Fertilized White Leghorn (Gallus gallus domesticus), SPF chicken eggs were obtained from commercial breeders (Prophyl-BIOVO Ltd., Hungary). The eggs were incubated at 37.5°C in a humidified incubator (Heka 1+ incubator Brutgerate, TS-7000C Rietberg, Germany) at 90% humidity, and the age of the embryos was determined by the number of embryonic (E) days. Hatched animals were housed in the aviary of the Department of Anatomy, Histology, and Embryology, Semmelweis University. All animal work was conducted according to relevant national and international guidelines and approved by the Animal Care and Use Committee at Semmelweis University, Budapest, Hungary.

For cryosections, tissue samples were fixed in 4% paraformaldehyde (PFA) for 1 hour at room temperature, then infiltrated with 15% sucrose overnight at 4°C, followed by 7.5% gelatine (Sigma, G-2625) in 15% sucrose for 1 hour at 37°C. Gelatine-impregnated tissues were rapidly frozen at −50°C in 2-methylbutane (Sigma, 78-78-4). 12 μm thick cryosections were labelled with primary antibodies ( Table 1 ) for 1 hour, followed by biotinylated secondary antibodies ( Table 2 ) (Vector Laboratories) and avidin-biotinylated peroxidase complex (Vectastain Elite ABC kit, Vector Laboratories, PK-6100). Endogenous peroxidase activity was quenched with 3% hydrogen peroxide (Sigma, H1009) for 10 minutes. The binding sites of the primary antibodies were visualized by 4-chloro-1-naphthol (Sigma, C8890).

12 μm-thick cryosections were stained with primary antibodies listed in Table 1 for 1 hour, followed by fluorescently labelled secondary antibodies ( Table 2 ) for 45 minutes. To assess cell proliferation, 10 µM 5-ethynyl-20-deoxyuridine (EdU) was added to the culture medium 3 hours before fixation with 4% paraformaldehyde (PFA). EdU incorporation was subsequently detected using the Click-iT EdU Imaging Kit (Thermo Fisher, C10337). Cell nuclei were stained with 4,6-diamidino-2-phenylindole (DAPI, Vector Laboratories), and sections were covered by aqueous Poly/Mount (Polyscience Inc., 18606). Sections were examined with fluorescent (Nikon Eclipse E800), confocal scanning (Zeiss LSM 780) and STED (Abberior Expert Line- Nikon Ti2) microscopy; section images were recorded using CellSens and ZEISS ZEN Imaging proprietary softwares. Image processing was performed using ImageJ and Adobe Photoshop CC 2020 programs.

4-week-old chicken bursa of Fabricius was fixed in 4% PFA for 24 hours at 4°C, washed in PBS and infiltrated with 10% sucrose overnight at 4°C, followed by 20% and 30% sucrose solutions. Tissues were impregnated with 7.5% gelatine for two hours at 37.5°C. Gelatine-impregnated tissues were rapidly frozen at −50°C in 2-methylbutane. 12 µm thick cryosections were collected on Epredia™ Superfrost Plus Adhesion Slides (Fisher Scientific, J1800AMNZ) and stored at -80°C until use. RNAscope was performed using the RNAscope Fluorescent Multiplex Detection Reagent Kit v2 (Bio-Techne, 323136) according to the manufacturer’s protocol. Hybridization was performed using the chicken-specific Gg-CXCL12-C2 probe (Bio-Techne, 458251). In case of double labelling, the RNAscope method was done first, followed by immunocytochemistry.

4-week-old chicken bursa of Fabricius were fixed in 4% buffered glutaraldehyde for 24 hours at 4°C, washed 3 times for 5 minutes in Milloning buffer (830ml 2.26% Na2HPO4 • 2H2O- Merck, 106462 and 170ml 2.52% NaOH - Merck, 1.06346.1000), then post-fixed with 1% osmium tetraoxide for 2 hours (Merck - B000970). Tissue samples were dehydrated in graded ethanol, placed in propylene oxide (Merck, 12492) twice for 5 minutes, followed by propylene oxide and EPON (Araldite Resin Grade 6005 - EMS-10920; Poly-bed 812 - EMS-14900; DDSA-Dodenyl Succinic Anhydride- EMS-13710; NMA-Nadic Methyl Anhydride-EMS-19000; DMS 30-2,4,6-Tris Dimethylaminomethylphenol-EMS-13600). Tissue samples were embedded in Polybed/Araldite 6500 (Polysciences Inc.) oriented and polymerized at 56°C for 2 days. 50-70 nm ultra-thin sections were contrasted with a mixture of uranyl acetate (10 min) and lead citrate (12 min). Sections were examined with a Jeol JEM-1200EX transmission electron microscope.

For the embryonic B cell migration assay, bursa of Fabricius was isolated from E13 chick embryos and cultured in the presence of CXCL12 (100 ng/ml; R&D Systems, 6448-SD-025) in RPMI 1640 (Sigma, R8758) culture media. To determine how different matrix molecules affect B cell migration, we coated Petri-dishes with 10 μg/ml laminin (Sigma, L2020), 10 μg/ml tenascin-C (Sigma, CC115) or 20 μg/ml fibronectin (Sigma, F1141) diluted in RPMI, containing 5% Penicillin/Streptomycin (Pen/Strep, Sigma, P0781) at 37°C for 2 hours. Petri-dishes were washed twice with PBS, and E13 bursal folds (n=6/experimental setup) were explanted on the different ECM-coated surfaces. Organ cultures were maintained at 37°C, 5% CO₂ for 24 hours.

For the B cell adhesion assay, BF was dissected from E18 chicken embryos and enzymatically digested in RPMI 1640 medium (Sigma, R8758-500ML) containing 0.03% collagenase IX (Sigma, C9722) and 2% dispase II (Stem Cell Technologies, 07913) for 30 minutes at 37°C with gentle agitation. The resulting cell suspension was filtered through a 40 µm cell strainer, washed twice in RPMI supplemented with 10% FBS (fetal bovine serum, Capricorn Scientific, FBS-12B) and 5% Penicillin/Streptomycin. 35 mm Petri dishes were coated for 1 hour at 37°C with chicken tenascin‐C (10 µg/ml) and CXCL12 (100 µg/ml; PeproTech, 250-20A). Double-coated dishes were washed twice with PBS, and BF cells were seeded at a density of 2 × 10⁵ cells per dish and incubated at 37°C, 5% CO₂. Inhibition of the CXCR4 receptor was achieved by adding the CXCR4-antagonist AMD3100 at a concentration of 200 µM (Sigma, 239820). Control cells were cultured in RPMI. After incubation for 6 hours, non-adhesive cells were removed by gentle aspiration using a pipette, the adhered cells fixed in ice-cold 100% methanol for 20 minutes and anti-chB6 immunostaining was performed. All cell adhesion assays were repeated 3 times.

We used replication-competent avian retrovirus (RCAS) expressing chick Shh (37) to induce ectopic expression of tenascin-C (38). Shh misexpression in the developing chick bursa of Fabricius was performed following the method of Nagy et al. (21, 39). Briefly, 1 μl Shh-RCAS virus was injected into the mesenchymal layer of the bursa of Fabricius of E9 chicken embryos using a Hamilton syringe attached to a Narishige brand microinjector. Shh-RCAS injected E9 bursa primordia were transplanted onto the chorioallantoic membrane of E9 chicken embryos and cultured for 8 days. CAM grafts were excised and fixed in 4% paraformaldehyde for 1 hour at RT, embedded in gelatin for cryosectioning and analyzed for Shh and tenascin-C expression. 3C2 immunocytochemistry was used to detect the presence of the viral vector. Control grafts were injected with PBS. At least three replicates were collected and analyzed for each experimental condition.

Statistical analysis was performed using the Kruskal-Wallis test or ANOVA with a post-hoc Dunn’s test or Tukey test (R Core Team). P-values were adjusted with Holm correction. P<0.05 was considered significant, and the following significance levels were used: **p<0.01; ***p<0.001; ****p<0.0001. Error bars represent the standard error of the mean.

Avian B cell development occurs within the specialized microenvironment of the lymphoid follicles of the bursa of Fabricius. While the stromal cells of the follicular medulla have been extensively characterized (4, 7, 40), the specific roles of resident cell populations and the cellular microenvironment of the cortex, which emerges later during ontogeny, remain largely unknown. Histological analysis of 8-week-old chicken bursal sections shows the follicular medulla as a central, loosely packed region with large lymphocytes, while the cortex surrounds the medulla as a dense, outer compartment. Hematoxylin and eosin staining highlights this clear distinction, with the cortex showing darker staining due to its higher cell density ( Figure 1A ). A thin connective tissue layer separates the individual follicles. Electron microscopy further highlights that the cortex contains numerous lymphocytes localized between cortical reticular cells ( Figure 1B ). The cortical reticular cells were identified by their prominent euchromatic nuclei and multiple cellular processes ( Figure 1C ).

In our previous studies addressing the origin of medullary stromal cells, we described that ontogenetically the medulla and the cortex have different origins (2, 41). Medullary epithelial cells, which form the supportive reticular network, are derived from the ectoderm and express cytokeratin intermediate filaments ( Figure 1D ). In contrast, cortical reticular cells arise from the mesoderm and express both desmin and vimentin intermediate filaments ( Figures 1D, E-E ). Importantly, while desmin is a specific marker for cortical reticular cells, vimentin is expressed in both cortical reticular cells and the medullary bursal secretory dendritic cells (BSDCs) (42). RNAscope in situ hybridization showed that CXCL12 expression is restricted to the cortex and is specifically produced by vimentin immunoreactive cortical reticular cells ( Figure 1F ). Recently, we have demonstrated that cortical chB6+ B cells selectively express the CXCR4 receptor (21, 32). Double immunofluorescence shows that CXCR4^high B cells in the cortex are restricted to the outer cortical region while CXCR4^low/dim B cells are localized near the cortico-medullary border, concentrated around laminin immunoreactive cortical capillaries ( Figure 1G ). The laminin+ basement membrane separates the cortex from the medulla at the cortico-medullary border.

Although collagen III, fibronectin, and laminin are known to be expressed in the cortex of bursal follicles (4, 29) a detailed analysis of extracellular matrix (ECM) protein expression has not been performed. In order to characterize the immunophenotype of the ECM produced by cortical reticular cells, comparative immunostainings were performed using a panel of commercially available collagen-, proteoglycan-, and glycoprotein-specific markers. All major ECM molecule families are highly expressed within the bursa, with distinct expression patterns. Collagens (type I, III, IV, and VI) are expressed by the interfollicular connective tissue septae, the basement membrane of the cortico-medullary epithelium, and the basement membrane of endothelial cells of cortical capillaries ( Figures 2A–D ). Heparan-sulphate proteoglycans (collagen XVIII, perlecan, and agrin) are specifically expressed in basement membranes ( Figures 2E–G ), while versican, a chondroitin-sulphate proteoglycan is restricted to the interfollicular connective tissue ( Figure 2H ). Among glycoproteins, laminin expression is restricted to basement membranes ( Figure 2I ), while fibronectin and fibrillin are highly expressed in the cortex, showing intense fibrillar staining from the cortico-medullary border to the interfollicular connective tissue ( Figures 2J, K ). In contrast, tenascin-C expression is restricted to the cortex, showing strong immunoreactivity in the inner cortical region and a gradient-like expression pattern from the cortico-medullary border towards the interfollicular septae. Compared to other ECM proteins, tenascin-C was not expressed in the interfollicular connective tissue ( Figure 2L ).

Recently, we demonstrated that subpopulations of chB6+ cortical B cells can be distinguished by their selective expression of the CXCR4 receptor (21). Immunostaining of adult chicken BF shows that CXCR4^high B cells are concentrated in the outer cortical region while CXCR4^low/dim B cells are localized near the cortico-medullary border ( Figure 3A ). The expression of tenascin-C in the cortex exhibits a complementary pattern, with its gradient highest around cortical capillaries ( Figure 3A inset) near the E-cadherin immunoreactive cortico-medullary epithelium and gradually decreases towards the outer cortical regions ( Figures 3A, B ). Interestingly, alternating with the tenascin-C-rich capillary region, gaps with low expression level of tenascin-C can also be observed in the inner cortical environment ( Figure 3B ).

Previous studies demonstrated that B cells leave the bursa directly from the follicular cortex through cortical capillaries and that bursal emigrants originate from the rapidly dividing cortical B cells (33–35). EdU incorporation confirmed the presence of proliferating cells within the cortex, CXCR4^high cortical cells representing highly proliferative regions, compared to low proliferative CXCR4^low/dim cells in tenascin-C rich zones adjacent to the cortico-medullary border ( Figures 3C–E ).

In recent years, several aspects of embryonic bursal cell migration have been uncovered, highlighting the role of the mesenchymal-derived CXCL12 chemokine in guiding CXCR4+ B cell precursors to colonize the developing lymphoid follicles (21, 32, 43). In addition to CXCL12 signaling, it has been reported that the ECM plays a crucial role in coordinating cell migration through ECM-cell and ECM-chemokine interactions. Although fibronectin is known to be expressed in the embryonic bursa mesenchyme (29), a detailed analysis of ECM expression patterns during bursa of Fabricius development has not been performed.

In chicken embryos, colonization of the bursal epithelial-mesenchymal rudiment by chB6+/CXCR4+ B cell precursors occur within a restricted developmental window between embryonic days 10 and 15 (E10-E15). To investigate the role of ECM proteins during this critical stage, we examined their distribution at E11, when the first wave of chB6+ B cells enters the bursal mesenchyme. At this developmental stage, the BF mesenchyme shows intense fibronectin immunoreactivity ( Figure 4A ). ECM-specific immunostaining on consecutive sections confirmed that all tested collagens, proteoglycans, and glycoproteins are broadly expressed in the bursal primordium with the exception of tenascin-C ( Figure 4B ). Basement membranes are specifically labeled by laminin, agrin, and collagen IV (data not shown).

The earliest sign of tenascin-C expression is seen in the connective tissue capsule of the BF primordium at developmental stage E13 ( Figure 4C ). At E18, as bursal development progresses and B cell colonization is completed, the outer mesenchymal compartment of developing bursal follicles shows intense fibronectin expression, but the epithelial follicle buds are fibronectin negative ( Figure 4D ). In contrast, tenascin-C expression is localized in the mesenchymal region surrounding larger developing follicles corresponding to the prospective cortical region, which begins to develop shortly after hatching ( Figure 4E ). Expression of tenascin-C around E17-E18 correlates with the loss of the follicle bud colonization potential of bursal B cell precursors (2).

To determine whether tenascin-C influences B cell migration in vivo, we performed embryo manipulation to experimentally modify the expression of tenascin-C within the bursal mesenchyme. Recent studies have uncovered signaling pathways that coordinate the expression pattern of mesenchymal growth factors and extracellular matrix proteins during vertebrate development. Sonic hedgehog (Shh) is an important morphogen that orchestrates key processes in the vertebrate nervous system, somite, gastrointestinal tract and limb development (39, 44–46). Importantly, Shh is not expressed in the BF at any developmental stage (2). According to Ting-Berreth et al., experimental activation of the Shh signaling in the mesenchyme of chicken feather buds, results in ectopic tenascin-C expression (38). Building on this data, we used the replication-competent retroviral vector (RCAS) system to experimentally overexpress Shh within the bursa mesenchyme to assess how ectopic expression of tenascin-C affects embryonic B cell migration. E9 bursa primordia were microinjected with RCAS-Shh retrovirus, transplanted onto the chorioallantoic membrane (CAM) of E9 chicken embryos and cultured for an additional 8 days in ovo ( Figure 5A ). Control bursa grafts showed normal B cell colonization, characterized by well-developed chB6+ lymphoid follicles along the bursal folds. In these CAM grafts, Shh expression was absent, and tenascin-C expression was restricted to the bursal connective tissue capsule ( Figures 5B–E ). In contrast, RCAS-Shh injected grafts showed effective viral infection of the bursal mesenchyme, as confirmed by anti-3C2 immunostaining specific for the gag protein of RCAS virus ( Figure 5F ), supporting the robust Shh protein expression in the bursal mesenchyme ( Figure 5G ). The ectopic activation of Shh signaling resulted in the mesenchymal expression of tenascin-C ( Figure 5H ), associated with disrupted B cell colonization and only rudimentary chB6+ lymphoid follicle development (n=6, Figure 5I ). These findings suggest that the exclusion of tenascin-C from the embryonic bursal mesenchyme is essential for the proper colonization of B cells and the development of lymphoid follicles in the BF.

To confirm the inhibitory effect of tenascin-C on B cell precursor migration, E13 BF folds were cultured on fibronectin or tenascin-C coated plastic surfaces in the presence of CXCL12 ( Figure 6A ), a chemokine known to induce robust migration of bursal B cells (21). At this developmental stage, large number of CXCR4-expressing B cells are localized within the bursal mesenchyme ( Figure 6B ). When 100 ng/ml CXCL12 was added to the culture medium, limited migration of CXCR4+ cells was observed on non-coated Petri dishes ( Figure 6C ). Coating the surface with 20 µg/ml fibronectin significantly enhanced migration, as indicated by the pronounced movement of CXCR4+ cells away from the explants and the appearance of numerous membrane protrusions on their cell surface ( Figure 6D , inset). In contrast, a 10 µg/ml tenascin-C coating markedly inhibited cell migration, with CXCR4+ cells exhibiting round morphology and a reduced migratory distance (220.6 ± 17.81 µm) compared to fibronectin (1098 ± 29.63 µm) ( Figures 6E, F ). Despite the initial response to CXCL12, characterized by the emigration of CXCR4+ cells from the explants in all conditions, further migration was effectively blocked on tenascin-C-coated surfaces. Morphological analysis revealed that cells on the tenascin-C surface adopted a round, non-polarized morphology ( Figure 6E , inset). We conclude that the presence of tenascin-C is inhibitory for B cell migration.

At the time when pre-bursal B cells stop colonizing the developing lymphoid follicles, tenascin-C expression appears around the follicle buds ( Figure 4E ). Around hatching, the pattern of CXCL12 expression changes: both molecules are strongly produced in the developing cortex. In the fully mature bursal follicles, tenascin-C and CXCL12 are exclusively detected within the cortex ( Figures 1F , 3A, B ). Because tenascin-C can effectively bind CXCL12 through conserved motifs (47, 48), we also analyzed the effect of CXCL12 in immobilized form adsorbed sequentially onto the tenascin-C coated surface. After 6 hours of culture, we observed that E18 bursal cells plated onto tenascin-CXCL12 double-coated dishes displayed two distinct morphologies depending on CXCR4 signaling activity. In the presence of culture media with no additives, B cells adopted an elongated morphology, with long cell protrusions (filopodia) spreading along the CXCL12+tenascin-C double-coated surface ( Figures 6G, I ). However, inhibition of CXCR4 using 200 µM CXCR4 antagonist AMD3100 abolished CXCL12-induced polarization, and B cells exhibited round, non-polarized morphology ( Figures 6H, I ) similar to B cell precursors analyzed in the migration assay ( Figure 6E ). Immunostaining with the chB6 antibody confirms their relative morphology in the tenascin-CXCL12 double-coated surfaces. These findings highlight the importance of ECM components such as tenascin-C in modulating CXCL12 signaling and suggest a potential mechanism by which the extracellular environment influences B cell behavior during BF development.