Edinburgh Research Explorer Characterization of conventional dendritic cells and macrophages in the spleen using the CSF1R-reporter transgenic chickens

major the immunological to blood-borne macrophages enriched for red pulp macrophages. However, our analysis could not deﬁne the cell lineage of the heterogeneous CSF1R -tg dim subpopulation. This detailed overview of the MPS in the chicken spleen will contribute to future research on their role in antigen uptake and presentation.

The spleen is a major site for the immunological responses to blood-borne antigens that is coordinated by cells of the mononuclear phagocyte system (MPS). The chicken spleen is populated with a number of different macrophages while the presence of conventional dendritic cells (cDC) has been described. However, a detailed characterization of the phenotype and function of different macrophage subsets and cDC in the chicken spleen is limited. Using the CSF1R-reporter transgenic chickens (CSF1R-tg), in which cells of the MPS express a transgene under the control elements of the chicken CSF1R, we carried out an in-depth characterization of these cells in the spleen. Immunohistological analysis demonstrated differential expression of MRC1L-B by periarteriolar lymphoid sheaths (PALS)-associated CSF1R-tg + cells. In the chicken's equivalent of the mammalian marginal zone, the peri-ellipsoid white-pulp (PWP), we identified high expression of putative CD11c by ellipsoid-associated cells compared to ellipsoid-associated macrophages. In addition, we identified a novel ellipsoid macrophage subset that expressed MHCII, CD11c, MRC1L-B, and CSF1R but not the CSF1R-tg. In flow cytometric analysis, diverse expression of the CSF1R-tg and MHCII was observed leading to the categorization of CSF1R-tg cells into CSF1R-tg dim MHCII inter−hi , CSF1R-tg hi MHCII hi , and CSF1R-tg hi MHCII inter subpopulations. Low levels of CD80, CD40, MHCI, CD44, and Ch74.2 were expressed by the CSF1R-tg hi MHCII inter cells. Functionally, in vivo fluorescent bead uptake was significantly higher in the CSF1R-tg hi MHCII hi MRC1L-B + cells compared to the CSF1R-tg dim and CSF1R-tg hi MHCII inter MRC1L-B + subpopulations while LPS enhanced phagocytosis by the CSF1R-tg hi MHCII inter subpopulation. The analysis of bead localization in the spleen suggests the presence of ellipsoid-associated macrophage subsets. In addition, we demonstrated the functionality of ex vivo derived CSF1R-tg + MRC1L-B neg cDC. Finally, RNA-seq analysis of the CSF1R-tg subpopulations demonstrated that separating the CSF1R-tg hi subpopulation into CD11c hi and CD11c dim cells enriched for cDC and macrophage lineages, respectively, while the CSF1R-tg hi MHCII inter subpopulation was

INTRODUCTION
The vertebrate spleen is a secondary lymphoid organ and despite differences in the immune system among vertebrates, the spleens basic histological architecture and its role in filtering the blood for antigens and damaged, aged cells has been conserved during evolution from fish, amphibians, birds, and mammals. In the mammalian spleen, the afferent splenic artery branches into the central artery which is sheathed by the white pulp areas that include the T cell zones known as the periarteriolar lymphoid sheath (PALS), arterioles, and B cell follicles. The white pulp and red pulp are extensively populated with macrophages and dendritic cells (DC) with distinct origins and functions that contribute to the control of blood-borne pathogens and homeostatic processes (1). The venous system of the red pulp gives it the capacity to filter the blood and remove old erythrocytes by red pulp macrophages. Red pulp macrophages are defined by ADGRE1 hi (ERM1, F4/80) CD11b lo /CD18 or ADGRE1 + CD11b/CD18 hi expression (2,3). The murine splenic red pulp also contains reservoirs of monocytes that are transcriptionally similar to their blood counterparts, Ly-6C + and Ly-6C ++ cells (ADGRE1 −/lo CD11b + CD11c − ) that egress to sites of tissue damage (4,5).
The main difference between rodents and primates spleen is the structure of the marginal zone (MZ) that surrounds the white pulp. In contrast to mice, the human MZ has an inner and outer region that is surrounded by a large perifollicular zone [reviewed by (6,7)]. The MZ consists of two resident macrophages subsets, metallophilic macrophages (MMM) located at the outer margins of the MZ and MZ macrophages (MZM) located in the inner regions of the MZ. Both MZ-resident macrophages express the scavenger receptor, MARCO while MZM also express SIGN-R1 (CD209b) (8). The MMM can uptake particular blood-borne antigens and present them to DC for cross-presentation to CD8 + cytotoxic T cells but they are not necessary for direct T cell activation (9).
Mammalian DC are composed of two major populations: conventional DC and plasmacytoid DC (pDC). Mammalian conventional DC are further subdivided into two major populations, cDC1 and cDC2. cDC1 develop in a BATF3 and IRF8 dependent manner and are identified by their surface expression of XCR1 and CD103, respectively (10)(11)(12). In the mammalian spleen, cDC1 are located within the T cell zones of the PALS and the outer regions of the MZ (13). The IRF4dependent cDC2 population express CD11b, SIRPα, and DEC-205 (CD205). cDC2 are a heterogeneous cell population, the definition of these is yet to be fully resolved but they have been implicated in driving Th1, Th2, and Th17 polarization (14). The third group of DC are the IRF8-, E2-2-, and SpiBdependent CD123 + pDC which are the primary source of type I IFN production following viral infection (15,16). In humans, CD141 + (BDCA3) and CD1c + (BDCA1), and the CD303 + (BDCA2) cells are the equivalent of murine cDC1, cDC2, and pDC subsets, respectively (17). In veterinary species where reagents are limited for the study of cDC and macrophage subsets, the identification of MPS cell ontogeny and cell type identity has been determined by specific gene expression patterns. For example, comparative genome analysis has identify homologous of cDC1 and cDC2 in sheep and pigs (18,19), and cattle (20,21). In the chicken, gene expression patterns have identified subsets of monocytes in the blood, resident tissue macrophages and cDC in the liver and spleen (22,23).
The basic structure of the chicken spleen resembles it mammalian counterpart, with morphologically defined white pulp and red pulp regions. The white pulp contains the PALS and peri-ellipsoid white-pulp (PWP) which surround the penicillar capillaries. The PWP is analogous to the mammalian marginal zone. The ellipsoid, more commonly known as the capsule of Schweigger-Seidel (CSS) in mammals, is embedded within the PWP. The PWP contains three distinct cell populations, the ellipsoid-associated cells (EAC) and B cells intermingled with ellipsoid-associated macrophages (EAM). Studies have demonstrated the ability of EAC to retain antigen at their cell surface to either provide the antigen to the EAM or for transport to the PALS where they develop into FDC (24)(25)(26)(27)(28). Mannose receptor C1-like B (MRC1L-B, recognized by KUL01) is one of five orthologs of mammalian MRC1 (206) and a key mononuclear phagocytic cell marker in lymphoid and non-lymphoid tissues in chickens (29,30). A previous comparative genomic analysis showed that splenic cDC lack MRC1L-B expression (23). However, there is limited research on the function and location of cDC, and the phenotype, function, and transcriptome of PALS-, PWP-, and red pulp-associated cells of the MPS in the chicken spleen.
The development, maintenance, and proliferation of cells of the MPS are dependent upon signals from CSF1 through its receptor, CSF1R (31) while DC require signals from FLT3L and its receptor, FLT3 (32). The tools to study the chicken MPS have been improved by the generation of the CSF1R-reporter transgenic chickens, in which a transgene is expressed under the control of the CSF1R-promoter and Fms-intronic regulatory element (FIRE) (33). These reporter chickens have contributed to the characterization of yolk-sac-derived macrophages, monocytes, and resident tissue mononuclear phagocytic cells (33)(34)(35)(36). In this study, we utilize the CSF1R-reporter transgenic chickens to demonstrate the heterogeneity of the MPS in the spleen. We show that the spleen contains at least three CSF1Rtg + subpopulations that diversely express CSF1R-tg, MHCII, MRC1L-B, and CD11c along with other macrophage markers. We also identified a CSF1R-tg neg macrophage population that express CSF1R protein in the PWP and demonstrate their ability to uptake beads in vivo. Furthermore, in vivo antigen uptake lead to the discovery of the diverse expression of the CSF1R-tg in EAM in the PWP. In addition, we demonstrate the phagocytic ability of cDC ex vivo. Finally, using RNA-seq we demonstrate that separation of the CSF1R-tg hi MHCII hi subpopulation into CD11c hi and CD11c dim subsets, enriches for genes associated with cDC and macrophage lineages while the gene signatures characteristic of red pulp macrophages are enriched in CSF1Rtg hi MHCII inter subpopulation. Future experiments will address the cDC-T cell interactions and the role of various cells of the splenic MPS in antigen-uptake and presentation.

MATERIALS AND METHODS
Chickens CSF1R-reporter transgenic (33) and non-transgenic Hy-Line birds (Hy-Line Brown) between the ages of 4-8 weeks of age were provided by the National Avian Research Facility (NARF), Edinburgh, U.K. Unvaccinated chickens were reared in floor pens with access to water and food ad-libitum.

Isolation of Splenocytes
Spleens were aseptically removed and stored in PBS on ice. Spleens were homogenized using PYREX R Potter-Elvehjem glass tissue grinder and passed through 70 µm cell strainer (Corning R ). Cells were overlaid on Histopaque (1.077 g, Sigma-Aldrich) and the interface and cell layer above were retrieved and washed twice with calcium and magnesium free PBS (pH 7.4). Cells were counted and stained at 1 × 10 6 cells for flow cytometric analysis.

In vitro Cell Cultures
Splenocytes were cultured at 20 × 10 6 cells/mL in complete media [RPMI supplemented with 10% heat-inactivated fetal bovine serum (FBS), 100 IU of penicillin, 100 µg of streptomycin (Gibco), 2 mM/mL L-glutamine (Life Technologies), and 200 ng/mL of recombinant chicken CSF1-Fc (35)]. After 24 h, non-adherent cells were removed and adherent cells were treated with TrypLE express (ThermoFisher) and reseeded at 5 × 10 5 cells in 24-well plates (Corning) in complete media. Cells were incubated at 41 • C, 5% CO 2 . Twenty-four hours later, cells were stimulated with pHrodo TM Red labeled Zymosan or E. coli bioparticles TM , unlabeled or LPS labeled 1 µm FluoSpheres R beads (ThermoFisher) (10:1 ratio of particles/beads to cells) at 41 or 4 • C. To inhibit actin polymerization, cells were treated with Cytochalasin D (Cayman Chemical) for 10 min prior to addition of the model antigens. Three hours post-treatment, ice-cold PBS was added to each well and cells were collected using TrypLE express (Gibco) for flow cytometric analysis. For immunostaining of in vitro cultured cells, 13 mm circular cover slips (ThermoFisher) were placed in 24-well plates and cultured as described above.

Immunostaining of Tissue Sections
For histological analysis, spleens were immediately flash frozen over liquid nitrogen or fixed in 4% paraformaldehyde (PFA/PBS) for 3 h and infused overnight with 30% sucrose (Sigma-Aldrich) at 4 • C. All samples were stored at −80 • C until use. Tissue sections were cut at 7 µm (Leica CM1900) and mounted on SuperFrost Plus TM slides (ThermoFisher) and air-dried overnight at RT after which samples were stained immediately or stored at 4 • C until use. Sections were blocked with goat serum (Chrondex) or heat inactivated horse serum (Sigma-Aldrich) for 30 min diluted in PBS supplemented with 0.01% Triton-X and 0.5% BSA (Sigma-Aldrich). Tissue samples were stained with antibodies outlined in Table 1 for 2 h at RT followed by secondary antibodies, rabbit poly anti-GFP-Alexa Fluor 488 R , goat anti-mouse Alexa Fluor 568 R IgG1, goat anti-mouse Alexa Fluor 647 R IgG1/IgG2a, goat anti-mouse Alexa Fluor 568 R IgG1 (ThermoFisher) and counterstained with 4 ′ ,6-diamidino-2-phenylindole (DAPI) or Hoechst 33258 (Sigma-Aldrich) for 1 h at RT. Slides were mounted using ProLong TM Diamond Antifade mountant (ThermoFisher) and analyzed using an inverted LSM710 or 880 Zeiss confocal microscope using 405, 488, 594, and 614 nm lasers, 40X or 63X Nikon oil lenses and captured using ZEN 2012 (Black Edition) software and analyzed using ZEN 2012 (Blue Edition).

Flow Cytometry and Flow Activated Cell Sorting
Splenocytes were stained using antibodies described in Table 1. Briefly, cells were incubated for 20 min on ice with primary antibodies diluted in Flow Buffer (PBS supplemented with 0.5% BSA and 0.01% sodium azide Sigma-Aldrich). Cells were then incubated with secondary antibodies for 20 min on ice. For multi-color flow cytometry, antibodies were conjugated using Lightning-Link TM (Novus Biologicals) with Alexa Fluor R 647, APC-Cy TM 7, PerCp-Cy TM 5.5, and PE-Cy TM 7 according to the manufacturer's instructions. All conjugated antibodies were titrated and tested for similar staining patterns observed using the unconjugated antibodies. For each experiment, live/dead cells were discriminated by the addition of SYTOX R Blue Dead Cell Stain (ThermoFisher) before cell analysis. Immuno-phenotyping of cells was carried out using BD LSRFortessa TM (BD Biosciences, UK) configured with 405, 488, 561, and 640 nm laser with 16 filters. Cells were sorted using BD FACS ARIA IIIu (BD Biosciences). All data was captured via BD FACSDiva TM Software v8.0.1 (BD Biosciences) and analyzed post-flow using FlowJo (Tree Star v10.0.2). Compensation was achieved using BD antimouse IgGκ compensation beads (BD Biosciences) for each flurochrome. Cell gating was carried out using cells from nontransgenic Hy-Line chickens and fluorescence minus one (FMO) controls. At least 50.000-100.000 live, single cell events were captured for analysis. Chickens between the ages of 5-8 weeks were intravenously administered with 100 µl of beads in PBS containing 3.6 × 10 9 0.5 µm or unlabeled or LPS labeled 1 µm carboxylate FluoSpheres R beads (ThermoFisher) or 100 µl of PBS-or clondronate-encapsulated liposomes (5 mg Liposoma, Europe) via the wing vein. Chickens were culled 3 h (beads) or 3 days (liposome) post-administration by cervical dislocation. Half of the spleen was prepared for immunohistology and the other half prepared for flow cytometric analysis. Ten thousand live, single cell events in the bead + gate were captured for analysis.

RNA-Seq Sample Preparation
Splenocytes were prepared as described above and sorted from five CSF1R-eGFP chickens at 7 weeks of age. Cells were sorted under low speed using an 80 µm diameter nozzle into 1.5 mL Eppendorf tubes at 4 • C. Following cell sort, samples were pelleted at 400 g for 3 min and lysed using RLT buffer (Qiagen) supplemented with β-mercaptoethanol and stored at −20 • C until use. RNA was purified using RNA Micro extraction kit according to the manufacturer's instructions (Qiagen). RNA concentration and quality was measured using High Sensitivity RNA Screen Tape System (Agilent). All samples were used with a RIN ≥7. cDNA was transcribed using Ovation R RNA-Seq System V2 (NuGen) according to the manufacturer's instructions. cDNA quality and quantity was analyzed using D1000 Screen Tape System (Agilent). RNA-seq was carried using TruSeq DNA Nano (350 bp insert) library preparation, sequencing using Illumina HiSeq4000 and data generation and preliminary analysis was carried out by Edinburgh Genomics, The University of Edinburgh, U.K.

RNA-Seq Analysis
Reads were trimmed using Cutadapt (version 1.9.dev2) (49) for quality at the 3' end using a quality threshold of 30 and for adapter sequences of the TruSeq DNA Nano kit. Reads after trimming were required to have a minimum length of 50 bp. Reads were aligned against the Gallus gallus 5.0 genome from Ensembl using STAR (50). Raw counts for each annotated gene were obtained using the feature counts software [version 1.5.2; (51)]. Reads were normalized using the weighted trimmed mean of M-values method (52). Differential gene expression analysis was performed within the Bioconductor edgeR package [version 3.16.5; (52)]. For differential analysis the raw counts table was filtered to remove genes consisting predominantly of near-zero counts, filtering on counts per million (CPM) to avoid artifacts due to library depth. Statistical assessment of differential expression was carried out with the likelihoodratio test. Contrasts specified were each pair of sorted cell populations. Differentially expressed genes were defined as those with a false-discovery rate (FDR) <0.05 and log2 fold-change (FC) >2. Heatmaps were constructed using normalized gene counts in R using the pheatmap package (v. 1.0.10; https:// CRAN.R-project.org/package=pheatmap). To statistically test whether mammalian immune cell transcriptional signatures were enriched in each chicken cell population we used a pairwise gene set enrichment analysis (GSEA) approach using the BubbleGum software for statistical calculations and graphical output (53). The chicken data was compared to mouse data as more spleen specific gene sets were available than for other species. We used mouse immune cell-specific gene sets developed previously (23,54), which are comprised of genes that show at least 1.5-fold higher expression in the cell population of interest compared to other immune cells. The gene sets used were for splenic T cells, B

Statistical Analysis
All data was analyzed for normality using Shapiro-Wilk test and non-parametric analysis was performed using GraphPad Prism 8 (GraphPad, USA). Differences between the groups were statistically evaluated by the non-parametric Mann-Whitney or Kruskal-Wallis test adjusted for post-hoc analysis. For all statistical analysis, P < 0.05 were considered significant.

RESULTS
Splenic PALS-Associated CSF1R-tg + Cells Do Not Uniformly Express MRC1L-B Whole mount microscopic analysis of the CSF1R-tg chicken spleen demonstrates the high density of CSF1R-tg + cell sheaths within the splenic PALS and the PWP ( Figure 1A).  Figure 1E). ChB6 + B cells are located within the GC and in the PWP where they form a ring around the ellipsoid ( Figure 1F). PWP-associated chB6 + B cells are intermingled with ellipsoid-associated macrophages (EAM) that express the CSF1R-tg, Ch74.2, MHCII, and MRC1L-B (Figures 1G,H). Closer analysis of the red pulp demonstrated that not all red pulp-associated CSF1R-tg + cells express Ch74.2 ( Figure 1G image insert) but they strongly express MRC1L-B ( Figure 1H). Within the inner regions of the PWP are EAC that express the CSF1R-tg (33) and can also be identified using the monoclonal Ch68.2 which binds to an unknown epitope [ Figure 1I; (25)]. The CSF1Rtg + expression demonstrates that cells of the splenic MPS differentially express MRC1L-B and help to distinguish cDC in the chicken spleen.
A CSF1R-tg neg Subpopulation Is Located in the Ellipsoid and Express CSF1R Protein The PWP is structurally defined as the region from the discontinuous basement membrane of the penicilliform capillary that harbors supportive cells surrounded by the capsule of the Schweigger-Seidel sheath [ Figure 2A; (28)]. The ellipsoid, embedded within the PWP, is the filtering apparatus of the spleen and represents a key component of chicken splenic lymphoid system. Using high-resolution confocal analysis, the CSF1R-tg + EAC were observed anchored to the ellipsoid, positioned to be the primary cells to encounter blood-borne antigens ( Figure 2B). Double staining of spleen sections with Ch68.2 and MHCII, lead to the identification of a ring of Ch68.2 neg MHCII + cells within the ellipsoid that were found to be CSF1R-tg neg ( Figure 2C). Although these cells lack transgene expression, they were found to express the CSF1R protein ( Figure 2D) and also expressed MRC1L-B and CD11c (Figures 2E,F Figure 2). We also observed lower levels of CD11c expression by the EAM compared to the EAC ( Figure 2D). LAMP1 expression was observed throughout the PWP and its expression was found within the EAM, EAC, and the EM cell populations ( Figure 2G).
Overall, our analysis show that a network of EM reside within the inner margins of the ellipsoid that express a number of chicken macrophage markers and the CSF1R protein but not the CSF1R-tg.
The Chicken Spleen Contains Cells of the MPS That Diversely Express the CSF1R-tg,

MHCII MRC1L-B, and CD11c
We next analyzed the phenotype of splenic cells of MPS from the CSF1R-tg chickens using flow cytometry. After gating the cells based on size/granularity and the removal of cell debris, doublets, and dead cells (Supplementary Figure 3), typically five subpopulations were observed based on their CSF1R-tg and MHCII expression patterns ( Figure 3A). Firstly, the CSF1R-tg neg population was gated into MHCII dim (Gate 1) and MHCII hi subpopulations (Gate 2). The CSF1R-tg + subpopulation made up ∼5% of the total live cell population in which three distinct subpopulations were observed based on their CSF1R-tg + and MHCII expression levels (Gates 4-6). Histogram overlays of the CSF1R-tg + subpopulations demonstrated diversity of CSF1Rtg and MHCII expression ( Figure 3B). The median fluorescent intensity (MFI) of CSF1R-tg and MHCII expression levels indicated that cells within Gate 4 had significantly lower levels of CSF1R-tg expression compared to cells in Gates 5 and 6. This subpopulation expressed intermediate to high levels of MHCII ( Figure 3C) and were named CSF1R-tg dim MHCII inter−hi subpopulation. Cells within Gates 5-6 expressed high levels of the CSF1R-tg and differ in their MHCII expression levels. Cells in Gate 5 expressed significantly higher levels of MHCII compared to cells in Gate 6 and hence named the CSF1Rtg hi MHCII hi (Gate 5) and CSF1R-tg hi MHCII inter (Gate 6) subpopulations, respectively. In terms of size and granularity, the CSF1R-tg hi MHCII inter subpopulation was more homogenous than the CSF1R-tg dim MHCII inter−hi and CSF1R-tg hi MHCII hi subpopulations ( Figure 3D).
Using the gating strategy outlined in Figure 3A, we next analyzed the expression of MRC1L-B and CD11c across each subpopulation (Gates 1-6, Figure 3E). We found no MRC1L-B + cells within the CSF1R-tg neg MHCII dim subpopulation and they were largely CD11c + (Figure 3E). Within the CSF1Rtg neg MHCII hi subpopulation a small percentage of MRC1L-B hi CD11c hi cells were observed (Figure 3E), which may represent the CSF1R-tg neg EM observed in the ellipsoid via immunohistology. The CSF1R-tg dim MHCII inter−hi and CSF1R-tg hi MHCII hi subpopulations were very heterogeneous, consisting of MRC1L-B hi CD11c hi , MRC1L-B inter CD11c inter−hi cells while a third subset expressed low level or no MRC1L-B and high levels of CD11c, making up around 30-35% of the cell population in each of these subpopulations, respectively ( Figure 3E and Supplementary Figure 4). The CSF1R-tg hi MHCII inter subpopulation were homogenous in their MRC1L-B and CD11c (88%) expression and lacked the MRC1L-B inter cells observed in the CSF1R-tg dim MHCII inter−hi and CSF1R-tg hi MHCII hi subpopulations ( Figure 3E). Histogram overlays of MRC1L-B and CD11c expression in each CSF1R-tg subpopulations shows the uniform level of expression of these markers in the CSF1R-tg hi MHCII inter subpopulation compared to the CSFR-tg dim MHCII inter−hi and CSF1R-tg hi MHCII hi subpopulations (Supplementary Figure 4). Although numerous extraction approaches were carried out, we did not observed a CSF1R-tg + MHCII neg/low MRC1L-B neg CD11c + phenotype that would represent the EAC and FDC subpopulations in the chicken spleen. This may reflect the low survival rates of these cells post-extraction or they may express low levels of MHCII and are represented in the CSF1R-tg + subpopulations. Even with these limitations, our complex multicolor flow cytometric analysis demonstrates the existence of at least three CSF1R-tg + subpopulations in the chicken spleen that diversely express the CSF1R-tg, MHCII, MRC1L-B, and CD11c.
subpopulation has lower levels of co-stimulatory molecule, CD80 and CD40, and MHCI and CD44 expression, although these differences were only significant for CD40 ( Figure 4A). We next analyzed the expression of non-classical MHC molecule, CD1.1, the chicken monocyte/thrombocyte marker, K1, and Ch74.2 across each CSF1R-tg + subpopulation ( Figure 4B). A subpopulation of CD1.1 hi cells were present in the CSF1Rtg dim and CSF1R-tg hi MHCII hi subpopulations whereas CSF1Rtg hi MHCII inter cells were negative for CD1.1 expression (Figure 4B). The K1 marker (43) was expressed at low levels across all CSF1R-tg + subpopulations with a small population of K1 hi cells within CSF1R-tg dim MHCII inter−hi FIGURE 4 | The CSF1R-tg hi MHCII inter subpopulation express low levels of CD80, CD40, MHCI, CD44, and Ch74.2. The three CSF1R-tg + subpopulations were gated as outlined in Figure 3A and analyzed for a number of chicken mononuclear phagocytic cell markers by flow cytometry. (A) Representative histograms and MFI expression levels of CD80, CD40, MHCI, and CD44, across the three CSF1R-tg + subpopulations; (B) representative histograms of CD1.1, K1, and Ch74.2 expression across the three CSF1R-tg + subpopulations; (C) representative flow plots of chB6, CD3, and CD8α expression across the CSF1R-tg + subpopulations indicates that the CSF1R-tg dim MHCII inter−hi subpopulation express B cell and T cell markers along with macrophage markers. Gates were applied to size/granularity, live, single cells, CSF1R-tg + MHCII + gates using FMO controls. Data is representative of 3-6 independent experiments from 4 to 6 week old chickens. Error bars indicate 95% confidence intervals. Asterisks indicate statistical significance (P < 0.05).
subpopulation. The CSF1R-tg hi MHCII inter cells contained Ch74.2 hi and Ch74.2 neg cell populations whereas the CSF1Rtg dim MHCII inter−hi and CSF1R-tg hi MHCII inter cells had a more uniformed expression pattern of Ch74.2 ( Figure 4B). The chicken B cell marker, chB6, binds to a subset of splenic macrophages (55). We therefore analyzed the expression of chicken T/NK cells and B cell markers within the subpopulations and found that ChB6 label cells within the CSF1R-tg dim subpopulation while a small population expressed CD3 and CD8α, possibly indicating the presence of T cells and NK cells along with macrophages ( Figure 4C) The CSF1R-tg neg MHCII hi MRC1L-B + subpopulation expressed high levels of CD80, CD40, and K1 (Supplementary Figure 5A). We also verified the lack of CD41/61 expression in CSF1R-tg + cells indicating that thrombocyte contamination was not altering the expression levels of the cell markers (Supplementary Figure 5B). Overall, our multicolor flow cytometry analysis indicates that the CSF1R-tg hi MHCII inter subpopulation expressed lower levels of CD80, CD40, MHCI, CD44, CD1.1, and Ch74.2 compared to the CSF1R-tg dim MHCII inter−hi and CSF1R-tg hi MHCII hi subpopulations.

The Phagocytosis Abilities of the CSF1R-tg hi MHCII inter Cells Are Enhanced by LPS
To determine the function of the CSF1R-tg + subpopulations, 5-6 week old CSF1R-eGFP transgenic chickens were intravenously administered 0.5 µm or unlabeled and LPS labeled 1 µm fluorescent beads. Three hours post-IV administration, the phenotype of the splenic phagocytic cells was analyzed by applying the gating strategy outlined in Figure 3A on the bead neg cell population which was subsequently applied to the bead + cell population ( Figure 5A). Due to low number of bead + events in the CSF1R-tg neg MHCII inter subpopulation, these cells were analyzed with the CSF1R-tg neg MHCII hi subpopulation. In the 0.5 µm bead + gate, the cell percentages of the CSF1R-tg neg , CSF1R-tg dim MHCII inter−hi , and CSF1R-tg hi MHCII hi subpopulations were significantly higher compared to the CSF1R-tg hi MHCII inter subpopulation. When the MFI of the 0.5 µm beads was analyzed in each subpopulation, all CSF1R-tg subpopulations had statistically significant higher bead uptake compared to the CSF1R-tg neg subpopulation ( Figure 5B).
Next, the ability of the CSF1R-tg cells to phagocytose 1 µm beads was analyzed. After 3 h, the percentage of CSF1R-tg neg cells within the 1 µm bead + gate was significantly higher compared to the CSF1R-tg dim MHCII inter−hi and CSF1R-tg hi subpopulations ( Figure 5C). Between the CSF1R-tg + subpopulations, the percentage of the CSF1R-tg hi MHCII hi subpopulation was significantly higher compared to the CSF1R-tg dim and CSF1Rtg hi MHCII inter subpopulations. The MFI analysis of the 1 µm beads show that although the CSF1R-tg dim subpopulation was significantly underrepresented in the bead + gate, they had significantly higher bead uptake compared to the CSF1R-tg hi MHCII inter subpopulation. In addition, the CSF1R-tg neg and CSF1R-tg hi MHCII hi subpopulations phagocytosed significantly higher number of beads compared to the CSF1R-tg hi MHCII inter subpopulation ( Figure 5C).
To determine the phagocytosis abilities of cells exposed to bacterial antigens, 1 µm beads were labeled with LPS and phagocytosis was analyzed 3 h post-IV administration. The percentages of both CSF1R-tg neg and CSF1R-tg dim MHCII inter−hi subpopulations in the bead + gate were significantly higher compared to the CSF1R-tg hi MHCII hi and CSF1R-tg hi MHCII inter subpopulations ( Figure 5D). The MFI levels of the LPS labeled 1 µm beads were significantly higher in the CSF1R-tg hi MHCII hi and CSF1R-tg hi MHCII inter subpopulations compared to the CSF1R-tg dim MHCII inter−hi subpopulation. All CSF1R-tg + subpopulations had significantly higher bead uptake compared to the CSF1R-tg neg subpopulation ( Figure 5D). In vivo functional analysis indicates that exposure to LPS enhances the phagocytosis of 1 µm beads by the CSF1R-tg hi MHCII inter subpopulation.

In vivo Antigen Uptake Demonstrates Diversity in the EAM in the PWP
Flow cytometric analysis demonstrated the differential expression of MRC1L-B across each of the CSF1R-tg + subpopulations and the existence of the CSF1R-tg neg MRC1L-B + subpopulation ( Figure 3D). We analyzed the percentage of MRC1L-B expressing cells in the bead + CSF1R-tg + and CSF1Rtg neg subpopulations and used immunohistochemistry to analyze the location of the beads in the spleen 3 h post-administration ( Figure 6). Firstly, we found a majority of 0.5 µm bead + CSF1Rtg + cells consisted of MRC1L-B expressing cells (∼95%) which was significantly higher compared to the percentage of MRC1L-B + in both the 0.5 µm bead + CSF1R-tg dim MHCII inter−hi and CSF1R-tg hi MHCII inter subpopulations (∼75-70%). In the 0.5 µm bead + CSF1R-tg neg subpopulation, we found that ∼10% of cells expressed MRC1L-B. In the unlabeled 1 µm bead + CSF1R-tg hi MHCII hi subpopulation (∼98%), the percentage of MRC1L-B expressing cells was significantly higher compared to the CSF1R-tg dim MHCII inter−hi and CSF1R-tg hi MHCII inter MRC1L-B subpopulations (∼90-85%) while no significance difference between the CSF1R-tg + cells was observed after phagocytosis of LPS labeled 1 µm beads. In the CSF1R-tg neg unlabeled and LPS labeled 1 µm bead + subpopulation, MRC1L-B expressing cells made up ∼50% of the cell population demonstrating the phagocytosis abilities of the EM (Figure 6A).
We next analyzed the location of the bead + cells in the spleen 3 h post-administration. Whole mount microscopic analysis shows that irrespective of bead size or LPS coating, beads were localized to the PWP (Supplementary Figure 6). Additional in depth analysis found that the 0.5 µm beads were located within the CSF1R-tg + Ch68.2 + cells but high concentrations of beads were observed in the CSF1R-tg + Ch68.2 neg cells, which represent the EAM (Figure 6Bi). Uptake of 1 µm beads was observed in the CSF1R-tg + Ch68.2 + EAC but a majority of beads were associated with the CSF1R-tg + Ch68.2 neg EAM (Figure 6Bii). Beads were also found closely associated with chB6 + cells in the ellipsoid representing the CSF1R-tg neg MRCL1-B neg bead + cell phenotype observed in flow cytometry (Supplementary Figure 6). Large deposits of LPS labeled 1 µm beads were also located in the CSF1R-tg + MRC1L-B + EAM compared to the CSF1R-tg + MRC1L-B neg EAC (Figure 6Biii). Due to their location in the PWP the uptake of beads by the different MRC1L-B + CSF1R-tg + subpopulations indicates the variable expression of the CSF1R-tg by EAM. Thus, in vivo antigen uptake demonstrates the existence of EAM subsets that differ in their CSF1R-tg expression and antigen uptake but not in their MRC1L-B expression.
The CSF1R-tg + MRC1L-B neg cDC Can Phagocytose Model Antigens We observed a vast majority of CSF1R-tg + MRC1L-B neg cells in the splenic PALS and very little in the PWP. These cells were found to be a minor population in CSF1R-tg + bead + cells in the in vivo uptake analysis by flow cytometry while immunohistological analysis demonstrated the lack of beads in the PALS 3 h post-administration (Supplementary Figure 6B). (i) few 0.5 µm beads were located in CSF1R-tg + Ch68.2 + EAC cells (yellow arrow) whereas CSF1R-tg + Ch68.2 neg EAM contain numerous beads (yellow arrow); (ii) the 1 µm beads co-localize with the CSF1R-tg + Ch68.2 + EAC (white arrow) but numerous beads were found in the CSF1R-tg + Ch68.2 neg EAM (insert image); (iii) LPS labeled 1 µm beads were located throughout the PWP in CSF1R-tg + MRC1L-B neg EAC (white arrow) while high concentrations of beads were located in the CSF1R-tg + MRC1L-B + EAM (yellow arrows). Data is representative of analysis of three individual chickens.
As previously mentioned, cells with the MHCII, MRC1L-B neg CD11c + phenotype consist of cDC while MRC1L-B expression is associated with cells of the macrophage lineage (23). Therefore, we used in vitro splenic cultures to determine if the low levels of phagocytosis observed in the CSF1R-tg + MRC1L-B neg cells was due to their location in the spleen rather than a functional aspect of these cells. Firstly, we analyzed the MRC1L-B phenotype of splenocytes grown in the presence of CSF1 by flow cytometry (Figure 7A). After 48 h, CSF1R-tg + MRC1L-B + cells made up around 75% of the total cell population while the CSF1R-tg + MRC1L-B neg subpopulation made up ∼4% (Figure 7A and gating strategy Supplementary Figure 7A). We did observe the presence of the CSF1R-tg dim MRC1L-B + subpopulation but not the CSF1Rtg − MRC1L-B + subpopulation in the cultures. As we were interested in analyzing the difference between the MRC1L-B +/neg subpopulations, we focused our analysis on the CSF1Rtg + MRC1L-B + and MRC1L-B neg subpopulations. Firstly, cells were treated with unlabeled and LPS-labeled 1 µm beads at 41 or 4 • C for 3 h and the phenotype of the bead + cells followed by the MFI of beads within the subpopulations were analyzed (Supplementary Figure 7B). After 3 h, the CSF1R-tg + MRC1L-B + subpopulation phagocytosed significantly more beads than the CSF1R-tg + MRC1L-B neg subpopulation, a similar pattern observed in vivo ( Figure 7A). Confocal analysis demonstrated the adherence of both MRC1L-B + and MRC1L-B neg subpopulations in the cell cultures (Figure 7Bi) and the high phagocytic abilities of the MRC1L-B + cells (Figure 7Bii). We also observed the co-localization of MRC1L-B expression with internalized beads indicting its internalization with beads from the cell surface to intracellular vesicles. In addition we observed the co-localization of LAMP1 expression with the phagocytosed beads (Figure 7Bii,iii).
We next analyzed whether there was a difference in the ability of the subpopulations to uptake model antigen for transport to intracellular vesicles, such as lysosomes and phagosomes, using E. coli and Zymosan bioparticles labeled with the pH-sensitive pHrodo red. As a control, cells were treated with the actin-polymerization inhibitor, Cytochalasin D, 10 min prior to exposure to the bioparticles and the data was normalize by subtracting background fluorescence (Supplementary Figure 7). After 3 h, the CSF1Rtg + MRC1L-B + subpopulation engulfed and transported significantly more bioparticles compared to the CSF1R-tg +  MRC1L-B neg subpopulation. However, the data demonstrates the ability of the CSF1R-tg + MRC1L-B neg to phagocytose models antigens better than latex beads and even with their low numbers in the cell culture these cells were highly phagocytic ( Figure 7C). Microcopy analysis demonstrated the inhibition of actin polymerization prevented the uptake of the bioparticles but it did not affect MRC1L-B expression (Figure 7Di-iii). In vitro functional analysis demonstrates the ability of the CSF1R-tg + MRC1L-B neg subpopulation to phagocytose model antigen at higher levels than latex beads indicating a more specific pattern of antigen-uptake compared to splenic macrophages.
The CSF1R-tg hi MHCII hi CD11c hi and CD11c dim Subpopulations Are Enriched With Cells of the cDC and Macrophage Lineages We carried out RNA-seq analysis on FACS sorted splenic subpopulations as shown in Figure 8A. To determine if the level of putative CD11c expression was specific to a cell subset within the CSF1R-tg + subpopulation, we further split the CSF1R-tg hi MHCII hi subpopulation into CD11c hi and CD11c dim subpopulations due its lower phenotypical levels of heterogeneity compared to the CSF1R-tg dim MHCII inter−hi subpopulation ( Figure 8A). The separation of the CSF1R-tg neg MRC1L-B + subpopulation from the CSF1R-tg neg MRC1L-B neg cells lead to insufficient cell numbers for their analysis via RNA-seq and therefore these cells were collected with the CSF1R-tg neg MHCII hi subpopulation. Pairwise comparisons of all subpopulations are shown in Supplementary Table 1.
Using pairwise GSEA, we analyzed whether the transcriptome signatures of murine immune cells were enriched in the sorted splenic subpopulations (23,54). The CSF1R-tg neg MHCII dim and CSF1R-tg neg MHCII hi subpopulations were significantly enriched for T cell and B cell gene fingerprints compared to the other cell populations (Figure 8B). Fingerprints of pDC genes were enriched in the CSF1R-tg neg MHCII hi subpopulation when compared to the CSF1R-tg dim MHCII inter−hi subpopulation but at the lowest level of significance. When comparing the CSF1Rtg dim subpopulation, signatures of different cells lineages were enriched, such as T cells, B cells, cDC1, and moDC. The CSF1R-tg hi MHCII hi CD11c dim subpopulation was enrichment of fingerprints of cDC1 and cDC vs. Mo/Mp (Figure 8B). The CSF1R-tg hi MHCII hi CD11c hi subpopulation expressed   (Supplementary Figure 8).
The CSF1R-tg dim MHCII inter−hi subpopulation expressed higher levels of genes encoding transcripts characteristic of cDC and macrophages (ETSI, ID2, IRF4, IRF8, SpiC, and SLA2). This observation, along with the expression of B cell and T cell related transcripts, indicates that this subpopulation is very heterogeneous. Overall the CSF1R-tg hi MHCII hi CD11c hi and the CSF1R-tg hi MHCII hi CD11c dim subpopulations express higher levels of genes characteristic of cDC and macrophage lineages across warm blooded vertebrate species (18,19,54). Thus, it can be deduced from this data that the CSF1R-tg hi CD11c hi and the CSF1R-tg hi CD11c dim subpopulations must primarily encompass cDC and macrophages, respectively.
We next analyzed the enrichment of cell-lineage related surface receptors involved in a number of cellular processes, such as T cell co-stimulation, antigen presentation and migration. cDC related markers FLT3, CADM1, and XCR1 (10, 13), were significantly enriched in the CSF1R-tg hi MHCII hi CD11c hi subpopulation whereas macrophage-associated genes, CSF1R, CD14, CX3CR1, CCR2, CXCR4, and TNFSF1A were enriched in the CSF1R-tg hi MHCII hi CD11c dim subpopulation ( Figure 9A). The CSF1R-tg hi MHCII hi CD11c dim subpopulation were enriched for migratory and adhesion related transcripts such as, CCR7, CDH2, C3AR1, and CDH2 (Supplementary Figure 9). The integrin ITGAD was enriched in the CSF1R-tg hi MHCII hi CD11c hi and CSF1R-tg hi MHCII inter subpopulations whereas FIGURE 9 | The CSF1R-tg hi MHCII hi and CSF1R-tg hi MHCII inter cells are enriched with transcripts related to cDC, macrophage and red pulp macrophage function. FACS sorted CSF1R-tg + subpopulations were analyzed by RNA-seq for lineage-specific transcription factors and genes related to function. The CSF1R-tg hi MHCII hi CD11c hi subpopulation was enriched for cDC transcription factors (EGR1, IRF8, ID2, NR4A3, and ZBTB46) and the CSF1R-tg hi MHCII hi CD11c dim subpopulation was enriched for macrophage signature genes (MAFB, MITF, SPI1, TFEC, ZEB2). The CSF1R-tg dim and CSF1R-tg hi MHCII inter subpopulations were enriched for transcripts associated with red pulp macrophages (SpiC) and regulation of CSF1R expression (ETS1, SLA2). The CSF1R-tg hi MHCII hi CD11c hi cells were enriched with cells expressing gene related to the cDC lineage (CADM1, FLT3, XCR1) and CSF1R-tg hi MHCII hi CD11c dim subpopulation were enriched macrophage related genes (CCR2, CD14, CX3CR1, CXCR4, TNFSF1A). The CSF1R-tg hi MHCII inter subpopulation was enriched for cells expressing red pulp macrophage-related genes (C3, LRG1, and MARCO). The CSF1R-tg dim subpopulation were enrich with a mixture of different cell lineages expressing genes related to the cDC and macrophage lineages. Heatmaps of log10 transformed normalized expression of genes. Heatmaps coloring is based on counts per million with red indicating high counts and blue indicating low counts.
ITGAV transcript was enriched in CSF1R-tg hi MHCII hi CD11c dim subpopulation. Co-stimulatory molecules, CD80 and CD86 transcripts were significantly expressed by cells within the CSF1R-tg hi MHCII hi CD11c hi and CD11c dim subpopulations. DPP4 (CD26) is highly expressed in the CSF1R-tg hi MHCII hi CD11c dim subpopulation (not CD11c hi ), which is not consistent with DPP4 being expressed to high levels selectively on cDC1 in humans, mice, and sheep (19). We next analyzed the expression of Toll-like receptors (TLR) transcripts across the CSF1Rtg + subpopulations and found that the CSF1R-tg hi MHCII hi CD11c dim subpopulation was enriched for cells expressing TLR4, TLR2A and co-receptor, CD14. The CSF1R-tg hi MHCII inter subpopulation expressed high levels of C-type lectin receptors, FCER2 (CLEC17AL-A), and CLEC17A (CLEC17AL-B) (68). The apoptotic cell recognition receptor, TIMD4 (69) transcript expression was enriched in both the CSF1R-tg hi MHCII hi CD11c hi and CD11c dim subpopulations ( Figure 9A).
The CSF1R-tg hi MHCII hi CD11c hi subpopulation significantly expressed genes involved in cross-presentation and MHCII-associated genes [ACVRL1 (70), CD74, LAMP5 (71), and WDFY4 (72)]. Key macrophage related genes, MERTK, VCAM1, and genes related to red pulp and alternatively activated macrophages, C1QA, C1QB, and PPARG (73), and suppressor of cytokine signaling molecules (SOCS2, SOCS4, and SOCS6) were highly expressed in the CSF1R-tg hi MHCII hi CD11c dim subpopulation compared to the CD11c hi subpopulation. An array of transcripts involved in antigen processing [CD74, CTSS, LGMN, LAMP1, LAMP3, LAMP5, MERTK, and PSAP (71)] were also strongly expressed by the CSF1R-tg hi MHCII hi CD11c dim subpopulation ( Figure 9B). The CSF1R-tg hi MHCII inter subpopulation expressed the highest levels of genes characteristic of the red pulp macrophage lineage, MARCO, C3, and LRG1 (73). Chicken heterophils express the CSF1R-tg at low levels (33,34). Therefore, we analyzed the expression of core heterophil gene transcripts across the CSF1R-tg + subpopulations (74) and found no significant enrichment of heterophil related gene transcripts in any of the CSF1R-tg + subpopulations (Supplementary Table 2). The CSF1R-tg dim MHCII inter−hi subpopulation were enriched for cathepsins family members (CTSC, CTSH, and CTSZ) but no clear definition of what enriched macrophage cell type this subpopulation represented was identified. To analyze whether the CSF1R-tg dim subpopulation were true phagocytic cells and not an artifact of transgene expression or cells undergoing physiological stress, we carried out an in vivo liposome depletion study, where phagocytic cells were targeted for depletion by apoptosis induction after the uptake of clondronate liposomes (75). Three days after a single IV-administration of liposomes, we observed a severe depletion of the CSF1R-tg dim MRC1L-B + subpopulation in chickens treated with the clondronate liposomes compared to those treated with PBS liposomes indicating that this subpopulation contains phagocytic cells belonging to the MPS (Supplementary Figure 10). Although our sorting approach would not have led to pure cell populations, we were able to determine the enrichment of cDC in the CSF1Rtg hi MHCII hi CD11c hi subpopulations as previously described (23). Furthermore, we demonstrated the existence of at least three subsets of resident tissue macrophages in the chicken spleen. Based on the RNA-seq and immunohistological analysis, the CSF1R-tg hi MHCII hi CD11c low subpopulation are enriched with macrophages from the PWP and PALS while the CSF1R-tg hi MHCII inter cells are enriched with red pulp macrophages. Although we cannot demonstrate which splenic cell they represent, the CSF1R-tg dim MHCII inter−hi MRC1L-B + are a subset of the chicken splenic MPS.

DISCUSSION
Mononuclear phagocytes reside within distinct niches of the spleen where they partake in the number of cellular processes from homeostasis, removal of aged, or damaged erythrocytes to killing of pathogens and induction of adaptive immune responses. The development, proliferation, and survival of monocytes, macrophages, and cDC are controlled by CSF1R [reviewed by (76)]. Lineage specific expression of CSF1R has allowed for the detailed study of the MPS using the CSF1R-reporter transgenic chickens. Since their generation, transcriptome analysis of bursal macrophages, liver macrophages, and to some extent cDC, along with subsets of blood monocytes have been described (22). Hu et al., observed the lack of CSF1R-tg expression in chicken Kupffer cells that express similar CSF1R mRNA levels as tissue macrophages. In certain cell types in mice, CSF1R expression is under controlled of a 150 bp distal promoter not included in the generation of the CSF1R-reporter chickens and thus is inferred as the reason for the lack of CSF1R-tg expression in chicken Kupffer cells (22,77). The splenic MPS cells has been described based on their MHCII, MRC1L-B, Ch74.2 expression, and acid-phosphatase staining and to a certain extent using the CSF1R-reporter transgenic chickens (24,25,29,33,(78)(79)(80). However, the separation of PALS and PWP macrophages has not been described while comparative gene profiling has identified the existence of the XCR1 cDC in the chicken spleen but not their location (23). Using the CSF1Rreporter transgenic chickens we carried out a comprehensive analysis of the splenic MPS.
In the chicken spleen, interdigitating cells (also known as interdigitating DC) and PALS-associated macrophages express MHCII and MRC1L-B while MHCII hi CD11c + MRC1L-B neg cells are enriched for cDC (28,29). As the CD11c, 8F2 antibody is not commercially available, progress in the field of chicken DC research has been limited. Using the CSF1R-reporter transgenic chickens, we observed a CSF1R-tg + subpopulation lacking MRC1L-B expression in the PALS (Figure 1) demonstrating that these chicken models can be used to study cDC without requiring specific cDC antibodies such as FLT3. In the PWP, the equivalent of the mammalian MZ, two subpopulations have previously been shown to express the CSF1R-tg (22,33). On the outer margins of the PWP, intermingled with B cells are the CSF1R-tg + EAM while the CSF1R-tg + EAC are anchored into the ellipsoid (81). In contrast to EAM, EAC lack MRC1L-B and MHCII expression and specifically stain for Ch68.2, a chicken monoclonal which binds to an unknown epitope (25,28). In this study, we show that EAC express higher levels of CD11c compared to the EAM [ Figure 1; (23,42)]. Due to their location in the PWP, the EAM resemble the mammalian MZM which function to trap and display antibody-coated antigens from the blood for transfer to the MZ B cells which is subsequently deposit to FDC in the follicles (82). However, in chickens, surface antigen display has been observed in EAC but not EAM (28,83). The high level of CD11c expression by the EAC may correlate with their ability to detach from the ellipsoid upon antigen exposure and their subsequent migration to the PALS to develop into FDC (28). We found intracellular punctate expression of MHCII by EAC but could not fully determine whether these cells express surface MHCII (Figure 2). These cells were found to express LAMP1, possibly demonstrating their ability to digest substances delivered by endocytosis and phagocytosis (84). In the red pulp, we found that the CSF1R-tg + cells express high levels of MRC1L-B but not all stained for the chicken macrophage marker, Ch74.2 as previously described (24). A previous study has described the presence of CD45 + Ch68.2 neg cells within the inner regions of the ellipsoid (28). We observed cells lacking CSF1R-tg expression in a similar location in the ellipsoid and found they express high levels of MHCII, MRC1L-B, CD11c, LAMP1, and CSF1R (22). We have called this cell population "ellipsoid macrophages" (EM). The EM may represent incoming monocytes or ellipsoid resident macrophages and requires further study in the future. Through flow cytometry we show that CSF1R-tg neg MRC1L-B + cells express high levels of CD80, CD40, and the chicken monocyte/thrombocyte marker, K1 (Supplementary Figure 4).
When analyzing the phenotype of cells via flow cytometry we found that splenic MPS diversely express the CSF1R-tg and MHCII. One subpopulation could be distinguished based on their CSF1R-tg dim MHCII inter−hi phenotype while two subpopulations expressed similar levels of the CSF1R-tg and were separated based on MHCII hi and MHCII inter levels (Figure 3). We found that the CSF1R-tg dim and CSF1R-tg hi MHCII hi subpopulations were very heterogenic in their MRC1L-B and CD11c expression levels while the CSF1R-tg hi MHCII inter subpopulation expressed these markers more uniformly (Figure 3). We did not identify a cell phenotype resembling EAC or FDC which have the CSF1R-tg + MHCII neg/low MRC1L-B neg CD11c + phenotype (28,33). This may be due to their lack of survival during cell isolation or they express low or dynamic levels of MHCII. Although there is an EACspecific surface marker, Ch68.2, this antibody fails to provide consistent staining in flow cytometry (data not shown). Further phenotypical analysis demonstrated low expression levels of CD80, CD40, MHCI, and CD44 in the CSF1R-tg hi MHCII inter subpopulation compared to the CSF1R-tg dim and CSF1R-tg hi MHCII hi subpopulations (Figure 4). The CSF1R-tg hi MHCII inter subpopulation express low levels of Ch74.2 which was also observed for red pulp macrophages via immunohistology (Figure 4). Mammalian red pulp macrophages express low level of co-stimulatory molecules and MHCII due to their reduced role in T cell antigen presentation (2). These observations would suggest that the CSF1R-tg hi MHCII inter subpopulation may consist of red pulp macrophages.
In vivo functional analysis demonstrated that the CSF1Rtg hi MHCII hi subpopulation phagocytosed significantly more beads compared to the CSF1R-tg dim and CSF1R-tg hi MHCII inter subpopulations while labeling beads with LPS increased the phagocytosis of the CSF1R-tg hi MHCII inter subpopulation ( Figure 5). LPS has been shown to enhance antigen uptake by chicken red pulp macrophages (24). In a recent study, ex vivo isolated splenic MRC1L-B hi MHCII lo cells showed higher phagocytosis of dead cells than MRC1L-B lo MHCII hi cells with the latter cell population suggested to represent red pulp macrophages. Another study has also demonstrated the expression of the apoptotic cell receptor, Tim4, by EAM which we show express high levels of MRC1L-B [ Figure 1; (22,85)]. Therefore, the previous characterized, MRC1L-B hi MHCII lo cells may represent a heterogeneous population of red pulp macrophages and EAM. In vivo functional analysis indicated that the majority of the bead + CSF1R-tg + subpopulations expressed MRC1L-B. When we analyzed location of the beads in the spleen we observed large deposits in the EAM while few beads were observed associated with EAC (Figure 6). Both flow cytometric and immunohistological analysis suggests that a subsets of CSF1R-tg dim MHCII inter−hi MRC1L-B + and CSF1Rtg hi MHCII hi MRC1L-B + subpopulations are located in the PWP demonstrating that EAM differ in their CSF1R-tg expression levels but not MRC1L-B. For the bead + CSF1R-tg hi MHCII inter subpopulation, we speculate that these cells are either a subset of EAM or red pulp macrophages residing in close proximity to the EAM, hence their ability to capture antigens.
We observed very little in vivo bead uptake by the CSF1R-tg + MHCII + MRC1L-B neg cells which may be due to their higher numbers in the PALS and not the PWP where antigen capture primarily takes place. In vitro functional analysis demonstrated that the CSF1R-tg + MRC1L-B neg subpopulation were capable of phagocytosis (Figure 7). Mammalian macrophages express large intracellular pools of MRC1 within their early endosomes, which undergo continual rapid recycling to the cell surface to endocytose bound antigens (86). Due the high recycling nature of MRC1 in mammals and the fixture of adherent cells with PFA in this study, we only observed intracellular punctate expression of MRC1L-B (Figure 7). However, we found the co-localization of MRC1L-B and LAMP1 expression with the internalized beads in vitro which for the first time demonstrates the ability of MRC1L-B to traffic with antigens to lysosomes and endosomes similar to mammalian MRC1 (86). Our in vivo functional analysis provided the first demonstration of antigen uptake by the CSF1Rtg + MRC1L-B neg cDC subpopulation in the spleen further demonstrating the ability to study cDC ex vivo using the CSF1Rtransgenic chickens. In addition, future studies using the CSF1Rtransgenic chickens and pathogens would allow for the in vivo visualization of cDC-T cell interactions in the chicken spleen.
Transcriptomic analysis has shown the conservation of MPS lineage specific gene signatures throughout distantly related mammals (18,19,23,54). A previous comparative analysis of chicken splenic cells demonstrated the existence of cDC which expressed the MHCII + MRC1L-B neg CD11c + phenotype by their enrichment for XCR1, FLT3, and key cDC lineage gene transcripts. The authors also described the enrichment of macrophage related genes in MRC1L-B + cells (23). We expanded on this study by separating the CSF1R-tg hi MHCII hi subpopulation into CD11c hi and CD11c dim subsets and analyzing two additional subsets of splenic MPS via RNA-seq analysis (Figure 8). It should be noted that MRC1L-B + cells were not excluded from the CSF1R-tg + subpopulations as performed in the Vu Manh et al., study (23). Primarily we analyzed the presence of murine splenic immune cell gene signatures using GSEA across all the cell subpopulations (Figure 8). We initially found an enrichment of the cDC1 gene signatures in the CSF1Rtg hi MHCII hi CD11c hi subpopulation and macrophage gene signatures in the CSF1R-tg hi MHCII hi CD11c dim subpopulation. We found that the CSF1R-tg dim MHCII inter−hi subpopulations contained cells expressing B cell, T cell, cDC, and macrophage gene signatures which correlates with the immuno-phenotyping analysis (Figures 3, 4). Upon closer analysis of the enrichment of genes associated with cell lineage development and function, we found that the CSF1R-tg hi MHCII hi CD11c hi subpopulation were enriched for genes characteristic of the cDC lineage (IRF8, ID2, ZEB2, FLT3, XCR1, and CAMD1) although they contain ∼30% of MRC1L-B neg−hi cells which may suggest that a significant fraction of chicken cDC express MRC1L-B contrary what was previously reported or this population contains non-cDCs that express MRC1L-B (23). This subpopulation contained cells enriched for genes also involved in cross-presentation and MHCII processing [WDFY4, ACVRL1, CD74, and LAMP5; (72)]. However, in contrast to human, mouse, and sheep cDC, we found strong expression of DPP4 in chicken cells corresponding to macrophages, CSF1R-tg hi MHCII hi CD11c dim cells, discussed later, and not chicken cDC (19,87). This suggests that chicken macrophages may have a role in T cell-stimulation (88). The hallmark of cDC is their ability to present to and activate T cells, however, numerous attempts to demonstrate DC-T cell interaction through MLR were to no avail, with numerous cells of the MPS engulfing/phagocytosing labeled CD4 + T cells (data not shown).
The CSF1R-tg hi MHCII hi CD11c dim subpopulation were enriched for genes characteristic of macrophages [CCR2, CD14, CD36, CSF1R, CXCR4, CX3CR1, HIF1A, MAFB, MITF, STAT1, SPI1, TLR4, and TNFSF1A; (22,76)]. This subpopulation was also enriched for M2 anti-inflammatory cell phenotype and red pulp macrophage related gene transcripts (C1QA C1QB, and PPARG) which are associated with tissue repair and homeostasis (89). Based on their CD11c dim phenotype these macrophages likely represent the EAM and PAL-associated macrophages in the spleen. Distinguishing the difference between the EAM and PAL-associated macrophages requires further investigation. The CSF1R-tg hi MHCII inter subpopulation was enriched for genes associated with red pulp macrophages [HEPHL1, LRP1, SpiC, and TF; (67)]. Similar to the CSF1R-tg hi MHCII hi subpopulation, we cannot rule out the possibility of the CSF1R-tg hi MHCII inter subpopulation containing more than one macrophage cell lineage, such as PALS-associated macrophages. The CSF1Rtg dim subpopulation were phenotypically heterogeneous and may represent cells undergoing different levels of development or activation in the spleen and require further analysis. In the murine spleen, pDC make up around 0.5% of cells (90). In our analysis, we found low levels of pDC gene signatures in the CSF1R-tg neg MHCII hi subpopulation which may be  Figure 3E). However, more analysis is required to confirm the existence of this cell lineage in chickens.
As of yet, no definitive study has identified different chicken cDC subsets as observed in mammals possible indicating the existence of just one cDC subset in chickens. We did observed the cDC2 related transcription factor, IRF4, in the CSF1R-tg dim subpopulation but this may be due to the mixture of B cell, cDC, and macrophage lineages in this subset (91)(92)(93). Future single cell RNA-seq analysis of splenocytes has the potential to further contribute to delineating the cells of the chicken splenic MPS.
In conclusion, we have shown that the CSF1R-reporter transgenic chickens can be used to distinguish different cell subsets of the MPS in the spleen. We have demonstrated that the PALS-associated CSF1R-tg + cells do not uniformly express MRC1L-B helping to differentiate PALS-associated cDC (MRC1L-B neg ) from macrophages (MRC1L-B + ) in the chicken spleen. In the PWP, a novel CSF1R-tg neg cell population, called ellipsoid macrophages were also characterized. We demonstrated that there are at least three distinctive cell subpopulations expressing diverse levels of CSF1R-tg and MHCII in the chicken spleen. In addition, we demonstrated that the CSF1Rtg hi MHCII inter subpopulation, phenotypically, and functionally resemble red pulp macrophages and identified functional diversity in EAM. For the first time we demonstrated the functionality of CSF1R-tg + MRC1L-B neg cDC in vitro. The phenotype and function of chicken splenic MPS are provided as an overview in Table 2. This is the first description of the function and transcriptome of a number of MPS cell lineages in chicken spleen that will contribute to future studies to dissect their roles in antigen presentation, immune responses, and resolve the EAC and FDC cell lineages and functions.

DATA AVAILABILITY STATEMENT
The datasets presented in this study can be found in online repositories. The names of the repository/repositories and accession number(s) can be found in the article/Materials and Methods.

ETHICS STATEMENT
The animal study was conducted under Home Office project license PE263A4FA according to the requirements of the Animal (Scientific Procedures) Act 1986, with the approval of the Roslin Institute's ethical review committees. Birds were humanely culled in accordance with Schedule 1 of the Animals (Scientific Procedures) Act 1986.

AUTHOR CONTRIBUTIONS
KS, LV, and AB conceptualized the study. KS, DB, and LV performed or assisted with the bird experiments. KM performed the transcriptomic data analysis. KS performed all experiments and analysis. LV, PK, and HS secured the funding. KS wrote the manuscript. All authors contributed to read and approved the final manuscript.

ACKNOWLEDGMENTS
We would like to dedicate this paper to our late colleague and friend PK. We express our gratitude to Dr. Sonja Härtle for providing the CD11c and K1 antibodies. We would like to thank Graeme Robertson and Bob Fleming for support with the cell sorting and confocal imaging. We wish to thank the animal caretakers of the National Avian Research Facility for the supply of birds and support during the studies. Sequencing work (library preparation, sequencing, and data generation and initial analysis) was carried out by Edinburgh Genomics at the University of Edinburgh.