In vitro Chicken Bone Marrow-Derived Dendritic Cells Comprise Subsets at Different States of Maturation

Research in chickens has been fundamental for the discovery of basic aspects of the immune system and has led to an interest in the in-depth characterization of avian immune cell types including dendritic cells (DCs). The in vitro generation and expansion of chicken bone marrow-derived DCs (chBMDCs) in the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF) has provided a way to study chicken DCs, which are only present at limited cell numbers in vivo. This method has been employed to study the interactions between chicken DCs and pathogens or vaccines. However, a detailed characterization of the chBMDC culture is still lacking. In the present study, we performed an elaborate phenotypical and functional analysis of the chBMDC culture and addressed its heterogeneity. After 8 days of culture, chBMDCs comprised major histocompatibility complex class II (MHC-II)low and MHC-IIhigh subsets with different morphologies. Compared with MHC-IIlow chBMDCs, the MHC-IIhigh subset showed a more mature phenotype, with higher expressions of CD1.1, CD40, CD80, CCR7, and CD83, and a relatively low opsonophagocytic capacity. Nevertheless, MHC-IIhigh chBMDCs did not show an increased capacity to induce T-cell proliferation. Therefore, MHC-IIhigh chBMDCs were found to be semi-mature. Interestingly, the presence of the semi-mature MHC-IIhigh chBMDC subset reduced when cells were cultured in the presence of IL-4. Finally, prolonged cell culture after fluorescence-activated cell sorting (FACS) converted the semi-mature MHC-IIhigh subset back into the immature phenotype of the MHC-IIlow subset, demonstrating plasticity of their maturation state. This detailed characterization explained the heterogeneity of the chBMDC culture by the simultaneous presence of immature and semi-mature chBMDC subsets, in addition to cells without features of antigen-presenting cells. Our findings are instrumental for the interpretation of experiments using the chBMDC culture in past and future research by providing insights into its phenotypically and functionally distinct cell types.


INTRODUCTION
Dendritic cells (DCs) are important innate immune cells that capture and process antigens to present them to cells of the adaptive immune system (1). Adaptive immune responses result in pathogen-specific and long-lasting immunological memory, which enables the immune system to act more rapidly upon a second encounter with the pathogen. Vaccination against pathogens critically depends on DCs, which, respectively, support and fine-tune antigen presentation by co-stimulatory molecules and cytokines. The expression pattern of costimulatory molecules and cytokines by DCs depends on signals from their environment, including pathogen-associated molecular patterns (PAMPs) (1,2), damage-associated molecular patterns (DAMPs) (3), and cytokines (1,2,4).
DCs are rare cells in all tissues and hard to isolate, which makes the use of primary DCs in functional assays challenging. To overcome this problem, granulocyte-macrophage colony-stimulating factor (GM-CSF)-differentiated bone marrow-derived DC (BMDC) culture methods have been developed to generate DCs in large numbers (5). The availability of cultured BMDCs facilitates their use in in vitro screening methods, including immunogenicity tests for vaccines and toxicity tests for allergens (6)(7)(8). In addition to well-characterized murine BMDC culture methods, such methods have been developed for veterinary species, including dogs [Ricklin (9)], cats (10), cattle (11), sheep (12), pigs (13), and chickens (14).
Chicken BMDCs (chBMDCs) have been cultured in the presence of both recombinant chicken GM-CSF and interleukin-4 (IL-4) and were defined as DCs because of their typical stellate morphology and high expression of both major histocompatibility complex class II (MHC-II) and CD11b/c (14). This chBMDC culture method has led to several studies into the role of chicken DCs in infection and vaccination. Maturation of chBMDCs has been observed after stimulation with lipopolysaccharide (LPS) or CD40L, as demonstrated by increased surface expression of co-stimulatory molecules CD40, CD83, and CD86; reduced phagocytosis and endocytosis; and an increased ability to induce a mixed lymphocyte reaction (14). Similarly, chBMDCs have been found to mature upon exposure to avian influenza virus (15,16), infectious bursal disease virus (17), or Salmonella enteritidis and Salmonella gallinarum vaccine candidates (18,19).
Despite the widespread use of BMDCs originating from chickens and other species, a recent transcriptome study showed that murine GM-CSF-differentiated BMDCs differ phenotypically from murine DC populations in vivo (20). Moreover, this study revealed that murine BMDC cultures comprise both CD11b high MHC-II low macrophage-like and CD11b low MHC-II high DC-like subsets that are closely related, but still phenotypically and functionally different. These findings had implications for conclusions drawn using in vitro murine BMDC cultures as a model for DC biology in vivo and are part of the ongoing discussion on how to distinguish DCs and macrophages (20)(21)(22)(23)(24)(25). In addition, these findings stressed the importance of thorough characterization of the cellular subsets present in in vitro BMDC cultures and triggered us to explore in depth the nature of chBMDCs raised in vitro with GM-CSF and to determine whether these indeed represent DC-like cells.
The initial results of the present study showed that the chBMDC culture was heterogeneous and comprised MHC-II low and MHC-II high subsets, similar to observations in murine BMDC cultures. Therefore, we hypothesized that chBMDC culture comprised MHC-II low macrophage-like and MHC-II high DC-like subsets. However, in contrast to murine BMDC cultures, the MHC-II low and MHC-II high subsets of the chBMDC culture were found to reflect different maturation states rather than distinct cell types. MHC-II high chBMDCs were found to exhibit increased expression of costimulatory molecules, also in the absence of stimuli. These findings on chBMDCs may have important consequences for conclusions drawn in past and future studies that make use of the chBMDC culture as a model for in vivo DC biology in chickens, in particular studies that assess chBMDC maturation.

Bone Marrow Isolation
Eighteen-day-old embryonated NOVOgen Brown eggs were obtained from a commercial breeder (Verbeek Broederij, Zeewolde, the Netherlands). Chicken embryos were removed from the eggs and euthanized by decapitation. Next, the tibiae and femurs were collected, bone heads were removed, and bone marrow was harvested by flushing the bones with RPMI-1640 cell culture medium supplemented with GlutaMAX TM -I, phenol red, and HEPES (Gibco TM , Life Technologies Limited, Paisley, UK) under sterile conditions using a Plastipak TM 10-ml syringe with a Microlance TM 3 21-G needle (both from BD Biosciences, Pharmingen, San Diego, CA, USA). Bones and bone marrow cells were kept on ice during the whole procedure. Bone marrow cells from 200 embryos were pooled, gently squeezed through a Falcon R 70-µm cell strainer (Corning R , Corning B.V. Life Sciences, Amsterdam, the Netherlands), and stored at −140 • C in RPMI, 50% chicken serum (Gibco TM , Life Technologies Limited, Paisley, UK), and 10% DMSO (Honeywell, Bucharest, Romania). This procedure resulted in batches comprising 1.3-2.3 × 10 9 bone marrow cells, which were frozen at a concentration of 2.5-5 × 10 7 cells per cryotube.

chBMDC Culture
As previously described by others (26), chBMDCs were cultured from isolated bone marrow cells in RPMI-1640 cell culture medium supplemented with 5% chicken serum and 50 U/ml of penicillin-streptomycin (all from Gibco TM , Life Technologies Limited, Paisley, UK) in the presence of recombinant GM-CSF (and IL-4) at 41 • C, 5% CO 2 . Recombinant GM-CSF and IL-4 were produced using COS-7 cells transfected with pCI-neo (Promega Corporation, Madison, Wisconsin, USA) expressing the relevant cytokine, which were a kind gift from P. Kaiser and L. Rothwell (Roslin Institute, Edinburgh, UK). The concentrations of the recombinant cytokines are given as a dilution of supernatant from transfected COS-7 cultures in accordance with a previous study (27). GM-CSF was used at the titrated concentration (2 µl/ml) that resulted in the highest percentage of MHC-II + CD40 + CD80 + cells. In one experiment, the chBMDC culture was supplemented with GM-CSF and titrated concentrations of IL-4. Bone marrow cells were seeded at 2.5 × 10 6 cells per milliliter in 75-cm 2 cell culture flasks in 15 ml of RPMI-1640 medium per flask, in 25-cm 2 cell culture flasks in 5 ml of RPMI-1640 medium per flask, in Costar R sixwell plates in 2 ml of RPMI-1640 medium per well, or in Costar R 24-well plates in 0.5 ml of RPMI-1640 medium per well (all from Corning R , Corning B.V. Life Sciences, Amsterdam, the Netherlands) depending on the required sample size. Early in the morning at day 3, culture medium with non-adherent cells was removed, and fresh RPMI-1640 medium with GM-CSF (and IL-4) was added. Late in the afternoon at day 4, the cultures received another volume of RPMI-1640 medium with GM-CSF (and IL-4). The morphology of chBMDCs was examined by light microscopy using an EVOS FL microscope (AMG, Mill Creek, Washington, USA). In selected experiments, chBMDC cultures were matured by 100 ng/ml of LPS O127:B8 (Sigma-Aldrich, Saint Louis, MO, USA) stimulation for 24 h at day 7. To harvest the cultures at day 8, the medium with non-adherent cells was first collected. Subsequently, loosely adherent cells were washed and collected with Dulbecco's phosphate-buffered saline (DPBS) without calcium and magnesium (DPBS −/− ; Lonza, Basel, Switzerland). Finally, the remaining adherent cells were incubated in DPBS −/− supplemented with 5 mM UltraPure EDTA (Invitrogen TM , Life Technologies Europe BV, Bleiswijk, the Netherlands) for 10 min at room temperature (RT) before being collected as well. All cell-containing fluids (cell culture medium, DPBS −/− , and DPBS −/− 5 mM EDTA) obtained during the harvest procedure were pooled for subsequent experiments.

Immunofluorescence Microscopy Analyses
Ethanol-cleaned 12-mm glass coverslips (Waldemar Knittel Glasbearbeitungs GmbH, Brunswick, Germany) were placed into the wells of a 24-well cell culture plate. Next, bone marrow cells were seeded and cultured in complete RPMI culture medium in the presence of recombinant GM-CSF at 41 • C, 5% CO 2 . At day 8, differentiated chBMDCs on glass coverslips were either washed three times with DPBS −/− or first stained with fluorescently labeled lectin wheat germ agglutinin (WGA)-Alexa Fluor 488 (Invitrogen TM , Life Technologies Europe BV, Bleiswijk, the Netherlands). Staining with WGA-Alexa Fluor 488 was performed by washing the cells twice with cold DPBS with calcium and magnesium, followed by staining with WGA-Alexa Fluor 488 diluted in DPBS with calcium and magnesium for 20 min at 4 • C. Fixation was performed in DPBS −/− with 4% paraformaldehyde ( Table 1). In addition, a nuclear staining was performed in 25 µl of blocking buffer at 10 µg/ml with 4 ′ ,6diamidino-2-phenylindole (DAPI) (Sigma-Aldrich, Saint Louis, MO, USA) for 5 min at RT. Between staining steps, the samples were washed three times with DPBS −/− with 0.05% Tween-20. The last wash step was performed in distilled water, before mounting the samples on Polysine R microscope slides (Menzel Glaser GmbH & Co KG, Braunschweig, Germany) in a FluorSave reagent (Calbiochem R , Merck Millipore, Burlington, MA, USA). The samples were captured using a TCS-SPE-II confocal microscope (Leica Microsystems B.V., Amsterdam, the Netherlands) equipped with 405-, 488-, 561-, and 635-nm diode lasers and processed using Fiji software (28).

Phagocytosis of IgY-Opsonized Beads by chBMDC Subsets
Chicken serum IgY fraction (Agrisera AB, Vännäs, Sweden) was added at 14.4 mg/ml to 1.44 × 10 10 beads per milliliter of 1µm crimson carboxylate-modified FluoSpheres (Invitrogen TM , Life Technologies Europe BV, Bleiswijk, the Netherlands) and incubated overnight on an orbital shaker at 4 • C to create IgYopsonized beads. The next day, the beads were washed twice and resuspended in DPBS −/− with centrifugation steps at 3,000 g for 20 min at 4 • C in between. To confirm IgY coupling, the beads were stained in FACS buffer with MαCh IgY-PE (SouthernBiotech, Birmingham, AL, USA) and analyzed on the CytoFLEX LX flow cytometer (data not shown). Next, the IgYcoupled beads were used in a phagocytosis assay to assess bead uptake by chBMDCs. After 8 days of culture in a 24-well plate, chBMDCs from one well were harvested and counted to determine the number of IgY-opsonized beads needed to obtain a 1:1 bead-to-cell ratio. Next, crimson beads were added to the remaining wells followed by 4-h incubation at 41 • C, 5% CO 2 , to allow phagocytosis by chBMDCs. Subsequently, chBMDCs were harvested and stained for flow cytometry or confocal microscopy according to the methods described above. For flow cytometry, the cells were stained for MHC-II expression and viability, using Zombie Aqua Fixable Viability Dye, and analyzed using the CytoFLEX LX flow cytometer. For confocal microscopy, the cells were stained with WGA-Alexa Fluor 488, MαCh-Ia BIOT, and streptavidin Alexa Fluor 405 and analyzed using the TCS-SPE-II confocal microscope.

IL-4 Bioactivity Assessment by 3 H-Thymidine Incorporation by PBMCs
The 3 H-thymidine incorporation assay to measure IL-4 bioactivity was modified from a published method to assess peripheral blood mononuclear cell (PBMC) proliferation (27). Heparinized blood was collected from healthy chickens (under registration number AVD108002016642-1 from the Dutch Central Authority for Scientific Procedures on Animals). The chickens were daily monitored by animal caretakers for signs and symptoms of disease, which were absent for the chickens used in this study. PBMCs were isolated from heparinized blood by density gradient separation using Ficoll-Paque PLUS (GE Healthcare, Chicago, IL, USA) according to standard procedure.

Separation of chBMDC Subsets by FACS
For sorting, chBMDCs were stained with antibodies specific for MHC-II, colony-stimulating factor 1 receptor (CSF1R), and CD80 as before. In addition, the dye 7-aminoactinomycin D (7-AAD; BD Biosciences, Pharmingen, San Diego, CA, USA) was added to the cells for viability assessment. Next, the cells were resuspended in DPBS −/− , 1% FBS, 2 mM EDTA, and 0.005% NaN 3 and flushed through a 70-µm cell strainer to create single-cell suspensions. MHC-II high CSF1R low and MHC-II low CSF1R high chBMDC subsets were sorted by FACS with a BD Influx cell sorter, equipped with 405-, 488-, 561-, and 635-nm lasers (BD Biosciences, Pharmingen, San Diego, CA, USA). Each chBMDC subset constituted close to 25% of the original sample. Approximately 2 × 10 6 cells were sorted for both subsets to perform quantitative real-time PCR (RT-qPCR). In parallel, the cells were analyzed before and after FACS using the CytoFLEX LX flow cytometer. Moreover, some cells were sorted onto a Polysine R microscope slide, fixed with 4% paraformaldehyde, and analyzed by confocal microscopy using the TCS-SPE-II microscope to confirm the expression patterns of MHC-II, CD80, and CSF1R by the chBMDC subsets. In addition, chBMDCs were sorted to evaluate the phenotypic stability of the cells by

Statistical Analysis
Statistical analysis was performed using GraphPad Prism 7 software (GraphPad Software, La Jolla, CA, USA). The data were tested for the assumptions of normally distributed data. Flow cytometry expression data showing the geometric mean fluorescent intensity (gMFI) was log-transformed in order to generate normally distributed data. Paired t-tests were used to test for statistically significant differences between MHC-II low and MHC-II high chBMDC subsets. A p-value of <0.05 was considered statistically significant.  Figure 1A). After 7 days of culturing, nearly all cells were found to be CD45 + and thus hematopoietic (Supplementary Figure 1B). In agreement with previous studies, chBMDCs formed clusters holding veiled cells (Figures 1A,C,  arrows), typical for DC morphology (14), and highly expressing MHC-II ( Figure 1B). MHC-II-expressing cell clusters did not appear in the absence of GM-CSF (Supplementary Figure 2). Beyond the clusters, cells with different morphologies were visible. These were large, round, without protrusions, and highly granular (Figure 1A, stars). Upon LPS stimulation, cell clusters became less dense, and many individual small cells with elongated protrusions became visible, a feature typical for mature DCs (Figure 1D, arrows). Next, a phenotypic analysis of the chBMDC culture by flow cytometry identified three subsets, distinguished by forward scatter (FSC) vs. MHC-II expression: FSC low with no or low expression of MHC-II (FSC low ), FSC int with high expression of MHC-II (MHC-II high ), and FSC high with low expression of MHC-II (MHC-II low ) (Figure 2A). These chBMDC subsets were evaluated for the expression of myeloid markers including integrin CD11b/c, costimulatory receptors CD40 and CD80, CSF1R, stem cell growth factor receptor c-kit, mannose receptor C-type 1-like B (MRC1L-B), non-classical MHC molecule CD1.1, and MHC class I component β 2 -microglobulin (β2m). FSC low cells showed high expression of CD11b/c, but no expression of CD40 and CD1.1 (Figure 2B). CD80, c-kit, MRC1L-B, and β2m were expressed at moderate levels. CSF1R and c-kit were expressed by some FSC low cells, but not by others, showing further heterogeneity within this subset. Since FSC low cells were largely positive for MRC1L-B but showed no or low expression of MHC molecules and costimulatory molecules, these were likely to represent undifferentiated monocytes. Both MHC-II high and MHC-II low cells showed expression of MHC molecules and costimulatory molecules. Compared to MHC-II low cells, MHC-II high cells expressed higher levels of costimulatory receptors CD40 and CD80, CD1.1, and MRC1L-B (Figures 2B,C and Supplementary Table 1). Conversely, MHC-II low cells expressed higher levels of CD11b/c, CSF1R, c-kit, and β2m. Taken together, MHC-II low and MHC-II high chBMDC subsets both showed a phenotype of antigen-presenting cells, but differentially expressed many myeloid markers.

RESULTS
Next, the expression patterns of CSF1R, MRC1L-B, and CD80 by MHC-II low and MHC-II high chBMDC subsets were evaluated by immunofluorescent confocal microscopy. The MHC-II low subset expressed higher levels of CSF1R than the MHC-II high subset (Figure 3A), in accordance with the flow cytometry data (Figure 2C). Therefore, CSF1R could be used as an additional marker to discriminate between chBMDC subsets. MHC-II low CSF1R high cells were found to be large and round and to have few protrusions, indicative of a macrophage-like morphology. Similar cells were observed by light microscopy (Figure 1A, stars). In contrast, MHC-II high CSF1R low cells showed irregular shapes with many protrusions, indicative of a DC-like morphology, and resemble the veiled cells that were observed by light microscopy (Figure 1A). MRC1L-B and CD80 expression levels were found to be highest on MHC-II high chBMDCs (Figures 3B,C), in accordance with the flow cytometry data (Figure 2C). Nevertheless, MRC1L-B and CD80 did not colocalize with MHC-II. MRC1L-B and CD80 were mainly found in intracellular compartments, whereas MHC-II was found more on the cellular surface of MHC-II high -expressing chBMDCs. A stronger colocalization was observed between MRC1L-B and CD80 ( Figure 3D).

MHC-II low chBMDCs Have a Higher Capacity to Phagocytose Fluorescent Beads Compared to MHC-II high Cells
To assess whether the phenotypic distinction between MHC-II low and MHC-II high chBMDCs was functionally relevant, the subsets were assessed for their ability to phagocytose chicken IgY-coated fluorescent latex beads. First, the uptake of IgYcoated crimson fluorescent beads by chBMDCs was confirmed by showing that the beads localize beneath the surface of the plasma membrane, which was visualized using WGA ( Figure 4A and Supplementary Video 1). chBMDCs were stained for MHC-II to identify the MHC-II high subpopulation. Both MHC-II low and MHC-II high chBMDCs were found to take up beads as determined by confocal immunofluorescent microscopy. Next, the bead content of the chBMDC subsets was quantified by flow cytometry (Figure 4B). On average, MHC-II low cells (0.54 beads per cell) contained 2.4 times more beads than MHC-II high cells (0.23 beads per cell), which shows that the MHC-II low and MHC-II high chBMDC subsets differ in opsonophagocytic capacity ( Figure 4C). A major part of IgY beads actually bound to the cells instead of being taken up, as shown in a separate experiment performed at 4 • C (Supplementary Figure 3). However, when bound IgY beads from the experiment performed at 4 • C were subtracted from the experiment performed at 41 • C, MHC-II low cells were still found to take up 2.7 times more beads than MHC-II high cells. The FSC low chBMDC subset showed little uptake of beads (0.048 beads per cell) ( Figure 4B).

Addition of Recombinant IL-4 to chBMDC Cultures Leads to a Smaller Proportion of MHC-II high Cells
The effect of recombinant IL-4 on the generation of chBMDC subsets was investigated, since this cytokine has been used to generate chBMDCs by others (14). To confirm that recombinant IL-4, produced in COS-7 cells, was biologically active, its ability to induce PBMC proliferation was demonstrated (Supplementary Figure 4). Next, IL-4 was given to chBMDCs alone or in combination with GM-CSF. IL-4 alone led to few MHC-II-expressing clusters of chBMDCs (Supplementary Figure 2). IL-4 in combination with GM-CSF led to many MHC-II-expressing clusters, similar to GM-CSF alone. Next, the proportion of MHC-II low and MHC-II high cells was quantified by flow cytometry. The addition of IL-4 to the standard chBMDC culture with GM-CSF was found to    increase the proportion of the MHC-II low from 32.6 to 49.7% at the highest administered dose (1/25 dilution) ( Figure 5A). The increase in the proportion of MHC-II low cells occurred largely at the expense of the MHC-II high subset, which changed proportionally from 35.0 to 23.0% at the highest administered dose of IL-4 ( Figure 5B). The proportion of FSC low cells remained fairly stable (data not shown).

LPS Stimulation Affects the Difference in Phenotype Between MHC-II low and MHC-II high chBMDCs
To determine the effect of LPS, commonly used to induce BMDC maturation, the cells cultured for 7 days were stimulated with 100 ng/ml of LPS for 24 h. MHC-II low and MHC-II high chBMDC subsets were still detected ( Figure 6A). Both MHC-II low and MHC-II high chBMDCs upregulated CD40, CD1.1, and β2m expression, whereas the subsets downregulated c-kit and MRC1L-B expression (Figures 6B,C). CD11b/c, CSF1R, and ckit expression on MHC-II low cells decreased to levels similar to those on the MHC-II high subset, while CD80 expression on MHC-II high cells decreased to a level similar to that on the MHC-II low subset. These expression patterns suggest that the phenotypes of the MHC-II low and MHC-II high subsets partially converged. Convergence in expression levels was also observed for CD40 and β2m, but not for MRC1L-B and CD1.1.

chBMDC Subsets Differ in Maturation Status Rather Than Cell Type
MHC-II low and MHC-II high chBMDC subsets were sorted by FACS to determine the differential expression of macrophageand DC-related genes by RT-qPCR, since well-characterized monoclonal antibodies for these cell surface markers in chickens are scarce. Since MHC-II and CSF1R showed good discrimination between the subsets in confocal microscopy (Figure 3), these markers were used to separate the subsets by FACS (Figure 7A). In addition, chBMDCs were stained for CD80, which was found to be more highly expressed by the MHC-II high CSF1R low subset than the MHC-II low CSF1R high subset (Figure 7A), in accordance with previous results (Figure 2C). Both subsets were sorted to above 90% purity as determined by flow cytometric reanalysis after each sort ( Figure 7B). Sorted cells were analyzed by fluorescent microscopy to confirm surface expression patterns of MHC-II, CSF1R, and CD80 ( Figure 7C). In accordance with the flow cytometry data, MHC-II and CD80 were expressed by the MHC-II high CSF1R low sorted subset, but not by the MHC-II high CSF1R low subset. In contrast, CSF1R was shown to be present on both sorted subsets. Next, RNA was isolated from the sorted subsets, and RT-qPCR was performed to study the gene expression patterns of macrophage-and DC-enriched genes ( Figure 7D and Supplementary Table 2). Tyrosine-protein kinase Mer (MERTK), toll-like receptor 4 (TLR4), TLR4 coreceptor CD14, and inducible nitric oxide synthase (iNOS) were used as macrophage-enriched genes, whereas zinc finger and BTB domain-containing protein 46 (ZBTB46), C-type lectins DEC205 and DC-SIGN, chemokine receptors C-C chemokine receptor type 6 (CCR6) and 7 (CCR7), and costimulatory receptor CD83 were used as DC- Thus, the differences between MHC-II low cells and MHC-II high cells were mainly found for DC-enriched genes, especially for CCR7 and CD83, which have been used by others before as maturation markers of chBMDCs (14,30). In contrast, differences in gene expression were hardly found for macrophage-enriched genes. Combined with the flow cytometric data (Figure 2C) that showed higher expressions of MHC-II, CD40, and CD80 by MHC-II high chBMDCs compared to the MHC-II low chBMDCs, the subsets seem to be DCs at different maturation states rather than different cell types.

MHC-II high chBMDCs Become MHC-II low During Prolonged Incubation After Sorting
MHC-II low and MHC-II high chBMDCs were found to differ in the expression of DC maturation markers, at both the protein and gene expression levels, but were similar in their ability to induce PBMCs to proliferate. To gain additional proof that MHC-II low and MHC-II high chBMDCs represent DCs at different maturation states rather than different cell types, sorted chBMDCs were reseeded for another 1 or 3 days of cell culture and stained again for MHC-II, CSF1R, and CD80. The MHC-II low subset showed only minor changes in the expression of abovementioned markers (Figures 9A,B). In contrast, the MHC-II high subset showed higher CSF1R, lower MHC-II, and lower CD80 expression and consequently became phenotypically more similar to the MHC-II low subset. Therefore, both chBMDC subsets seem to consist of a DC-like cell type, but these appear to be different states of maturation with MHC-II high chBMDCs being at a more mature but reversible state. In addition, the mRNA expression levels of CCR7 and CD83 were determined and found to be decreased for both subsets after prolonged cell culture ( Figure 9C).

DISCUSSION
The present study aimed to characterize GM-CSF-induced chBMDC cultures and address their heterogeneity. Despite the widespread use of in vitro grown DCs, there is still an ongoing debate about the representativeness of in vitro grown DCs for their in vivo counterparts (20-22, 24, 25, 31). In research so far, discussion focused mainly on BMDCs from mice and less on those of farm and companion animals. Early studies already described the presence of macrophage and granulocyte "contaminants" in the murine BMDC culture (5,32). Moreover, murine (20,33) and ovine (12) BMDC cultures, as well as bovine monocyte-derived DC cultures (34), have been shown to include CD11b low MHC-II high cells with a DC-like phenotype and CD11b high MHC-II low cells with a macrophage-like phenotype. The first study by Wu et al. describing the chBMDC culture already recognized its heterogeneity (14). However, this study excluded adherent and relatively small cells from analysis, which may, respectively, represent the MHC-II low chBMDCs and FSC low undifferentiated monocytes of the current study. In our opinion, these neglected cells should be characterized to interpret responses of the chBMDC culture correctly when analyzed in bulk. This has been illustrated by studies using LPSstimulated murine BMDCs, in which individual cell subsets (20) or individual cells (35) were shown to respond very differently in their maturation, cytokine expression profile, and capacity to induce T-cell proliferation. Moreover, virus infection studies with chBMDCs, including avian influenza virus (16,26) and infectious bursal disease virus (36) were analyzed in bulk, while the different chBMDC subsets might vary in their susceptibility for viruses and influence the outcome of these studies.
In the present study, the chBMDC culture was found to contain MHC-II low and MHC-II high subsets, similar to BMDC and monocyte-derived DC cultures in other species (12,20,33,34). Expression of myeloid markers, morphology, and phagocytosis capacity differed between MHC-II low and MHC-II high subsets. Compared to the MHC-II high subset, the MHC-II low subset showed higher expression of CSF1R, contained larger and more granular cells based on flow cytometry scatter profile and light microscopy, and showed a higher capacity to phagocytose IgY-opsonized beads. These results suggested that the MHC-II low chBMDC subset consists of macrophagelike cells, in agreement with studies in murine (20, 33) and FIGURE 9 | Sorted major histocompatibility complex class II (MHC-II) low chicken bone marrow-derived dendritic cells (chBMDCs) maintain their phenotype, while sorted MHC-II high chBMDCs become MHC-II low . (A) The MHC-II low and MHC-II high chBMDC subsets were sorted using the gating strategy shown in Figure 7A. The cells were either reanalyzed directly after sorting or seeded for another 1 or 3 days of prolonged cell culture in the presence of granulocyte-macrophage colony-stimulating factor (GM-CSF). After these periods, the cells were stained again using the same protocol that was used before the sort. The contour plots that are shown are representative for three independent technical replicates. (B) Surface expression levels of MHC-II, CD80, and colony-stimulating factor 1 receptor (CSF1R) are shown for MHC-II low and MHC-II high chBMDCs directly after sorting (0 h) and after the 3 days of prolonged cell culture (72 h) (n = 3). (C) Similarly, messenger RNA (mRNA) expression levels of CD83 and CCR7 are shown for both subsets at 0 and 72 h (one experiment, performed in triplicate). ovine (12) BMDCs, as well as bovine monocyte-derived DCs (34). In these studies, macrophage-like MHC-II low cells showed high expression of CD11b, which contributed to the distinction between the MHC-II low and MHC-II high subsets of these BMDC cultures. Unfortunately, this was not possible for the chicken CD11b/c antigen, since the corresponding antibody has not yet been confirmed to recognize either CD11b or CD11c (26). Compared to the MHC-II low chBMDC subset, the MHC-II high subset showed higher expression of costimulatory molecules CD40 and CD80, in agreement with the DC-like phenotype shown for MHC-II high cells in murine BMDC (20,33) and bovine monocyte-derived DC cultures (34). In addition, the MHC-II high chBMDC subset showed a relatively high expression of the non-classical MHC molecule CD1.1, which is also indicative of a DC-like phenotype (37). Our phenotypical findings of the chBMDC subsets suggested that MHC-II low and MHC-II high subsets, respectively, represent macrophage-and DC-like cells, in agreement with studies that used murine, ovine, and bovine in vitro DC cultures (12,20,33,34). However, there is some discrepancy in the literature about murine BMDCs, since its MHC-II low and MHC-II high subsets have also been suggested to, respectively, represent an immature and mature phenotype of the same cell type (3,38).
To further explore this alternative hypothesis, we investigated whether the chBMDC subsets were different cell types or DCs at different maturation states; MHC-II low and MHC-II high subsets were sorted to perform RT-qPCR for macrophage-and DCspecific markers. Both subsets showed similar expression of the macrophage-specific markers MERTK, TLR4, and CD14, which is in contrast to studies of murine BMDCs (20,33,39). Moreover, no difference in expression of the DC-specific marker DEC205 was observed, while ZBTB46 and CCR6 were only moderately more highly expressed by the MHC-II high subset. The most striking differences between the chBMDC subsets were higher CD83 and CCR7 expressions and lower DC-SIGN expression for the MHC-II high subset. Increased CD83 and CCR7 expression and decreased DC-SIGN expression have been reported as maturation signatures of human and murine monocyte-derived DCs (20,40,41), suggesting that the MHC-II high chBMDC subset represents a mature DC phenotype. Moreover, it has been shown that CCR7 expression is upregulated by chBMDCs shortly after LPS stimulation (36). Taken together, these results suggest that, rather than being different cell types, MHC-II low and MHC-II high chBMDC subsets are DCs at different maturation states with MHC-II high chBMDC being more mature. Nonetheless, MHC-II low and MHC-II high chBMDCs induced similar levels of PBMC proliferation in an allogeneic mixed lymphocyte reaction. Therefore, we hypothesize that of MHC-II high chBMDCs are in a semi-mature state, which is described in literature as the state at which DCs express high levels of MHC-II and costimulatory molecules but do not produce elevated levels of pro-inflammatory cytokines or optimally stimulate Tcell proliferation (42,43). Reseeding sorted chBMDCs showed that the immature phenotype of the MHC-II low subset remained stable over time, whereas the semi-mature MHC-II high subset decreased MHC-II and CD80 expressions and increased CSF1R expression, indicating the plasticity and reversibility of this semi-mature phenotype. The chBMDC subsets differed in phagocytosis capacity, with the MHC-II low subset being more efficient in bead uptake. This finding can also be explained by a different maturation status of the subsets, since mature DCs generally have a lower phagocytotic capacity (44,45). Others observed that LPS-induced maturation diminished the phagocytosis capacity of chBMDCs (14), which was not the case for the semi-mature MHC-II high chBMDC subset of the present study.
The semi-mature phenotype of the MHC-II high chBMDCs must have been induced by the culture conditions that were used, since the cells were not intentionally stimulated. Since it is common practice to include IL-4 cytokine in BMDC differentiation protocols of different species, including chBMDCs (10-12, 14, 46, 47), the effect of IL-4 on the development of chBMDC subsets was investigated. When IL-4 alone was added to the culture, we observed that few chBMDC aggregates appeared, in agreement with previous studies (17). Surprisingly, addition of IL-4 led to a lower proportion of MHC-II high cells, inhibiting chBMDC maturation, though the number of observed chBMDC aggregates remained unaffected. To the best of our knowledge, there are no earlier reports that show an inhibitory effect of IL-4 on maturation of the chBMDC culture in terms of MHC-II expression. Previous studies have only observed the occurrence of chBMDC aggregates when investigating the effects of IL-4 (14,17). Studies with murine and rat BMDCs have reported that IL-4 supplementation leads to proportionally larger MHC-II high subsets (20,47), in contrast to our findings for the chBMDC culture. Another study has shown that IL-4 has no effect on ovine BMDC yield or phenotype (12). Therefore, the effect of IL-4 on BMDC cultures seems to differ between species. Another parameter affecting BMDCs was the source of the serum used in the culture. FBS was found to lead to a large MHC-II high CD80 + chBMDC population (Supplementary Figure 5), whereas chicken serum led to the immature MHC-II low and semi-mature MHC-II high chBMDCs of the present study.
BMDCs are often stimulated by LPS to induce maturation. In the chBMDC culture, LPS stimulation led to a striking increase in CD40 and CD1.1 expression by both chBMDC subsets, whereas expression of maturation marker CD80 by the MHC-II high subset was unexpectedly decreased. Of note, LPS stimulation led to smaller differences between the subsets in their expression of MHC-II, CD80, and CSF1R. Overall, MHC-II low and MHC-II high chBMDCs responded similarly to the LPS stimulus, which favored the hypothesis that the subsets reflected one cell type at different states. In contrast, murine MHC-II low and MHC-II high BMDC subsets were shown to maintain differential gene expression profiles after LPS stimulation, which provided additional proof that the murine subsets were truly different cell types (20).
In conclusion, this study describes the heterogeneity of the GM-CSF-differentiated chBMDC culture, which comprised MHC-II low and MHC-II high subsets that both possess features of antigen-presenting cells. These populations were found to differ in phenotype, morphology, and their phagocytosis capacity, whereas their ability to induce PBMC proliferation was similar. Based on higher expressions of maturation markers MHC-II, CD40, CD80, CD83, and CCR7 by MHC-II high chBMDCs compared to MHC-II low chBMDCs, the MHC-II low and MHC-II high subsets were found to, respectively, represent immature and semi-mature chBMDCs. The semi-mature phenotype of the MHC-II high subset was found to be reversible, since reseeding and prolonged culture of these cells led to a transition toward the immature phenotype of the MHC-II low cells. Taken together, these results yield a thorough characterization of the chBMDC culture and explain its heterogeneity by the simultaneous presence of immature and mature subsets. Our findings are instrumental for the interpretation of experiments that use this culture in future research.

DATA AVAILABILITY STATEMENT
All datasets generated for this study are included in the article/Supplementary Material.

ETHICS STATEMENT
The animal study was reviewed and approved by the Dutch Central Authority for Scientific Procedures on Animals.

AUTHOR CONTRIBUTIONS
RB and CJ designed the research. RB and GA performed the research. RB analyzed the data. VR, WE, and CJ supervised the work. RB, VR, WE, and CJ wrote the paper.