed three times with identical results.

Close modal
FIGURE 4.

IKDC15 lose their MHC class II –restricted Ag presenting capacities in vitro. A, Down-regulation of MHC class II mRNA in IKDC15. Quantitative RT-PCR was performed on freshly sorted IKDC and on IKDC15 (at day 7 of expansion). B and C, Activation of OTII Tg T cells in contact with IKDC but not IKDC15 in vitro. FACS sorted 105 CD4+ resting OTII lymphocytes purified from naive OTII Tg mice were added at various effector/T cell ratios (as indicated) to different effector cells (either resting IKDC, B220NK, immature BMDC (C), or trans-IL-15-activated IKDC or NK cells (B)) or to a control without APC (B) after a 24-h coculture with B16 tumor cells in the presence of 1 mg/ml OVA protein. After a 20-h incubation period, cocultures were stained with anti-CD3, anti-CD4, anti-Vα2, and anti-CD69 Ab and analyzed by FACS. A representative experiment is depicted of three yielding identical results. The mean + SEM of the % of CD69 OTII cells is indicated on the graphs.

FIGURE 4.

IKDC15 lose their MHC class II –restricted Ag presenting capacities in vitro. A, Down-regulation of MHC class II mRNA in IKDC15. Quantitative RT-PCR was performed on freshly sorted IKDC and on IKDC15 (at day 7 of expansion). B and C, Activation of OTII Tg T cells in contact with IKDC but not IKDC15 in vitro. FACS sorted 105 CD4+ resting OTII lymphocytes purified from naive OTII Tg mice were added at various effector/T cell ratios (as indicated) to different effector cells (either resting IKDC, B220NK, immature BMDC (C), or trans-IL-15-activated IKDC or NK cells (B)) or to a control without APC (B) after a 24-h coculture with B16 tumor cells in the presence of 1 mg/ml OVA protein. After a 20-h incubation period, cocultures were stained with anti-CD3, anti-CD4, anti-Vα2, and anti-CD69 Ab and analyzed by FACS. A representative experiment is depicted of three yielding identical results. The mean + SEM of the % of CD69 OTII cells is indicated on the graphs.

Close modal

Altogether, IL-15 trans-presentation allowed the selective ex vivo expansion of IKDC. However, IKDC15 seem to be poor APCs in both classical MHC pathways.

Resting IKDC did not express basal levels of mRNA encoding any of the 11 mouse TLR and could not respond to TLR stimuli (Fig. 5,A and our unpublished data). However, trans-presentation of IL-15 significantly up-regulated the transcription levels of TLR3 and TLR4 in IKDC (Fig. 5,A). Thus, IKDC15 acquired the capacity to respond to TLR4 ligands (LPS) by producing high levels of CCL2 (MCP-1) and CXCL1 (KC-GROα) (Fig. 5,B). It is important to note that this responsiveness of IKDC required trans-presentation of IL-15 (not shown). Importantly, IKDC15 acquired the capacity to produce CCL2 even after 2–3 days culturing in IL-15/MS-5 (not shown). In addition, IKDC15 responded to TLR3 ligands (poly(I:C)) for the production of CCL5 (RANTES, Fig. 5 B). Moreover, trans-presentation of IL-15 lead to responsiveness of IKDC to IL-2 and IFNα for the secretion of high amounts of CCL2 and CCL5 (23).

FIGURE 5.

Trans-presentation of IL-15 licenses IKDC to respond to TLR3 and TLR4 ligands. A, IL-15 trans-presentation induced TLR3 and TLR4 expression in IKDC. Quantitative RT-PCR was performed on freshly sorted IKDC and on IKDC15 (at day 7 of expansion). B, Chemokine release by IKDC and IKDC15. After 24-h stimulation with medium, TLR3L (poly (I:C)), TLR4L (LPS), or TLR9L (CpG) multiplex analysis of chemokine release were performed on IKDC stimulated or not with trans-IL-15 presentation. IKDC0 meant freshly sorted cells without ex vivo stimulation. The experiments were performed at least twice with identical results.

FIGURE 5.

Trans-presentation of IL-15 licenses IKDC to respond to TLR3 and TLR4 ligands. A, IL-15 trans-presentation induced TLR3 and TLR4 expression in IKDC. Quantitative RT-PCR was performed on freshly sorted IKDC and on IKDC15 (at day 7 of expansion). B, Chemokine release by IKDC and IKDC15. After 24-h stimulation with medium, TLR3L (poly (I:C)), TLR4L (LPS), or TLR9L (CpG) multiplex analysis of chemokine release were performed on IKDC stimulated or not with trans-IL-15 presentation. IKDC0 meant freshly sorted cells without ex vivo stimulation. The experiments were performed at least twice with identical results.

Close modal

Trans-presentation of IL-15 allowed IKDC to respond to TLR3 and TLR4 ligation.

The basal transcription level of the killing machinery (perforin/granzyme B/FasL), which was detectable in resting NK cells (not shown), was absent in resting IKDC (Fig. 6,A). However, upon IL-15 trans-presentation, the transcription of perforin, granzyme B, CD95L, and TRAIL was dramatically induced in IKDC (Fig. 6, A and B). Interestingly, the transcription levels of perforin increased by 1000-fold in ex vivo expanded IKDC15 compared with freshly sorted IKDC. At the protein level, similar conclusions could be drawn in that IKDC15 contained high amounts of granules of perforin and granzyme B compared with resting IKDC as observed in confocal microscopy (Fig. 6,B) or flow cytometry (not shown). Accordingly, the lytic activity of IKDC against B16F10 was markedly enhanced by trans-presentation of IL-15 (Fig. 6,C). The lytic activity of IKDC15 was mainly dependent on TRAIL molecules (Fig. 6,E). The side by side comparisons between IKDC and NK cells both stimulated for short (36 h) or long (7 days) periods of time with MS-5/IL-15 revealed qualitative but not quantitative differences. Although NK cell-mediated killing was dependent on granule exocytosis, IKDC lytic functions mostly rely on TRAIL molecules (Fig. 6, D and E).

FIGURE 6.

IL-15 triggers the TRAIL-dependent killing machinery selectively in IKDC. A, Transcription profile of IKDC and IKDC15 for perforin, granzyme B, TRAIL, and FasL (CD95L) encoding genes. Quantitative RT-PCR was performed on freshly sorted IKDC and on IKDC15 (at day 7 of expansion). B, Perforin, granzyme B protein expression as assessed by confocal microscopy. Confocal microscopy was performed on the same cells after intracellular staining using anti-mouse perforin, granzyme B primary Ab, Alexa 488 secondary Ab and DNA staining with Topro. Isotype control Abs were used in parallel. Background staining was not significant. A representative cell is shown in each condition and the percentage of cells containing positive granules is depicted on the right panel. C, IKDC15 exerted a more potent lysis of B16F10 compared with freshly sorted IKDC. 51Cr release assays were performed 12 h at different ratios of IKDC or IKDC15 effector cells on 51Cr-labeled B16F10 cells allowing calculation of the percentages of B16F10 lysis as described in Materials and Methods. D and E, Study of the TRAIL and perforin/granzyme B-dependent lytic pathway in NK15 and IKDC15. Cytotoxicity assays were performed at 36 h (D) or 7 days (D and E) after IL15/IL15Rα-stimulation of IKDC or B220NK cells. Crystal Violet assays were performed in 24 h at a ratio of 30:1, 15:1, and 5:1 of effector:B16OVA cells. Lysis of B16OVA was blocked by treatment with anti-CD95L Ab, anti-TRAIL (N2B2) Ab, or concanamycin A (CMA). The experiments were at least performed three times.

FIGURE 6.

IL-15 triggers the TRAIL-dependent killing machinery selectively in IKDC. A, Transcription profile of IKDC and IKDC15 for perforin, granzyme B, TRAIL, and FasL (CD95L) encoding genes. Quantitative RT-PCR was performed on freshly sorted IKDC and on IKDC15 (at day 7 of expansion). B, Perforin, granzyme B protein expression as assessed by confocal microscopy. Confocal microscopy was performed on the same cells after intracellular staining using anti-mouse perforin, granzyme B primary Ab, Alexa 488 secondary Ab and DNA staining with Topro. Isotype control Abs were used in parallel. Background staining was not significant. A representative cell is shown in each condition and the percentage of cells containing positive granules is depicted on the right panel. C, IKDC15 exerted a more potent lysis of B16F10 compared with freshly sorted IKDC. 51Cr release assays were performed 12 h at different ratios of IKDC or IKDC15 effector cells on 51Cr-labeled B16F10 cells allowing calculation of the percentages of B16F10 lysis as described in Materials and Methods. D and E, Study of the TRAIL and perforin/granzyme B-dependent lytic pathway in NK15 and IKDC15. Cytotoxicity assays were performed at 36 h (D) or 7 days (D and E) after IL15/IL15Rα-stimulation of IKDC or B220NK cells. Crystal Violet assays were performed in 24 h at a ratio of 30:1, 15:1, and 5:1 of effector:B16OVA cells. Lysis of B16OVA was blocked by treatment with anti-CD95L Ab, anti-TRAIL (N2B2) Ab, or concanamycin A (CMA). The experiments were at least performed three times.

Close modal

Therefore, trans-presentation of IL-15 endowed IKDC with TRAIL-dependent killing capacities, a biological attribute not shared by conventional NK cells.

We previously reported that IKDC invade tumor beds and were necessary and sufficient to prevent tumor outgrowth after adoptive cell transfer in Rag−/−×IL-2Rγ−/−-deficient hosts (2). However, the immunizing potential of IKDC in nonimmunocompromized animals remained to be assessed. In as much as NK and IKDC diverge in their mechanisms of killing tumor cells, we addressed the differential immunizing potential of both innate effectors. We used B16OVA as target cells incubated with IKDC15 or B220NK cells (equally activated in MS-5/IL-15) for 16 h before s.c. inoculation as immunization protocols. The ex vivo killing of B16OVA was comparable to that of B16F10 (Figs. 6 and 7,C, inset). When mice were rechallenged 10 days later with a lethal dose of B16OVA, only those vaccinated with IKDC15, but not with 24 h or 7 day IL-15/IL-15Rα stimulated B220NK cells or 1 × 105 dying tumor cells (24, 25, 26), exhibited delayed tumor outgrowth associated with a significantly prolonged survival compared with untreated animals (Fig. 7, A and B). It is noteworthy that inoculation of an increased number of at least 3 × 106 doxorubicin-treated B16OVA tumor cells could confer a significant protection after rechallenge (24). Because IKDC invade tumor beds and could theoretically be subjected to TGF-β-induced immunosuppression, we analyzed the effects of recombinant human (rh) TGF-β on their killing potential and their immunogenicity in vivo. TGF-β could substantially reduce the killing potential of IKDC against B16OVA in vitro (Inset, Fig. 7,C), but did not abrogate their protective activity against tumor challenge in vivo (Fig. 7 C). The prophylactic effects of IKDC15 treated with TGF-β were not observed in Nude counterparts, suggesting that IKDC15/TGF-β mediated T cell-based antitumor immunity (not shown).

FIGURE 7.

IKDC15 have immunizing potential and resist to TGF-β-induced immunosuppression. A total of 106 IKDC15 or NK15 cells (maintained 7 days in MS-5 + IL-15) were incubated with B16OVA at a 10:1 E:T ratio for 16 h before inoculation into the footpad of C57BL/6 mice. Rechallenge was performed 10 days later with a lethal tumorigenic dose of B16OVA (3 × 105 cells) s.c. in the flank of mice (A and B). Controls included untreated mice (PBS) or immunization with 105 B16OVA tumor cells incubated 15 h with 5 μM doxorubicin (23 ) (apoptotic B16OVA). The same experiments were performed adding TGF-β to IKDC15 or apoptotic tumor cells for 6 h before washing and injection into the footpads (C). Graphs show survival curves of animals (n = 8–12 per group) from at least four independent experiments. Statistical analyses were performed using Mantel Cox test. The inset of C depicts the TGF-β reduced killing of IKDC15 in vitro. IKDC15 incubated for 24 h with B16OVA in medium alone or with TGF-β (2 ng/ml) were subjected to a 24-h crystal violet assay against B16OVA. Results of a representative experiment of three are depicted as means ± SEM.

FIGURE 7.

IKDC15 have immunizing potential and resist to TGF-β-induced immunosuppression. A total of 106 IKDC15 or NK15 cells (maintained 7 days in MS-5 + IL-15) were incubated with B16OVA at a 10:1 E:T ratio for 16 h before inoculation into the footpad of C57BL/6 mice. Rechallenge was performed 10 days later with a lethal tumorigenic dose of B16OVA (3 × 105 cells) s.c. in the flank of mice (A and B). Controls included untreated mice (PBS) or immunization with 105 B16OVA tumor cells incubated 15 h with 5 μM doxorubicin (23 ) (apoptotic B16OVA). The same experiments were performed adding TGF-β to IKDC15 or apoptotic tumor cells for 6 h before washing and injection into the footpads (C). Graphs show survival curves of animals (n = 8–12 per group) from at least four independent experiments. Statistical analyses were performed using Mantel Cox test. The inset of C depicts the TGF-β reduced killing of IKDC15 in vitro. IKDC15 incubated for 24 h with B16OVA in medium alone or with TGF-β (2 ng/ml) were subjected to a 24-h crystal violet assay against B16OVA. Results of a representative experiment of three are depicted as means ± SEM.

Close modal

In this study, we demonstrate that ex vivo expanded IKDC15 not only gained lytic capacity in vitro (Fig. 6), but also protective antitumor function in vivo (Fig. 7 A) that even resists to immunosuppressive TGF-β. These data support the hypothesis that IKDC15 could link innate and cognate immunity and, therefore, would be capable of inducing an antitumor immune response resistant to tumor-induced tolerance.

This manuscript describes for the first time the pivotal role of IL-15 trans-presentation in the biology of IKDC, a novel subset of innate effectors sharing markers of both NK cells and conventional DC (1, 2). We initially reported that IKDC were B220+CD11c+NK1.1+ cells expressing MHC class II molecules during treatment with IM + IL-2 and invading tumor beds to kill in a TRAIL-dependent fashion, whereas Chan et al. (1) described that lymph node IKDC in BALB/c mice were endowed with MHC class II-restricted Ag presenting function in vitro. Therefore, IKDC may be considered as a MHC class II expressing NK cell subset or alternatively as a DC endowed with TRAIL-dependent killing capacities. This view has been recently challenged by several authors supporting the notion that IKDC, rather, represent an activated state of conventional NK cells (12, 13, 27). This manuscript aimed at clarifying the functional differences between IKDC and NK cells.

First, IKDC exhibited marked proliferative potential in vitro and in vivo following IL-15/IL-15Rα-driven stimulation (Figs. 1 and 2). Indeed, we showed that CpG ODN, rIL-15 (Fig. 1), and even IM + IL-2 (23) all drove IKDC proliferation in vivo in an IL-15Rα-dependent manner. Interestingly, despite their cloning expansion capacity, IKDC appeared to represent fully differentiated cells because they did not convert into bona fide B220CD11cNK cells after adoptive transfer into congenic animals (Fig. 1,B). In sharp contrast, B220NK cells failed to proliferate in vitro during stimulation with IL-15/IL-15Rα and their CpG driven-proliferation in vivo was IL-15Rα-independent. Although harboring intracytosolic IL-15Rα (Fig. 2 A), why did B220NK cells fail to respond to IL-15 for ex vivo proliferation? Several hypotheses can be drawn to account for the IL-15-driven proliferation of IKDC and not NK cells. There are several isoforms of IL-15Rα. The full-length sIL-15Rα ectodomain resulting from the proteolytic degradation of IL-15Rα is inhibitory when binding to IL-15. In contrast, some isoforms, such as the sushi sIL-15Rα resulting from an alternative splicing of the mRNA of IL-15Rα, are agonists (28). Moreover, there is a reciprocal activation of IL-15Rα with a tyrosine kinase receptor Axl leading to the phosphorylation of both receptors upon binding of IL-15 or Gas6 (the ligand for Axl) and survival effects of the transduced cell type (29). Hence, it is plausible that NK cells might secrete the antagonist form of IL-15Rα and would not be able to benefit from IL-15 and/or that IKDC do secrete a sushi-like isoform of IL-15Rα. Likewise, it is unlikely that Axl plays a dominant role because Axl was not found in Western blot analyses, neither in IKDC nor NK cells (not shown).

Second, following trans-presentation of IL-15, IKDC acquired high lytic capacities (against B16F10 (Fig. 6,C) and B16OVA (Fig. 7)) that were fully abrogated in the presence of anti-TRAIL neutralizing Ab (N2B2, Fig. 6). In contrast, B220NK cells exhibited high basal transcription levels of perforine (in contrast to IKDC, not shown) and killed target cells using secretory granules and not TRAIL molecules (Fig. 6, D and E). TRAIL-dependent cytotoxicity was shown to play a dominant role in the prevention and treatment of neoplasia (30, 31). IKDC15 became capable of sensing and killing tumor cells mainly through TRAIL molecules while also up-regulating their levels of perforine and granzyme B (Fig. 6). Although previous observations tend to demonstrate that IL-15 can up-regulate TRAIL and boost TRAIL-dependent cytotoxicity of murine NK cells in vitro, it remains to be determined whether the IKDC component of the mouse NK cell pool was in fact mediating these TRAIL-dependent effects (32).

Third, following trans-presentation of IL-15, B16OVA-lysing IKDC mediated T cell-dependent protective effects in vivo, even in the presence of TGF-β. Such prophylactic immunization properties were not found with IL-15/IL-15Rα-stimulated NK cells (displaying equivalent quantitative killing capacities as IKDC). One of the main issues remains whether IKDC could not only play a scavenger role by mediating tissue destruction but also a role in T cell priming. Because we have shown that IL-15 trans-presentation skews IKDC toward cytotoxic effector cells rather than APC, we suggest that footpad inoculation of IKDC15 encountering B16OVA may indirectly promote recruitment and activation of conventional DC that will prime naive T lymphocytes. Given that IKDC15 differ from NK15 in their TRAIL-dependent killing of targets, we anticipate that programmed cell death triggered by the extrinsic (membrane bound, TRAIL-mediated) as opposed to the intrinsic (mitochondrial perforine/granzyme-mediated) cell death pathways could matter in the outcome of the prophylactic potential of both effectors. Indeed, our group has reported that apoptosis mediated by anthracyclines, oxaliplatinum, or X Rays was immunogenic, whereas other cytotoxic agents failed to promote an immunogenic cell demise. This was due to the ability of some cytotoxic compounds to induce ecto-calreticulin (CRT) at the plasma membrane of dying cells (26) and to release HMGB1 alarmins to interact with TLR4 harbored on DC (33). Although ecto-calreticulin was required for phagocytosis by DC of dying tumor cells, HMGB1 was involved in the processing of apoptotic material by DC. Therefore, whether the immunogenicity of IKDC15-mediated cell death is TRAIL-, HMGB1-, and/or CRT-dependent needs to be addressed.

Fourth, as recently demonstrated trans-presentation of IL-15 allowed CCR2 expression on IKDC but not on B220NK cells, likely contributing to their CCL2-dependent intratumoral trafficking (23).

It is interesting to note that B220 and CD11c molecules were not acquired by conventional NK cells after 24–48 h of stimulation with a variety of cytokines or DC (Table I), presumably because such NK cells do not enter cell cycle in vitro. Moreover, IL-15 or IL-2 down-regulated MHC class II transcription levels on IKDC, supporting the notion that B220, CD11c, and MHC class II unlikely correspond to activation markers because they were differentially modulated by these activating cytokines.

IL-15 is a pivotal cytokine for the development and function of innate immune cells such as NK, NKT, and TCRγδ intestinal intraepithelial lymphocytes and DC (34). IL-15 also affects acquired immunity by stimulating the proliferation and survival of naive and memory CD8+ T cells (35). T cell-dependent delayed type hypersensitivity responses are impaired in IL-15−/− mice but restored by injection of IL-15 producing WT DC in vivo (36). Furthermore, IL-15 could mediate deleterious effects and has been involved in the exacerbation of numerous inflammatory processes, such as rheumatoid arthritis (37, 38), inflammatory bowel disease (39, 44, 45), type C chronic hepatitis (40), sarcoidosis (41), multiple sclerosis (42), and celiac disease (43). Elevated IL-15 production and IL-15 producing cells were identified as potential initiators of the inflammation. IL-15 can be produced by DC, macrophages, monocytes, and endothelial cells (34). Recently, Ohteki et al. (46) could identify DC derived-IL-15 as the initiator for the development of liver inflammatory diseases. The authors showed that DC-derived IL-15 could stimulate an autocrine loop leading to IL-12 and IFN-γ production and a cascade of inflammatory processes involving CCL2, CCL3, and CCL4 culminating in granuloma formation and liver injury (46). Although their results suggested that asialo-GM1 expressing cells (which include IKDC, our unpublished data) were not involved in the inflammatory cascade, the role of IKDC as a master regulator of the initial steps of granuloma formation and/or at later stages during hepatic injury has yet to be defined. This hypothesis is also driven by other results suggesting the crucial role of CCL2 in granuloma formation promoted by Propionibacterium acnes, zymosan, or Mycobacterium tuberculosis (47). Therefore, one of the major challenges will be to delineate the relevance of IKDC in the sequential events where IL-15 is beneficial or deleterious.

Another unsolved question remains the identification of the human counterpart for IKDC. Some authors have discussed the possibility that IKDC represent the mouse ortholog of human CD56brightNK cells (4, 13). Human CD56brightNK cells are mostly localized in lymph nodes and were considered as potential precursors of more mature CD56dimNK (48). Although indeed IKDC preferentially home and accumulate in lymph nodes (1), they do not appear to convert into B220NK cells in vitro (Table I) nor in vivo upon adoptive transfer in irradiated hosts (Fig. 1 B). It is clear that the identification of more specific IKDC markers will allow not only to characterize the human IKDC ortholog but also to delineate the biological significance of IKDC in pathophysiology.

We thank W. Vainchenker, A. Mackensen, A. Caignard, L. da Costa, and B. Azzarone for discussing our data and H. Yagita, E. Tomasello, E. Vivier, and T. Walzer for providing reagents and mice. We thank O. Lantz for the kind gift of Tg OTII mice. We thank Y. Lecluse, D. Métivier, and P. Rameau for cell sorting, G. Elain, N. Brunel, and B. Besson-Lescure for technical assistance, and the animal facility of the Institut Gustave Roussy under the direction of P. Gonin.

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 European Union grants (ALLOSTEM, DC THERA), Association pour la Recherche sur le Cancer, and Ligue Nationale contre le Cancer (équipes labelisées de G. Kroemer and L.Z.). E.U. received a fellowship from the Deutsche Forschungsgemeinschaft and from the Fondation pour la Recherche Médicale, M.B. was supported by the Poste d’Accueil Institut National de la Santé et de la Recherche Médicale, and G.M. by the Association pour la Recherche sur le Cancer.

4

Abbreviations used in this paper: IKDC, interferon-producing killer dendritic cells; BMDC, bone marrow-derived DC; IM, imatinib mesylate; rm, recombinant murine; ODN, oligodeoxynucleotide; rh, recombinant human.

1
Chan, C. W., E. Crafton, H. N. Fan, J. Flook, K. Yoshimura, M. Skarica, D. Brockstedt, T. W. Dubensky, M. F. Stins, L. L. Lanier, et al
2006
. Interferon-producing killer dendritic cells provide a link between innate and adaptive immunity.
Nat. Med.
12
:
207
-213.
2
Taieb, J., N. Chaput, C. Menard, L. Apetoh, E. Ullrich, M. Bonmort, M. Pequignot, N. Casares, M. Terme, C. Flament, et al
2006
. A novel dendritic cell subset involved in tumor immunosurveillance.
Nat. Med.
12
:
214
-219.
3
Ullrich, E., M. Bonmort, G. Mignot, N. Chaput, J. Taieb, C. Menard, S. Viaud, T. Tursz, G. Kroemer, L. Zitvogel.
2007
. Therapy-induced tumor immunosurveillance involves IFN-producing killer dendritic cells.
Cancer Res.
67
:
851
-853.
4
Spits, H., L. L. Lanier.
2007
. Natural killer or dendritic: what’s in a name?.
Immunity
26
:
11
-16.
5
Shortman, K., J. A. Villadangos.
2006
. Is it a DC, is it an NK? no, it’s an IKDC.
Nat. Med.
12
:
167
-168.
6
Vremec, D., M. O'Keeffe, H. Hochrein, M. Fuchsberger, I. Caminschi, M. Lahoud, K. Shortman.
2006
. Production of interferons by dendritic cells, plasmacytoid cells, natural killer cells and interferon-producing killer dendritic cells.
Blood
109
:
1165
-1173.
7
Welner, R. S., R. Pelayo, K. P. Garrett, Y. Chen, S. Scott Perry, X.-H. Sun, B. L. Kee, P. W. Kincade.
2007
. Interferon-producing killer dendritic cells (IKDC) arise via a unique differentiation pathway from primitive c-kitHiCD62L+ lymphoid progenitors.
Blood
109
:
4825
-4931.
8
Moretta, L., C. Bottino, D. Pende, R. Castriconi, M. C. Mingari, A. Moretta.
2006
. Surface NK receptors and their ligands on tumor cells.
Semin. Immunol.
18
:
151
-158.
9
Diefenbach, A., D. H. Raulet.
2003
. Innate immune recognition by stimulatory immunoreceptors.
Curr. Opin. Immunol.
15
:
37
-44.
10
Vivier, E..
2006
. What is natural in natural killer cells?.
Immunol. Lett.
107
:
1
-7.
11
Walzer, T., E. Vivier.
2006
. NK cell development: gas matters.
Nat. Immunol.
7
:
702
-704.
12
Vosshenrich, C. A., S. Lesjean-Pottier, M. Hasan, O. Richard-Le Goff, E. Corcuff, O. Mandelboim, J. P. Di Santo.
2007
. CD11cloB220+ interferon-producing killer dendritic cells are activated natural killer cells.
J. Exp. Med.
204
:
2569
-2578.
13
Blasius, A. L., W. Barchet, M. Cella, M. Colonna.
2007
. Development and function of murine B220+CD11c+NK1.1+ cells identify them as a subset of NK cells.
J. Exp. Med.
204
:
2561
-2568.
14
Lanier, L. L..
2006
. Natural killer cells: roundup.
Immunol. Rev.
214
:
5
-8.
15
Colucci, F., J. P. Di Santo.
2000
. The receptor tyrosine kinase c-kit provides a critical signal for survival, expansion, and maturation of mouse natural killer cells.
Blood
95
:
984
-991.
16
Berardi, A. C., E. Meffre, F. Pflumio, A. Katz, W. Vainchenker, C. Schiff, L. Coulombel.
1997
. Individual CD34+CD38lowCD19.
Blood
89
:
3554
-3564.
17
Bonnotte, B., N. Larmonier, N. Favre, A. Fromentin, M. Moutet, M. Martin, S. Gurbuxani, E. Solary, B. Chauffert, F. Martin.
2001
. Identification of tumor-infiltrating macrophages as the killers of tumor cells after immunization in a rat model system.
J. Immunol.
167
:
5077
-5083.
18
Walzer, T., M. Bléry, J. Chaix, N. Fuseri, L. Chasson, S. H. Robbins, S. Jaeger, P. André, L. Gauthier, L. Daniel, et al
2007
. Identification, activation, and selective in vivo ablation of mouse NK cells via NKp46.
Proc. Natl. Acad. Sci. USA
104
:
3384
-3389.
19
Huntington, N. D., C. A. Vosshenrich, J. P. Di Santo.
2007
. Developmental pathways that generate natural-killer-cell diversity in mice and humans.
Nat. Rev. Immunol.
7
:
703
-714.
20
Pillarisetty, V. G., S. C. Katz, J. I. Bleier, A. B. Shah, R. P. Dematteo.
2005
. Natural killer dendritic cells have both antigen presenting and lytic function and in response to CpG produce IFN-γ via autocrine IL-12.
J. Immunol.
174
:
2612
-2618.
21
Chaudhry, U. I., G. Plitas, B. M. Burt, T. P. Kingham, J. R. Raab, R. P. Dematteo.
2007
. NK dendritic cells expanded in IL-15 exhibit antitumor responses in vivo.
J. Immunol.
179
:
4654
-4660.
22
Dubois, S., J. Mariner, T. A. Waldmann, Y. Tagaya.
2002
. IL-15Rα recycles and presents IL-15 in trans to neighboring cells.
Immunity
17
:
537
-547.
23
Mignot, G., E. Ullrich, M. Bonmort, C. Menard, L. Apetoh, J. Taieb, D. Bosisio, S. Sozzani, M. Ferrantini, J. Schmitz, et al
2008
. The critical role of IL-15 in the antitumor effects mediated by the combination therapy imatinib and IL-2.
J. Immunol.
180
:
6477
-6483.
24
Casares, N., M. O. Pequignot, A. Tesniere, F. Ghiringhelli, S. Roux, N. Chaput, E. Schmitt, A. Hamai, S. Hervas-Stubbs, M. Obeid, et al
2005
. Caspase-dependent immunogenicity of doxorubicin-induced tumor cell death.
J. Exp. Med.
202
:
1691
-1701.
25
Apetoh, L., F. Ghiringhelli, A. Tesniere, A. Criollo, C. Ortiz, R. Lidereau, C. Mariette, N. Chaput, J. P. Mira, S. Delaloge, et al
2007
. The interaction between HMGB1 and TLR4 dictates the outcome of anti-cancer chemotherapy and radiotherapy.
Immunol. Rev.
220
:
47
-59.
26
Obeid, M., A. Tesniere, F. Ghiringhelli, G. M. Fimia, L. Apetoh, J. L. Perfettini, M. Castedo, G. Mignot, T. Panaretakis, N. Casares, et al
2007
. Calreticulin exposure dictates the immunogenicity of cancer cell death.
Nat. Med.
13
:
54
-61.
27
Caminschi, I., F. Ahmet, K. Heger, J. Brady, S. L. Nutt, D. Vremec, S. Pietersz, M. H. Lahoud, L. Schofield, D. S. Hansen, et al
2007
. Putative IKDCs are functionally and developmentally similar to natural killer cells, but not to dendritic cells.
J. Exp. Med.
204
:
2579
-2590.
28
Bulanova, E., V. Budagian, E. Duitman, Z. Orinska, H. Krause, R. Ruckert, N. Reiling, S. Bulfone-Paus.
2007
. Soluble interleukin IL-15Rα is generated by alternative splicing or proteolytic cleavage and forms functional complexes with IL-15.
J. Biol. Chem.
282
:
13167
-13179.
29
Budagian, V., E. Bulanova, Z. Orinska, L. Thon, U. Mamat, P. Bellosta, C. Basilico, D. Adam, R. Paus, S. Bulfone-Paus.
2005
. A promiscuous liaison between IL-15 receptor and Axl receptor tyrosine kinase in cell death control.
EMBO J.
24
:
4260
-4270.
30
Cretney, E., K. Takeda, H. Yagita, M. Glaccum, J. J. Peschon, M. J. Smyth.
2002
. Increased susceptibility to tumor initiation and metastasis in TNF-related apoptosis-inducing ligand-deficient mice.
J. Immunol.
168
:
1356
-1361.
31
Cretney, E., K. Takeda, M. J. Smyth.
2007
. Cancer: novel therapeutic strategies that exploit the TNF-related apoptosis-inducing ligand (TRAIL)/TRAIL receptor pathway.
Int. J. Biochem. Cell Biol.
39
:
280
-286.
32
Kayagaki, N., N. Yamaguchi, M. Nakayama, K. Takeda, H. Akiba, H. Tsutsui, H. Okamura, K. Nakanishi, K. Okumura, H. Yagita.
1999
. Expression and function of TNF-related apoptosis-inducing ligand on murine activated NK cells.
J. Immunol.
163
:
1906
-1913.
33
Apetoh, L., F. Ghiringhelli, A. Tesniere, M. Obeid, C. Ortiz, A. Criollo, G. Mignot, M. C. Maiuri, E. Ullrich, P. Saulnier, et al
2007
. Toll-like receptor 4-dependent contribution of the immune system to anticancer chemotherapy and radiotherapy.
Nat. Med.
13
:
1050
-1059.
34
Waldmann, T. A., Y. Tagaya.
1999
. The multifaceted regulation of interleukin-15 expression and the role of this cytokine in NK cell differentiation and host response to intracellular pathogens.
Annu. Rev. Immunol.
17
:
19
-49.
35
Kennedy, M. K., M. Glaccum, S. N. Brown, E. A. Butz, J. L. Viney, M. Embers, N. Matsuki, K. Charrier, L. Sedger, C. R. Willis, et al
2000
. Reversible defects in natural killer and memory CD8 T cell lineages in interleukin 15-deficient mice.
J. Exp. Med.
191
:
771
-780.
36
Ruckert, R., K. Brandt, E. Bulanova, F. Mirghomizadeh, R. Paus, S. Bulfone-Paus.
2003
. Dendritic cell-derived IL-15 controls the induction of CD8 T cell immune responses.
Eur. J. Immunol.
33
:
3493
-3503.
37
McInnes, I. B., J. al-Mughales, M. Field, B. P. Leung, F. P. Huang, R. Dixon, R. D. Sturrock, P. C. Wilkinson, F. Y. Liew.
1996
. The role of interleukin-15 in T-cell migration and activation in rheumatoid arthritis.
Nat. Med.
2
:
175
-182.
38
McInnes, I. B., B. P. Leung, R. D. Sturrock, M. Field, F. Y. Liew.
1997
. Interleukin-15 mediates T cell-dependent regulation of tumor necrosis factor-α production in rheumatoid arthritis.
Nat. Med.
3
:
189
-195.
39
Kirman, I., O. H. Nielsen.
1996
. Increased numbers of interleukin-15-expressing cells in active ulcerative colitis.
Am. J. Gastroenterol.
91
:
1789
-1794.
40
Kakumu, S., A. Okumura, T. Ishikawa, M. Yano, A. Enomoto, H. Nishimura, K. Yoshioka, Y. Yoshika.
1997
. Serum levels of IL-10, IL-15 and soluble tumour necrosis factor-α (TNF-α) receptors in type C chronic liver disease.
Clin. Exp. Immunol.
109
:
458
-463.
41
Agostini, C., L. Trentin, M. Facco, R. Sancetta, A. Cerutti, C. Tassinari, L. Cimarosto, F. Adami, A. Cipriani, R. Zambello, G. Semenzato.
1996
. Role of IL-15, IL-2, and their receptors in the development of T cell alveolitis in pulmonary sarcoidosis.
J. Immunol.
157
:
910
-918.
42
Kivisakk, P., D. Matusevicius, B. He, M. Soderstrom, S. Fredrikson, H. Link.
1998
. IL-15 mRNA expression is up-regulated in blood and cerebrospinal fluid mononuclear cells in multiple sclerosis (MS).
Clin. Exp. Immunol.
111
:
193
-197.
43
Di Sabatino, A., R. Ciccocioppo, F. Cupelli, B. Cinque, D. Millimaggi, M. M. Clarkson, M. Paulli, M. G. Cifone, G. R. Corazza.
2006
. Epithelium derived interleukin 15 regulates intraepithelial lymphocyte Th1 cytokine production, cytotoxicity, and survival in coeliac disease.
Gut
55
:
469
-477.
44
Yoshihara, K., T. Yajima, C. Kubo, Y. Yoshikai.
2006
. Role of interleukin 15 in colitis induced by dextran sulphate sodium in mice.
Gut
55
:
334
-341.
45
van Heel, D. A..
2006
. Interleukin 15: its role in intestinal inflammation.
Gut
55
:
444
-445.
46
Ohteki, T., H. Tada, K. Ishida, T. Sato, C. Maki, T. Yamada, J. Hamuro, S. Koyasu.
2006
. Essential roles of DC-derived IL-15 as a mediator of inflammatory responses in vivo.
J. Exp. Med.
203
:
2329
-2338.
47
Peters, W., H. M. Scott, H. F. Chambers, J. L. Flynn, I. F. Charo, J. D. Ernst.
2001
. Chemokine receptor 2 serves an early and essential role in resistance to Mycobacterium tuberculosis.
Proc. Natl. Acad. Sci. USA
98
:
7958
-7963.
48
Chan, A., D. L. Hong, A. Atzberger, S. Kollnberger, A. D. Filer, C. D. Buckley, A. McMichael, T. Enver, P. Bowness.
2007
. CD56bright human NK cells differentiate into CD56dim cells: role of contact with peripheral fibroblasts.
J. Immunol.
179
:
89
-94.