Discrimination of distinct chicken M cell subsets based on CSF1R expression

In mammals, a subset of follicle-associated epithelial (FAE) cells, known as M cells, conduct the transcytosis of antigens across the epithelium into the underlying lymphoid tissues. We previously revealed that M cells in the FAE of the chicken lung, bursa of Fabricius (bursa), and caecum based on the expression of CSF1R. Here, we applied RNA-seq analysis on highly enriched CSF1R-expressing bursal M cells to investigate their transcriptome and identify novel chicken M cell-associated genes. Our data show that, like mammalian M cells, those in the FAE of the chicken bursa also express SOX8, MARCKSL1, TNFAIP2 and PRNP. Immunohistochemical analysis also confirmed the expression of SOX8 in CSF1R-expressing cells in the lung, bursa, and caecum. However, we found that many other mammalian M cell-associated genes such as SPIB and GP2 were not expressed by chicken M cells or represented in the chicken genome. Instead, we show bursal M cells express high levels of related genes such as SPI1. Whereas our data show that bursal M cells expressed CSF1R-highly, the M cells in the small intestine lacked CSF1R and both expressed SOX8. This study offers insights into the transcriptome of chicken M cells, revealing the expression of CSF1R in M cells is tissue-specific.


Phenotypic characterisation of CSF1R-expressing bursal M cells by flow cytometry
CSF1R-transgene expression distinguishes the bursal FAE from IFE 24 .In CSF1R-reporter transgenic chickens, M cells in bursal FAE express CSF1R-transgene and lack the expression of CD45 24 .We utilised the CSFR-reporter transgenic chickens to examine the phenotype of bursal CSF1R-trangene POS FAE cells (M cells) and CSF1Rtrangene NEG IFE cells.After gating the bursal cells based on size/granularity and the removal of doublets, and dead cells (Gates P1-2), two subpopulations were observed in the live cell channel (Fig. 1A).Cells within Gate P3 exhibited auto-fluorescence while cells in Gate P4 were low/negative in the violet channel (Fig. 1A).Cells expressing CD45 were excluded from the cells within Gates 3 and 4 to remove leukocytes from the analysis (Fig. 1B).The cells in Gate P3, consisted of 16% of CD45 NEG cells compared to 4% in Gate P4.Next, we examined CSF1R-eGFP expression within the CD45 NEG subpopulations (Gates P5 and P6), and found 32% of the cells in Gate P5 were CSF1R-eGFP POS compared to 8% in Gate P6 (Fig. 1B).Based on the CD45 NEG CSF1R-eGFP POS phenotype these cells represent chicken M cells.A vast majority of cells in the CD45 NEG gates (Gates P5 and 6) were also negative for CSF1R-eGFP expression and were considered to represent IFE cells.This gating strategy confirmed the ability to isolate populations of live CD45 NEG CSF1R-eGFP POS M cells and CD45 NEG CSF1R-eGFP NEG IFE cells from the bursa.
Further phenotypic analysis of CD45 NEG CSF1R-eGFP POS M cells demonstrated that they lack surface expression of MHC II and chicken monocyte/macrophage markers including MRC1L-B, TIM4, and putative CD11(Fig.1C).CD44 has been reported as a putative pan-M cell marker 21,24 .CD44 staining of bursal M cells demonstrated heterogenous expression of this marker with two subpopulations, CD44 NEG and CD44 POS .The absence of immune cell markers confirmed that these populations of bursal CD45 NEG CSF1R-eGFP POS cells and CD45 NEG CSF1R-eGFP NEG cells were highly enriched in bursal M cells and IFE cells, respectively.

Stable transcriptional profile of bursal M cells during the post-hatch period
At the time of hatch, a bursal follicle segregates into defined cortical and medullary regions while the FAE forms epithelial tufts that transport particulate antigens from the lumen into the bursal follicle 26 .To identify the gene expression profile and assess whether the post-hatch development of the bursa and exposure to gut-derived molecules affect the transcriptome of M cells and IFE cells, we isolated these cells from 1-and 4-week-old chickens.Using a similar flow cytometric gating strategy (Fig. 1A), CD45 NEG CSF1R-eGFP POS M cells and CD45 NEG CSF1R-eGFP NEG IFE cells were FAC sorted and transcriptome analysis performed by mRNA-seq (Fig. 2A).PC analysis of the global transcriptional profiles of each mRNA-seq dataset identified distinct clusters of M cells and IFE cells where the cell-type difference accounted for 46% and age 10% of the variance in the data (Fig. 2B; PC1, PC2 respectively).
We next compared the number of differentially expressed genes (DEGs) between the age-matched M cells and IFE cells, and the influence of age within each cell population.Analysis of cells from 1-week-old birds revealed 2260 upregulated and 3246 downregulated DEGs in M cells compared to IFE cells (Fig. 2C), whereas analysis of cells from 4-week-old birds revealed 2047 upregulated and 2335 downregulated DEGs in M cells compared to IFE cells (Fig. 2C).Comparison of the transcriptomes between M cells enriched from 1-and 4-week-old chickens found no DEGs that reached the cut-off threshold (FDR < 0.05, log 2 FC > 1 or < − 1, Fig. 2C).This implied that 1-and 4-week-old M cells share a similar transcriptional profile.In contrast, a similar comparison in the global transcriptomes of IFE cells identified 290 upregulated and 60 downregulated DEGs in 1-week-old IFE cells compared to IFE cells from 4-week-old birds.Together, this analysis suggests that the transcriptional profiles of bursal M cells are very stable during the first four weeks of post-hatch development, whereas changes in the transcriptional profiles of IFE may suggest a potential maturation of the cells.

Transcriptional profile of bursal M cells differs from murine M cells
To gain insights into the fundamental biology of bursal M cells, we searched upregulated DEGs for key regulatory genes and pathways involved in M cell development and function.With the stability between 1 and 4-week-old M cell transcriptome, we utilised the 1-week-old DEG dataset for further analysis.This data set contains 2260 upregulated and 3246 downregulated DEGs.Functional enrichment analysis of the downregulated DEGs of bursal M cells indicated that the genes upregulated in IFE are involved in pathways related to importing nutrients and their digestion, cell division and DNA replication in line with the IFE being composed of columnar epithelial cells and mucus-producing cells (data not shown).Here we focused on the upregulated DEGs in chicken bursal M cells and compared the genes with a previous study that identified a set of 28 M cell-associated genes conserved across murine PP and chicken FAE cells 21 .The analysis of the 2260 upregulated DEGs identified in 1-week-old M cells compared to IFE cells found that 13 of the 28 genes, originally identified by Nakato et al. 21, were represented in our dataset (Supplementary Table 1).However, of the missing genes, such as CLU and ANXA10, have been implicated in mammalian and chicken M cell biology 21,24,27 while genes not present in our DEG dataset, such as FLI1, CXCR4 and VIM, implied that our enrichment approach excluded immune cells and fibroblasts 28,29 .
A similar approach was applied to other murine M cell-associated genes that were missing from the current chicken genome (GRCg6a) to identify closely-related family members amongst our list of 2260 chicken M-cell upregulated DEGs (Supplementary File 1).This suggested that the genes encoding two annexin family members ANXA5 and ANXA10, were not represented in the chicken M cell DEGs although ANXA10 has been reported to be expressed in chicken and murine M cells 21,24,35 .However, the genes encoding two other annexin family www.nature.com/scientificreports/members, ANXA4 and ANXA11, were upregulated.β1 integrin (encoded by Itgb1) is an adhesion molecule that is expressed at the basolateral surface of small intestinal enterocytes 38 and relocated to the apical surface of M cells during Yersinia enterocolitica infection and interacts with invasin leading to bacterial uptake through micropinocytosis 16 .Itgb1 was not present in our dataset but fellow integrin family members integrin β8 (ITGB8) and integrin β2 (ITGB2) were present (Table 1).A similar observation was made for the tight junction protein CLD4 39 which was not present, but a close family member CLDN3 was detected.Aif1, is not currently annotated in the chicken genome but a similar gene, allograft inflammatory factor 1 like (AIF1L) 40 , was present in the chicken M cell dataset and its expression in M cells of the bursal FAE was confirmed by immunohistochemical staining (Fig. 3A).Both GP2 and Uromodulin (UMOD) are expressed by mature mouse M cells and considered as uptake receptors for certain bacterial species 15,41 .At the time of writing (January 2024), the gene encoding GP2 was not annotated in the GRCg6a build of the chicken genome and UMOD was not present in the dataset.The chemokine genes, chemokine (C-C motif) ligand 9 (Ccl9), Ccl20 and Cxcl11, are highly expressed in the FAE of mice, working to recruit leukocytes 34,42,43 .While the expression of CCL20 was downregulated in bursal M cells, CCL9 and CXCL11 are not annotated in chicken genome.Instead, CCL19 was the highest upregulated chemokine in our dataset (Supplementary File 1-2) which may function to attract CCR7 expressing macrophages and dendritic cells (DC) 44 .The expression levels of identified M cell-associated genes (Supplementary Table 1) were similar in bursal M cells derived from 1-and 4-week-old chickens (Fig. 3B).To further explore conserved M cell-associated genes in mouse and chicken, we compared the upregulated DEGs identified from single cell RNA-seq data of mouse intestinal M cells 45 with those of 1-week-old bursal M cells, resulting in the identification of 9 conserved genes (Supplementary Table 2).Taken together, we have identified the conserved expression of some murine M-cell related genes in chicken M cells, including expression of the key transcriptional regulator SOX8 14 .While some mouse M cell-associated genes, such as those encoding GP2 were not represented in the chicken genome, we identified a set of close family members of other mouse M cell associated-genes that were also highly expressed in chicken M cells.However, this analysis also identified many genes such as the transcription factor SPI1 that were highly expressed in chicken M cells, but not mouse M cells.

Identification of candidate transcription factor encoding genes involved in bursal M cell development
We next searched for the presence of other transcriptional regulators that may provide insights into the differentiation and functional maturation of these cells.Filtering this list using the 'DNA-binding transcription factor activity' GO term (GO:0003700) identified transcription factors, belonging to the following families; winged helix/forkhead (ELF1, ELF5, ETV3, IRF7, SPI1); HMG box (SOX8, SOX13, SOX21); DNA-binding (BBX, ID2, STAT3); C2H2 zinc finger (KLF5, KLF9, ZNF292, ZBTB33); and basic helix-loop-helix (ASCL2, HIF1A, OLIG2) (Supplementary File 3).Analysis of top 20 highly expressed transcription factors in M cells compared to IFE cells revealed two transcription factors, SOX8 (conserved in murine M cells 14 ), and SPI1 (associated with chicken M cells), highly enriched in bursal M cells (Fig. 4A).Immunostaining using an anti-human SOX8 antibody confirmed the specific expression of SOX8 protein in chicken bursal M cells (Fig. 4B).Although CTNNB1 was the second highest gene in chicken M cells, it was also highly expressed in IFE cells and thus unlikely to have a specific role in chicken M-cell development.Since SPI1 expression was unique to chicken M cells, we used gProfiler to determine whether the chicken M cell genes were enriched in binding site motif for this transcription factor (Supplementary File 4).This analysis suggested that 472 (~ 20%) genes have promoter regions with SPI1 binding motifs, indicating an important role of SPI1 in M cell biology.RANKL-RANK signal through the NFκB pathway which is essential for M cell differentiation in mice 30,46,47 .NFκB binding site motifs were present in the promoter regions of 1293 (57%) DEGs in chicken M cells (Supplementary File 4), indicating the conserved requirement of this pathway for M cell differentiation in mice and chickens.This analysis supports the hypothesis that the chicken CSF1R POS M cells have a distinct transcriptional profile compared to murine M cells.
Table 1.Assessment of the transcriptome of sorted CSF1R-transgene POS bursal M cells for the presence (green) and absence (red) of conserved murine M cell-associated genes and alternative related gene family members for non-annotated genes (N/A) in GRCg6a chicken genome and genes not upregulated in chicken M cells.

Expression of chemokine/cytokine encoding genes by bursal M cells
To further identify candidate genes involved in M cell development and function, our list of 2260 genes were filtered using the 'cytokine-mediated signalling pathway' GO term (GO:0019221) and 'transmembrane signal receptor' Panther protein classification (Supplementary File 5).In mouse, M cell differentiation is critically dependent on stimulation from RANKL produced by subepithelial stromal cells beneath the FAE 10 .This cytokine signals via its receptor, RANK, which is expressed throughout the gut epithelium 10 .Here, higher levels of the expression of TNFRSF11A (encoding RANK) were detected in M cells compared to IFE (Fig. 5A), consistent with studies in mice 9 .In addition to CSF1R, we also found an enrichment of interleukin (IL) receptor genes in bursal M cells, such as IL17REL, IL4R, IL22RA2, and IL10RA at both ages (Fig. 5A).The chemokine genes, Ccl9, Ccl20 and Cxcl11 are highly expressed in the FAE of mice, working to recruit leukocytes.None of these genes were upregulated in chicken M cells (Supplementary File 1).The most highly expressed chemokine in our dataset was CCL19 (Supplementary File 2) which may function to attract CCR7 expressing macrophages and DC 44 .Collectively, these data imply that chicken M cells express a distinct array of cytokine receptors, suggesting a potential role of their signalling in the development or function of M cells.

Chicken bursal M cells are enriched with genes associated with antigen sampling
M cells actively transport luminal antigens across the epithelial barrier via transepithelial pathway known as transcytosis.Functional enrichment (GO terms) and KEGG signalling pathways analysis were performed on the upregulated gene dataset using gProfiler (Supplementary File 4).Murine M cells express GP2, UMOD, and PRNP on their apical surface to transport the luminal antigens across the epithelial barrier 15,21,32 .Whereas PRNP was enriched in chicken M cells implying a similar role in the uptake of certain microorganisms or antigens into these cells, UMOD was not present and the gene encoding GP2 is not annotated in the chicken genome.
To identify potential genes associated with transcytosis by chicken M cells, we filtered the upregulated genes of bursal M cells for endocytic, G-protein-coupled, and surface receptors genes using gProfiler and UniProt.A full list of 376 genes is provided in Supplementary File 6 and the top 20 genes are shown in Fig. 5B.The genes were largely categorised in following groups; transmembrane signal receptor (FGFR1, IGF1R, PLXNB2, SDC4), scaffold/adaptor protein (CD151, CD63, CD81, TSPAN4), membrane traffic protein (CAV2, DNM1, LAMP1-2, LY75, PLA2R1), integrin (ITGA6, ITGAV, ITGB2, ITGB8), G-protein (ADIPOR2, GNAI2, GNB1, GNB4), non-motor actin binding protein (CTNNA1, CTNNB1, CTTN), and transporter (ATP11C, ATP2B4, ATP6V1H, SLC9A3).ITGAV was highly upregulated in M cells compared to IFE cells.ITGAV forms a heterodimer with ITGB8, which also found highly expressed in chicken M cells (Fig. 3A).LY75 (DEC-205 or CD205) is a C-type lectin found to be enriched in chicken M cells.LY75 is expressed on murine DC, B and T cells, and thymic epithelial cells and has endocytic capacity 48,49 .In addition, genes involved in ion channel regulation (STOM) 50 , immune cell attraction (ATRN) 51 , vesicular trafficking (MAL) 52 , and transmembrane glycoprotein, NRP2 were expressed at substantially higher levels in M cells compared to IFE cells and may indicate their role in subcellar transport of molecules in M cells.In contrast to murine M cells, which express lower levels of LAMP1 compared to FAE enterocytes 53 , LAMP1 was highly expressed in chicken bursal M cells compared to IFE cells which was also observed at a protein level 24 .Members of Annexin family have been implicated in antigen transportation and cell morphology 54 .Studies suggest that ANXA5 in murine M cells is associated with cytoskeletal regulation and endocytic pathways 11,43 .ANXA4 and ANXA11 induce membrane curvature and promote vesicle aggregation [55][56][57] , and are highly enriched in chicken M cells (Fig. 3A).Therefore, it is likely that ANXA4 and ANXA11 play analogous roles in chicken M cells.Similarly, AIF1 is involved in particulate uptake in murine M cells 36 .However, it is not present in current chicken genome and instead AIF1L, a gene that shares 60% homology with AIF1, was found to be highly enriched in bursal M cells.Together these data suggest chicken M cells express many genes associated with antigen-uptake and subcellular transport.

M cells in the chicken small intestine lack CSF1R expression
Transmission electron microscopical analysis on the FAE overlying the PP in the jejunum confirmed the existence of M cells exhibiting horseshoe-shaped morphology and an intraepithelial pocket harbouring immune cells (Fig. 6A).Our studies using CSF1R-reporter chickens have shown that M cells within the FAE of the BALT 23 , bursa, caecal tonsils (Fig. 6B-D), and proctodaeum 24 express CSF1R 23,24 .However, whether M cells in the FAE of the chicken small intestine similarly express CSF1R was unknown.We therefore analysed the expression CSF1R-eGFP in the PP regions of the duodenum, jejunum, and ileum of 4-week-old CSF1R-reporter transgenic chickens (Fig. 6E-G).In contrast to the bursa, expression of CSF1R-eGFP POS cells and CSF1R protein was undetectable in the FAE throughout the small intestine (Fig. 6H).Although occasional CSF1R-eGFP POS cells were detected in the epithelial layer of the jejunum and ileum, these cells lacked the characteristic epithelial cell shape and may represent intra-epithelial cells.As anticipated, the expression of CSF1R protein was detected in macrophages/DC in the underlying lamina propria (Fig. 6H-I).This analysis also supported the idea that the CSF1R-eGFP POS cells detected within the epithelial layer were filopodia protruding through the epithelial layer (Fig. 6I, Supplementary Video 1).Together, these data show that in contrast to the lung, large intestine, and the bursa, M cells in the FAE of the small intestine do not express CSF1R.

SOX8 as common M cell gene marker
Finally, we determine whether the CSF1R NEG M cells in the small intestine expressed the key transcriptional regulator SOX8 (Fig. 7).Consistent with data from the bursa (Fig. 3A), SOX8 expression was detected in epithelial  cells in the BALT and small intestine (Supplementary Fig. 1).Immunofluorescent analysis identified strong SOX8 expression in CSF1R-eGFP POS M cells in the lung and bursa (Fig. 7A).In caecal tonsils, both SOX8 POS CSF1R-eGFP POS and SOX8 POS CSF1R-eGFP NEG cells were observed in the epithelium (Fig. 7Bi-ii), with the latter being confined to the lower crypt regions.In the jejunum and ileum, SOX8 staining was observed in CSF1R-eGFP NEG cells within the FAE regions (Fig. 7C).Together, these data reveal that whereas the M cells in the BALT, bursa and caecal tonsils and small intestine of chickens all express SOX8, two subpopulations were specifically identified in the small intestine that differed in their expression of CSF1R.

Discussion
Here, we utilised CSF1R-eGFP reporter transgenic chickens to specifically enrich CSF1R-expressing bursal M cells and analysed their transcriptome using mRNA-seq.Our analysis identified genes conserved between mouse and chicken M cells while also identifying genes specific to chicken bursal M cells.In turn, we demonstrate that chickens may have two subsets of M cells in the small intestine based on their expression of CSF1R.
In mammals, the regulation of M cell development is orchestrated by the RANK-RANKL signalling and intrinsic expression of the transcription factors Spi-B and SOX8, leading to the differentiation of functionally mature GP2 POS M cells 13,32 .There are no current annotations for the genes that encode SPIB and GP2 in the chicken genome.Chicken bursal M cells express CSF1R highly, suggesting their ontogeny and/or developmental pathway may differ from mammalian M cells which do not express CSF1R 22 .Through transcriptome analysis we identified 2260 DEGs which were specifically upregulated in chicken M cells compared to IFE cells.Within this dataset elevated transcript levels of murine M-cell-associated genes, MARCKSL1, NFKB2, PRNP, SOX8, TNFAIP2, RELB, TNFRSF11B, and TNFRSF11A, were present suggesting the conservation of mammalian M cell developmental or functional pathways in chickens 12,14,21,31,37 .Moreover, chicken M cells expressed a significant abundance of many gene transcripts from closely related mouse M cell-associated gene families of the absent genes, notably those implicated in lineage determination, for example SPI1, and essential cellular functions as discussed below.
Cell differentiation in the intestinal epithelium is governed by transcription factors, some of which exhibit highly lineage-restricted expression patterns.The induction of M cells relies on the SOX8 and Spi-B transcription factors 13,14,32 .In this study, SOX8 and a paralog of SPIB, SPI1, were found to be highly upregulated in bursal M cells compared to IFE cells.Indeed, our analysis suggested that 450 of the M cell-associated DEGs contained SPI1 binding motifs in their promoter regions 58,59 .In mice, Spi1 has the capacity to compensate for the loss of SPIB, and this compensatory mechanism may suggest a similar role for SPI1 in the functional differentiation of chicken M cells 60 .The origin of bursal FAE cells is currently uncertain with evidence supporting their mesenchymal and haematopoietic origins 24,[61][62][63] .The expression of PU.1 (encoded by SPI1) is consistent throughout the hematopoietic hierarchy and orchestrates the differentiation of myeloid cells in concert with CEBPA 64 .In mammals, cell differentiation beyond the pre-granulocyte-macrophage (GM)/common myeloid progenitor stage is blocked in the absence of CEBPA, enabling the cells to follow a non-GM lineage developmental pathway 65 .We found no enrichment of CEBPA or CEBPB gene transcripts in chicken M cells but did observe elevated transcript levels of genes associated with haematopoiesis, such as CSF1R, HIF1A, and STAT3 [66][67][68] .Although the elevated transcript levels of SPI1 and CSF1R would suggest chicken M cells have a similar transcriptional profile as macrophages, however, no enrichment of macrophage gene signatures 69,70 suggest that bursal M cells like mammalian M cell do not follow a similar transcriptional pathway as macrophages.
Although bursal M cells are efficient at the uptake of 0.02-0.1 µm latex beads, they were impaired in the uptake of larger particles and bacteria, such as Salmonella bacteria 24 .Few known transcripts of surface receptors related to endocytosis activity and conserved pathogen capture receptors were upregulated in bursal M cells.Those that were detected included LY75, NRP2, CD44, CD151, CD81 and the conserved murine M cell gene, PRNP 21,[71][72][73][74][75] .Genes associated with actin remodelling processes, including AIF1L, ANAX4, ANAX11, CD63, exhibited significant expression levels in bursal M cells, suggesting their capability to modify cellular structure for particle uptake 76 .The most upregulated integrins in bursal M cells were ITGB8 and ITGAV, which form the αvβ8 complex.This strictly expressed complex mediates TGFβ activation in murine lung epithelial cells leading to the recruitment of immune cells 77 .Along with its ability to control TGFβ activation, αvβ8 may play a role in regulating immune-M cell interaction.A number of chemokines transcripts were elevated in bursal M cells compared to IFE cells.Murine M cells express CCL9 and CCL20 which attract B cells and DC towards the M cells 78 , but both genes were not present in our dataset.Both CCL20 and Psg18 expression in the FAE of mouse are dependent on RANK-TRAF6 mediated NF-κB signalling 30 .We observed no upregulation of TRAF6 in chicken M cells which may explain the lack enrichment of these genes.We did find CCL19 enriched in chicken M cells which may attract CCR7 POS DC 44 .In addition, low transcript levels of CCL26 were present in chicken M cells.In mice the expression of CCL26 in M cells mediates the recruitment of CX3CR1 POS DC 79 .Chicken bursal M cells exhibited a notable enrichment of receptors that regulate the bioactivity of cytokines and inflammation, such as IL17REL, IL22RA2, IL1R2, IL1RAP, and IL10RA 80 .In mice, the negative regulation of IL-22 signalling in the FAE by IL-22-binding protein-expressing DC is important for facilitating the transcytosis of antigens in PP 81 .In addition, bursal M cells also exhibited moderate transcripts for IFNGR2 and IFNAR2, indicating their ability to respond to IFNα/γ supported by their enrichment of transcriptional regulators of anti-viral responses, ELF1 and ELF5 82,83 .Gut-derived molecules, are critical for post-hatch B cell development and believed to be transported through the bursal M cells 84 .Thus, the restricted expression of endocytic or pathogen capture receptors on bursal M cells may have a role in controlling entry of gut-derived particulates to regulate the proliferation and maturation of B cells within the underlying follicles.
The histological examination of the FAE cells throughout the intestine revealed that small intestinal M cells lacked expression of CSF1R mRNA and protein, and while SOX8 is highly conserved in all chicken M cells, the expression of CSF1R is tissue-specific.This implies that there are key differences the transcriptional programs between these two types of M cells.Further work is required to determine the relationship between chicken CSF1R POS SOX8 POS and CSF1R NEG SOX8 POS M cell lineages.A schematic overview of the conserved and unique genes identified in chicken bursal M cells is provided in Fig. 8.
In summary, the gene expression profile of CSF1R POS bursal M cells demonstrates the conservation of murine M-cell associated genes but also chicken-specific gene signatures.In addition, we have identified diversity between chicken M cells located in the small and large intestine based on CSF1R expression.Future studies will resolve the developmental difference between CSF1R POS and CSF1R NEG M cells and in defining the functionality of M cells in the small intestine to help improve mucosal immune responses through vaccine targeting.

Experimental animals and ethics statement
CSF1R-eGFP reporter transgenic and non-transgenic, Hy-line Brown, chickens were provided by the National Avian Research Facility, The Roslin Institute, University of Edinburgh, UK.Birds received food and water ad libitum and were not vaccinated.Breeding of transgenic chickens was carried out under the authority of Project License PPL 70/8940 with the consent of The Roslin Institute Animal Welfare and Ethical Review Board and experiments were carried out in accordance with the UK Home Office project licence (PP8571233) and in compliance with the UK Home Office Animals Act 1986 (Scientific Procedures), and these ensured they were carried out in compliance with the ARRIVE guidelines and recommendations.

Immunohistology
The bursa, PP from the duodenum, jejunum and ileum, caecal tonsils, and lung of CSF1R-eGFP reporter transgenic and non-transgenic chickens were collected and fixed in 10% neutral buffered formalin (NBF, Cell Path) for 24 h at RT and embedded in paraffin.Tissues were sectioned at 8 µm thickness and mounted on Superfrost Plus slides (Thermo Fisher Scientific (TFS)) from a 42 °C water bath and incubated at 50 °C overnight.For IHC, formalin-fixed paraffin-embedded (FFPE) sections were dewaxed and antigen retrieval was performed for SOX8 and AIF1L antibodies in 10 mM sodium citrate (Sigma-Aldrich) buffer at 121 °C for 15 min.Slides were washed in staining buffer (PBS, 0.1% BSA and 0.01% Tween-20, Sigma) for 5 min at RT and incubated with 5% horse (Sigma) or goat serum (Chondrex) for 1 h.Sections were stained with primary antibodies (Supplementary Table 3) for 3 h at RT.For biotin-conjugated antibodies, ABC solution (Vector Laboratories) was applied to the sections for 30 min.Sections were visualised using diaminobenzidine (DAB; Sigma-Aldrich), counterstained with Haematoxylin (Sigma-Aldrich) and mounted using Pertex mounting medium (Cell Path).For immunofluorescence staining, after incubation with primary antibodies, sections were incubated with fluorophoreconjugated secondary antibodies and Hoechst 33258 (Sigma-Aldrich) for 1 h at RT. Slides were mounted using Prolong Diamond Anti-fade mountant (Invitrogen).Brightfield and fluorescence images were captured using Nikon light microscope and Zeiss LSM 880 confocal microscope, respectively.Images were analysed using ZEN software (Carl Zeiss).

Transmission electron microscopy (TEM)
PP regions from the jejunum of 4-week-old chickens were isolated based on CSF1R-eGFP expression and fixed in 3% glutaraldehyde in 0.1 M sodium cacodylate buffer, pH 7.3, for 2 h and processed as described previously 85 .Ultrathin sections (60 nm) were stained in uranyl acetate and lead citrate and imaged using a JEOL JEM-1400 Plus TEM and analysed using ImageJ v1.52e (Fiji).

Isolation of bursal cells
Isolation of bursal epithelium was performed as previously described 21 .In brief, bursas were removed and stored in ice-cold PBS until use.Bursas were opened to reveal the epithelial surface and digested in 50 mM EDTA (in-house) in Mg 2+ Ca 2+ free Hanks' balanced salt solution (HBSS, Corning) for 30 min at 400 g at 37 °C.The supernatant was collected, filtered through a 70 µm sieve and centrifuged at 400 g at 4 °C for 10 min.The number of live cells was determined by Trypan blue (Corning) exclusion using a Haemocytometer and the cell number was adjusted to 10 7 cells/mL for flow cytometric analysis.

Flow cytometry
One million isolated bursal cells were stained with primary and secondary antibodies, outlined in Supplementary Table 3, on a 96-well U-bottom plate (TFS), on ice, in the dark for 30 min.Cells were pelleted at 190 g for 3 min and washed twice with FACS buffer (PBS, 0.5% BSA and 0.001% sodium azide, Sigma).Antibodies were diluted in FACS buffer.Compensation was achieved using BD™ CompBeads anti-mouse Igκ particles (BD Bioscience) for each flurochrome.Fluorescence minus one (FMO) and secondary, isotype controls were prepared as described previously 69 .Phenotypical analysis was performed on the LSRFortessa™ with 4 lasers and 16 filters (BD Biosciences).Sytox™ Blue Dead Cell Stain (TFS) was added 5 min prior to flow analysis to differentiate dead and live cells.Cell gating was carried out using cells from non-transgenic chickens and FMO controls.At least 10,000 to 100,000 live, single cell events were captured for analysis.Data were analysed using the software program FlowJo (Tree Star v10).

Fluorescence-activated cell sorting (FACS)
To enrich for bursal M cells and IFE cells for RNA-seq, 3 bursas from 1-week-old birds and two bursas from 4-week-old birds were pooled for each biological replicate (1-week-old, n = 4 biological replicates and 4-weeksold, n = 3 biological replicates).For each sorting experiment, an age-matched, non-transgenic chicken was used to apply cell gates.Bursal segments were digested in 50 mM EDTA in HBSS containing RNasin® Plus RNase inhibitor (1:100, Promega) as outlined above.Cells were stained with mouse anti-chicken CD45-APC as described in the previous section.Cell numbers were adjusted to 8 × 10

RNA extraction
Sorted bursal M cells and IFE cells were pelleted at 8000 g at 4 °C for 10 min immediately after sorting and lysed with RLT Plus buffer (Qiagen) supplemented with 2-Mercaptoethanol (2-ME, Sigma-Aldrich) and stored at − 80 °C until use.Total RNA was purified using RNeasy Plus Micro Kit (Qiagen) with a genomic-DNA eliminator column according to the manufacturer's instructions.The quality and integrity of the RNA was analysed using high sensitivity RNA ScreenTape Kit (Agilent) according to the manufacturer's instructions.Samples were analysed on the Agilent 2200 TapeStation instrument (Agilent) and samples with RNA integrity number (RIN) of 9 or more were selected for RNA sequencing.

RNA sequencing and data analysis
Reads were trimmed for quality at the 3' end with a threshold of 30 and for adapter sequences using Cutadapt [version cutadapt-1.9.dev2 86 ].A minimum length of 50 was retained after trimming.Reads were aligned to Gallus gallus genome (GRCg6a) from Ensembl using STAR [version 2.5.2b 87].Genes with near-zero counts were filtered based on counts per million (CPM), where a row of the expression matrix was required to have values greater than 0.1 in at least four samples.Reads were normalised using the weighted trimmed mean of M-values (TMM) method 88 .Differential gene expression analysis and principal component (PC) analysis was carried out using edgeR (version 3.36.0),where maximum false discovery rate (FDR) < 0.05 and minimum Log 2 fold change (FC) > 1 for upregulated DEGs and Log 2 FC < − 1 for downregulated DEGs.Statistical assessment of differential expression was analysed using quasi-likelihood F-test.Functional enrichment analyses including Gene Ontology (GO) and Kyoto encyclopedia of genes and genomes (KEGG) pathway were performed using gProfiler.Transcription factors, cytokine receptors and cell surface receptors were obtained from the UniProt database and AmiGO database and categorised using Panther Classification (System v.14.0).

Figure 1 .
Figure 1.Flow cytometric analysis of isolated chicken bursal cells.(A) After debris removal (Gate P1), isolated bursal cells were gated on single cells (Gate P2).Live cells (Sytox Blue NEG ) were gated based on auto-fluorescence (Gate P3) or negative auto-fluorescence (Gate P4) in the ultraviolet channel.(B) To exclude leukocytes, CD45 NEG gates were applied to Gates P3 and 4 and analysed for CSF1R-eGFP POS cells as putative bursal M cells and CSF1R-eGFP NEG cells as putative IFE cells (Gates P5 and P6).Data is representative of 5 biological replicates from 1-week-old CSF1R-eGFP reporter transgenic chickens.(C) CD45 NEG CSF1R-eGFP POS M cells lack the surface expression of chicken monocyte/macrophage markers, MHC II, MRC1L-B, CD11, TIM4 while consist of both CD44 POS and CD44 NEG populations.Data is representative of two independent experiments from 2-week-old CSF1R-eGFP reporter transgenic chickens.

Figure 2 .
Figure 2. Gene expression profiling of chicken bursal M cells.(A) Representative diagram illustrating the isolation and FAC sorting of bursal single, live auto-fluorescent, CD45 NEG CSF1R-eGFP POS M cells and CSF1R-eGFP NEG IFE cells from 1-and 4-week-old chickens for RNA-seq analysis.(B) PCA demonstrates the association of sample clustering with the cell type (PC1) and age (PC2).(C) Venn diagram demonstrating the number of genes up or downregulated in comparison groups based on cell types (M cells and IFE cells) and age (1-and 4-week-old chickens).DEG were based on FDR < 0.05 with log 2 FC > 1 for up-regulated and Log 2 FC < − 1 for down-regulated genes.

Figure 3 .
Figure 3. Enrichment of mouse M cell-associated genes in bursal M cells.(A) FFPE bursal sections were stained with rabbit anti-human AIF1L (brown) and haematoxylin (blue).Data is representative of four independent experiments using 4-week-old CSF1R-eGFP reporter transgenic chickens.Scale bars = 100 and 10 µm.(B) Heatmap analysis of the expression levels of murine M cell-associated genes in FAC sorted bursal M cells and IFE cells from 1-and 4-week-old chickens.Genes were identified based on FDR < 0.05 with log 2 FC > 1 for upregulated and log 2 FC < − 1 for downregulated genes between M cells and IFE cells of 1-week-old chickens.Data represent n = 4 for 1-week-old and n = 3 for 4-week-old chickens.Heatmaps colouring shows the relative expression level log 2 CPM (counts per million) of genes (rows) across cells (columns) with red indicating high counts and blue indicating low counts.

Figure 4 .
Figure 4. Identification of transcription factors, signalling receptors and putative functional molecules associated with bursal M cells.(A) The top 20 upregulated transcription factors and signalling receptors (associated with 'DNA-binding transcription factor activity' (GO:0003700) and 'cytokine-mediated signalling pathway' (GO:0019221) GO terms, respectively) in sorted bursal M cells and IFE cells from 1-and 4-week-old chickens.(B) FFPE bursal sections stained with rabbit anti-human SOX8 (brown) and haematoxylin (blue).Data is representative of at least four independent experiments using (i) 4-week-old CSF1R-reporter transgenic chickens and (ii) 18-day-old embryos.Scale bars = 100 and 10 µm.Heatmaps colouring shows the relative expression level log 2 CPM (counts per million) of genes (rows) across cells (columns) with red indicating high counts and blue indicating low counts.

Figure 5 .
Figure 5. Identification of cytokine signalling receptors and putative functional molecules associated with bursal M cells.(A) Top 20 upregulated genes in sorted bursal M cells and IFE cells from 1-week and 4-weekold chickens associated with (A) cytokine receptor signalling, (B) endocytic, G-protein-coupled, and surface receptors using gProfiler and UniProt.Genes were identified based between 1-week-old M cells and IFE cells.Data represent n = 4 for 1-week and n = 3 for 4-week-old chickens.Heatmaps colouring shows the relative expression level log 2 CPM (counts per million) of genes (rows) across cells (columns) with red indicating high counts and blue indicating low counts.

Figure 6 .
Figure 6.FAE of the small intestine lack CSF1R-eGFP and CSF1R expression.(A) TEM analysis of the jejunal PP reveals the presence of M cells with the characteristic horse-shoe shape morphology and intraepithelial pocket harbouring a lymphocyte.Immunofluorescent analysis of the FAE in the (B) lung, (C) bursa, and (D) caecal tonsil (CT) demonstrates the cuboidal morphology, characteristic of an epithelial cell, of CSF1R-eGFP POS M cells.Immunofluorescent analysis of the FAE in the PP regions of (E) the duodenum (DD), (F) jejunum (JJ) and (G) ileum (IL) demonstrates the lack of CSF1R-eGFP POS epithelial cells (yellow dashed line).CSF1R-eGFP POS macrophages and DC (white arrows) can be observed near the epithelial cell layer.Scale bars = 50 µm.(H) FAE cells of the jejunal PP do not express CSF1R protein (red), demonstrating lack of CSF1R-eGFP (green) in small intestinal M cells is not due to the promoter regions used for the development of the CSF1R transgenic chickens.(I) The CSF1R-eGFP POS CSF1R POS cells within the epithelial cell layer represent extended dendrites of mononuclear phagocytes and not M cells.Scale bars = 100 and 50 µm.Data is representative of four independent experiments using 4-week-old CSF1R-eGFP reporter transgenic chickens.

Figure 7 .
Figure 7. Chickens exhibit diversity in their M cells characterised by CSF1R expression.(A) SOX8 expression co-localises with the CSF1R-eGFP POS FAE cells in the BALT and the bursa.(B) CSF1R-eGFP is differentially expressed in SOX8 POS M cells in the caeca.(Bi) CSF1R-eGFP POS SOX8 POS cells are evident in the higher regions of the villi.(Bii) CSF1R-eGFP NEG SOX8 POS cells are predominantly located in the crypt regions.(C) SOX8 expression in the FAE regions of the jejunum (JJ) and ileum (IL) indicating diversity in chicken M cells associated with CSF1R expression.Data is representative of three independent experiments using 4-week-old CSF1R-eGFP reporter transgenic chickens.Scale bars = 100 and 10 µm.

Figure 8 .
Figure 8. Schematic overview of M cell-associated genes in mouse PP M cell and chicken bursal M cell.Image created using Biorender (www.biore nder.com).
7cells/tube in 500 µL of FACS buffer supplemented with RNase inhibitor.Prior to the cell sort, Sytox™ Blue Dead Cell Stain (TFS 1:1000) was added to the sample.Cells were sorted using FACS ARIA™ III cell sorter, with 4 lasers and 11 detectors (BD Biosciences).CD45 NEG CSF1R-eGFP POS (M cells) and CD45 NEG CSF1R-eGFP NEG cells (IFE) were collected into ice-cold 1.5 mL tubes (TFS), containing 200 µL of FACS buffer supplemented with RNase inhibitor.