Published online before print March 1, 2007
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,1
* Department of Pathology and
Comprehensive Cancer Center, Ohio State University Medical Center, Columbus, Ohio, USA; and
Laboratory of Molecular Growth Regulation, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland, USA
1 Correspondence: Department of Pathology, The Ohio State University Medical Center, 129 Hamilton Hall, 1645 Neil Avenue, Columbus, OH 43210, USA. E-mail: jian-xin.gao{at}osumc.edu
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: differentiation • IFN-producing killer dendritic cells (IKDC) • Ly6C+ monocytes
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
/ß upon activation [9
]. A DC subset expressing CD11clowCD45RBhigh has been shown to regulate adaptive immune responses [10
, 11
], and another less-characterized DC subset exhibits cytolytic activity [12
]. However, the phenotypes and functions between some DC subsets, such as pDCs versus cDCs [13
], overlap and/or are interchangeable with each other [3
, 13
14
15
], suggesting that a developmental relationship between DC subsets remains elusive [7
].
Recently, a novel population of cells, which express the NK cell marker NK1.1/DX5 and an intermediate level of DC marker CD11c, has been identified in humans and rodents [16
17
18
19
20
21
]. These cells have been shown to express NK and DC markers and to exhibit spontaneous cytolytic and antigen presentation functions and thus, are referred to as NKDCs [12
, 16
17
18
19
20
]. They have been found in peripheral blood [18
, 22
], lymphoid tissues [16
, 17
, 20
], the liver [20
, 23
, 24
], and peritoneal cavity [21
]. In humans, cytolytic activity was demonstrated in monocyte-derived CD11c+ DCs stimulated by IFN- or -
[12
, 18
], and it may be mediated by TRAIL [18
]. CD4– but not CD4+ DCs, freshly isolated from rat spleen, expressed the NK cell receptor protein 1, NKR-P1 [16
]. Moreover, they were able to kill NK-sensitive yeast artificial chromosome-1 (YAC-1) cells in a perforin-, Fas-, TRAIL-, TNF--, as well as Ca2+-independent mechanism [17
]. In the murine system, NKDCs have been identified in the spleen, lymph nodes, and liver [19
, 20
, 23
, 24
]. Although viral infection-induced CD11c+/DX5+ cells can potently kill NK-sensitive cells, CD40 ligand blockade-induced CD11c+/DX5+ cells play a regulatory role in autoimmunity [19
]. More recently, a population of DCs, expressing NK1.1 and B220, was characterized as IFN-producing killer DCs (IKDCs) [25
, 26
]. These findings suggest that NKDCs might have a developmental relationship to NK1.1– DCs.
To test this hypothesis, we examined the developmental relationship of NKDCs to cDCs by systematic analysis of NKDCs in the thymus, spleen, lymph nodes, and bone marrow (BM) of mice in the steady state. We report herein that NKDCs exist exclusively within a compartment of CD11c+MHC II– cells, which we and others [27
, 28
] have reported previously, representing 
50% of the CD11c+MHC II– cells. They are heterogeneous, as they express variable levels of DC subset markers, such as CD8 and B220, in a tissue-dependent manner. It is more important that NKDCs can differentiate into cDCs, accompanied with the up-regulation of the MHC Class II molecules and down-regulation of the NK cell marker. NK cells did not develop into NKDCs upon adoptive transfer. Ly6Chigh BM cells can be a potential progenitor of NKDCs, as they differentiated into CD11c+NK1.1+ as well as CD11c+NK1.1– cells upon adoptive transfer. In the steady state, NKDCs have the strong capacity to lyse tumor cells but little ability to present allogeneic antigens. Our studies suggest that NKDCs are a distinct subset of DCs, which are characterized by the unique phenotype of CD11c+NK1.1+MHC II–, great cytotoxic capability, and poor ability to present antigen, and act as an intermediate of developing DCs. Thus, NKDCs may play a critical role in linking innate and adaptive immunity.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Cloned monocytic cell line 3B11 was developed in our laboratory [28 , 29 ] and maintained in the culture of R10F (RPMI-1640 medium supplemented with 10% FBS, 50 µM 2-ME, 5 mM glutamine, and antibiotics).
The fluorescent dye-conjugated antibodies were purchased from BD PharMingen (San Diego, CA, USA), including mAb to CD11c, NK1.1, I-Aß, CD8, B220, CD11b, Ly6C, CD16/32, CD80, CD86, B220, CD40, CD45RB, Ly49A, and Ly49D, except for the PE-conjugated mAb to F4/80 antigen, which was purchased from Serotec (UK).
Isolation of BM Ly6C+ monocytes
CD11b+Ly6C+ BM monocytes were isolated using MACS beads as described [30
] with necessary modifications. Briefly, BM cells were first depleted of CD11c+ and NK1.1+ cells using anti-FITC antibody-coated MACS beads following staining of BM cells with FITC-conjugated anti-CD11c and NK1.1 antibodies, as described by the manufacturer (Miltenyi Biotec, Auburn, CA, USA). The DC-depleted BM cells were then positively selected for Ly6C+ cells, which also express CD11b, with FITC-conjugated anti-Ly6C mAb and anti-FITC antibody-coated MACS beads. The purity of Ly6C+ cells was 100%; and >98% were CD11b+Ly6C+ cells.
Enrichment of CD11c+ and NK1.1+ cells
Pooled single cells from spleen, thymus, lymph nodes, or BM of five to 10 mice were stained with FITC-conjugated anti-CD11c or anti-NK1.1 mAb, followed by incubation with anti-FITC mAb-coated MACS beads, as instructed by the manufacturer (Miltenyi Biotec). In some experiments, NK1.1+ cells were enriched with PE-conjugated mAb to NK1.1, followed by MACS beads coated by anti-PE antibody (Miltenyi Biotec). The enriched CD11c+ or NK1.1+ cells were used to analyze phenotype and function or sort subsets as required.
Flow cytometry and cell sorting
CD11c-enriched cells were stained with fluorescent dye-conjugated antibodies in appropriate combination and analyzed by multicolor flow cytometry (BD Biosciences, San Jose, CA, USA). To purify NKDCs, CD11c-enriched cells were stained with APC-NK1.1 mAb and sorted for CD11c+NK1.1+ and CD11c+NK1.1– subpopulations using FACS Aria (BD Biosciences). In some experiments, NK1.1-enriched cells were stained with APC-CD11c mAb and sorted for NK1.1+CD11c– and NK1.1+CD11c+ subpopulations or as where indicated. The purity of sorted cells was >95%.
T cell proliferation assay
Splenic T cells from BALB/c (H-2d) were purified by negative selection through incubation with biotinylated mAb to CD11b (M1/70), CD45RB/220 (RA3-6B2), and Ly6G/C (RB6-8C5; BD PharMingen), followed by MACS beads coated with antibiotin antibody (Miltenyi Biotec). The purified T cells were cocultured with or without the sorted CD11+NK1.1+ or CD11c+NK1.1– DCs from the spleen of C57BL/6 mice (H-2b) for 5 days in the presence of IL-2 (5 ng/ml). Cell proliferation was measured by pulsation with 3H-TdR 16 h before harvest and expressed as stimulatory index (SI): cpm of T cells coincubated with DCs/cpm of T cell alone.
Western blot
Monocytic cells (Clone 3B11) maintained in R10F were lysed, and the whole cell lysates were analyzed by Western blot for the expression of IFN consensus sequence-binding protein (ICSBP), PU.1, and GAPDH, as described [31
]. Total protein extracts (50 µg/lane) were fractionated on a 12% SDS-PAGE and transferred to Hybond-P membrane (Amersham Biosciences, Buckinghamshire, UK). The transferred membrane was blotted by goat anti-ICSBP, PU.1, or GAPDH antibodies (1:4000; Santa Cruz Biotechnology, Santa Cruz, CA, USA) followed by alkaline phosphatase-conjugated rabbit antigoat IgG antibody (1:100,000; Sigma-Aldrich, St. Louis, MO, USA). Anti-GAPDH was used as an internal blotting control. The blots were developed by adding chemiluminescent substrate (Cat No. B21901, Molecular Probes, Eugene, OR, USA).
RT-PCR
Total RNA was extracted from 3B11 cells, cDNA was generated by RT using Superscriptase II (Invitrogen, San Diego, CA, USA), and PCR was performed as described [31
]. The primers used for PCR were as follows: ICSBP forward primer: 5'-CGAACAGATCGACAGCAGCATGTA-3', and reverse primer: 5'-GGACCGGTCAGTCACTTCTTCAAA-3' (253 bp); PU.1 forward primer: 5'-CCTCCATCGGATGACTTGGTTACT-3', and reverse primer: 5'-GAGTATCGAGGACGTGCATCTGTT-3' (256 bp); the ribosome L-19 forward primer: 5'-CTGAAGGTCAAAGGGAATGTG-3', and reverse primer: 5'-GGACAGAGTCTTGTGATCTC-3' (200 bp). The ribosome L-19 primers were used as an internal PCR control [32
].
Cytotoxicity assay
Cytotoxicity was measured by 51Cr-release assay, as described previously [33
]. Sorted NK1.1+CD11c+ and NK1.1+CD11c– cells were used as effector cells, and 51Cr-labeled, NK-sensitive tumor cell line YAC-1 was used as target cells. The effector and target cells were coincubated for 6 h in R10F medium, and the percent specific lysis was calculated based on the following formula: Specific lysis % = 100 x (cpmsample–cpmmedim)/(cpmmax–cpmmedium).
Adoptive transfer
In adoptive transfer experiments, lethally (9 Gy) or sublethally (3.5 Gy) irradiated, congenic mice were used as recipients. The donor cells were injected i.v. into recipients along with 1 x 105 recipient-type BM cells to prevent potential death of the lethally irradiated recipients. No recipient-type BM cells were injected into the sublethally irradiated animals. The mice were killed 7 days post-transfer. The splenocytes were isolated and analyzed by flow cytometry.
Electron microscopy
Cell pellets were fixed in 3% glutaraldehyde in 0.2 M sodium cacodylate (pH 7.2) for 1 h at 4°C. After initial fixation, pellets were washed with two changes of sodium cacodylate buffer (10 min each) and then postfixed in 1% osmium tetroxide in sym-collidine buffer (pH 7.6) for 1 h at room temperature. Following two washes with sym-collidine buffer (10 min each), the pellets were enbloc-stained with a saturated aqueous uranyl acetate solution (pH 3.3) for 1 h. Pellets were dehydrated in a graded ethanol series up to absolute (10 min each), followed by two, 10-min acetone rinses. The cell pellets were infiltrated overnight with a 1:1 mixture of acetone and Spurr’s epoxy resin (Electron Microscopy Sciences, Fort Washington, PA, USA). Finally, the pellets were placed into conical tip BEEMTM embedding capsules containing 100% Spurr’s resin. Polymerization of epoxy blocks was carried out at 70°C overnight. Polymerized blocks were sectioned with a Leica UCT-125 ultramicrotome (Leica, Wein, Austria). Methylene blue-basic fuschin-stained semithin (750 nm) sections were evaluated, and two representative areas were thin-sectioned for ultrastructural examination. Ultrathin (80 nm) sections were collected on 200 mesh copper grids and poststained with uranyl acetate (15 min) and lead citrate (3 min). Electron micrographs were generated with a Zeiss EM 900 (Carl Zeiss SMT Inc., Thornwood, NY, USA) transmission electron microscopy (TEM) equipped with a MegaView III digital camera (Soft Imaging System, GmbH, M
nster, Germany).
Statistical analysis
Data were analyzed statistically by a two-tailed Student’s t-test. P 
0.05 was considered significant and P 0.01, highly significant. All data shown are mean ± SD.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
5% of all the NK1.1+ cells and 15–25% of all the CD11c+ cells in the spleen (Fig. 1B
; Table 1
). As NKDCs are rare in the steady state, we enriched CD11c+ cells using MACS beads and flow cytometry to determine the level of MHC Class II molecules on the surface of NKDCs, the hallmarker of DC maturation [7
]. Among CD11c+ cells, 30% of the cells did not express surface MHC Class II (I-Aß) molecules. It is interesting that NK1.1+ cells were found exclusively in CD11c+MHC II–rather than in the CD11c+MHC II+ population and represented 45% of the CD11c+MHC II– cells (Fig. 1C)
. The results indicate that NKDCs are a major compartment of CD11c+MHC II– cells and that the MHC II+ (mature or fully differentiated) DCs do not express NK1.1.
 View larger version (47K): [in a new window] |
Figure 1. NKDCs exist in the compartment of the CD11c+MHC II– cells. (A and B) Identification of NKDCs in the spleen. Splenocytes isolated from B6 mice were stained with mAb to NK1.1 and CD11c and analyzed for NK1.1+CD11c– (NK), NK1.1+CD11c+ (NKDC), and NK1.1–CD11c+ (DC) cells by gating on live cells using flow cytometry (A), and the percent of NKDCs among total NK1.1+ or CD11c+ cells was calculated using the following formula: % NKDC of CD11c cells = % NKDC of splenocytes/(% NKDC of splenocytes + % NK1.1+CD11c– or NK1.1–CD11c+ cells of splenocytes) x 100 (B). (C) The relationship of NKDCs to CD11c+MHC II– cells. CD11c+ cells were enriched from splenocytes of five to 10 mice using MACS beads, stained with PE-conjugated mAb to I-Aß and APC-conjugated mAb to NK1.1, and then analyzed by three-color flow cytometry. APC-conjugated isotype mAb was used as a control for APC-NK1.1 mAb. The CD11c+MHC II+ and CD11c+MHC II– cells were gated and analyzed for NK1.1 expression in each population. (A and B) The results shown are from five mice (NKDCs were 3.0±0.8% of NK cells and 17.0±2.3% of CD11c+ cells; n=5). (C) The data shown are a representative of at least three reproducible experiments.
|
|
View this table: [in a new window] |
Table 1. Distribution and Phenotype of NKDCs (CD11c+NK1.1+) in Various Lymphoid Organsa
|
and B220. To this end, we examined the expression of CD8 and B220 on NK1.1+ and NK1.1– DCs using flow cytometry. As shown in Figure 2A
, CD11c+NK1.1+ cells (NKDCs) expressed a low-to-intermediate level of CD8 with <1000 units of MFI, and CD11c+NK1.1– cells contained a distinct subset of CD8+ cells with >1000 units of MFI. In contrast, NKDCs expressed a higher level of B220 and 10–20% of them with >1000 units of MFI, comparable with the B220+ subset among CD11c+NK1.1– DCs (Fig. 2A)
. The CD11c+NK1.1+B220+ cells have recently been characterized as IKDCs [25
, 26
]. These results suggest that IKDCs may be a subset of NKDCs.
 View larger version (34K): [in a new window] |
Figure 2. The phenotype of NKDCs in lymphoid tissues. CD11c+ cells were enriched from the spleen, BM, thymus, and lymph nodes of B6 mice as described in Figure 1
, stained with APC-conjugated mAb to NK1.1 and PE-conjugated mAb to I-Aß, CD8, B220, CD80, CD86, CD40, CD45RB, F4/80, CD16/32, CD11b, Ly49A, Ly49D, or isotype mAb for flow cytometric analysis. The CD11c+NK1.1+ and CD11c+NK1.1– populations were gated, respectively, to analyze expression of subset markers and other DC-related markers as indicated. Green lines in the histograms represent the PE-conjugated isotype mAb control, and red lines represent corresponding surface antigens tested. The data shown are a representative from the spleen (A), BM (B), thymus (C), and lymph nodes (D). The experiments were repeated four times.
|
50% of them expressed the NK cell-activating receptor Ly49D, which is required for NK cell-mediated cytolysis [33
]. Conversely, the NKDCs expressed monocyte/macrophage markers, including F4/80, CD11b, and CD16/32 (Fig. 2A)
. It is interesting that they homogeneously expressed a higher level of CD45RB compared with NK1.1– DCs (Fig. 2A)
, a phenotype resembling CD11clowCD45RBhigh regulatory DCs [10
, 11
].
Differential distribution and development of NKDCs in lymphoid tissues
If NKDCs represent a stage of developing DCs, their distribution and subset commitment should vary within different tissue environments. To test this assumption, we examined the distribution of NKDCs in various lymphoid tissues and their phenotypes under the physiological or steady state. As shown in Table 1
, the distribution of the percentage of NKDCs varied within the specific lymphoid tissues, where they resided. In the spleen, NKDCs represented 15–25% of total CD11c+ cells, comparable with their distribution in the BM (15–20%). However, the proportion of NKDCs in the lymph nodes was relatively lower than that in the spleen (10–15%). In the thymus, only 3–5% of the CD11c+ cells expressed NK1.1. The results suggest that the tissue environment may affect the development and/or recruitment of NKDCs. This notion is supported by the differential phenotypes of NKDCs in various lymphoid tissues (Fig. 2)
. As in the spleen, a low-to-moderate level of CD8 (<1000 units of MFI) was detected on NKDCs in the BM, thymus, and lymph nodes (Fig. 2B
2C
2D
, and Table 1
). Likewise, the B220 was also detected in the BM, thymus, and lymph nodes, and a small, indistinct population expressed >1000 units MFI of B220, which was comparable with the level of CD11c+NK1.1– DCs in the same tissues (Fig. 2B
2C
2D)
. Unlike in the spleen and BM, MHC Class II molecules were detectable on the NKDCs in the thymus and lymph nodes with levels varying in individual experiments (Fig. 2D
and data not shown). This may be related to their activation in these tissues [19
]. It is interesting that few MHC Class II molecules were detected on the CD11c+NK1.1– DCs of BM (Fig. 2B)
, suggesting that the BM environment did not provide an environmental cue for DC maturation. Taken together, the results suggest that NKDCs are differentially distributed and developed in lymphoid tissues, and they might have a developmental relationship to cDCs.
It should be noted that the detection of CD8 on NKDCs from the spleen, thymus, and lymph nodes was substantial compared with isotype mAb staining, but its MFI was always less than 1000 units when compared with the CD8 subset of NK1.1– DCs (Fig. 2A
and 2D)
. The moderate staining of CD8 on the NKDCs may be related to the uptake of CD8 molecules from T cells [36
]. To exclude this possibility, we stained NKDCs with the anti-ß-chain of CD8 and found that although the CD8ß-chain was detectable on NKDCs, its level of MFI was significantly lower than CD8 (less than 100 MFI; data not shown). These results suggest that the CD8 may be expressed in a low level on NKDCs.
NKDCs are capable of differentiation into MHC II+NK1.1– DCs in vivo
To examine whether NKDCs can be a precursor of cDCs, we sort-purified CD11c+NK1.1+ and CD11c+NK1.1– cells from the spleens of CD45.2 B6 mice. To obtain highly purified NKDCs, we sorted them, using flow cytometry, following the enrichment of CD11c+ cells by MACS beads (Fig. 3A
). The purified CD11c+NK1.1+ and CD11c+NK1.1– cells, along with 1 x 105 recipient-type BM cells to prevent radiation-induced death, were adoptively transferred i.v. into the lethally irradiated CD45.1 congenic B6 mice. The recipients were killed 7 days after transfer, and the splenocytes were harvested for flow cytometric analysis. Donor cells were identified by mAb to CD45.2. As shown in Figure 3B
, CD11c+NK1.1+ (NKDCs) cells differentiated into CD11c+NK1.1– DCs, as NK1.1 was down-regulated in 85% of NKDCs within 7 days post-transfer. The differentiation of NKDCs into CD11c+NK1.1– DCs appeared to be accompanied by a proportional up-regulation of MHC Class II molecules (Fig. 3B)
. This up-regulation was evident, as the NKDCs did not express a significant level of MHC Class II molecules before transfer (Figs. 1B
and 2B)
. In contrast, CD11c+NK1.1– DCs did not differentiate into CD11c+NK1.1+ cells, although 2.5% donor-type CD11c+NK1.1+ cells were detected in the mice receiving CD11c+NK1.1– cells. This is not surprising, and it may be explained by the contamination of CD11c+MHC II–NK1.1– cells in the CD11c+NK1.1– DC preparation (Fig. 1)
, which might differentiate into NKDCs. It would be interesting to address this issue in the future. In addition, 20% of the donor cells in the mice receiving NKDCs expressed B220, which was expressed similarly in donor cells from mice receiving CD11c+NK1.1– DCs (Fig. 3B)
. Taken together, the results support the hypothesis that NKDCs can differentiate into NK1.1– DCs.
 View larger version (43K): [in a new window] |
Figure 3. NKDCs can differentiate into NK1.1– DCs. The enriched CD11c+ cells from the spleens of CD45.2 B6 mice were sorted for CD11c+NK1.1+ and CD11c+NK1.1– subsets, respectively, by FACSorter Aria (A). Each subset was injected (5–10x105/mouse) i.v. into the lethally irradiated (9 Gy), congenic CD45.1 B6 mice. The recipients were killed 7 days post-transfer. The splenocytes were isolated, stained with various fluorescent dye-conjugated mAb to CD11c and CD45.2 in combination with mAb to NK1.1, I-Aß or B220, and analyzed by three-color flow cytometry (B). Fluorescent dye-conjugated, isotype mAb were used as a staining control. The numbers in the quadrants indicate the percentage of each subset. The data shown are a representative of three experiments. FSC, Forward-scatter.
|
 View larger version (35K): [in a new window] |
Figure 4. NKDCs are not developed from NK cells. Then, the enriched NK1.1+ cells, as described in Materials and Methods, were stained with mAb to CD11c followed by FACS sorting for NK1.1+CD11c– and NK1.1+CD11c+ cells (>98% pure). The sort-purified cells (0.5–1.5x106/mouse) were adoptively transferred i.v. into CD45.1 congenic B6 mice, which had been irradiated sublethally (3.5 Gy). The irradiated mice receiving vehicle alone were used as negative controls. The recipients were killed at Day 7 after transfer. The spleen cells were isolated, stained with mAb to CD45.2, CD11c, and NK1.1 or in combination with mAb to I-Aß, and then analyzed by three- or four-color flow cytometry. CD45.2+ cells of each group were gated and analyzed for CD11c and NK1.1 expression (A, bottom panel), and the CD45.2+ NKDC-derived subsets (CD11c+NK1.1+ and CD11c+NK1.1–) were analyzed further for I-Aß expression (B). The donor cell repopulation activity (numbers per million injected cells) in the spleen of recipients was calculated using the following formula: number of injected donor cells/106 x number of recipient spleen cells x percent of CD45.2+ donor cells in the recipient spleen (C). The data shown are a representative of two experiments.
|
 View larger version (43K): [in a new window] |
Figure 5. Monocytes have the capability to differentiate into NKDCs. CD11b+Ly6C+CD11c–NK1.1– BM cells were isolated from CD45.2 B6 mice using MACS beads, as described in Materials and Methods, and were verified by flow cytometry (A). The Ly6C+CD11b+ cells were injected i.v. (5–10x106/mouse) into lethally irradiated CD45.1 B6 mice. The control group received 1 x 105 Ly6C-depleted BM cells per mouse. The recipients were killed 7 days post-transfer. The splenocytes were harvested, stained with mAb to CD45.2, NK1.1, and CD11c or Ly6C, and analyzed for NKDC differentiation by gating on CD45.2+ cells using three-color flow cytometry (B). The number of total donor-derived DCs, NKDCs, and NK1.1– DCs was determined in percentage, respectively (C). The data shown are one of two experiments (n=3/group; *, P<0.05; **, P<0.01). To verify the results from BM monocytes at a clonal level, the cloned monocytic cells were used (D and E). (D) Western blot and RT-PCR analysis of the expression of PU.1 and ICSBP in 3B11 cells; GAPDH and L19 were used as internal controls for Western blot and RT-PCR, respectively. (E) 3B11 cells (CD45.2+) were injected i.v. (3–5x106/mouse) into the lethally irradiated CD45.1 B6 mice as described above. The recipients were killed on Day 7 post-transfer. The pooled splenocytes of three mice were stained with mAb to CD45.2 and CD11c in combination with isotype mAb or mAb to NK1.1, CD8, B220, or I-Aß and analyzed for DC subsets by gating on CD45.2+ cells using three-color flow cytometry. The data shown are one of three reproducible experiments.
|
20%) of the NKDCs was detected. The proportion of NKDCs among the monocyte-derived CD11c+ cells was 15–25% (data not shown), which was similar to the results from the CD11b+Ly6C+ BM monocytes. In addition to CD11c+CD8+ and CD11c+B220+ DCs, we observed a CD11c+I-Aß– population, approximately equivalent to the CD11c+MHC II– population observed in the blood [27
] and spleen [28
]. It is interesting that NKDCs appeared to be dominant over other DC subsets, suggesting that the inflammatory environment in irradiated mice favors 3B11 cells differentiating into NKDCs, reminiscent of bitypic NK/DC regulatory cells, which increased dramatically in response to a viral infection [19
]. These results confirm that monocytes can be a potential precursor of NKDCs at the clonal level [18
].
NKDCs have the great capacity to kill tumor cells but little capability to prime naïve T cells
As 50% of the NKDCs expressed NK cell-activating receptor Ly49D (Fig. 3A)
, we examined whether they have the capacity of NK cells to kill tumor cells spontaneously. To this end, we first enriched NK1.1+ cells using MACS beads and then sorted NK1.1+CD11c+ and NK1.1+CD11c– populations from spleens using FACS Aria. Each population was subjected to cytological analysis and cytolytic assay by incubation with 51Cr-labeled, NK cell-sensitive cell line YAC-1 for 6 h.
As shown in Figure 6A , the NK1.1+CD11c+ cells demonstrated monocytic morphology, whereas NK1.1+CD11c– cells exhibited lymphoid morphology, which is much smaller in size than NKDCs (Fig. 6A and 6B) . Moreover, both populations exhibited a distinct cytochemistry characteristic. NKDCs were acidophilic, and NK cells were basophilic (Fig. 6A) . Ultrastructure analysis by TEM revealed that NKDCs exhibited the morphology similar to that of CD11c+NK1.1– DCs, which were characterized by irregular nuclei, increased vacuoles, ample mitochondria, and higher cytoplasm:nucleus ratio, as compared with NK1.1+CD11c– NK cells (Fig. 6B) .
 View larger version (37K): [in a new window] |
Figure 6. Cytological and functional characterization of NKDCs. (A and B, right panels, and C) NK1.1+ cells were enriched from spleen of B6 mice with FITC-NK1.1 mAb and anti-FITC-coated MACS beads stained with APC-CD11c mAb and sorted by FACS Aria for NK1.1+CD11c+ and NK1.1+CD11c– cells. Each subset was used for cytological analysis by light microscopy (A, Giemsa/Wright staining) and TEM analysis as well (B, right panel) or used for killing assay by incubation with 51Cr-labled YAC-1 cells for 5 h (C). (B, left and middle panels, and D) CD11c+ cells were enriched with FITC-CD11c mAb and anti-FITC antibody-coated MACS beads, stained with APC-NK1.1 mAb. The CD11c+NK1.1+ and CD11c+NK1.1– cells were sorted by FACSorter. Each population (1.25x104/well) was used to stimulate allogeneic, naïve T cells (1x105/well), which were purified by negative selection using MACS beads in the presence of IL-2 (5 ng/ml). Control cultures were naïve T cells or DCs alone. The cultures were incubated for 5 days and pulsed with 3H-TdR 16 h before harvest. The data shown are the SI relative to the cultures of T cell alone (D). In some experiments, CD11c+NK1.1+ and CD11c+NK1.1– cells were used for TEM analysis (B, left and middle panels). The original magnification of TEM (B) is as follows: upper left, 12,000x; upper middle, 12,000x; upper right, 20,000x; lower left, 20,000x; lower middle, 20,000x; lower right, 30,000x. Each experiment was repeated two or three times. **, P < 0.01, when NKDCs were compared with NK1.1– DCs (D).
|
50% of NKDCs expressed a high level of Ly49D (Fig. 2A)
, an activating receptor critical for in vivo cytolysis by NK cells [33
]. In contrast, pooled CD11c+ DCs, as previously reported [20
], did not exhibit significant killing activity (data not shown).
As NKDCs did not express MHC Class II molecules, we assumed that they had little capacity to present allogeneic antigens. We sorted CD11c+NK1.1+ and CD11c+NK1.1– cells and cocultured them with allogeneic T cells. As shown in Figure 6D
, NKDCs showed poor capability to activate naïve, allogeneic T cells. The SI was <2 and was only 1/3 of the SI of NK1.1– DCs (Fig. 6D)
. The results positively correlated with the level of MHC Class II molecules expressed on each population (Fig. 1)
. Thus, NKDCs have acquired the cytolytic activity of NK cells but not the sufficient capacity of antigen presentation. These findings are consistent with a recent report by Pillarisetty et al. [20
], demonstrating that NKDCs killed tumor cells more potently than NK cells but had less capacity (about one-third) to present antigens than cDCs. It is likely that NKDCs from viral-infected mice can potently prime antigen-specific T cells, as maturing NKDCs, in response to inflammatory stimuli, may express a high level of MHC Class II molecules [19
].
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The developmental pathways for various DC subsets have been an enigma because of the overlapping phenotypes and/or functions between DC subsets [3 , 13 14 15 ]. For example, CD8+ DCs have the capacity to lyse activated T cells [35 ]; B220+ DCs may differentiate into CD8+ DCs and vice versa [38 , 45 46 47 48 49 ]. Recently, a population of CD11c+NK1.1/DX5+B220+ cells has been identified as IKDCs [25 , 26 ]. In this study, we demonstrate further that 10–20% of NKDCs were B220-positive cells, suggesting that IKDCs represent only a subset of NKDCs. The developmental relationship of IKDCs to B220– NKDCs remains to be elucidated further. As NKDCs can differentiate into NK1.1– DCs, it would be of great interest to investigate the developmental relationship of IKDCs to B220+ pDCs.
At least three lines of evidence support the notion that NKDCs may be an intermediate of developing DCs. First, NKDCs exist exclusively in the compartment of CD11c+MHC II– cells, which have been shown to differentiate into CD8+, as well as B220+ DC subsets [27 ]. Second, NKDCs can differentiate into NK1.1– DCs upon adoptive transfer, which is accompanied by up-regulation of MHC Class II molecules and down-regulation of the NK cell marker. Finally, the developmental relationship is supported by evidence that the B220 expression pattern of the NKDCs is comparable with NK1.1– DCs in the same lymphoid tissue. As B220+ pDCs can also be converted into cDCs [13 ], it remains to be elucidated whether pDCs and NKDCs are developmentally related.
Monocytes are a well-known precursor for myeloid DCs [50 ]. In this study, we demonstrated that the monocytes can be a precursor of NKDCs. As B220 is detectable on IKDCs [25 , 26 ] and monocyte-derived DCs [28 , 39 ], it is likely that monocytes can differentiate into NKDCs, which can subsequently develop into NK1.1– DCs. In fact, the BM-derived CD11b+Ly6C+ cells can differentiate into NKDCs in the recipients. These results were confirmed at a clonal level. The transfer of the monocytic Clone 3B11 into the irradiated, adoptive recipients led to their differentiation into CD11c+MHC II–, CD11c+NK1.1+, CD11c+CD8+, CD11c+CD8–, and CD11c+B220+ cells. The comparable proportion of NKDCs among total CD11c+ cells produced in these two experimental systems suggests a homeostatic differentiation of NKDCs. Thus, BM-derived CD11b+Ly6C+ cells and monocytic cell lines have the potential to develop into NKDCs.
The capacity of 3B11 cells to differentiate into CD8 and B220 cells is probably associated with their constitutive expression of ICSBP mRNA and protein, which is required for the development of CD8+ and B220+ DCs [41
, 44
]. However, the relationship of monocytes to various DC subsets remains controversial. Although we and others [28
, 39
] have shown that not only can monocytes develop into CD8+, CD8–, and B220+ DCs, but they can also differentiate into cytotoxic DCs [18
, 51
, 52
]. However, the hypothesis that monocytes serve as a precursor of CD8+ and CD8– DCs has been questioned by recent publications [34
, 40
]. This may be a result of the use of different experimental systems [34
, 40
]. Although 3B11 monocytic cells, in our hands, can differentiate into a substantial population of CD8+ DCs after s.c. injection into normal mice [28
], their differentiation into CD8+ cells was marginal in the lethally irradiated mice after i.v. injection. This suggests that the environment is important for monocyte differentiation. Moreover, whether the CD8+ cells are CD8 DCs or CD8-bearing pDCs needs to be elucidated further.
It has been controversial whether NKDCs are actually derived from NK cells. In our model, we did not detect donor-derived NKDCs in the recipients of NK cells, suggesting that NK cells are unlikely a precursor of NKDCs.
Despite the observation that the NK activity of NKDCs appears to be consistent with that of NK cells, their DC-related function may vary with the experimental systems used [19 , 20 ]. CD11c+DX5+ NKDCs activated by viral infection or CD40 blockade exhibited cytotoxic or regulatory activity [19 ]. Pillarisetty et al. [20 ] reported that the capacity of NKDCs to present allogeneic antigens was about one-third of the cDCs. This is consistent with our finding that NKDCs have little ability to stimulate naïve, allogeneic T cells. The capacity of activated NKDCs to present antigens may be enhanced by the up-regulation of surface MHC Class II molecules, e.g., in the virus-infected animals [19 ]. Although NKDCs express low levels of B7-1 and B7-2, a substantial population of NKDCs in the spleen expressed a high level of CD40. The constitutively expressed CD40 may be important for NKDCs differentiating into activating DCs [19 ]. In the absence of CD40-mediated signaling, MHC II– NKDCs are likely to differentiate into tolerogenic [53 54 55 56 ] or regulatory DCs [19 ].
Taken together, we have comprehensively characterized the recently discovered NKDCs with a unique phenotype of CD11c+NK1.1+MHC II–. The freshly isolated NKDCs have the capability to lyse tumor cells but have little capability to stimulate allogeneic T cells, probably as a result of the low level of MHC Class II and costimulatory molecule expression. However, it is most important that although we could not conclude exclusively at this stage that NKDCs represent an essential and necessary intermediate stage for all of the developing DCs, NKDCs are at least one of the precursors for NK1.1– DCs. Based on their distribution pattern in lymphoid tissues, their capability to differentiate into NK1.1– DCs, and their potential origin from BM-derived Ly6C+ monocytes, we propose a putative, developmental pathway for DCs (Fig. 7
): Ly6C+ cells 
CD11c+MHC II–NK1.1– cells [precursor (pre)DCs] CD11c+MHC II–NK1.1+ cells (NKDCs) CD11c+MHC II–NK1.1+CD8/B220+ cells (NKDC subsets) CD11c+MHC II+NK1.1–CD8/B220+ cells (DC subsets). In this pathway, NKDCs likely play an important role in linking innate and adaptive immunity. The cytolytic activity of NKDCs endows them with the capacity to kill cancer and/or viral-infected cells, ingest dead cells, and present antigens to naïve T cells [16
, 17
, 19
, 20
]. These processes are accompanied by differentiation and maturation [57
]. Further elucidation of the developmental and functional relationship between NKDCs and NK1.1– DCs may lead to a novel strategy for the immunotherapy of cancer as well as autoimmune diseases.
 View larger version (29K): [in a new window] |
Figure 7. A schematic pathway of NKDC development. Monocytes (CD11b+Ly6C+) in the BM commit to NK1.1– precursors DCs (CD11c+MHC II–NK1.1–) and then to NKDCs (CD11c+MHC II–NK1.1+), which might migrate to the peripheral lymphoid organs and commit to the CD8+ and/or B220+ NKDC subsets. In the steady state, NKDCs may differentiate into NK1.1– DCs. Alternatively, NKDCs may phagocytose tumor or virus-infected cells accompanied by a process of activation. The consequence of NKDC activation may lead to immunity or tolerance, depending on their activation status.
|
Received November 13, 2006; revised January 12, 2007; accepted January 29, 2007.
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
+ and CD8– subclasses of dendritic cells direct the development of distinct T helper cells in vivo J. Exp. Med. 189,587-592 via autocrine Il-12 J. Immunol. 174,2612-2618} in IL-12/IL-18-stimulated mouse macrophage populations Blood 105,1319-1328(+) dendritic cells (DC) suggests that monocytes can be direct precursors for both CD8(+) and CD8(–) DC in the mouse J. Immunol. 170,5927-5935 Nat. Immunol. 6,600-607[CrossRef][Medline]+ B220+ dendritic cells endowed with type 1 interferon production capacity and tolerogenic potential Blood 100,383-390 + dendritic cells Blood 101,305-310(+) dendritic cells J. Exp. Med. 196,1415-1425+ DC, with a plasmacytoid phenotype, induce differentiation and support function of T cells with regulatory properties Immunology 108,481-492[CrossRef][Medline], and IFN- by mouse dendritic cell subsets J. Immunol. 166,5448-5455 J. Exp. Med. 179,1109-1118This article has been cited by other articles:
 |
|
 |
|
  B. M. Burt, G. Plitas, J. A. Stableford, H. M. Nguyen, Z. M. Bamboat, V. G. Pillarisetty, and R. P. DeMatteo CD11c identifies a subset of murine liver natural killer cells that responds to adenoviral hepatitis J. Leukoc. Biol., October 1, 2008; 84(4): 1039 - 1046. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 U. I. Chaudhry, G. Plitas, B. M. Burt, T. P. Kingham, J. R. Raab, and R. P. DeMatteo NK Dendritic Cells Expanded in IL-15 Exhibit Antitumor Responses In Vivo J. Immunol., October 1, 2007; 179(7): 4654 - 4660. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Nicolas, D. Cathelin, N. Larmonier, J. Fraszczak, P.-E. Puig, A. Bouchot, A. Bateman, E. Solary, and B. Bonnotte Dendritic Cells Trigger Tumor Cell Death by a Nitric Oxide-Dependent Mechanism J. Immunol., July 15, 2007; 179(2): 812 - 818. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
