Identification of dendritic cell precursor from the CD11c+ cells expressing high levels of MHC class II molecules in the culture of bone marrow with FLT3 ligand

Dendritic cells (DCs) are readily generated from the culture of mouse bone marrow (BM) treated with either granulocyte macrophage-colony stimulating factor (GM-CSF) or FMS-like tyrosine kinase 3 ligand (FLT3L). CD11c+MHCII+ or CD11c+MHCIIhi cells are routinely isolated from those BM cultures and generally used as in vitro-generated DCs for a variety of experiments and therapies. Here, we examined CD11c+ cells in the BM culture with GM-CSF or FLT3L by staining with a monoclonal antibody 2A1 that is known to recognize mature or activated DCs. Most of the cells within the CD11c+MHCIIhi DC gate were 2A1+ in the BM culture with GM-CSF (GM-BM culture). In the BM culture with FLT3L (FL-BM culture), almost of all the CD11c+MHCIIhi cells were within the classical DC2 (cDC2) gate. The analysis of FL-BM culture revealed that a majority of cDC2-gated CD11c+MHCIIhi cells exhibited a 2A1-CD83-CD115+CX3CR1+ phenotype, and the others consisted of 2A1+CD83+CD115-CX3CR1- and 2A1-CD83-CD115-CX3CR1- cells. According to the antigen uptake and presentation, morphologies, and gene expression profiles, 2A1-CD83-CD115-CX3CR1- cells were immature cDC2s and 2A1+CD83+CD115-CX3CR1- cells were mature cDC2s. Unexpectedly, however, 2A1-CD83-CD115+CX3CR1+ cells, the most abundant cDC2-gated MHCIIhi cell subset in FL-BM culture, were non-DCs. Adoptive cell transfer experiments in the FL-BM culture confirmed that the cDC2-gated MHCIIhi non-DCs were precursors to cDC2s, i.e., MHCIIhi pre-cDC2s. MHCIIhi pre-cDC2s also expressed the higher level of DC-specific transcription factor Zbtb46 as similarly as immature cDC2s. Besides, MHCIIhi pre-cDC2s were generated only from pre-cDCs and common DC progenitor (CDP) cells but not from monocytes and common monocyte progenitor (cMoP) cells, verifying that MHCIIhi pre-cDC2s are close lineage to cDCs. All in all, our study identified and characterized a new cDC precursor, exhibiting a CD11c+MHCIIhiCD115+CX3CR1+ phenotype, in FL-BM culture.

CD11c + DCs efficiently develop after culture of bone marrow (BM) with GM-CSF or FLT3 ligand (FLT3L) for 6 to 10 days (8-10), which has been widely used as in vitro-generated BM-derived DCs (BM-DCs) in various experiments.Apparently, however, not all the CD11c + cells generated from those BM cultures are a homogeneous population.In the BM culture with GM-CSF (GM-BM culture), DCs and Macs are abundantly found among the CD11c + cells.While MDPs, cMoPs, and monocytes differentiate into both DCs and Macs, CDPs preferentially generate DCs in GM-BM culture (11).Meanwhile, in the CD11c + cells generated from the BM culture with FLT3L (FL-BM culture), a variety of DC subsets such as pDCs, classical type 1 DCs (cDC1s) and cDC2s are found (10).
2A1 is one of the monoclonal antibodies (mAbs) generated by early efforts to produce new reagents that specifically detected DCs in GM-BM culture as well as in lymphoid tissues in situ (8,(12)(13)(14)(15).The cells detected by mAb 2A1, i.e., 2A1 + cells are abundantly present in the T cell rich regions of lymph nodes, spleen, and Peyer's patch, suggesting that 2A1 marks a subset of DCs closely located near T cells in lymphoid tissues (16)(17)(18).Analysis of GM-BM culture also revealed that only the MHCII hi cells were detected by mAb 2A1 (8,14).Based on those observations, 2A1 + DCs have been considered as activated or mature DCs (19).However, unlike other well-characterized surface markers for DCs, publications examining 2A1 + DCs have been rare because mAb 2A1 recognizes an unknown antigen localized intracellularly in DCs.As a result, the function of 2A1 antigen (hereafter 2A1) in DCs is still unknown.
Here, we set out to examine the expression of 2A1 in CD11c + cells from GM-BM and FL-BM cultures.All the DCs in GM-BM culture (GM-DCs) highly expressed 2A1, whereas only small fractions of cDC-gated cell subsets in FL-BM culture were 2A1 + .In FL-BM culture, most of those 2A1 + cells were found within the cells gated for cDC2 subset expressing high levels of MHC class II molecules (MHCII hi ), and more than a half of CD11c + MHCII hi cDC2-gated cells were 2A1 -.Up to now, together with GM-DCs, CD11c + MHCII + cells in FL-BM culture have been widely used as in vitro-generated DCs (20 -22).Surprisingly, however, we found only a half of the CD11c + MHCII hi cells in FL-BM culture were DCs.The remaining half of the CD11c + MHCII hi cells in FL-BM culture were not DCs but DC precursor cells which were poor at antigen presentation.In this report, we identified and characterized a novel DC precursor population, which d isplayed a CD11c + MHCII hi CD115 + CX 3 CR1 + phenotype, in FL-BM culture.

Antibody purification and labeling
Supernatants were collected from the mAb hybridoma cultures of 2A1 and L5 (an isotype control for 2A1) (25) in DMC7 medium and filtrated before purification by affinity chromatography with protein G bead (Pierce, Rockford, IL; GE Healthcare, Piscataway, NJ).For those hybridoma cultures, DMC7 medium was prepared with ultra-low IgG FBS (Thermo Fisher Scientific) to remove bovine IgG and increase the purity of those mAbs (Supplementary Figure 1).Then, mAbs were concentrated using Amicon ® Ultra-4 Centrifugal Filter (Millipore, Burlington, MA) before quantification.Purified mAbs were fluorescently labeled using Alexa Fluor ™ 488 Antibody Labeling Kit (Invitrogen) or Alexa Fluor ™ 647 Antibody Labeling Kit (Invitrogen) according to the manufacturer's instructions.

Bone marrow culture
Mice were euthanized by asphyxiation in a CO 2 chamber.Bone marrow cells were isolated from femurs and tibias of C57BL/6 mice at 8-14 weeks of age under sterile conditions, as described (9,(26)(27)(28), with some modifications.Briefly, femurs and tibias were detached from mice and purified from the surrounding muscle tissue by rubbing with sterile gauze followed by disinfecting in 70% ethanol for 3 minutes and washing with fresh DMEM (8).Next, the marrow was isolated from the bones by flushing using a syringe with a 23-gauge needle and disintegrated by vigorous pipetting.After lysing erythrocytes with RBC lysis buffer (BioLegend), single cell suspension of BM cells was generated by sieving through cell strainers (100 mm; SPL Life Sciences, Pocheon, Korea), counted, and cultured at 1×10 6 cells per well for GM-BM culture or at 2×10 6 cells per well for FL-BM culture in DMC7 medium containing GM-CSF or FLT3L with 24-well tissue culture plates for 8-10 days, as described previously (26,27,29).Specifically, 0.5 mg/ml of GM-CSF and 10 mg/ml of FLT3L were used for GM-BM and FL-BM cultures, respectively.During the culture, half of the medium in each well was carefully removed and replenished with fresh DMC7 medium containing GM-CSF or FLT3L every 2 days until harvest for use in subsequent experiments.

Flow cytometry
Single cell suspension was prepared from mouse tissues or cultures and incubated with Fc receptor blocking mAb 2.4G2 for 20 minutes at 4°C followed by washing with FACS buffer (DPBS containing 2% FBS, 2mM EDTA, and 0.1% sodium azide).Then, cells were incubated with appropriate cocktails of fluorochromeconjugated mAbs and dead cell staining dye for 30 minutes at 4°C.For intracellular staining, cells were stained for surface markers before fixation and permeabilization followed by staining intracellular molecules according to the manufacturer's i n s t r u c t i o n s ( F i x a t i o n B u ff e r / I n t r a c e l l u l a r S t a i n i n g Permeabilization Wash Buffer, BioLegend).Multiparameter analysis of each sample was performed on LSRFortessa ™ flow cytometer (BD Biosciences) and flow cytometric isolation of cells was performed on BD FACSAria ™ II cell sorter (BD Biosciences) at the Flow Cytometry Core Facility of the Yonsei University College of Medicine.Collected data were analyzed with FlowJo software (BD Biosciences).

Treatment of TLR agonists
Graded doses of LPS (Sigma-Aldrich) were added into BM cultures at culture day 8.After 18 hours of incubation, cells were harvested, washed, and analyzed using flow cytometry.LPS at a final concentration of 1 mg/ml was used for in vitro stimulation of DCs unless otherwise noted.BM cultures at day 8 were also treated with CpG (1 mM of ODN 2216; TLRGRADE ® , Enzo Life Science, Farmingdale, NY) or poly I:C HMW (1 mg/ml; InvivoGen) for 18 hours.

Antigen uptake
Cultures of BM were treated with FITC-labeled OVA (100 mg/ ml) or YGM beads (0.000675%) for 1 hour at 37°C or at 4°C as control.Then, each sample was washed twice with cold DMEM before being stained with appropriate reagents for flow cytometric analysis as described above.

Generation of CHO cells expressing OVA antigens intracellularly
CHO/OT.EGFP cells, stable CHO cell-lines expressing OVA antigens, were generated as described previously (30)(31)(32).In brief, a cDNA construct encoding the open reading frame of OT.EGFP (GenBank accession number OR405547) expressing a fusion protein between enhanced green fluorescent protein (EGFP) and OVA antigen epitope peptides (i.e., OT-1 and OT-2) were cloned by PCR, sequenced, and inserted into the pCMV mammalian expression vector (Clontech, TaKaRa Bio, Shiga, Japan).Then, the stable CHO cells were generated by transfection with pCMV-OT.EGFP by Lipofectamine ™ 2000 reagent (Thermo Fisher Scientific) followed by selection of G418 (1.5 mg/ml) in DMC7 medium.

Antigen presentation and T cell proliferation
Assays for antigen presentation and T cell proliferation were performed as described previously (33).In brief, APCs were pulsed with 100 mg/ml of OVA on culture plates at 37°C for 30 minutes before being harvested by gentle pipetting and washing.Then, specific populations of APCs were isolated with flow cytometric sorter as described above.The enrichment of naive CD8 + OT-1 and CD4 + OT-2 T cells from the splenocytes of OT-1 and OT-2 mice respectively were described previously (33,34).Naive OT-1 and OT-2 T cells were labeled with CFSE or CTV at 37°C for 10 minutes and washed with DPBS.Then, T cells (2.5×10 4 ) were cultured with graded doses of APCs in 96-well round-bottom cell culture plates with DMC7 medium containing 57.2 mM b-mercaptoethanol (Sigma-Aldrich) for 3-5 days.The proliferation of T cells was assessed by the dilution of CFSE or CTV.Unpulsed APCs were co-cultured with OT-1 or OT-2 T cells in the presence of OVA (10 mg/ml) with/without LPS (1 mg/ml).For the cross-presentation assays with cell-associated antigens, 2.5×10 4 CTV-labeled naive CD8 + OT-1 T cells were co-cultured with APCs plus heat-treated control CHO cells or CHO/OT.EGFP cells (1×10 5 cells/well) for 4 days.Apoptotic CHO cells were prepared by treatment with heat at 45°C for 30 minutes.Then, the heat-treated CHO cells were plated in 24-well plates overnight before being used in the subsequent coculture experiments.For the maturation of FL-cDC1s, media containing 20 ng/ml of GM-CSF were used for the co-culture of APCs and T cells.

Microscopic analysis
On the day 8 or 9 of BM culture, cells were harvested by gentle pipetting and stained with appropriate fluorochrome-conjugated antibodies.Each subset of the cells was isolated according to the appropriate gating strategies.Then, the isolated cells were cultured with DMC7 medium containing GM-CSF or FLT3L in 96-or 48well flat-bottom cell culture plates overnight.Phase-contrast observations of cultures were made by means of eclipse TS100 microscope (Nikon, Tokyo, Japan) at 600× magnification.Before staining, 1×10 5 cells were prepared on slides by cytocentrifugation (Cytospin 4, Thermo scientific) at 650 rpm for 5 minutes.For Giemsa staining, cells were visualized by Diff-Quik (Sysmex, Kobe, Japan) staining of cytospin preparations.The images were acquired with IX73 fluorescent inverted microscopes (Olympus, Tokyo, Japan).Data analysis was conducted using ImageJ.

Transmission electron microscopy
Cells were fixed with 2% Glutaraldehyde and 2% Paraformaldehyde in 0.1M phosphate buffer (pH 7.4) for 12 hours; washed with 0.1M phosphate buffer; post-fixed with 1% OsO4 in 0.1M phosphate buffer for 2 hours; dehydrated with a series of ascending ethanol (50, 60, 70, 80, 90, 95, 100, and 100%) for 10 minutes each; infiltrated with propylene oxide for 10 minutes.Then, specimens were embedded with a Poly/Bed 812 kit (Polysciences) and polymerized in an electron microscope oven (TD-700, DOSAKA, Japan) at 65°C for 12 hours.The block was cut with a diamond knife in the ultramicrotome into 200 nm semi-thin section and stained toluidine blue for observation of optical microscope.The region of interest was then cut into 80 nm thin sections using the ultramicrotome, placed on copper grids, double stained with 3% uranyl acetate for 30 minutes and 3% lead citrate for 7 minutes, and imaged with a transmission electron microscopy (JEM-1011, JEOL, Tokyo, Japan) at the acceleration voltage of 80 kV equipped with a Megaview III CCD camera (Soft Imaging System, Germany).

RNA sequencing and analysis
Cell subsets in the BM cultures were stained and sorted according to the suitable gating strategies as described above.Total RNA was extracted twice for each cell subset by MiniBEST universal RNA extraction kit (TaKaRa Bio) from at least 1×10 5 cells.However, while cDC1-gated cells were fewer than 20% of the total cDC-gated cells in FL-BM culture, the number of CD83 + cDC1s were significantly less than 10% of the cDC1-gated cells.As a result, we were able to obtain an enough amount and high quality of RNAs only once after many trials and errors of sorting the live CD83 + cDC1s from FL-BM culture.Subsequent RNA sequencing (RNAseq) procedure and analysis were performed by Macrogen (Seoul, Korea) as follows.Reverse transcription of mRNA and generation of cDNA library were carried out with SMARTer Ultra low input RNA library kit.Sequencing was performed with Illumina NovaSeq (Illumina, San Diego, CA).The raw reads from the sequencer were preprocessed to remove low quality and adapter sequence before analysis to align the processed reads to the Mus musculus (mm10) using HISAT v2.1.0.After alignment, StringTie v1.3.4d was used to measure the relative abundances of genes in FPKM (Fragments Per Kilobase of exon per Million mapped fragments).Multidimensional scaling method was used to visualize the similarities among samples.The larger the dissimilarity between two samples, the further apart the points representing the experiments in the picture should be.Euclidean distance was applied as the measure of dissimilarity.Hierarchical clustering analysis was also performed using complete linkage and Euclidean distance as a measure of similarity to display the expression patterns of differentially expressed transcripts which were satisfied with |fold change|≥2.Data analysis and visualization of differentially expressed genes was conducted using GraphPad Prism 8 (GraphPad Software, La Jolla, CA) R 3.5.1 (https://www.r-project.org),Multiple Experiment Viewer software (MeV), and Morpheus software (https:// software.broadinstitute.org/morpheus).Venn diagram analysis was performed using the InteractiVenn website (http:// www.interactivenn.net).Enrichment analyses were performed with GSEA software (Broad Institute) on the gene set curated in the Molecular Signature Database (MSigDb).

Statistical analysis
Experiments with multiplicate samples were analyzed for statistical comparisons between different groups using with ANOVA and unpaired Student's t-test using GraphPad Prism 5 (GraphPad Software, La Jolla, CA).Statistical significance is denoted by the p values equal or below 0.05 (*), 0.01 (**), and 0.001 (***).Data were plotted for graphs with GraphPad Prism 5 and GraphPad Prism 8.

2A1 + cells are significantly generated in both GM-BM and FL-BM cultures
Highly pure mAb 2A1 was prepared (Supplementary Figure 1) and the detection of 2A1 required the permeabilization of cells to stain with the mAb (Supplementary Figure 2) indicating the intracellular localization of 2A1, as reported previously (8).We examined the expression of 2A1 among the cells in GM-BM and FL-BM cultures at various time points (Figures 1A, B; Supplementary Figure 3).2A1 + cells were hardly found in the BM, but became readily detectable among the MHCII hi cells of both GM-BM and FL-BM cultures in 2 to 3 days after the culture.After a week of culture, around 15% of the CD11c + cells in GM-BM culture and around 10% of CD11c + cells in FL-BM culture became 2A1 + .
2A1 + Cells are within the DC gate and cells within the DC gate are 2A1 + in GM-BM culture CD11c + cells in GM-BM culture (Supplementary Figure 4) are grouped into CD11b hi MHCII int/lo GM-Macs (Macs in GM-BM culture) and CD11b int MHCII hi GM-DCs (11).When the 2A1 + cells were backgated on those two populations in GM-BM culture, virtually all the 2A1 + cells were within the GM-DC gate (Figure 2A).Conversely, virtually all the cells within the GM-DC gate in GM-BM culture were 2A1 + (Figure 2B; Supplementary Figure 5).In addition to 2A1, GM-DCs expressed higher levels of CD80, CD83, and CD86; GM-Macs exhibited higher levels of CD115, CX 3 CR1, and CCR9 (Figure 2B; Supplementary Figure 5).

MHCII hi cDC2-gated cells in FL-BM culture are functionally and morphologically heterogeneous
Currently, CD11c + MHCII + or CD11c + MHCII hi cells in FL-BM culture are widely used as in vitro-generated DCs (20-22).Since most of the CD11c + MHCII hi cells in FL-BM culture were within the cDC2 gate (Figures 3A, B) and those CD11c + MHCII hi cells consisted of CD83 -CD115 + CX 3 CR1 + , CD83 -CD115 -CX 3 CR1 -, and CD83 + CD115 - CX 3 CR1 -subpopulations (Figure 3F), we set out to characterize the DC features of each subpopulation within the MHCII hi cDC2 gate.As described above, the expression of 2A1 is faithfully correlated with the expression of CD83 among the cDC2-gated CD11c + MHCII hi cell subsets in FL-BM culture.Besides, the detection of 2A1 requires the permeabilization of cells and intracellular staining which prevents and complicates the analysis of live cells.Henceforth, we utilized the cellsurface staining markers for the identification of each cell subsets in FL-BM culture.To assess the functional capacity of DCs to prime and expand T cells, naive T cells were isolated from OVA-specific TCR (i.e., CD8 + OT-1 and CD4 + OT-2) transgenic mice and used as responders to each MHCII hi cDC2-gated subpopulation isolated from the OVA-pulsed FL-BM culture.In the co-cultures of graded doses of each MHCII hi subpopulation with naive OVA-specific T cells, MHCII hi CD83 + CD115 -CX 3 CR1 -cells were far superior APCs for CD8 + OT-1 T cells to MHCII hi CD83 -CD115 + CX 3 CR1 + and MHCII hi CD83 -CD115 -CX 3 CR1 -cells (Figure 4A), while both MHCII hi CD83 + CD115 -CX 3 CR1 -and MHCII hi CD83 -CD115 - CX 3 CR1 -cells were much better APCs for CD4 + OT-2 T cells than MHCII hi CD83 -CD115 + CX 3 CR1 + cells (Figure 4B).Therefore, only the MHCII hi CD83 -CD115 + CX 3 CR1 + cDC2-gated cells exhibited the poor antigen presentation to naive T cells on both MHCI and MHCII pathways.Also, MHCII hi CD83 -CD115 + CX 3 CR1 + cDC2-gated cells survived most poorly in the co-culture with OVA-specific T cells (Supplementary Figure 8).Meanwhile, unlike cDC1s isolated from the spleen (Supplementary Figure 9B), cDC1s in FL-BM culture were poor at cross-presenting both soluble OVA and cell-associated OVA antigens to CD8 + OT-1 T cells (Figure 4A; Supplementary Figure 9A) indicating that cDC1s in FL-BM culture require an additional maturation stimulus, such as treatment with GM-CSF (Supplementary Figure 9C), to acquire cross-presentation capacity (40,41).
Particular morphology is one of the key features that characterize distinct types of hematopoietic cells (24, 26, 29, 33).In the overnight culture of MHCII hi cDC2-gated cells isolated from FL-BM culture, the individual morphologies of each subpopulation were examined.As expected, CD115 -CX 3 CR1 -cDC2s illustrated the large clusters of non-adherent cells with noticeably long cellular processes or dendrites ( Figures 4E, F), a typical morphology of DCs (7,33,45).In contrast, CD115 + CX 3 CR1 + non-DCs were devoid of the morphology and clustering of DCs ( Figures 4E, F).Besides, CD115 -CX 3 CR1 -cDC2s exhibited the typical electron microscopic features of DCs (12,(46)(47)(48) that the irregularly shaped nuclei possessed a dense rim of heterochromatin and small nucleoli; the cytoplasm comprised many scattered mitochondria, short slips of rough endoplasmic reticulum, few electron-dense membranebound lysosomes but many electron-lucent vesicles including multivesicular bodies of varying size in the Golgi region (Supplementary Figure 11), while CD115 + CX 3 CR1 + non-DCs displayed more frequent electron-dense membrane-bound lysosomes (Supplementary Figure 12).

Gene expression profiles indicate that MHCII hi cDC2-gated non-DCs in FL-BM culture are closely related to MHCII int GM-Macs
Bulk RNA sequencing (RNA-Seq) was performed to check the distinctive gene expression profile of each cell subset in GM-BM and FL-BM cultures, which are GM-DCs and MHCII int GM-Macs from GM-BM culture; MHCII hi FL-cDC2-gated non-DCs, mature FL-cDC2s, immature FL-cDC2s, mature FL-cDC1s, and immature FL-cDC1s from FL-BM culture.Comparative analysis of the transcriptomes revealed the transcriptional differences and similarities among those cell subsets in GM-BM and FL-BM cultures.The principal components analysis (PCA) of the transcriptomes showed that the clusters of CD83 + cell subsets (i.e., GM-DCs, mature FL-cDC2s, and mature FL-cDC1s) were located close to each other; so were those of CD83 -cDC subsets (i.e., immature FL-cDC2s and immature FL-cDC1s) and those of CD83 - non-DC subsets (i.e., MHCII int GM-Macs and MHCII hi FL-cDC2gated non-DCs), respectively (Figure 5A).Also, the hierarchical cluster analysis (HCA) of the transcriptomes revealed comparable outcomes (Figure 5B).In FL-BM culture, CD83 + mature cDC2s elevated the expression of marker genes for mature DCs such as CCR7 and CCL22 (49-51); CD83 -immature cDC2s augmented the expression of DCs-related genes including Kmo, Flt3, and CD209a (DC-SIGN) (29, 52, 53); CD83 -cDC2-gated non-DCs exhibited higher levels of gene expression for macrophage signatures such as Lyz2, Apoe, Csf1r, Clec4e, Mmp12 (54) (Supplementary Figure 13).Select gene expression profiles among cell subsets in FL-BM and GM-BM cultures were compared to those among MHCII + mononuclear phagocytes in mice such as splenic cDC subsets, peritoneal DC subsets, and small peritoneal macrophages (SPMs) (Supplementary Figure 14).While each cell subset in FL-BM and GM-BM cultures was clustered close to the related cell subset in the spleen and peritoneal cavity, it was notable that FL-cDC2-gated non-DCs and MHCII int GM-Macs showed close clustering with splenic cDC2s and peritoneal DC2s in the expression profiles of DC signature genes (Supplementary Figure 14D).

Differentiation of MHCII hi cDC2-gated non-DCs in FL-BM culture
CD11c + cells in the gate of MHCII int GM-Macs, exhibiting a similar gene expression profile as that of CD11c + MHCII hi cDC2gated non-DCs in FL-BM culture (Figures 5A, B), have been suggested to contain mixed populations of precursors or immature DCs in addition to the Macs or Mac-like cells (55).MHCII hi cDC2gated non-DCs were found to express higher levels of various toll-like receptors (TLRs) than other cells within the cDC2 gate in FL-BM culture (Supplementary Figure 15).Because the treatment with TLR agonists is commonly employed to induce the maturation or differentiation of DCs ( 7), we investigated the phenotype change of MHCII hi cells in FL-BM culture following overnight treatment with graded doses of LPS.Treatment of FL-BM culture with more than 10 ng/ml of LPS sharply reduced the number of MHCII hi cDC2-gated non-DCs and largely augmented the population of MHCII hi mature Transcriptome analysis reveals the close relationship between MHCII int GM-Macs and MHCII hi CD115 + CX3CR1 + cDC2-gated non-DCs in FL-BM culture.(A) Principal component analysis (PCA) of transcriptomes of cell subsets in GM-BM and FL-BM cultures.(B) Heatmap displays the one-way hierarchical clustering using z-score for normalized value of randomly selected genes that are differentially expressed among cell subsets in GM-BM and FL-BM cultures.
In et al. 10.3389/fimmu.2023.1179981 Frontiers in Immunology frontiersin.orgcDC2s (Figure 6A; Supplementary Figure 16).Similarly, upon stimulation of FL-BM culture with TLR3 or TLR9 ligands, MHCII hi cDC2-gated non-DCs disappeared and MHCII hi mature cDC2s increased (Supplementary Figure 17).Meanwhile, we found that the MHCII hi cDC2-gated cell subsets in FL-BM culture were not MHCII + RORgt + APCs such as ILC3s, Janus cells, or Thetis cells (56-58) (Supplementary Figure 18).Then, we tested whether the MHCII hi cDC2-gated non-DCs, vanishing by stimulation of FL-BM culture with TLR ligands, could differentiate into MHCII hi mature cDC2s.After isolation from the FL-BM culture at day 8, MHCII hi cDC2-gated non-DCs were further co-cultured with BM filler cells from congenic mice (Figures 6B, C).After more than 4 days of additional co-culture with FLT3L, a significant fraction of the MHCII hi cDC2-gated non-DCs became MHCII hi cDC2s (Figure 6C).As expected, in the FL-BM co-culture, the MHCII hi cDC2s which had differentiated from the MHCII hi cDC2-gated non-DCs also expressed the higher levels of 2A1 and CD83 than the undifferentiated MHCII hi cDC2-gated non-DCs (Figure 6D).Therefore, the MHCII hi cDC2-gated non-DCs in FL-BM culture are precursors to the MHCII hi cDC2s, i.e., MHCII hi pre-cDC2s.
MHCII hi Pre-cDC2s in FL-BM culture are derived from pre-cDCs and CDPs but not from monocytes and cMoPs In the development of DC lineage cells, MDPs can differentiate either into cDCs via CDPs and pre-cDCs (35, 37) or into Mo-DCs via cMoPs and monocytes (11,38).Hence, we examined the differentiation potential of various DC progenitors and precursors into the MHCII hi pre-cDC2s in FL-BM culture.Lineage - CD11c + CD135 + MHCII -pre-cDCs, immediate precursors to cDCs (36), were isolated from the BM, and further co-cultured in FL-BM culture of congenic mice (Supplementary Figure 19).After 6 days of the co-culture with BM fillers and FLT3L, pre-cDCs differentiated mostly into the cells within the MHCII hi cDC2 gate, which consisted of MHCII hi pre-cDC2s and MHCII hi cDC2s in a similar ratio as Frontiers in Immunology frontiersin.orgfound in the MHCII hi cDC2-gated cells derived from the BM fillers (Figure 7A).Next, lineage -CD115 + CD117 + CD135 + MHCII -CDPs, progenitors to cDCs and some pDCs (37), were isolated from the BM, and co-cultured with the BM fillers of congenic mice and FLT3L (Supplementary Figure 20).After 8 days of the co-culture, CDPs were able to differentiate into a variety of CD11c + cells in FL-BM culture where the MHCII hi cDC2-gated cells comprised MHCII hi pre-cDC2s and MHCII hi cDC2s in a similar ratio as found in the MHCII hi cDC2-gated cells derived from the BM fillers (Figure 7B).Monocytes, more abundant than DCs in the blood and BM, are precursors to MHCII hi GM-DCs and MHCII int/lo GM-Macs in GM-BM culture and can rapidly differentiate into MHCII hi Mo-DCs in vivo during infection and inflammation (7,11,26).Next, we have tested whether monocytes could differentiate into the MHCII hi pre-cDC2s in FL-BM culture.Lineage -Ly6C hi CD11b + CD115 + MHCII -monocytes were isolated from the BM, and then co-cultured with the BM fillers of congenic mice and FLT3L (Supplementary Figure 21).After 6 days of the co-culture, monocytes have differentiated mainly into CD11c + MHCII int and CD11c + MHCII lo cells but hardly into the CD11c + MHCII hi cells in FL-BM culture (Figure 7C).Then, lineage - Ly6C + CD16/32 + CD34 + CD115 + CD117 + MHCII -cMoPs, progenitors to monocytes (5), were also isolated from the BM, and co-cultured with the BM fillers of congenic mice and FLT3L (Supplementary Figure 22).After 8 days of the co-culture, cMoPs have rarely survived and mostly differentiated into CD11c + MHCII int and CD11c + MHCII lo cells but not into the CD11c + MHCII hi cells in FL-BM culture (Figure 7D).Therefore, in FL-BM culture, MHCII hi pre-cDC2s are derived from pre-cDCs and CDPs but not from monocytes and cMoPs (Figures 7E, F).
When we tested the co-culture of pre-cDCs and BM fillers with FLT3L to be further treated with LPS, most of the MHCII hi pre-cDC2s derived from both pre-cDCs and BM fillers differentiated similarly into MHCII hi cDC2s (Figure 8A).We also examined the expression of DC-specific transcription factor Zbtb46 and found that MHCII hi pre-cDC2s expressed Zbtb46 as highly as MHCII hi immature cDC2s in FL-BM culture (Figure 8B), indicating that MHCII hi pre-cDC2s are closely related to the DC lineage.

A B D E F C
MHCII hi CD115 + CX3CR1 + cDC2-gated pre-cDC2s in FL-BM culture are derived from pre-cDCs and CDPs but not from monocytes and cMoPs.

Discussion
DCs have been typically identified as CD11c + MHCII hi cells in various mouse tissues and organs (59).Likewise, from GM-BM culture as well as FL-BM culture, CD11c + MHCII hi cells (11), if not CD11c + cells (60), have been routinely isolated and used as in vitro-generated DCs for the numerous experiments.Our present study, however, demonstrated that only a half of the CD11c + MHCII hi cells in FL-BM culture were DCs whereas the remaining half were precursor cells with poor antigenpresenting capacity.
Considering that Helft and others (11) had shown that GM-BM culture comprised heterogeneous populations of CD11c + CD11b hi MHCII int GM-Macs and CD11c + CD11b int MHCII hi GM-DCs, we began to investigate the heterogeneity of CD11c + cells in GM-BM culture in terms of expression of 2A1, an intracellular antigen of mature or activated DCs (12,14).Subsequently, in GM-BM culture, we found that CD11c + CD11b hi MHCII int GM-Macs are 2A1 -and CD11c + CD11b int MHCII hi GM-DCs are 2A1 + , indicating that 2A1 could be a valuable marker to distinguish DCs from non-DCs among the CD11c + MHCII + cell population.
Then, we examined the heterogeneity of CD11c + cells in FL-BM culture in terms of expression of 2A1.Surprisingly, in FL-BM culture, CD11c + MHCII hi cDC2-gated CD115 + CX 3 CR1 + pre-cDC2s are a close precursor to cDC2s.we discovered that less than a half of the CD11c + MHCII hi cells were 2A1 + , but the remaining cells were 2A1 -.Flow cytometric analysis revealed that almost of all the CD11c + MHCII hi cells in FL-BM culture were within the cDC2 gate, and categorized into three groups: two large majorities of 2A1 + CD83 + CD115 -CX 3 CR1 -and 2A1 -CD83 - CD115 + CX 3 CR1 + cells and a smaller number of 2A1 -CD83 -CD115 - CX 3 CR1 -cells.According to the assays on antigen uptake and presentation capacities, morphologies, and gene expression profiles of the CD11c + MHCII hi cDC2-gated cell subsets, 2A1 -CD83 -CD115 - CX 3 CR1 -cells were immature cDC2s; 2A1 + CD83 + CD115 -CX 3 CR1 - cells were mature cDC2s; and 2A1 -CD83 -CD115 + CX 3 CR1 + cells were non-DCs.The treatment with LPS and other TLR ligands greatly decreased both non-DCs and immature cDC2s but highly increased mature FL-cDC2s among the CD11c + MHCII hi cDC2-gated cells in FL-BM culture.Then, the FL-BM culture with adoptively transferred cells confirmed that the MHCII hi cDC2-gated non-DCs were precursors to cDC2s, i.e., MHCII hi pre-cDC2s.Besides, MHCII hi pre-cDC2s also showed the higher expression of DC-specific transcription factor Zbtb46 as similarly as immature cDC2s in FL-BM culture.In the meantime, the MHCII hi pre-cDC2s were found generated from pre-cDCs and CDPs but not from monocytes and cMoPs, corroborating that the MHCII hi pre-cDC2s have close lineage to cDCs but not Mo-DCs.All in all, our present study identified and characterized a novel cDC precursor population, exhibiting a CD11c + MHCII hi CD115 + CX 3 CR1 + phenotype, in FL-BM culture.
DCs are the most efficient professional PCs and typically form clusters under the in vitro culture condition.Meanwhile, DCs are much poorer at taking up antigens than other APCs such as monocytes and macrophages (7,26,43,44).For example, Mo-DCs are much poorer at taking up antigens than monocytes, i.e., their precursors.Therefore, the loss of the strong ability to take up antigens and other characteristics of monocytes and macrophages is considered as key features of fully differentiated Mo-DCs (7).In this study, we discovered that MHCII hi cDC2-gated CD115 + CX3CR1 + cells are precursors to cDC2s in FL-BM culture.The newly identified MHCII hi cDC2 precursor cells might possess certain characteristics similar to those of monocytes and macrophages prior to differentiating into cDC2s in FL-BM culture.Besides, there are still unresolved differences in view on the nature of MHCII int GM-Macs (11,55,61,62).Helft and others classified CD11c + CD11b hi MHCII int cells in GM-BM culture as Macs or GM-Macs on the basis of their poor antigen presentation capacity, morphology, progenitor/precursor, and gene expression profile (11), whereas Lutz and others argued that GM-Macs might be a mixture of macrophages and immature DCs (55).The features of MHCII hi pre-cDC2s that we have identified and characterized in the present study are quite similar to those of GM-Macs in many aspects: lack of dendritic morphology and cell clustering, poor antigen presentation, CD11c + 2A1 -CD83 -CD115 + CX3CR1 + phenotype, and similar gene expression profiles.Those two cell populations also have some notable differences: MHCII hi pre-cDC2s in FL-BM culture have a MHCII hi phenotype and are derived efficiently from pre-cDCs and CDPs but not from monocytes and cMoPs; GM-Macs have a MHCII int phenotype and are derived efficiently from monocytes and cMoPs.Therefore, it will be intriguing to investigate and compare the differentiation potentials of those two cell populations under a variety of conditions.
It is well known that FLT3L-induced BM-DCs do not fully recapitulate the phenotype of cDCs in vivo.It is also recognized that developing better phenotypes of FLT3L-induced BM-DCs in vitro requires the addition of still many unidentified factors into FL-BM culture (63).Our discovery of a large number of MHCII hi pre-cDC2s in FL-BM culture may reflect less than perfect in-vitro culture conditions for FLT3L-induced BM-DCs.If so, it is not likely that the in-vivo counterpart cells of the MHCII hi pre-cDC2s exist in large numbers.Nonetheless, in the future studies, it will be important to determine and characterize an in-vivo counterpart of the MHCII hi pre-cDC2s identified in FL-BM culture.

1
FIGURE 1 Expression of 2A1 antigen in GM-BM and FL-BM cultures.(A) Representative flow cytograms show kinetic change in the expression of 2A1 from CD11c + cells in GM-BM culture for 8 days (left panels).Distribution of 2A1 -and 2A1 + cells within CD11c + cells at day 8 of GM-BM culture (right graph).Flow cytograms are representative from 3 independent experiments.Graph is shown in mean ± SEM (n = 10).(B) Representative flow cytograms show kinetic change in the expression of 2A1 from CD11c + cells in FL-BM culture for 9 days (left panels).Distribution of 2A1 -and 2A1 + cells within CD11c + cells at day 9 of FL-BM culture (right graph).Flow cytograms are representative from 3 independent experiments.Graph is shown in mean ± SEM (n = 10).

2
FIGURE 2 Expression of 2A1 and surface molecules in GM-DCs and GM-Macs.(A) Gating strategy (left panel) identifies 2A1 + cells (red contours) in CD11c + cells at day 8 of GM-BM culture.2A1 + cells (red contours in left panel) are backgated in an MHCII versus CD11b contour plot of live-gated CD11c + cells (black contours in middle panel), indicating 2A1 + cells within the GM-DC gate.Graph (right panel) shows % distribution of 2A1 + cells between GM-Macs and GM-DCs.Representative flow cytograms are from 10 independent experiments.Graph is shown in mean ± SEM (n = 10).(B) Flow cytometric analysis of 2A1 and surface molecules in MHCII int GM-Macs (green) and GM-DCs (purple) at day 8 of GM-BM culture.Filled histograms for isotype controls of the mAbs against the respective molecules.Representative flow cytograms are from at least 4 independent experiments.

3
FIGURE 3 Expression of 2A1 reveals heterogeneous populations of the CD11c + MHCII + cDC-gated cells in FL-BM culture.(A) Gating strategy identifying 2A1 + cells (red contours) in CD11c + cells at day 9 of FL-BM culture (upper left panel).2A1 + cell population (red contours) is backgated in an MHCII versus B220 contour plot of live-gated CD11c + cells (black contours), indicating 2A1 + cells within the cDC1 and cDC2 gates (upper middle and right panels).Distribution of 2A1 + cells among DC subsets (lower left graph).Distribution of 2A1 -and 2A1 + cells within cDC1s or cDC2s (lower right graphs).Graphs are shown in mean ± SEM (n = 11) from 5 independent experiments.(B) Gating strategies and distributions indicating cDC1 and cDC2 populations within MHCII lo , MHCII int , or MHCII hi cDC subsets at day 9 of FL-BM culture (left panels and graphs).Distributions of 2A1 -and 2A1 + cells within cDC1s or cDC2s (right panels and graphs).Graphs are shown in mean ± SEM (n = 13) from 6 independent experiments.Flow cytometric analysis of 2A1 according to the expression levels of positively (C) or negatively (D) correlating surface molecules in the MHCII hi cDC gate.Flow cytometric analysis of MHCII correlating to the expression levels of 2A1, CD83, CD115, and CX 3 CR1 in the cDC1-gated (E) and cDC2-gated (F) cells from FL-BM culture.Distribution of cells among three subsets in the MHCII hi cDC2 gate [purple box in right histogram in (F)] are shown [lower graph in purple box in (F)].Representative flow cytograms are based on 3 independent experiments.Graph is shown in mean ± SEM (n = 8).***p ≤ 0.001; one-way ANOVA test.
(A) Gating strategies to analyze the co-culture of isolated pre-cDCs and BM fillers at day 6 (upper panels).Graph (lower) is shown in mean ± SEM (n = 4).NS, not significant; one-way ANOVA test.(B) Gating strategies to analyze the co-culture of isolated CDPs and BM fillers at day 8 (upper panels).Graph (lower) is shown in mean ± SEM (n = 3).NS, not significant; one-way ANOVA test.(C) Gating strategies to analyze the co-culture of isolated monocytes and BM fillers at day 6 (upper panels).Graph (lower) is shown in mean ± SEM (n = 4).**p ≤ 0.01; NS, not significant; one-way ANOVA test.(D) Gating strategies to analyze the co-culture of isolated cMoPs and BM fillers at day 8 (upper panels).Graph (lower) is shown in mean ± SEM (n = 5).***p ≤ 0.001; NS, not significant; one-way ANOVA test.(E, F) Resulting output numbers and fold changes of the CD45.2 + CD11c + MHCII hi cells generated from the respective CD45.2 + BM precursor/progenitor inputs are shown.The input numbers of CD45.2 + precursor/progenitor are 5×10 3 for CDPs, 2×10 4 for pre-cDCs, 1×10 4 for cMoPs, and 1×10 5 for monocytes, respectively, per well.Bars indicate the mean.Data are collected from more than 3 independent experiments (n = 3 for CDPs; n = 4 for pre-cDCs; n = 5 for cMoPs; n = 4 for monocytes).