ORIGINAL RESEARCH published: 08 October 2020 doi: 10.3389/fimmu.2020.553859 Development of Pig Conventional Dendritic Cells From Bone Marrow Hematopoietic Cells in vitro Yanli Li 1* , Lucinda Puebla-Clark 2 , Jesús Hernández 2 , Ivan Díaz 3 and Enric Mateu 1,3 1 Departament de Sanitat i Anatomia Animals, Facultat de Veterinària, Universitat Autònoma de Barcelona, Barcelona, Spain, 2 Laboratorio de Inmunología, Centro de Investigación en Alimentación y Desarrollo, Hermosillo, Mexico, 3 Centre de Recerca en Sanitat Animal, IRTA-UAB, Bellaterra, Spain In recent years, porcine dendritic cells (DCs) have been identified from pig tissues. However, studying the interaction of porcine DCs with pathogens is still difficult due to the scarcity of DCs in tissues. In the present work, the Flt3-ligand (Flt3L)- based in vitro derivation system was further characterized and compared with Edited by: Elodie Segura, other cytokine derivation models using a combination of factors: stem cell factor Institut Curie, France (SCF), GM-CSF, and IL-4. The method using Flt3L alone or combined with SCF Reviewed by: supported the development of pig bone marrow hematopoietic cells into in vivo Nicolas Bertho, INRA Biologie, Épidémiologie et equivalent conventional DCs (cDCs). The equivalent cDC1 (the minor population in Analyse de Risque en Santé Animale the cultures) were characterized as CADM1+ CD14− MHC-II+ CD172a−/lo CD1− CD163− (BIOEPAR), France DEC205+ CD11R3lo CD11R1+ CD33+ CD80/86+ . They expressed high levels of FLT3, Sreekumar Balan, Icahn School of Medicine at Mount ZBTB46, XCR1, and IRF8 mRNA, were efficient in endocytosing dextran and in Sinai, United States proliferating allogenic CD4+ CD8+ T cells, but were deficient in phagocyting inactivated *Correspondence: Staphylococcus aureus (S. aureus). Also, after poly I:C stimulation, they predominantly Yanli Li yanli.li@uab.cat produced IL-12p40a and matured as indicated by the increase of MHC-I, MHC-II, and CD80/86. The equivalent cDC2 (the main population) were CADM1+ CD14− MHC-II+ C Specialty section: D172a+ CD1+ CD163−/lo DEC205lo CD11R3+ CD11R1+ CD33+ CD80/86+ ; meanwhile, This article was submitted to Antigen Presenting Cell Biology, they overexpressed FcεR1α and IRF4 mRNA. They showed high efficiency a section of the journal in the endocytosis of dextran, but weak in phagocytosing bacteria. They Frontiers in Immunology supported allogenic CD4+ CD8− /CD4+ CD8+ T cell proliferation and were high Received: 20 April 2020 producers of IL-12p40 (upon TLR7 stimulation) and IL-10 (upon TLR7 stimulation). Accepted: 25 August 2020 Published: 08 October 2020 TLR ligand stimulation also induced their maturation. In addition, a CD14+ Citation: population was identified with the phenotype CADM1+ CD14+ MHC-II+ CD172a+ Li Y, Puebla-Clark L, Hernández J, CD1+ CD163+ DEC205− CD11R3+ CD11R1+ CD33−/lo CD80/86+ . They shared some Díaz I and Mateu E (2020) Development of Pig Conventional functional similarities with cDC2 and were distinguishable from macrophages. This Dendritic Cells From Bone Marrow CD14+ population was efficient in phagocyting S. aureus but showed less maturation Hematopoietic Cells in vitro. upon TLR ligand stimulation than cDC1 or cDC2. The alternative methods of DC Front. Immunol. 11:553859. doi: 10.3389/fimmu.2020.553859 derivation including GM-CSF and/or IL-4 produced mostly CADM1− cells that did Frontiers in Immunology | www.frontiersin.org 1 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC not fulfill the canonical phenotype of bona fide porcine DCs. Our study provides an exhaustive characterization of Flt3L-derived DCs with different methods that can help the in vitro study of the interaction of DCs with porcine-relevant pathogens. Keywords: Flt3L, cDC1, cDC2, CD14+ DC, pig INTRODUCTION CD163+ moDCs, the second generated a more heterogeneous population (16). Pig farming is an important source of animal protein. As of In human and mice, bona fide DCs are identified as cells January 2020, it was estimated that about 677 million pigs originating from a common macrophage and DC precursor were raised worldwide and are therefore an important source (MDP), which gives rise to a common DC precursor (CDP), of economic income in many countries (1). Additionally, pigs restricted to develop into cDCs (cDC1 and cDC2), and pDC are epidemiologically important hosts of pathogens that can (6, 17). Flt3 seems to be a key molecule in DC development. In affect humans such as the H1N1 “swine flu” pandemic in mice, Flt3 is broadly expressed on hematopoietic precursors and 2009 (2). Pigs are also models for studying several human exists throughout the whole process of DC development (18) diseases (3) and potential donors for xenotransplantation and during commitment to DCs; the expansion and migration (4). In all those aspects, a better understanding of the pig of precursor cells are essentially Flt3L-dependent processes (8). immune system and of the interaction of pathogens with it is In pigs, Guzylack-Piriou et al. (19) showed that addition of scientifically relevant. Flt3 ligand (Flt3L) to bone marrow cell cultures produced cells Dendritic cells (DCs) are sentinels of the immune system, with features of bona fide DCs. In the past years, DCs from essential for the activation of naïve T cells to promote the pig lymphoid (spleen, tonsil, and lymph nodes) (15, 20) and initiation of the adaptive immune response. Currently, there non-lymphoid organs (skin and lung) (14, 21) as well as blood are different types of DCs that are recognized based on their (22) have been successfully identified and characterized. From ontogeny (5, 6): the monocyte-derived DCs (moDCs), which the abovementioned studies, consensus has been reached on the are generated from activated monocytes in the tissues (7), and phenotype of cDC1 (CADM1+ CD14− MHC-II+ CD172a−/lo ), the conventional or bona fide DCs that are generated from cDC2 (CADM1+ CD14− MHC-II+ CD172a+ ), and pDC Flt3 (CD135)-expressing precursors (8) and can be subdivided (CADM1− CD14− CD172a+ CD4+ ). In the present study, into classical/conventional (c) cDC1 and cDC2, which can be we re-examined the features of pig DCs produced in an differentiated by gene expression and cytokine profiles (5, 6) Flt3L-dependent culture model regarding the phenotype, gene and their ability to cross-present antigens (9, 10). Additionally, expression, and functionality. Moreover, we compared the use of there are plasmacytoid DCs (pDCs), specialized in producing FLt3L alone with other cytokine combination [addition of stem high amounts of type I interferons (especially IFN-α) during cell factor (SCF), GM-CSF, and IL-4] in order to determine the antiviral responses (11, 12), and finally, Langerhans cells effects on the yield and type of DCs produced. are a special type of DCs generated from erythro-myeloid progenitors (6). One limitation when studying bona fide DCs in pigs MATERIALS AND METHODS or other species is the low frequency of these cells. For example, in pig blood, cDC1, cDC2, and pDC accounted for only 0.1, 0.1, and 0.5-1%, respectively, of the intact Obtention of Bone Marrow Cells Porcine bone marrow hematopoietic cells (BMHCs) were Peripheral Blood Mononuclear Cells (PBMCs) (13); in lung, aseptically obtained from femora and humeri from 4-week- 5.6 ± 5.1% and 11.8 ± 12.3% of the MHC-IIhi cells were old pigs as previously described (16). Animals were all identified as potential cDC1 and cDC2, respectively (14); in Landrace × Large White and came from the same source lymphoid tissues, 56.9 ± 3.1% and 30.1 ± 8.5% of cells were farm. BMHCs were tested by PCR to confirm that they were identified as cDC1 and cDC2, respectively, among lymphocyte- free of porcine reproductive and respiratory syndrome virus depleted MHC-IIhi CADM1hi cells (15). This scarcity makes (PRRSV; VetMAX PRRSV NA & EU Reagents; Thermo Fisher it extremely difficult to perform functional analysis or to Scientific, Spain), porcine circovirus type 2 (PCV2), Mycoplasma investigate DC functionality upon sensing pathogens. Apart hyopneumoniae, and torque teno sus virus (TTSuV) 1 and 2 from that, when obtaining DCs from tissues, the extraction (23–25). Cells were frozen in liquid nitrogen until used. process imposes a series of manipulation that may damage cell surface receptors or modify cell maturation status, which subsequently will deviate the DC functionality from in vivo Generation of DCs From conditions. Thus, in vitro production of DCs can be a convenient Flt3L-Dependent Bone Marrow Cultures alternative. A review of the literature shows that, for many For DC generation, BMHCs were seeded in 24-well plates at years, production of DCs in pigs was done by stimulating a density of 1 × 106 cells/600 µl in RPMI 1640 medium monocytes with GM-CSF and IL-4 or bone marrow precursors supplemented with 10% fetal calf serum (FCS), 2 mM glutamine, with GM-CSF (16). While the first method consistently produced 20 mM HEPES, 100 units/ml penicillin, and 100 µg/ml Frontiers in Immunology | www.frontiersin.org 2 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC streptomycin. 20 nanograms per milliliter of recombinant human analyzed using FCS Express 7 (de novo Software, Glendale, Flt3L (rhuFlt3L, Fisher Scientific, Spain) was added to promote CA, United States). expansion and differentiation of Flt3+ or relevant cells. Cells For phenotypic characterization of the generated DCs, a five- were incubated for 14 days at 37◦ C in a 5% CO2 and 90% color staining was performed for cell surface molecules CADM1, humidity environment. Half of the medium was replaced every CD14, MHC-II, and CD172a, plus a fifth marker targeting 3 days with fresh supplemented medium containing the same CD1/CD163/DEC205/CD11R3/CD11R1/CD33/CD80/86. Anti- concentration of Flt3L. CD11R3 and anti-CD11R1 were used to stain CD11b. These To assess the development of DCs, cultures were two antibodies were reported to be specific for CD11b but examined at regular intervals during the 2-week period of with different staining patterns in pigs (26). Cells were initially incubation. For this purpose, BMHCs were labeled with incubated with antibody anti-CADM1 followed by anti-chicken 5 µM CellTrace Violet (Fisher Scientific, Spain), washed, IgY biotin and the primary antibody of the fifth marker. Then, and seeded at 0.8 × 106 cells/600 µl into the 24-well plates. streptavidin PerCP-Cy5.5 and a secondary antibody coupled to Proliferation was assessed at 0, 1.5, 3, 6, 9, 12, and 14 days of BV421 were added. Finally, cells were stained with anti-CD14 incubation by flow cytometry. Proliferated CADM1+ MHC- FITC, anti-MHC-II conjugated to anti-mouse IgG2b Zenon II+ cells were considered as potential Flt3L-derived DCs. All Alexa Fluor 647 (Thermo Fisher Scientific, Spain), and anti- experiments in the present work were done at least in triplicate CD172a conjugated to LYNX Rapid RPE (Bio-Rad, Spain). except when indicated. For the combination with CD1, CD172a and CD1 were both indirectly labeled with secondary antibodies coupled to BV421 Obtention of Alveolar Macrophages, and RPE, respectively. CD80/86 was labeled by an antibody Isolation of Peripheral Blood anti-mouse CD80 (B7-1) Super Bright 436. Mononuclear Cells and Monocytes, and RNA Extraction and Quantitative Generation of Monocyte-Derived DCs Alveolar Macrophages (AMs) and PBMCs were obtained from Real-Time PCR the same pigs as described in “Obtention of Bone Marrow Total RNA from FACS-sorted DC subpopulations was extracted Cells” through, respectively, broncho-alveolar lavage and density with TRIzol (Thermo Fisher Scientific, Spain) according to gradient centrifugation with Histopaque 1.077 (Sigma-Aldrich, the manufacturer’s protocol and quantified with a Biodrop Spain). AMs and PBMCs were frozen in liquid nitrogen until spectrophotometer (BioDrop, United Kingdom). Genomic DNA used. The selected batches had viabilities >90% after thawing as was removed using a gDNA removal kit (ArcticZymes, Sweden). assessed by trypan blue staining. For PBMCs, the proportion of Then, 20 ng was used to amplify mRNA transcripts of FLT3, each T cell subset and the proliferation capability were also tested ZBTB46, XCR1, BATF3, IRF4, IRF8, FCεR1a, ITGAM (CD11b), in one sample per batch. Batches passing this quality control ITGAX (CD11c), CD64, CSF1R, MAFB, MERTK, KLF4, and (viability and proliferation capability) were used. Monocytes MGL2 genes using iTaq universal SYBR green one-step kit (Bio- were obtained through plastic adherence of PBMCs for 16 h Rad, Spain) and the adequate primers (Table 2) in a 7500 Real- at 37◦ C in 5% CO2 and used freshly. moDCs were generated Time PCR System (Applied Biosystems). The same pig origin of by differentiating blood monocytes with recombinant porcine AMs, monocytes, or GM-CSF/IL-4-differentiated moDCs were GM-CSF (rpGM-CSF, 20 ng/ml; R&D Systems, Spain) and rpIL- included for comparison as needed. Relative levels of mRNA were 4 (20 ng/ml; R&D Systems, Spain) according to the method assessed according to the 1Ct method and calculated with the described by Carrasco et al. (16). formula 2−1 Ct . Both Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and ribosomal protein 24 (RSP24) were used as reference genes as previously described (14). No significant Flow Cytometry Analysis difference was found when applying the normalization to the two Staining of cell surface molecules was performed in PBS with 5% reference genes; thus, the result to GAPDH was shown. FCS and 5% horse serum for 30 min on ice. Primary antibodies (conjugated or not) and their working dilutions are listed in Table 1. When necessary, secondary antibodies, anti-mouse Dextran-FITC Uptake IgG1, IgG2a, and IgG2b coupled to Brilliant Violet 421 (BV421), Sorted DC subpopulations (with 4000 to 6000 cells used in Alexa Fluor 488, R-phycoerythrin (R-PE), or Alexa Fluor 647, each test) were incubated with Dextran-FITC (molecular weight were used at a dilution of 1:800 to 1:1600. The antibodies included 40,000; Sigma-Aldrich, Spain) at 1 mg/ml in complete RPMI 1640 in the panel were previously titrated to determine the appropriate for 2.5 h either at 37◦ C or on ice (negative control). Free Dextran- dilution. Between each labeling step, cells were washed twice with FITC was washed away, and uptake activity at both temperatures PBS containing 2% FCS and centrifuged at 500 × g for 5 min. was determined by flow cytometry. The median fluorescence Samples were acquired on a MACSQuant Analyzer 10 (Miltenyi intensity (MFI) of cells incubated at 37◦ C was compared with the Biotec, Bergisch Gladbach, Germany) or sorted on BD FACSJazz corresponding cells incubated on ice. (BD Biosciences, Oxford, United Kingdom). Fluorescence minus one (FMO) controls, matched isotype controls for mouse IgG1, Phagocytosis of Staphylococcus aureus IgG2a, and IgG2b, and background from secondary antibodies Sorted DC subpopulations (10,000 cells in each test) were were used for analysis and gating. The acquired data were incubated with Alexa 488-labeled S. aureus (Thermo Fisher Frontiers in Immunology | www.frontiersin.org 3 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC TABLE 1 | Antibodies used for flow cytometry analysis. Antibody Clone Isotype Species produced Working dilution Supplier Anti-CD14 FITC conjugated MIL2 IgG2b Mouse 1/100 Bio-Rad Anti-CD172a BL1H7 IgG1 Mouse 1/500 Anti-SLA II DR 2E9/13 IgG2b Mouse 1/500 Anti-CD11R1 MIL4 IgG1 Mouse 1/500 Anti-CD11R3 2F4/11 IgG1 Mouse 1/500 Anti-CD1 76-7-4 IgG2a Mouse 1/50 Anti-CD163 2A10/11 IgG1 Mouse 1/250 Anti-CD3 PPT3 IgG1 Mouse 1/200 Anti-CD4a FITC conjugated 74-12-4 IgG2b Mouse 1/200 BD Pharmingen Anti-CD8a Alexa Fluor 647 conjugated 76-2-11 IgG2a Mouse 1/100 Anti-human CADM1 3E1 IgY Chicken 1/1000 MBL Anti-mouse CD80 (B7-1) Super Bright 436 conjugated 16-10A1 IgG2a Armenian hamster 1/150 eBioscience human CD152 (CTLA-4)-muIg fusion protein – IgG2a Mouse 1/50 Ancell Anti-DEC-205 (hybridoma supernatant) 9HZF7 IgG1 Mouse 1/1 Provided by Lab. de Inmunología, CIAD, A.C. Anti-CD33 (Siglec-3) (hybridoma supernatant) 5D5 IgG1 Mouse 1/1 Gift from Dr. J. Domiìnguez* Anti-MHC-I (hybridoma supernatant) 4B7/8 IgG2a Mouse 1/1 *Department of Biotechnology, INIA, Madrid, Spain. Antibodies without specific indications are anti-pig antibodies. Scientific, Spain) at 20 µg/ml (equivalent to a ratio of 60 particles plates (200 µl/well, equivalent as 50,000 cells/well) in the per cell) in complete RPMI 1640 for 1 h either at 37◦ C or presence of TLR agonists: 10 µg/ml poly I:C, 1 µg/ml LPS, on ice (negative control). Extracellular bound S. aureus were or 10 µg/ml gardiquimod (all from InVivogen, reconstituted quenched by addition of 1 mg/ml trypan blue before the analysis according to the manufacturer’s instructions). In parallel, control by flow cytometry. cells were cultured without the TLR agonists, and AMs were included as other cell-type reference. After 18 h of incubation, Mixed Lymphocyte Reaction Assay cells were stained for MHC-I, MHC-II, and CD80/86. For To test antigenic presentation capabilities of each DC CD80/86 staining, human CD152 (CTLA-4)-muIg (Ancell, subpopulation, allogeneic PBMCs were used as a source of Bayport, MN, United States), which binds to CD80/86, was responder cells. These PBMCs were obtained from 12-week-old used (see Table 1). Supernatants were harvested and frozen at pigs (free of PRRSV, PCV2, Mycoplasmas, TTSuV 1 and 2, and −80◦ C until used. influenza A) and then frozen and tested as described above. Presence of IFN-α, IFN-γ, IL-4, IL-6, IL-12p40, IL-10, IL- Then, they were labeled with CellTrace Violet (Fisher Scientific, 1β, TNF-α, and IL-8 (CXCL8) in the cell culture supernatants Spain) and re-suspended in medium containing 10% FCS at was measured by means of the Porcine ProcartaPlex Multiplex 1 × 106 cells/ml with 100 µl/well plated in the 96-well U-bottom Immunoassay (Labclinics, Spain) according to the manufacturer’s plate. The sorted DC subpopulations (cDC1, cDC2, and CD14+ instructions. Briefly, 50 µl of each sample was added to DCs) were added as stimulators at a ratio of 1:5 based on the a well of a 96-well plate containing fluorescent antibody- number of allogeneic PBMCs. After 5 days of culture, cells coated beads. After overnight incubation, 25 µl of biotinylated were harvested and stained with anti-CD3, anti-CD4, and detection antibodies followed by 50 µl of streptavidin-PE anti-CD8a antibodies (see Table 1 for details) before the flow antibodies were added. Finally, beads in each well were re- cytometry analysis. PBMCs without DCs added were used suspended with 120 µl of assay buffer and analyzed by a as the negative control; AMs and monocytes (the same pig MAGPIX system. Concentration of cytokines in each sample origin as the sorted DCs) were obtained as described above and was determined by using standard curves, which were fitted by used as other cell-type references. Results were reported as in a four-parameter logistic regression model. All samples were previous studies (14) using a relative proliferation index where run in duplicate. a value of 1 was assigned to the cell population producing the highest percentage of proliferating cells; all other subsets were Effect of SCF, GM-CSF, and IL-4 Addition normalized to it. to Flt3L-Based BMHC Cultures In subsequent experiments, additional cytokines that regulate Response to TLR Agonist Stimulation: DC lineage commitment and differentiation were added to Maturation Markers and Cytokine compare with the use of Flt3L alone. SCF (Kingfisher Biotech, Patterns Spain) was added to test its effect on the expansion and Sorted DC subpopulations (cDC1, cDC2, and CD14+ DCs) survival of progenitor cells, which afterward was combined with were re-suspended at a density of 250,000 cells/ml with medium the addition of GM-CSF (20 ng/ml; R&D Systems, Spain) or containing 10% FCS and were then seeded in 96-well U-bottom IL-4 (20 ng/ml; R&D Systems, Spain). The combination of Frontiers in Immunology | www.frontiersin.org 4 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC TABLE 2 | Primers used for RT-qPCR. Target mRNA Primer sequences (50 –30 ) Gene ID Product size (bp) References RPS24 F: CTTCAGCGGAAACAAATGGT 100155012 219 Designed R: CCTGTGTTTGGGCTCATTTT GAPDH F: CACCATCTTCCAGGAGCGAG 396823 154 (21) R: AGAAGGGGCAGAGATGATGA FLT3 F: CAACTGCACGGAAGAGATCA 100515445 132 Designed R: GCACAGCACTTCACCAGAAA XCR1 F: ACTGCTGGAAGAGGCTGTGT 414375 220 Designed R: ATAAATGCCGGCAGAAGATG FCeR1α F: ATTTACAGACCCACAGCCTAGCT 100152827 183 (21) R: TGCCAGTTACAGTTGGTTCAA ITGAM (CD11b) F: AGAAGGAGACACCCAGAGCA 397459 197 (56) R: GTAGGACAATGGGCGTCACT ITGAX (CD11c) F: GCTCCTTCGAGTTGGAGATG 100738907 229 Designed R: ACCAGGCTCTGTACCCCTTT IRF4 F: CCAAAAGCAAAGATGGCATT 100144625 181 Designed R: ACCTGGAGCTCTCATGCTGT IRF8 F: TCCTCCTTCTGAATCCGAACC 396645 202 Designed R: GAGCAGGACTTGAGCGGAAA BATF3 F: GAGGAGCTGAAGCACCTGAG 100621683 249 Designed R: CGAAGCTCTGAGGAGGTGAC ZBTB46 F: TGTCTCTCCTTTGGCTCGTT 110257361 182 Designed R: GACGGAATGTCAGTGTCACG CD64 (FCGR1A) F: CCACCAACTCCTGTCTGGTT 613130 155 Designed R: TTACCTTCTTCCCGTGACCA CSF1R F: CAGCAAGGACACGTACAGAT 100517086 138 Designed R: GGGTACAGAAGGGTGAGTTC MAFB F: GACTCCTGGCTTCCTGAACT 100518227 186 Designed R: TTCTCCGGACCCTAGACAGA MERTK F: GGACACAGGACCAAAGCATA 100519652 193 Designed R: ATCGCCGCTGTAAATCTTCT CX3CR1 F: GGGAGGGAAAGGGAAATACTC 100622126 157 Designed R: GAAGCCCATCGTAGTCAAAGT KLF4 F: AGTTCTCATCTCAAGGCACA 595111 63 Designed R: CCAGTCACAGTGGTAAGGTT MGL2 F: ACTTCTCTGGAGAGCACACT 216864 82 Designed R: GCTGTTGTACATGGTCGGTT cytokine cocktails included the following: (1) Flt3L (as standard Statistical tests applied to each data set are indicated in the method), (2) Flt3L + SCF, (3) Flt3L + SCF + GM-CSF (FSG), relevant figure legend. and (4) Flt3L + SCF + GM-CSF + IL-4 (FSG4). In cases (3) and (4), a two-step protocol was used. In the first step, BMHCs were cultured in the presence of Flt3L (either 20 RESULTS or 100 ng/ml) and SCF (either 20 or 100 ng/ml) to expand progenitors. After 6 days of culture with one medium changed Three Distinct Cell Populations Can Be at day 3, cells were re-plated in cultures containing cytokine Identified in Flt3L-Derived Bone Marrow cocktails Flt3L + SCF + GM-CSF (FSG) or Flt3L + SCF + GM- CSF + IL-4 (FSG4) to induce further cell differentiation. Cells Cell Cultures After 2 weeks of culture, stimulation of BMHCs with were kept for an additional 12-day period (18-day period Flt3L resulted in the development of a large subset of cells in total) with cocktail replenishment after 6 days. All the expressing CADM1 (59.5 ± 12.9%), among cells after gating cases were phenotypically analyzed by a four-color staining of singlets and excluding cell debris as shown in Figure 1A. CADM1, CD14, MHC-II, and CD172a as described under “flow Among those CADM1+ cells, three subpopulations could cytometry analysis.” be identified based on the cell surface expression of CD14, MHC-II, and CD172a. The major one (52.9 ± 4.0%, Statistical Analysis among cells of CADM1+ ) presented a phenotype of All data were analyzed using the GraphPad Prism 8.0 software CD14− MHC-IIhi CD172a+ that was compatible with cDC2. package (GraphPad Software, La Jolla, CA, United States). The second population was CD14− MHC-IIhi CD172a−/lo , Frontiers in Immunology | www.frontiersin.org 5 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC FIGURE 1 | Phenotype of Flt3 ligand-derived dendritic cells (Flt3L-DCs). Cells were stained for multi-color flow cytometry. (A) Illustrative density plots show the gating strategy: singlets → excluding cell debris → selecting CADM1+ population and divided by CD14 → identify cDC1-putative and cDC2-putative populations based on MHC-II+ and different expression of cD172a. Three putative populations were identified with cDC1-putative population gated as CADM1+ CD14- MHC-II+ CD172a- /lo , cDC2-putative population as CADM1+ CD14- MHC-II+ CD172a+ , and plus a CD14+ population (CADM1+ CD14+ MHC-II+ CD172a+ ) without classification. (B) Expression of phenotypic markers CD1, CD163, DEC205, CD11R1, CD11R3, CD33, and CD80/86 on the defined cDC1-putative, cDC1-putative, and CD14+ populations was assessed by flow cytometry. The filled histogram represents the FMO control. The density plots and histograms shown are illustrative for one but represent four pigs. which represented only 4.7 ± 1.6% of the CADM1+ CD11R1+ , CD33+ , and CD80/86+ , while putative cDC1 cells and was compatible with cDC1. Finally, a third were CD1− , CD163− , DEC205+ , CD11R3lo , CD11R1+ , population with phenotype CD14+ MHC-II+ CD172a+ CD33+ , and CD80/86+ . The CD14+ subset was predominantly accounted for 39.6 ± 3.8% of the CADM1+ cells. No CD1+ , CD163+ , DEC205− , CD11R3+ , CD11R1+ , CD33−/lo , CD4+ cells (not shown) were detected, indicating that and CD80/86+ . The latter subset (CD14+ ) phenotypically pDCs were probably not present in the Flt3L-derived resembled moDCs but could not be classified with absolute cultures or in a very low proportion. Further phenotypical certitude due to the high expression of CADM1 (16). characterization (Figure 1B) showed that putative cDC2 They will be designated as CD14+ DCs for the purpose of were predominantly CD1+ , CD163lo , DEC205lo , CD11R3+ , the present paper. Frontiers in Immunology | www.frontiersin.org 6 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC FIGURE 2 | Development of Flt3-ligand-derived dendritic cells (Flt3L-DCs) during a 14-day culture period. (A) Bone marrow hematopoietic cells (BMHCs) were labeled with CellTrace Violet before stimulation with Flt3L. Cells were collected on days 1.5, 3, 6, 9, 12, and 14 and stained with antibodies anti-MHC-II and anti-CADM1. Proliferating cells (Violetlo ) with a MHC-II+ CADM1+ phenotype were considered as potential DCs. The gating strategy was MHC-II+ CADM1+ Violetlo . (B) Total number of viable cells in the culture versus time of incubation (black line), and proportion of potential DCs among viable cells versus time (orange line). Results were obtained with cells from three pigs. Average and standard deviation of the proportion of potential DCs are shown. Development Process of Flt3L-Derived all viable cells that boosted at day 6 (19.0 ± 5.7% of viable cells) Bone Marrow Cell Cultures until day 12 (42.9 ± 4.3%) and finally declined slightly to 36.4% (Figure 2B, orange curve). Next, we examined the development of CADM1+ and MHC-II+ cells in Flt3L-stimulated bone marrow cultures during the 14-day period (Figure 2). At day 0, no CADM1+ cells were detected Expression of Conserved DC Markers in BMHC while MHC-II+ cells accounted for 16.0 ± 1.2% of Assessed by RT-qPCR the viable cells (data not shown). After 1.5 days of incubation, To assess the expression of relevant canonical genes for the the number of viable cells dropped sharply (62.5 ± 3.5% of identification of DCs, the putative cDC1, cDC2, and CD14+ reduction; Figure 2B, black curve), suggesting that most of non- DC populations were sorted as indicated in Figure 1A, and M8 Flt3L-primed progenitors and other non-precursor cells died were used for comparison. As shown in Figure 3, putative cDC1, during this time. From that day on, CADM1+ MHC-II+ cells cDC2, and CD14+ DCs expressed FLT3, with cDC1 showing started to develop. By day 6, an explosive increase of MHC- the highest expression followed by cDC2 and CD14+ DCs; the II+ cells was observed (35.5 ± 7.9% of total cells), among same expression pattern occurred on ZBTB46, a transcription which 53.0 ± 4.2% were CADM1+ Violetlo (Figure 2A). This factor used to distinguish cDCs from other immune lineages trend continued by day 12 with MHC-II+ reaching 56.3 ± 1.1% in mice (27). Beyond that, putative cDC1 exclusively expressed and CADM1+ Violetlo increasing to 76.3 ± 9.2% among MHC- XCR1 and IRF8; the assumed cDC2 overexpressed FcεR1a and II+ cells. A small decrease in the proportion of CADM1+ IRF4, while CD14+ DCs showed low or intermediate levels of was observed by day 14 (Figure 2A). The development of these genes. All three subsets expressed ITGAM. By contrast, CADM1+ MHC-II+ cells determined the potential DCs among AMs scarcely expressed these genes (FLT3, ZBTB46, XCR1, IRF8, Frontiers in Immunology | www.frontiersin.org 7 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC FIGURE 3 | mRNA expression levels of FLT3, ZBTB46, XCR1, IRF8, FcεR1a, IRF4, ITGAM, ITGAX, and CD64 in sorted putative cDC1, cDC2, and CD14+ DCs were assessed by quantitative Real-Time PCR (RT-qPCR). Alveolar macrophages (AMs) were used as a different cell-type reference. For each gene, data were normalized to the expression of the reference gene GAPDH (glyceraldehyde 3-phosphate dehydrogenase) and shown as relative value [arbitrary units (AUs)]. For each animal, the cell population with the highest expression for each gene was considered as 1.0 and the other populations were normalized to it. Each symbol represents one animal; median is shown for each data set. Statistical significance was calculated by the Kruskal–Wallis test with multiple comparisons performed by Dwass–Steel–Critchlow–Fligner. Significant difference was indicated with different letters a, b, c, . . . ; no letters were shown if no significance was found between groups. FcεR1a, IRF4, and ITGAM), but overexpressed CD64 as expected. To further characterize the CD14+ subset, we assessed the ITGAX (CD11c) was highly expressed on macrophages as well, expression of MAFB, CSF1R, MERTK, KLF4, and MLG2 on but lower on the putative cDC1, cDC2, and CD14+ DCs. This this cell population and compared them with cDC1, cDC2, is contrary to the result of pig blood cDCs (22) but somewhat monocytes, GM-CSF/IL-4 differentiated moDCs, and AMs. As consistent with the analysis of lung DCs (28). shown in Figure 4, CD14+ DCs highly expressed CSF1R and Frontiers in Immunology | www.frontiersin.org 8 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC FIGURE 4 | mRNA expression levels of CSF1R, MAFB, MERTK, KLF4, and MGL2 in sorted putative cDC1, cDC2, and CD14+ DCs were assessed by quantitative Real-Time PCR (RT-qPCR). Monocyte-derived DCs (moDCs), monocytes, and alveolar macrophages (AMs) were used as cell-type references. For each gene, data were normalized to the expression of the reference gene GAPDH (glyceraldehyde 3-phosphate dehydrogenase) and shown as relative value [arbitrary units (AUs)]. For each animal, the cell population with the highest expression for each gene was considered as 1.0, and the other populations were normalized to it. Each symbol represents one animal, median is shown for each data set. Statistical significance was calculated by the Kruskal–Wallis test with multiple comparisons performed by Dwass–Steel–Critchlow–Fligner. Significant difference was indicated with different letters a, b, c, . . . ; no letters were shown if no significance was found between groups. MAFB and expressed MERTK at an intermediate level. They also CD14+ DCs showed a clear phagocytic activity (15.6 ± 7.2% of expressed KLF4 but not MGL2. positive cells). Next, we examined whether putative cDC1, cDC2, and Functional Analysis of Putative cDC1, CD14+ DCs possessed the capacity to activate allogeneic T cells in a mixed lymphocyte reaction (Figure 5C). The three DC cDC2, and CD14+ DCs populations did not show any significant difference in inducing Two of the main properties of DCs are their capacities to CD4 or CD8 T cell proliferation. Of note, both the putative cDC1 capture antigens and then present them to activate naïve T cells. and cDC2 were more efficient than CD14+ DCs or macrophages Thus, we tested the endocytic and phagocytic capabilities of the in inducing the proliferation of CD4+ CD8+ T cells, with cDC1 putative cDC1, cDC2, and CD14+ DCs. Dextran-FITC uptake being the most efficient type, although no significant difference and S. aureus phagocytosis experiments were set up for that was shown. Interestingly, cDC1 and cDC2 were less efficient in purpose. All three populations presented an efficient uptake of inducing proliferation of CD4− CD8− T cells. dextran-FITC, with cDC1 and cDC2 showing higher capabilities than CD14+ DCs (Figure 5A). Of note, the histogram of cDC1 showed a bimodal distribution with two distinct subsets based Response to TLR Agonist Stimulation: on their uptake (high and moderate, Figure 5A). Although Maturation Markers and Cytokine cDC1 contained cells of CD172a− and CD172alo , the two Patterns subsets with distinct uptake capabilities were not correlated Response of each DC subpopulation to TLR3, TLR4, and TLR7 with this molecule (Supplementary Figure 1) or with MHC- ligands was assessed by determining the expression of MHC- II (Supplementary Figure 1). cDC2 and CD14+ DCs showed I, MHC-II, and CD80/86 (Figure 6), and the cytokine profiles a unimodal distribution. With regard to the phagocytosis of (Figure 7) after stimulation with poly I:C, LPS, and gardiquimod, S. aureus, only a small fraction of cDC1 (0.9 ± 0.2%) and respectively. The results showed that in cDC1, MHC-I, MHC- cDC2 (1.5 ± 1.8%) were phagocytic (Figure 5B); in contrast, II, and CD80/86 were highly increased when stimulated by poly Frontiers in Immunology | www.frontiersin.org 9 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC FIGURE 5 | Assessment of antigen uptake and antigen presentation capabilities of putative cDC1, cDC2, and CD14+ DCs. (A) Endocytic capability of different DC populations was evaluated by incubating cells with dextran-FITC for 2.5 h either at 37◦ C or on ice with FITC fluorescence assessed by flow cytometry. Results are depicted as Mmedian fluorescence intensity (1MFI) by subtracting MFI of cells incubated on ice from MFI of cells incubated at 37◦ C (data are the means of two pigs with each animal tested in triplicate). Overlapped histograms showed the fluorescence intensity of cells incubated at 37◦ C (open histograms) and cells incubated on ice (closed histograms). (B) Phagocytic activity of different DC populations was evaluated by incubating cells with Alexa 488-labeled inactivated S. aureus for 1 h either at 37◦ C or on ice and analyzed by flow cytometry. Results are depicted as cells positive for S. aureus (data are the means of four pigs with each animal tested in triplicate). Overlapped histograms showed the fluorescence intensity of cells incubated at 37◦ C (open histograms) and cells incubated on ice (closed histograms). (C) Sorted DCs (putative cDC1, cDC2, and CD14+ DC) and PBMCs (stained with CellTrace Violet) from allogeneic pigs were mixed at an APC:T cell ratio of 1:5 before being cultured for 5 days at 37◦ C. Alveolar macrophages (AMs) and monocytes were used as other cell-type references. Proliferation of CD4+ CD8α- , CD4- CD8α+ , CD4+ CD8α+ , and CD4- CD8α- was assessed by CellTrace Violet dilution (see Supplementary Figure 2 for gating strategy). Each symbol corresponds to one type of DC reacting with one batch of PBMCs. The median is shown for each data set (horizontal bar). For each animal, the highest proliferation value was considered as 1.0, and the other populations were normalized to it. Data are from three independent experiments and a minimum of two animals for each cell type. Statistical significance was calculated by the Kruskal-Wallis test with multiple comparisons performed by Dwass-Steel-Critchlow-Fligner. ***p < 0.001, **p < 0.01, and *p < 0.05. I:C or gardiquimod. Nevertheless, the cell viability was largely Regarding the cytokine patterns (Figure 7), no cytokines were damaged by using gardiquimod or LPS, but not by poly I:C. boosted by LPS or gardiquimod in cDC1. IFN-α (<40 pg/ml), In cDC2, MHC-I, MHC-II, and CD80/86 were increased by the IFN-γ (<5 pg/ml), or IL-4 (<5 pg/ml) were rarely detected in addition of any of the three TLR ligands, but LPS was less effective any DC subset supernatant regardless of the TLR agonists used in increasing MHC-I or CD80/86. On the contrary, CD14+ DCs (data not shown). had only a slight increase of MHC-I, but not MHC-II or CD80/86 Beyond that, the putative cDC1 produced substantial amounts when stimulated. of IL-12p40 (3284 ± 4429 pg/ml) when stimulated by poly Frontiers in Immunology | www.frontiersin.org 10 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC FIGURE 6 | Expression of MHC-I, MHC-II, and CD80/86 on putative cDC1, cDC2, and CD14+ DCs after stimulation by TLR3, TLR4, and TLR7 agonists (10 µg/ml poly I:C, 1 µg/ml LPS, and 10 µg/ml gardiquimod, respectively) for 18 h or unstimulated as controls. Unstimulated control, gray histogram; poly I:C, red histogram; LPS, blue histogram; gardiquimod, green histogram. Data are representative of three pigs. I:C. Putative cDC2 and CD14+ DCs only produced IL12p40 cells were stimulated by poly I:C, indicating that there might be in cells stimulated with gardiquimod (4292 ± 4599 pg/ml and an inhibition of IL-8 secretion by the activation of TLR3 signaling 1218 ± 1686 pg/ml, respectively; Figure 7). AMs did not pathway. cDC1 did not produce this cytokine (Figure 7). produce this cytokine. With respect to IL-10, only putative cDC2 and CD14+ DCs produced this cytokine with CD14+ DCs making larger amount Addition of SCF, GM-CSF, or IL-4 Does following the stimulation by gardiquimod (1083 ± 996.4 pg/ml) Not Improve the Production of cDCs or poly I:C (666 ± 722 pg/ml), while cDC2 produced an Once the cell populations originated after Flt3L stimulation were intermediate amount (485 ± 509 pg/ml) in response to identified and characterized, we tried to optimize the production gardiquimod (Figure 7). of DCs. Firstly, we compared the use of 20 ng/ml versus Regarding the pro-inflammatory cytokines (IL-1β and TNF- 100 ng/ml. This had no evident effect on the phenotype, the yield, α), none was detected in the supernatant of cDC1 regardless of or the relative proportions of different DC populations within the the TLR agonist used. In contrast, moderate amounts of IL-1β Flt3L-derived bone marrow cell cultures (not shown). Secondly, was detected in the supernatant of cDC2 (299 ± 277 pg/ml) SCF was added to Flt3L-stimulated cultures at 20 ng/ml. The only or CD14+ DCs (215 ± 184 pg/ml) in both cases, through noticeable effect was an increase in the yield of DC-compatible the TLR7-signaling pathway. Notably, while negligible TNF- cells at the end of the 14-day incubation period that, on average, α was produced by cDC2 or CD14+ DCs, AMs produced was 24% with no significant change in the type of cells produced a considerable amount, regardless of the TLR ligand used or their relative proportions. Increase of SCF to 100 ng/ml did not (19,814 ± 6655 pg/ml with poly I:C; 11,514 ± 3305 pg/ml with show significant changes. LPS; and 12,984 ± 5169 pg/ml with gardiquimod; Figure 7). In the next step, GM-CSF was added to the bone marrow In addition, a tiny amount of IL-6 was produced by cDC2 cultures after 6 days of the expansion of Flt3+ precursors, (20.5 ± 19.7 pg/ml) and CD14+ DCs (17.1 ± 22.0 pg/ml) upon culturing for an extra 12 days (18-day period in total). The gardiquimod treatment. IL-8 (CXCL8), an essential chemokine resulting cells on day 12 contained mostly (around 90%) secreted from DCs involved in both Th1 and Th2 responses, was CADM1− CD14+ CD172a+ , among which 30% were MHC- highly secreted by cDC2, CD14+ DCs, or AMs. It is noteworthy II+ . These CD14+ CD172a+ cells were further characterized that there was a decrease in cDC2 cytokine secretion when these as CD1− CD163+ DEC205+ . Nevertheless, by day 18, around Frontiers in Immunology | www.frontiersin.org 11 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC FIGURE 7 | Continued Frontiers in Immunology | www.frontiersin.org 12 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC FIGURE 7 | Cytokine production by sorted DC subsets (putative cDC1, cDC2, and CD14+ DCs) following TLR stimulation. Cells were stimulated for 18 h with 10 µg/ml poly I:C, 1 µg/ml LPS, 10 µg/ml gardiquimod, or left unstimulated as controls. The supernatants were harvested and tested for IFN-α, IFN-γ, IL-4, IL-6, IL-12p40, IL-10, IL-1β, TNF-α, and IL-8 (CXCL8) by multiplex immunoassay. Production of IL-12p40, IL-10, IL-1β, TNF-α, IL-6, and IL-8 (CXCL8) was shown, while IFN-α, IFN-γ, or IL-4 were not shown as rarely detected. Alveolar macrophages (AMs) were used as a different cell-type reference. The values were displayed with logarithmic y-axis. Symbols in different colors represent different animals. The median is shown for each data set. 90% of CD163+ cells transformed to CD163− . Increasing of cDC1 (20) and, in lung cDC1 after influenza A virus infection Flt3L up to 100 ng/ml did not substantially modify this profile (14). However, this differed from blood cDC1 where no IL- (Supplementary Figure 3). 12p40 was detected by multiplex cytokine assay or qRT-PCR. With the same scenario, addition of IL-4 produced mostly It is worth noting that stimulation with poly I:C increased the CADM1− CD14+ CD172a+ cells with MHC-II+ increasing viability of cDC1 in comparison to using other TLR ligands or from 15% on day 12 to 30% by day 18. However, in this just plain medium. Since poly I:C was a potent inducer of IL- case, the CD14+ CD172a+ population could be further 12p40 in cDC1, it is tempting to speculate that this cytokine subdivided based on the expression of CD1 (roughly a was the cause of the enhanced survival. Poor viability of cDC1 ratio of 1:0.8 of CD1+ /CD1− ) and CD163 (roughly a ratio (32) and protection from cell death of cDC1 by IL-12 were of 1:1.5 of CD163+ /CD163− ). Only a small proportion previously reported (33, 34); (5) they were unresponsive to LPS of the CADM1− CD14+ CD172a+ expressed DEC205 or gardiquimod, compatible with the negligible expression of (Supplementary Figure 3). TLR4 and TRL7 on blood cDC1 (22); (6) they were incapable of producing IL-1β, same as human blood cDC1 where the scarcely expressed caspase-1 was reasoned (35); (7) they showed DISCUSSION high endocytic capability. Interestingly, two subsets of cDC1 were identified in the dextran uptake experiments: low and The study of DCs and their interactions with pathogens is one high uptake, which did not correlate with CD172a or MHC- of the main areas of research in immunology. One obstacle for II expression levels; and (8) they showed very weak phagocytic these studies is the low proportion of DCs residing in tissues, activity, comparably to cDC2, but contrasting with the more which makes it difficult to obtain sufficient cell number for efficient CD14+ DCs. This is in agreement with the finding further analysis. Establishing an in vitro DC derivation model can in mice that the clearance of bulk particles was mostly by make it more accessible for research i.e. the interaction of DCs macrophages rather than cDCs (36). With respect to the yield, with pathogens, the preliminary screening or testing hypotheses this population was the minor one in the Flt3L-derived cells, without need to perform costly whole-animal challenges. Here, accounting for less than 5%. This fact suggests that either the we present a comprehensive characterization of cells produced in precursors of cDC1 are scarce or the expansion of precursors an Flt3L-based in vitro derivation method. and the survival of generated cDC1 require synergizing efforts The use of Flt3L alone to stimulate BMHCs produced from more factors. In human in vitro production models, cDC1 is three main populations, which shared a CADM1+ MHC-II+ also a minor population compared to cDC2 (17), which contrasts phenotype and a variable expression of CD14 and CD172a. with the more balanced proportion of cDC1 and cDC2 in mice The first population was CADM1+ CD14− MHC-II+ CD172a−/lo , models (37). which presented full characteristics of cDC1: (1) it was further The second, a major population, had a phenotype of phenotyped as CD1− , CD163− , CD11R3lo , CD11R1+ , and CADM1+ /CD14− /MHC-II+ /CD172a+ , representative of cDC2 DEC205+ , equal to the cDC1 described in porcine lymphoid (52% on average among CADM1+ cells). This population tissues (15, 20), skin (21), lung (14), and blood (22). Of note, expressed a high level of FcεR1α, which has been identified CD11b was positive on our cDC1, as revealed by the staining as an overexpressed factor on cDC2 in human and pig tissues of CD11R3 and CD11R1 and the Quantitative Real-Time PCR (14, 22, 38). IRF4, a driving factor of cDC2 differentiation (RT-qPCR) of ITGAM. This was compatible with the expression (39), was also highly expressed. Combined with additional pattern of CD11b on blood cDC1 (22) but in contrast with the phenotypic markers CD1+ , CD163lo , CD11R3+ , and CD11R1+ , findings for lung cDC1 (14). In mice, CD11b is considered a this population was assigned to cDC2. However, the level of specific marker of cDC2 (29); in pigs, according to the current DEC205 on this population was lower than on the putative result, CD11b expression on cDCs may depend on the cell cDC1, which might be explained by our previous work in type, as well as on the location of the cells, namely, blood or which expression of DEC205 was variable in cDCs obtained tissues; (2) in addition to FLT3 and ZBTB46, these cells highly from different lymphoid tissues (15). When we extended the expressed XCR1, which is a factor strictly associated with cDC1 culture of these cDC2 for an additional 48 h, more cells (30) as well as IRF8, a driving factor of cDC1 development in gained DEC205 expression but lost CADM1 (data not shown), human and mice (31); (3) they showed antigen presentation indicating that the lack of DEC205 expression might correspond capability, particularly in activating allogenic CD8+ CD4+ T cells, to a relatively immature developmental stage. More interestingly, indicating the potential of inducing CD8+ effector/memory T in our hands, the addition of different cytokines also had an cells; (4) they secreted a substantial amount of IL-12p40 following impact on DEC205 expression. For example, addition of GM- poly I:C stimulation, similarly to what was found in tonsil CSF induced higher expression of DEC205. In other models, Frontiers in Immunology | www.frontiersin.org 13 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC i.e., human DCs obtained ex vivo from skin, a similar variable adherence (data not shown). Other alternative hypotheses pattern was observed for CD1a+ DDC (40). Furthermore, can also be formulated. The CD14+ cells might represent a from a functional perspective, the putative cDC2 were weak subtype of cDC2, originated from the same precursors, but in phagocytosing killed S. aureus, but efficient in endocytosing diversified into a distinct cDC2 subtype dependent on variable dextran, and were able to induce allogenic T cell proliferation. transcription clusters; even they could be in the transitional They produced IL-12p40 after TLR7 stimulation, which differed stage to cDC2. The presence of IRF4 and ZBTB46, and from cDC1 upon TLR3 stimulation. The secretion of IL-12p40 the comparable expression of KLF4 was in line with the was not the specialty of one DC subtype in our model; this two hypotheses. is contrary with what is reported in human (cDC2) (41) and CD4+ pDCs were not consistently detected in the cultures of mice (cDC1) (42). Moreover, the putative cDC2 were efficient our study, indicating that no or very few pDCs were present. in producing IL-10 or IL-1β, but not TNF-α, as was described According to Guzylack-Piriou et al. (19), less than 1% of Flt3L- in human cDC2 (43, 44). Another thing worth mentioning stimulated cultures were compatible with pDC. This might be is the secretion of IL-6. Although it was a small amount due to the lack of stromal cells in the culturing system (50), since (20.5 ± 19.7 pg/ml), it has been reported that, in mice, low their presence has proved to be indispensable for human pDC amounts of IL-6 in the milieu can be trans-presented by the derivation in vitro (17). CD11b+ Sirpα+ DCs to initiate the transcriptional program The kinetic development of Flt3L-derived cells showed a and generate pathogenic Th17 cells (45). Thus, the cDC2 sharp drop in cell numbers during the first days of incubation, produced in our study might play a guiding role in Th17 cell suggesting death of non-stimulated Flt3− cells in the culture; differentiation through secreting IL-6 in pigs. CD33 (Siglec- meanwhile, the proportion of MHC-II+ cells increased and 3) was positive on both cDC1 and cDC2, consistent with the afterward the CADM1+ population appeared and expanded. study on human in which CD33 has a pan-distribution on Based on this kinetic, we attempted to increase the yield of myeloid pre-cDCs and cDCs (44). This is the first report of this DCs by introducing SCF to induce hematopoietic progenitor molecule on pig DCs. expansion and subsequent survival (51). Although this resulted The third identified population was CADM1+ CD14+ MHC- in a 24% increase in yield, the profile of cells produced II CD172a+ (accounting for roughly 40% of the CADM1+ + was not affected. In contrast, the addition of GM-CSF at cells), designated as CD14+ DCs in this paper. It was day 6 of incubation resulted in a proportion of CD14+ further phenotyped as CD1+ , CD163+ , CD11R3+ , CD11R1+ , cells exceeding 90%, in agreement with what was previously DEC205− , and CD33− . Based on the expression of FLT3, observed in GM-CSF-derived bone marrow cultures (16, 19). ZBTB46, FcεR1a, and IRF4 although low, they could be classified Moreover, the produced cells were mostly CADM1− . These as cDCs. Furthermore, this CD14+ subset shared functional facts indicate that addition of GM-CSF may produce a similarities with cDC2 in regard to T cell proliferation and different pattern of cDCs. In human cultures, omission of cytokine production in response to TLR stimulation (secretion GM-CSF results in very low expression of DNGR-1+ BDCA3+ of IL-12p40, IL-10, and IL-6 upon gardiquimod treatment DCs (cDC1 equivalent) (52), but in mice cultures, the lack and very low or no TNF-α secretion following poly I:C of GM-CSF or its receptor did not result in a critical or gardiquimod stimulation). This is in concordance with impairment (53). Another noticeable difference between several what was observed in CD14+ DCs in human interstitial cytokine combinations is the expression of CD1. Most of dermis (46, 47). However, the relatively high expression of the cells produced by Flt3L plus GM-CSF stimulation were CSF1R, MAFB, and, to a lesser extent, MERTK would point CD1− . CD1 expression was restored by addition of IL-4. The these CD14+ cells toward a myelomonocytic origin. The combination of four factors (Flt3L, SCF, GM-CSF, and plus CD14+ cells were also the ones with the highest phagocytic IL-4) generated a major population (>90%) with phenotype activity and were less responsive to TLR ligands in terms of CADM1− CD14+ CD172a+ , and a minor one (<10%) of of maturation, elements favoring the monomyelocytic origin CADM1− CD14− CD172a+ . These cells were heterogeneous hypothesis. Nevertheless, moDC differentiation is boosted for CD1 and CD163 expression, with CD163 exclusively by GM-CSF rather than Flt3L (48, 49). Actually, in the expressed by CD14+ cells (data not shown). Within CD1+ present study, addition of GM-CSF radically changed the compartment, the CD14− CD163− and CD14+ CD163+ subsets profile of cells obtained. Taken together, and considering may be analogous to, respectively, cDC2 and the recently that CADM1 is a cross-species, evolutionarily-conserved DC- identified human DC3 (54, 55). In our case, CD1 expression specific marker that is not present in moDCs (38, 39), in Flt3L-stimulated BMHCs had a similar expansion kinetic classification of the CD14+ subset as cDCs or moDCs to CADM1. The reasons why IL-4 restored that expression cannot be entirely precise. In the work of Soldevila et al. are not known yet. (20), a similar CD14+ subset was also identified in porcine Overall, the Flt3L-based in vitro derivation model produced tonsils, which displayed an intermediate attribute between in vivo equivalent cDC1 and cDC2 as well as a CD14+ population cDCs and moDCs in terms of phenotype, transcriptomics, that, at present, cannot be precisely classified. This can be a and pathogen clearance. To note, morphologically cDC2 suitable model to study the ontogeny of pig DCs, to assess were produced as clusters of 5–15 cells, while CD14+ potential functions of pig DCs as well as to study the interaction cells were seen as single cells, but both possessed typical of DCs with porcine-related pathogens. The latter application DC dendritic and veiled extensions and lack of plastic permits in vitro examination of DCs in responding to pathogens Frontiers in Immunology | www.frontiersin.org 14 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC without performing animal challenges, complying with the FUNDING ethical obligation of reducing animal use in experimentation. This work was funded by the Veterinary Infectious Diseases Diagnostic Laboratory of the UAB. LP-C was supported by DATA AVAILABILITY STATEMENT CONACYT to perform a research stay at UAB. All datasets presented in this study are included in the article/Supplementary Material. ACKNOWLEDGMENTS We gratefully thank Manuela Costa García (SCAC, UAB, ETHICS STATEMENT Spain) who assisted in cell sorting. We thank Esmeralda The animal study was reviewed and approved by IRTA-CReSA Cano and Núria Navarro (CReSA, IRTA-UAB, Spain) who Autonomous University of Barcelona. kindly provided technical assistance in the isolation of alveolar macrophages. We also thank Joan Pujols (CReSA, IRTA- UAB, Spain) and Mariano Domingo (CReSA, IRTA-UAB, AUTHOR CONTRIBUTIONS Spain) for help in the collection of femur and humerus. We thank Thomas Shelite (The University of Texas Medical YL contributed to the design and the execution of the whole Branch, Galveston, TX, United States) who helped in panel of experiments, the analysis and the interpretation of the English editing. data, and the writing of the manuscript. EM contributed to the design of the experiments, the analysis and the interpretation of the data, and the writing of the manuscript. JH contributed SUPPLEMENTARY MATERIAL to the design of the experiments and the interpretation of the data. LP-C contributed to the design and the execution of the The Supplementary Material for this article can be found phenotype characterization. ID assisted in the execution of part online at: https://www.frontiersin.org/articles/10.3389/fimmu. of the functional analysis. All authors reviewed the manuscript. 2020.553859/full#supplementary-material REFERENCES 10. Den Haan JMM, Lehar SM, Bevan MJ. CD8+ but not CD8- dendritic cells cross-prime cytotoxic T cells in vivo. J Exp Med. (2000) 192:1685–96. doi: 1. Statista Number of Pigs Worldwide from 2012 to 2020 (in Million Head). (2020). 10.1084/jem.192.12.1685 Available online at: https://www.statista.com/statistics/263963/number-of- 11. Siegal FP, Kadowaki N, Shodell M, Fitzgerald-Bocarsly PA, Shah K, Ho S, et al. pigs-worldwide-since-1990/. (accessed February 24, 2020). The nature of the principal Type 1 interferon-producing cells in human blood. 2. Smith GJD, Vijaykrishna D, Bahl J, Lycett SJ, Worobey M, Pybus OG, et al. Science. (1999) 284:1835–7. doi: 10.1126/science.284.5421.1835 Origins and evolutionary genomics of the 2009 swine-origin H1N1 influenza 12. Cella M, Jarrossay D, Faccheth F, Alebardi O, Nakajima H, Lanzavecchia A, a epidemic. Nature. (2009) 459:1122–5. doi: 10.1038/nature08182 et al. Plasmacytoid monocytes migrate to inflamed lymph nodes and produce 3. Meurens F, Summerfield A, Nauwynck H, Saif L, Gerdts V. The pig: a model large amounts of type I interferon. Nat Med. (1999) 5:919–23. doi: 10.1038/ for human infectious diseases. Trends Microbiol. (2012) 20:50–7. doi: 10.1016/ 11360 j.tim.2011.11.002 13. Edwards JC, Everett HE, Pedrera M, Mokhtar H, Marchi E, Soldevila F, et al. 4. Ibrahim Z, Busch J, Awwad M, Wagner R, Wells K, Cooper DKC. Selected CD1- and CD1+ porcine blood dendritic cells are enriched for the orthologues physiologic compatibilities and incompatibilities between human and porcine of the two major mammalian conventional subsets. Sci Rep. (2017) 7:40942. organ systems. Xenotransplantation. (2006) 13:488–99. doi: 10.1111/j.1399- doi: 10.1038/srep40942 3089.2006.00346.x 14. Maisonnasse P, Bouguyon E, Piton G, Ezquerra A, Urien C, Deloizy C, et al. 5. Merad M, Sathe P, Helft J, Miller J, Mortha A. The dendritic cell lineage: The respiratory DC/macrophage network at steady-state and upon influenza ontogeny and function of dendritic cells and their subsets in the steady state infection in the swine biomedical model. Mucosal Immunol. (2016) 9:835–49. and the inflamed setting. Annu Rev Immunol. (2013) 31:563–604. doi: 10.1146/ doi: 10.1038/mi.2015.105 annurev-immunol-020711-074950 15. Parra-Sánchez H, Puebla-Clark L, Reséndiz M, Valenzuela O, Hernández J. 6. Guilliams M, Ginhoux F, Jakubzick C, Naik SH, Onai N, Schraml BU, et al. Characterization and expression of DEC205 in the cDC1 and cDC2 subsets of Dendritic cells, monocytes and macrophages: a unified nomenclature based porcine dendritic cells from spleen, tonsil, and submaxillary and mesenteric on ontogeny. Nat Rev Immunol. (2014) 14:571–8. doi: 10.1038/nri3712 lymph nodes. Mol Immunol. (2018) 96:1–7. doi: 10.1016/j.molimm.2018.02. 7. León B, López-Bravo M, Ardavín C. Monocyte-derived dendritic cells formed 003 at the infection site control the induction of protective T helper 1 responses 16. Carrasco CP, Rigden RC, Schaffner R, Gerber H, Neuhaus V, Inumaru S, against leishmania. Immunity. (2007) 26:519–31. doi: 10.1016/j.immuni.2007. et al. Porcine dendritic cells generated in vitro: morphological, phenotypic and 01.017 functional properties. Immunology. (2001) 104:175–84. doi: 10.1046/j.1365- 8. Waskow C, Liu K, Darrasse-Jèze G, Guermonprez P, Ginhoux F, Merad 2567.2001.01299.x M, et al. The receptor tyrosine kinase Flt3 is required for dendritic cell 17. Lee J, Breton G, Oliveira TYK, Zhou YJ, Aljoufi A, Puhr S, et al. Restricted development in peripheral lymphoid tissues. Nat Immunol. (2008) 9:676–83. dendritic cell and monocyte progenitors in human cord blood and bone doi: 10.1038/ni.1615 marrow. J Exp Med. (2015) 212:385–99. doi: 10.1084/jem.20141442 9. Jongbloed SL, Kassianos AJ, McDonald KJ, Clark GJ, Ju X, Angel CE, et al. 18. Karsunky H, Merad M, Cozzio A, Weissman IL, Manz MG. Flt3 ligand Human CD141+ (BDCA-3)+ dendritic cells (DCs) represent a unique myeloid regulates dendritic cell development from Flt3+ lymphoid and myeloid- DC subset that cross-presents necrotic cell antigens. J Exp Med. (2010) committed progenitors to Flt3+ dendritic cells in vivo. J Exp Med. (2003) 207:1247–60. doi: 10.1084/jem.20092140 198:305–13. doi: 10.1084/jem.20030323 Frontiers in Immunology | www.frontiersin.org 15 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC 19. Guzylack-Piriou L, Alves MP, McCullough KC, Summerfield A. Porcine Flt3 marginal zone of mouse spleen. Proc Natl Acad Sci USA. (2009) 106:1524–9. ligand and its receptor: generation of dendritic cells and identification of a doi: 10.1073/pnas.0812247106 new marker for porcine dendritic cells. Dev Comp Immunol. (2010) 34:455–64. 37. Naik SH, Sathe P, Park HY, Metcalf D, Proietto AI, Dakic A, et al. doi: 10.1016/j.dci.2009.12.006 Development of plasmacytoid and conventional dendritic cell subtypes from 20. Soldevila F, Edwards JC, Graham SP, Stevens LM, Crudgington B, Crooke HR, single precursor cells derived in vitro and in vivo. Nat Immunol. (2007) et al. Characterization of the myeloid cell populations’ resident in the porcine 8:1217–26. doi: 10.1038/ni1522 palatine tonsil. Front Immunol. (2018) 9:1800. doi: 10.3389/fimmu.2018.01800 38. Segura E, Valladeau-Guilemond J, Donnadieu MH, Sastre-Garau X, Soumelis 21. Marquet F, Vu Manh T-P, Maisonnasse P, Elhmouzi-Younes J, Urien C, V, Amigorena S. Characterization of resident and migratory dendritic cells Bouguyon E, et al. Pig skin includes dendritic cell subsets transcriptomically in human lymph nodes. J Exp Med. (2012) 209:653–60. doi: 10.1084/jem.201 related to human CD1a and CD14 dendritic cells presenting different 11457 migrating behaviors and T cell activation capacities. J Immunol. (2014) 39. Schlitzer A, McGovern N, Teo P, Zelante T, Atarashi K, Low D, et al. IRF4 193:5883–93. doi: 10.4049/jimmunol.1303150 transcription factor-dependent CD11b+ dendritic cells in human and mouse 22. Auray G, Keller I, Python S, Gerber M, Bruggmann R, Ruggli N, et al. control mucosal IL-17 cytokine responses. Immunity. (2013) 38:970–83. doi: Characterization and transcriptomic analysis of porcine blood conventional 10.1016/j.immuni.2013.04.011 and plasmacytoid dendritic cells reveals striking species-specific differences. J 40. Harman AN, Bye CR, Nasr N, Sandgren KJ, Kim M, Mercier SK, et al. Immunol. (2016) 197:4791–806. doi: 10.4049/jimmunol.1600672 Identification of lineage relationships and novel markers of blood and skin 23. doi: 10.1128/jcm.33.4.893-897.1995Mattsson JG, Bergström K, Wallgren P, human dendritic cells. J Immunol. (2013) 190:66–79. doi: 10.4049/jimmunol. Johansson KE, Wallgren PER. Detection of Mycoplasma hyopneumoniae in 1200779 nose swabs from pigs by in vitro amplification of the 16S rRNA. J Clin 41. Nizzoli G, Krietsch J, Weick A, Steinfelder S, Facciotti F, Gruarin P, et al. Microbiol. (1995) 33:893–7. Human CD1c+ dendritic cells secrete high levels of IL-12 and potently prime 24. Quintana J, Balasch M, Segalés J, Calsamiglia M, Rodríguez-Arrioja GM, cytotoxic T-cell responses. Blood. (2013) 122:932–42. doi: 10.1182/blood- Plana-Durán J, et al. Experimental inoculation of porcine circoviruses type 1 2013-04-495424 (PCV1) and type 2 (PCV2) in rabbits and mice. Vet Res. (2002) 33:229–37. 42. Hochrein H, Shortman K, Vremec D, Scott B, Hertzog P, O’Keeffe M. doi: 10.1051/vetres:2002011 Differential production of IL-12, IFN-α, and IFN-γ by mouse dendritic 25. Segalés J, Martínez-Guinó L, Cortey M, Navarro N, Huerta E, Sibila M, et al. cell subsets. J Immunol. (2001) 166:5448–55. doi: 10.4049/jimmunol.166. Retrospective study on swine Torque teno virus genogroups 1 and 2 infection 9.5448 from 1985 to 2005 in Spain. Vet Microbiol. (2009) 134:199–207. doi: 10.1016/j. 43. Nizzoli G, Larghi P, Paroni M, Crosti MC, Moro M, Neddermann P, et al. vetmic.2008.08.002 IL-10 promotes homeostatic proliferation of human CD8+ memory T cells 26. Domínguez J, Alvarez B, Alonso F, Thacker E, Haverson K, McCullough and, when produced by CD1c+ DCs, shapes naive CD8+ T-cell priming. Eur J K, et al. Workshop studies on monoclonal antibodies in the myeloid panel Immunol. (2016) 46:1622–32. doi: 10.1002/eji.201546136 with CD11 specificity. Vet Immunol Immunopathol. (2001) 80:119–119. doi: 44. Collin M, Bigley V. Human dendritic cell subsets: an update. Immunology. 10.1016/S0165-2427(01)00286-0 (2018) 154:3–20. doi: 10.1111/imm.12888 27. Satpathy AT, Kc W, Albring JC, Edelson BT, Kretzer NM, Bhattacharya D, et al. 45. Heink S, Yogev N, Garbers C, Herwerth M, Aly L, Gasperi C, et al. Zbtb46 expression distinguishes classical dendritic cells and their committed Trans-presentation of IL-6 by dendritic cells is required for the priming progenitors from other immune lineages. J Exp Med. (2012) 209:1135–52. of pathogenic T H 17 cells. Nat Immunol. (2017) 18:74–85. doi: 10.1038/ doi: 10.1084/jem.20120030 ni.3632 28. Bordet E, Blanc F, Tiret M, Crisci E, Bouguyon E, Renson 46. Tamoutounour S, Guilliams M, MontananaSanchis F, Liu H, Terhorst D, P, et al. Porcine reproductive and respiratory syndrome virus Malosse C, et al. Origins and functional specialization of macrophages type 1.3 Lena triggers conventional dendritic cells 1 activation and of conventional and monocyte-derived dendritic cells in mouse and t helper 1 immune response without infecting dendritic skin. Immunity. (2013) 39:925–38. doi: 10.1016/j.immuni.2013. cells. Front Immunol. (2018) 9:2299. doi: 10.3389/fimmu.2018. 10.004 02299 47. Haniffa M, Shin A, Bigley V, McGovern N, Teo P, See P, et al. Human tissues 29. Schlitzer A, Ginhoux F. Organization of the mouse and human DC network. contain CD141 hi cross-presenting dendritic cells with functional homology Curr Opin Immunol. (2014) 26:90–9. doi: 10.1016/j.coi.2013.11.002 to mouse CD103 + non-lymphoid dendritic cells. Immunity. (2012) 37:60–73. 30. Crozat K, Guiton R, Contreras V, Feuillet V, Dutertre CA, Ventre E, et al. The doi: 10.1016/j.immuni.2012.04.012 XC chemokine receptor 1 is a conserved selective marker of mammalian cells 48. Mildner A, Yona S, Jung S. A close encounter of the third kind. monocyte- homologous to mouse CD8α+ dendritic cells. J Exp Med. (2010) 207:1283–92. derived cells. Adv Immunol. (2013) 120:69–103. doi: 10.1016/B978-0-12- doi: 10.1084/jem.20100223 417028-5.00003-X 31. Brown CC, Gudjonson H, Pritykin Y, Deep D, Lavallée VP, Mendoza A, et al. 49. Varol C, Vallon-Eberhard A, Elinav E, Aychek T, Shapira Y, Luche H, Transcriptional basis of mouse and human dendritic cell heterogeneity. Cell. et al. Intestinal lamina propria dendritic cell subsets have different origin (2019) 179:846–63.e24. doi: 10.1016/j.cell.2019.09.035 and functions. Immunity. (2009) 31:502–12. doi: 10.1016/j.immuni.2009. 32. Kronin V, Winkel K, Süss G, Kelso A, Heath W, Kirberg J, et al. A subclass 06.025 of dendritic cells regulates the response of naive CD8 T cells by limiting their 50. Olivier A, Lauret E, Gonin P, Galy A. The Notch ligand delta-1 is IL-2 production. J Immunol. (1996) 157:3819–27. a hematopoietic development cofactor for plasmacytoid dendritic 33. Ploemacher RE, van Soest PL, Voorwinden H, Boudewijn A. Interleukin-12 cells. Blood. (2006) 107:2694–701. doi: 10.1182/blood-2005-03- synergizes with interleukin-3 and steel factor to enhance recovery of murine 0970 hemopoietic stem cells in liquid culture. Leukemia. (1993) 7:1381–8. 51. Ho CCM, Chhabra A, Starkl P, Schnorr PJ, Wilmes S, Moraga I, 34. Pirtskhalaishvili G, Shurin GV, Esche C, Cai Q, Salup RR, Bykovskaia et al. Decoupling the functional pleiotropy of stem cell factor by SN, et al. Cytokine-mediated protection of human dendritic cells tuning c-Kit signaling. Cell. (2017) 168:1041–52.e18. doi: 10.1016/j.cell.2017. from prostate cancer-induced apoptosis is regulated by the Bcl-2 02.011 family of proteins. Br J Cancer. (2000) 83:506–13. doi: 10.1054/bjoc.20 52. Poulin LF, Salio M, Griessinger E, Anjos-Afonso F, Craciun L, Chen JL, 00.1289 et al. Characterization of human DNGR-1+ BDCA3+ leukocytes as putative 35. Worah K, Mathan TSM, Vu Manh TP, Keerthikumar S, Schreibelt G, Tel J, equivalents of mouse CD8α+ dendritic cells. J Exp Med. (2010) 207:1261–71. et al. Proteomics of human dendritic cell subsets reveals subset-specific surface doi: 10.1084/jem.20092618 markers and differential inflammasome function. Cell Rep. (2016) 16:2953–66. 53. Vremec D, Lieschke GJ, Dunn AR, Robb L, Metcalf D, Shortman K. The doi: 10.1016/j.celrep.2016.08.023 influence of granulocyte/macrophage colony-stimulating factor on dendritic 36. Idoyaga J, Suda N, Suda K, Park CG, Steinman RM. Antibody to cell levels in mouse lymphoid organs. Eur J Immunol. (1997) 27:40–4. doi: Langerin/CD207 localizes large numbers of CD8α + dendritic cells to the 10.1002/eji.1830270107 Frontiers in Immunology | www.frontiersin.org 16 October 2020 | Volume 11 | Article 553859 Li et al. Flt3L-Derived Pig Conventional DC 54. Cytlak U, Resteu A, Pagan S, Green K, Milne P, Maisuria S, et al. Differential Conflict of Interest: The authors declare that the research was conducted in the IRF8 requirement defines two pathways of dendritic cell development in absence of any commercial or financial relationships that could be construed as a humans. SSRN Electron J. (2019). 53:353-70.e8. doi: 10.2139/ssrn.3365028 potential conflict of interest. 55. Bourdely P, Anselmi G, Vaivode K, Ramos RN, Missolo-Koussou Y, Hidalgo S, et al. Transcriptional and functional analysis of CD1c+ Copyright © 2020 Li, Puebla-Clark, Hernández, Díaz and Mateu. This is an open- human dendritic cells identifies a CD163+ subset priming CD8+CD103+ access article distributed under the terms of the Creative Commons Attribution T cells. Immunity. (2020) 53:335–52.e8. doi: 10.1016/j.immuni.2020. License (CC BY). The use, distribution or reproduction in other forums is permitted, 06.002 provided the original author(s) and the copyright owner(s) are credited and that the 56. Hurst J, Maria A, Hofmann ML, Kuehn S, Herms F, Schnichels S, et al. iNOS- original publication in this journal is cited, in accordance with accepted academic inhibitor driven neuroprotection in a porcine retina organ culture model. J practice. No use, distribution or reproduction is permitted which does not comply Cell Mol Med. (2020) 24:4312–23. doi: 10.1111/jcmm.15091 with these terms. Frontiers in Immunology | www.frontiersin.org 17 October 2020 | Volume 11 | Article 553859