Characterization of CD56+ Dendritic-Like Cells: A Normal Counterpart of Blastic Plasmacytoid Dendritic Cell Neoplasm?

Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare hematological malignancy. Plasmacytoid DCs (pDCs), which are defined as lineage marker (Lin)−HLA−DR+CD56−CD123+CD11c− cells, are considered to be the normal counterpart of BPDCNs. However, BPDCN can be distinguished from pDCs by uniform expression of CD56. In this study, to identify a normal counterpart of BPDCN, we searched for a Lin−HLA−DR+CD56+ population and focused on a minor subpopulation of Lin−DR+CD56+CD123+CD11c− cells that we designated as pDC-like cells (pDLCs). pDLC constituted 0.03% of peripheral blood mononuclear cells (PBMCs), and the pDLC/pDC ratio was higher in bone marrow cells than in PBMCs. pDLC clearly expressed BDCA2, BDCA4, and myeloid antigens, which are frequently expressed by BPDCN. pDLCs exhibited modest expression of Toll-like receptors and produced less interferon-α after CpG stimulation, but presented very low endocytic ability unlike mDCs. These functional differences were attributed to the expression profile of transcriptional factors. After in vitro culture with Flt3-ligand and GM-CSF, pDLCs expressed CD11c and BDCA1. These data suggested that pDLCs are a distinct subpopulation, with an immunophenotype similar to BPDCNs. Moreover, our results indicate that pDLCs might be immature DCs and might contribute to the immunophenotypical diversity of BPDCNs.


Introduction
Blastic plasmacytoid dendritic cell neoplasm (BPDCN) is a rare type of myeloid neoplasm. Its diagnosis is based on its CD4 + CD56 + CD123 + HLA−DR + cell-surface marker profile and the absence of lineage marker (Lin) expression [1,2]. This neoplasm often involves the skin and bone marrow, and its clinical course is usually aggressive [3]. Although many studies have described BPDCN, descriptions of its phenotype and function, which are critical for accurate diagnosis, vary between studies [4,5]; thus, the features of BPDCN might be heterogeneous.
A normal counterpart of BPDCN is thought to be the plasmacytoid dendritic cell (pDC). pDCs are a subset of the heterogeneous dendritic cell (DC) family that serve a critical role in the immune system as pathogen sensors and activators of adaptive immunity [6]. Human peripheral blood DCs are phenotypically defined as leukocytes that lack markers of other leukocyte lineages, e.g., CD3 (T cell), CD14/16 (monocyte), CD19 (B cell), or CD56 (natural killer (NK) cell), and express high levels of major histocompatibility complex (MHC) class II molecules [6,7]. CD11c and CD123 expression can be used to categorize Lin − HLA−DR + blood DCs into CD123 + CD11c − pDCs and CD123 − CD11c + myeloid dendritic cells (mDCs) [8]. In addition, pDCs are distinguishable by the expression of CD303 (BDCA2) and CD304 (BDCA4), whereas mDCs express BDCA1 and BDCA3 [9]. pDCs produce interferon (IFN)-α and induce Th2 polarization of naïve CD4 + T cells in response to various stimuli [1]; both these functions are known to affect various disease outcomes [10,11]. Although DCs are thought to be derived from bone marrow, the pathways that control DC differentiation and molecular regulation are not completely understood. Furthermore, the common dendritic progenitor (CDP), which can directly differentiate into a pDC or mDC, has been identified in mice [12][13][14]. In addition, the neural cell adhesion molecule (NCAM), which corresponds to CD56 in humans, is known to mark immature DCs [15].
Although pDCs are a normal counterpart of BPDCN, BPDCNs differ from pDCs in several ways; for example, BPDCN expresses CD56, CD13, and CD33, and occasionally develops concomitantly with acute myeloid leukemia [3,16]. These observations suggest that the oncogenic origin of BPDCN is not equivalent to that of pDCs in the blood. Moreover, several studies have reported that a small population of CD56 + pDCs in the blood are immunophenotypically similar to BPDCN [17,18]. Although the immunophenotype, function, and differentiation of pDCs have been extensively investigated, this minor population of CD56 + pDCs has not been previously characterized.
In this study, we focused on the Lin − HLA−DR + CD56 + population in steady-state condition to identify a normal counterpart of BPDCN. In order to understand the relationship between BPDCN and its normal counterpart as well as its oncogenic origin, we investigated the characteristics, immunophenotype, function, and transcription factor expression patterns of CD56 + pDC, pDC, and mDC.

Cell culture
Sorted cells were cultured in 96-well round-bottom plates in RPMI1640 medium supplemented with 10% fetal bovine serum and containing 100 ng/mL recombinant human Flt3-ligand (Flt3-L; R&D Systems, Minneapolis, MN, USA) and 100 ng/mL granulocyte colony-stimulating factor (GM-CSF; R&D Systems) for six days. The medium was changed every three days.

RNA isolation, cDNA preparation, and quantitative realtime polymerase chain reaction (qRT-PCR)
RNA was isolated using the RNeasy Micro Kit (Qiagen, Hilden, Germany) and was transcribed to cDNA using the High Capacity RNA-to-cDNA Kit (Applied Biosystems, Carlsbad, CA, USA) as per the instructions provided by the manufacturer. qRT-PCR was performed using the 7300 Real-Time PCR System (Applied Biosystems). Taqman Gene Expression Assays (Applied Biosystems) were as follows: Toll-like receptor (

FITC-labeled dextran uptake
After negative selection, enriched PBMCs were incubated with 1mg/mL of FITC-labeled dextran (MW 40000; Sigma-Aldrich) for 1 h at 37°C or 4°C. The cells were washed four times with phosphate-buffered saline and was analyzed by flow cytometry.

Mixed-lymphocyte culture (MLC) assay
MLC assays were conducted in 96-well round-bottom plates. Carboxyfluorescein succinimidyl ester (CFSE)-labeled CD4 + T cells were used as responder cells, and the sorted DCs were used as effectors. CD4 + T cells were purified using anti-human CD4 microbeads (Miltenyi Biotec). Purified CD4 + T cells were CFSE-labeled according to the instructions provided by the manufacturer (Molecular Probes, Carlsbad, CA, USA). The final volume of each well was 200 μL. Responder cells (50,000 cells/well) and effector cells (10,000 cells/well) were incubated for 72 h at 37°C in 5% CO 2 , with or without CpG (1 μM) or phytohemagglutinin (PHA; 1 μM; Sigma-Aldrich) stimulation. Tcell proliferation, indicated by CFSE dilution, was analyzed by flow cytometry.

DC proliferation assay
DC proliferation was analyzed by flow cytometry. Sorted DCs were labeled by CFSE according to the protocol provided by the manufacturer. CFSE-labeled cells were cultured in RPMI medium supplemented with 10% fetal bovine serum and containing 100 ng/mL recombinant human FLT3-L (R&D Systems) and 100 ng/mL GM-CSF (R&D Systems) for six days. Proliferation, which was indicated by CFSE dilution, was analyzed by flow cytometry.

Statistical methods
Statistically significant differences were determined using one-way analysis of variance followed by Bonfferoni's post-hoc comparision tests. A P-value of <0.05 was considered to be statistically significant. Statistical analysis was performed using GraphPad Prism 6.0 software (GraphPad Software, Inc., San Diego, CA, USA).

pDLCs express TLRs at an intermediate level between pDC and mDC
TLRs are critical molecules for DC activation. This study investigated the TLR expression profile of three DC subpopulations (pDLC, pDC, and mDC) by qRT-PCR ( Figure  3A). TLRs were expressed in pDLCs at an intermediate level between that of pDCs and mDCs in pDLC. In case of TLR4, although we could detect mRNA at least but could not detect the protein with flow cytometry ( Figure 3A and Figure 2).

pDLC cytokine production patterns are distinct from pDC or mDC after TLR stimulation
Cytokine production is critical role of DCs. Especially pDCs are robust IFN-α-producing cells. Cytokine production among pDLC, pDC, and mDC after CpG or LPS stimulation were evaluated by qRT-PCR ( Figure 3B). Levels of IFN-α and TNF-α expression after CpG stimulation were abundant in pDCs but low in pDLCs and mDCs. Furthermore, IL-10 and IL-12 production after LPS stimulation were high in mDCs but low in pDLCs and pDCs. These results were consistent with the low expression of the CpG receptor TLR9 in pDLCs. In addition, IL-10 and IL-12 production in response to LPS was lower in pDLCs than in mDCs. Those were caused by low expression of TLR4 with flow cytometry in pDLCs.    pDLCs, expression of transcription factors, including Ets family, Spi-B, E-box protein, E2-2, and Irf8 was lower than that in pDCs but higher than that in mDCs ( Figure 3C). PU.1 expression, a gene involved in the differentiation of murine DCs precursors, was higher in pDLCs than in pDCs ( Figure 3C).

pDLCs have lower endocytic activity than mDCs and induce less proliferation of CD4 + T cells than that in pDCs
Endocytosis is a functional feature of mDCs. PBMCs were incubated at 37°C or 4°C as a control, with FITC-labeled dextran for 1 h, and the frequency of FITC-labeled dextranpositive cells was assessed using flow cytometry ( Figure 4A). Compared with 4°C controls, mDCs revealed higher uptake levels of FITC-labeled dextran than that in both pDLCs and pDCs.
To determine whether pDLCs could stimulate T-cell proliferation, sorted pDLCs, pDCs, or mDCs were stimulated with CpG and cocultured with allogeneic T cells ( Figure 4B). Furthermore, compared with pDCs, both pDLCs and mDCs induced fewer proliferating cells. These data indicated that compared with pDCs, pDLCs had reduced capacity of prime Tcell proliferation after CpG stimulation.

pDLC express mDC-specific antigens when stimulated with Flt3-Ligand and GM-CSF
pDLCs were cultured with Flt3-L (100 ng/mL) and GM-CSF (100 ng/mL) for 6 days. A CD123 dim CD11c high population and a CD123 high CD11c dim population appeared at day 6 ( Figure 5A). CD56 expression was decreased at day 6 compared to that of pDLC at day0. BDCA1 expression in pDLCs was absent at day 0 but abundant at day 6 especially in CD123 dim CD11c high population ( Figure 5B). On the other hand CD123 dim CD11c high population was not observed when pDCs were cultured with Flt3-L and GM-CSF for 6 days. In order to rule out the possibility that these BDCA1-expressing cells represented contaminating cells that had proliferated in the culture, the proliferation activity of purified pDLC cells was analyzed. CFSE-labeled cells were analyzed by flow cytometry after being cultured for six days with Flt3-Land GM-CSF; however, no CFSE dilution was observed. These data indicated that the expression of cell-surface markers by purified pDLCs changed after in vitro culture. Moreover, these populations were phenotypically similar to pDCs and mDCs.

Discussion
This study investigated the immunophenotype, function, and transcriptional factor expression of pDLCs and compared them with those of pDC, the putative normal counterpart of BPDCN. Initially, we had to exclude the possibility that this population was composed of a mixture of several cell lineages (e.g., pDC and mDC). Our finding that pDLCs had one large peak for many antigens by flow cytometry, such as BDCA2, CD33, and CD45RA, indicated that they represented a distinct population.
pDLCs expressed many antigens that were specific to pDCs, including BDCA2, BDCA4, CD40, CD45RA, and CD86, but they also expressed antigens that were atypical for pDCs and had been frequently reported for BPDCN, such as CD2, CD13, and CD33. These findings suggest that pDLCs are a distinct type of DC, which are more similar to pDCs than mDLCs and mDCs, particularly regarding BDCA antigen expression [9].
Furthermore, the complex immunophenotype resembles the heterogeneous immunophenotype of BPDCN. Although one study suggested that pDC in humans treated with Flt3-L changed in immunophenotype (e.g., CD56 expression), this population exhibited lower expressions of CD45RA and CD56 [19]. pDC can express CD56 in response to stimuli that promote maturation of immature DCs other than Flt3-L that promote differentiation from stem cell to DCs [18,20,21]. However, expression patterns of CD56 and other antigens in activated pDC are also different from those of pDLC in steadystate and for BPDCN.
The expression profile of transcriptional factors in pDLC was distinct from those of pDC and mDC. Furthermore, the CD11c and CD56 expressions were analyzed at day 0 and 6. CD123 high CD11c dim and CD123 dim CD11c high populations were observed at day 6. Results of one representative experiment are presented (left). The average frequencies of CD123 high CD11c dim and CD123 dim CD11c high populations are indicated (n = 5, middle). The average MFI of CD56 at day 0 and 6 are indicated (n = 5, right). * P < 0.05 compared to day0. (B) BDCA1 expression was compared by flow cytometry. pDLCs did not express BDCA1 at day 0, but they did express BDCA1 at day 6 especially CD123 dim CD11c high population. * P < 0.05 compared to day0 (n = 5). (C) Sorted cells were CFSE-labeled and incubated for six days with Flt3-L and GM-CSF. The filled histograms represent PHA-stimulated CFSElabeled CD4 + T cells that were included as a control for cell division (left). The average frequencies of divided cells are indicated (n = 5, right). doi: 10.1371/journal.pone.0081722.g005 expression levels of transcriptional factors, which are specific for pDC differentiation (E2-2 and SpiB) in pDLCs, were intermediate, between the levels in pDCs and mDCs. Id2 blocks pDC differentiation from CDP, and expression of Id2 in pDLC was intermediate as well, between the levels in pDC and mDC [22]. Intriguingly, PU.1 expression, which is directly regulated by Flt3 and is required for GM-CSF-induced DC differentiation in mice, was higher in pDLC than in both pDC and mDC [23]. Considering that these transcriptional factors directly regulate the expression of TLRs [24,25], low expression levels of E2-2 and SpiB suggest possible association with relatively low expression levels of TLRs and other functional molecules in pDLCs. Circulating pDCs mediate rapid and robust secretion of IFN-α following bacterial or virus infection through TLR7 and TLR9 by sensing either non-methylated DNA or RNA. Thus, low expression levels of TLR7 and TLR9 in pDLC accounts for the low IFN-α and TNF-α response after CpG stimulation. LPS stimulation induces IL-10 production in mDC but not in pDLC which express TLR4 at mRNA level but not at protein level. In fact, our data suggested that pDLC also have weak phagocytic function and T cell priming capacity [26,27]. These data indicate that pDLCs are not functionally equivalent to pDCs or BPDCNs, probably due to their distinct transcriptional factor expression profile.
Transcription factor expression pattern in pDLCs might be similar to that seen in dendritic cell precursors of mice. Moreover dendritic precursors appear to differentiate in the bone marrow [13], and our study revealed that pDLCs reside in relatively higher numbers in the bone marrow. NCAM, corresponding to CD56 in humans, marks immature mouse DCs [15]. On the basis of these data, we speculate that pDLCs could represent an immature stage of developing DCs. Finally, this study revealed that the cytokine combination of Flt3-L and GM-CSF, which can induce CD34 + cell differentiation of mDCs [28], could decrease CD56 expression and induce BDCA1 and CD11c expressions. Phenotypically CD123 high CD11c dim is similar to pDCs and CD123 dim CD11c high is also similar to mDCs. We observed that our cytokine combination was quite different from stimuli, which induce phenotypic and functional changes from mouse pDC to mDC at mature stage [29]. On the other hand BPDCN cell line, PMDC05, was reported to be consisted of several phenotypically different populations, and one of these populations, CD123 + BDCA1 -, changed its phenotype from BDCA1 negative to positive after GM-CSF stimulation [30]. This suggests that BPDCN contains potential, maybe immature, population that gives rise to mDC at least in phenotype. In our study, pDCs cultured with Flt3-L and GM-CSF did not show mDC differentiation even in their phenotype. Considering PMDC05 and pDLC phenotypical change and transcriptional factors expression profile, pDLCs could be another candidate of normal counterpart of BPDCN. Further studies will be required to identify the differentiation stage represented by pDLCs and the relation between BPDCN and pDLC.