Multiple cell populations generate macrophage progenitors in the early yolk sac

Abstract

Yolk sac (YS) CSF1 receptor positive (CSF1R⁺) cells are thought to be the progenitors for tissue-resident macrophages present in various tissues. The YS progenitors for tissue-resident macrophages are referred to as erythroid–myeloid progenitors (EMPs). However, diverse types of hematopoietic progenitors are present in the early YS, thus it is not precisely known which type of hematopoietic cell gives rise to the CSF1R⁺ lineage. In this study, an analysis was conducted to determine when CSF1R⁺ progenitors appeared in the early YS. It showed that CSF1R⁺ cells appeared in the YS as early as embryonic day 9 (E9) and that the earliest hematopoietic progenitors that differentiate into CSF1R⁺ cells were found in E8. Since these progenitors possessed the capability to generate primitive erythroid cells, it was likely that primitive erythroid lineages shared progenitors with the CSF1R⁺ lineage. Mutual antagonism appears to work between PU.1 and GATA1 when CSF1R⁺ cells appear in the early YS. One day later (E9), multiple progenitors, including myeloid-restricted progenitors and multipotent progenitors, in the YS could immediately generate CSF1R⁺ cells. These results suggest that EMPs are not an exclusive source for the CSF1R⁺ lineage; rather, multiple hematopoietic cell populations give rise to CSF1R⁺ lineage in the early YS.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00018-022-04203-7.

Introduction

Macrophages are distributed throughout the body and contribute to tissue homeostasis by ingesting and destroying dead cells, cell debris, foreign substances, and microbes. Macrophages are also critical for inducing the responses of other immune cells through antigen presentation and cytokine secretion [1, 2]. Most tissue-resident macrophages are thought to be derived from monocyte/macrophage lineages present in embryonic and fetal stages [3–6]. More specifically, these cells are considered to be derived from erythroid–myeloid progenitors (EMPs) present in the yolk sac (YS) of the embryos [7–9].

The term EMP has been used to describe the cells that give rise to definitive erythrocytes and the granulocyte/macrophage lineage, but not to primitive erythrocytes and lymphoid lineages [10, 11]. Although CD41 and KIT are often used to identify EMPs in the early YS, the cell populations fractionated by these two markers are still heterogeneous, thus the early YS contains more variations of hematopoietic cells than previously thought [12]. The cell fractionation method using several cell surface markers revealed that CD45⁺KIT^highAA4.1⁺ cells, which were also positive for CD41 in the embryonic day 9.5 (E9.5) YS, showed multipotent hematopoietic cell activity with definitive erythroid, myeloid, and lymphoid potentials but were devoid of primitive erythroid potency [12–14]. In contrast, CD45⁺KIT^highAA4.1^– cells within the CD41⁺ cell fraction highly resembled EMPs because they lacked primitive erythroid and lymphoid potentials but possessed myeloid and definitive erythroid potentials [13]. These progenitors are referred to as myeloid-restricted progenitors (MRPs) in this paper. In addition, the more immature committed hematopoietic cells with a CD45⁻CD41⁺AA4.1⁻KIT^med cell surface phenotype were identified in the E8.5–9.5 YS. These cells gave rise to the primitive and definitive erythrocytes and lympho-myeloid cells [12, 15, 16]. Therefore, these cells serve as common primitive-definitive precursors (CPDPs).

All of the above-mentioned multiple cell populations have the potential to differentiate into macrophage lineages. Therefore, EMPs might not be the sole source of embryonic macrophages, and the cell population that primarily supplies the first-wave embryonic macrophage lineages in the early YS remains unknown. In this study, CSF1 receptor-positive (CSF1R⁺) cells were detected in the early YS to determine which cell population is the initial source of embryonic macrophages.

Results

CSF1R⁺ cells appear on E9 in the YS

CSF1R is one of the specific markers expressed in monocyte-macrophage lineages and their progenitors [17, 18]. First, the early YS was examined to detect when CSF1R⁺ cells appeared. Although it was reported that Csf1r expression was observed as early as embryonic day 8.5 (E8.5) using the Cre or MerCreMer mouse lines under control of the Csf1r transgenic promoter [8]. It is unknown whether the transgenic reporter recreates the genuine expression from the Csf1r locus in the mouse genome. Therefore, flow cytometric analysis was performed on the early YS cells for CSF1R expression by using the antibody against it. Cell surface phenotype analysis and retrospective cell lineage analysis had previously identified various cell populations existing in the early YS [13, 15]. As shown in Supplementary Fig. 1A, B, CD45^–Ter119^– cell fractions, which contained the most immature committed hematopoietic cell fractions and non-hematopoietic cell fractions, did not include the apparent CSF1R⁺ cell population in the E9.5 YS. In contrast, CD45⁺ cell fractions contained many CSF1R⁺ cells. Approximately two-fifths of cells in the CD45⁺KIT^–/lowAA4.1^– cell fraction were positive for CSF1R, whereas CD45⁺KIT^highAA4.1^– and CD45⁺KIT^highAA4.1⁺ cell fractions, which represent MRPs and multipotent progenitors with lympho-myeloid differentiation potentials, respectively, contained a small fraction of CSF1R⁺ cells.

Within the whole CSF1R⁺ cell population, approximately 80% of CSF1R⁺ cells were found in the CD45⁺KIT^–/lowAA4.1^– cell fraction. There were some minor contributions from the CD45⁺KIT^highAA4.1^– cell fraction, which explained about one-seventh of CSF1R⁺ cells (Fig. 1A). Gating of CSF1R⁺ cells in the cell fraction excluded from the primitive erythroid lineage revealed that most of the CSF1R⁺ cells in the E9.5 YS had CD41⁺CD45⁺AA4.1^–KIT^–/low phenotypes (Fig. 1B, Supplementary Fig. 1C). Because the E8.5 YS virtually lacked CSF1R⁺ cells (Fig. 1C, Supplementary Fig. 2), it was considered that E9 was the time point when CSF1R⁺ cells appeared in the YS. The CSF1R⁺ cells isolated earliest from E9.5 YS displayed an immature morphology (Fig. 1D, Supplementary Fig. 3). Cells with mature macrophage morphology were not observed at this stage. It is thus likely that E9.5 CSF1R⁺ cells represent monocyte-macrophage progenitors.

Fig. 1.

Presence of CSF1R⁺ cells in the E9 YS. A Distribution of CSF1R⁺ cells through cell fractions in the E9.5 YS. Data are shown as mean ± SD of three independent experiments, each performed with pooled embryos (n = 7, 8, or 10) in the upper graph. The same data are displayed as a pie graph in the lower one. B Cell surface phenotypes of CSF1R⁺ cells in E9.5 YS. Representative flow cytometry plots are shown. In lower panels, gated CSF1R⁺ cells are shown as red dots. Gray dots are unstained controls, which are shown for comparison. C Representative flow cytometry plots of the E8.5 YS from two independent experiments, each carried out with pooled embryos [n = 9] are shown. The complete gating strategy is depicted in Supplementary Fig. 2. D Representative May-Grunwald/Giemsa-stained images of CSF1R⁺ cells obtained from E9.5 YS. Scale Bar: 20 μm. Morphologies of other hematopoietic cell populations from E9 YS and mature macrophages from E17 embryos are presented in Supplementary Fig. 3 for comparison

The earliest committed hematopoietic progenitors are the precursors of the first-wave CSF1R⁺ cells in the YS

The presence of CSF1R⁺ cells at E9 indicated that the precursors of CSF1R⁺ cells were already present before E9. MRPs (CD45⁺KIT^highAA4.1^–) and multipotent lympho-myeloid progenitors (CD45⁺KIT^highAA4.1⁺) were not considered to be the initial precursors for CSF1R⁺ cells because these progenitors first appear at E9 [13, 15]. Thus, the earliest committed hematopoietic progenitors, which have the cell surface phenotypes of CD45^–CD41⁺AA4.1^–KIT^med and are capable of generating primitive erythroid cells and CD45⁺ lympho-myeloid lineages, were hypothesized to be the first precursors of CSF1R⁺ cells. This is because these CPDPs are abundantly present in the YS at E8–9 [15, 16]. To test this hypothesis, CD45^–CD41⁺AA4.1^–KIT^med CPDPs were isolated from E9.5 YS (Fig. 2A), and cell culture was performed for one day. As shown in Fig. 2B, CPDPs gave rise to CSF1R⁺ cells within one day. Notably, these cells had the CD41⁺CD45⁺KIT^–/low phenotype, which was consistent with the first observation of CSF1R⁺ cells in the live embryos (Fig. 1B). Importantly, CPDPs isolated from E8.5 YS gave rise to CSF1R⁺ cells within one day (Fig. 2C), suggesting that the earliest CSF1R⁺ cells observed in the E9 YS were derived from CPDPs present in the E8 YS. Altogether, these data indicate that CPDPs are the earliest precursors for CSF1R⁺ cells in developing embryos.

Fig. 2.

CPDPs are the immediate precursors for CSF1R⁺ cells. A–C CPDPs were sorted from E9.5 (A, B) or E8.5 (C) YS and then cultured for one day and analyzed. Sort gates (A) and post cell culture flow cytometry plots B, C are shown. Small panels shown in bottom left in B or C are an unstained control or fluorescence minus one control, respectively. In the right two panels in B, gated CSF1R⁺ cells are shown as red dots. Blue dots are unstained controls, which are shown for comparison. All data shown are representative of two independent experiments, each carried out with pooled embryos [n = 6 or 8 (A, B); n = 8 or 9 (C)]

Multiple cell populations have the potential to generate CSF1R⁺ progenitors at the E9 YS

As mentioned above, MRPs and multipotent lympho-myeloid progenitors are present in the E9 YS in addition to CPDPs. Therefore, these additional hematopoietic progenitors were examined to determine if they can also give rise to CSF1R⁺ progenitors.

Before the experiments were performed, different ways to enrich MRPs (EMPs) were examined. Previously it has been shown that the presence or absence of AA4.1 expression within the CD45⁺KIT^high cell fraction distinguishes multipotent lympho-myeloid progenitors from AA4.1^– MRPs [13]. A more recent study from another group showed that cells expressing FcγRII (CD32) and/or FcγRIII (CD16) within the KIT^high cell fraction showed EMP readouts but did not generate primitive erythroid cells or lymphoid cells [11]. Therefore, AA4.1 and FcγRII/III were co-stained to identify an effective way to enrich MRPs (EMPs). The analysis revealed that FcγRII/III was broadly expressed in virtually all committed hematopoietic cell fractions present in the E9.5 YS, including CD45^–CD41⁺AA4.1^–KIT^med CPDPs, CD45⁺KIT^highAA4.1^– MRPs, and CD45⁺KIT^highAA4.1⁺ multipotent lympho-myeloid progenitors. FcγRII/III was also expressed in the CD45^–AA4.1⁺ endothelial lineage (Supplementary Fig. 4) [15]. Therefore, the AA4.1-based separation method was adopted for the following experiments.

As described in Supplementary Fig. 5, AA4^– MRPs and AA4.1⁺ multipotent progenitors within the CD45⁺KIT^highCSF1R^– cell fraction were isolated from the E9.5 YS and then cultured. Both cell populations generated abundant CSF1R⁺ cells within 2 days. These results suggest that MRPs and multipotent lympho-myeloid progenitors in the YS also contribute to the production of macrophage lineages, in addition to CPDPs, when mouse development reaches E9.

The earliest committed hematopoietic progenitors in the YS give rise to macrophages with tissue-resident phenotypes and osteoclasts

Early YS hematopoietic cells are thought to give rise to macrophages with tissue-resident phenotypes. These macrophages have different phenotypes from macrophages matured from circulating monocytes during the postnatal period [8, 9, 19, 20]. These observations from previous studies prompted us to examine the phenotype of macrophages derived from CPDPs in the early YS. YS CPDPs and MRPs were isolated simultaneously and cultured in the presence of M-CSF for 2 weeks. The phenotypes of the generated macrophages were then compared. GM-CSF was added to these cultures during the first 3 days to support the initial growth of CPDPs and MRPs [16]. Mature macrophages became dominant in the cells cultured for 2 weeks, although CPDPs and MRPs generated F4/80⁺ mature macrophages within 5 days of cell culture (Fig. 3A). As shown in Fig. 3B, CPDP-derived and MRP-derived macrophages were F4/80^highI-A/I-E^lowCD73⁺, similar to tissue-resident macrophages present in the peritoneal cavity [19, 20]. Cells with postnatal monocyte-derived macrophage phenotypes (F4/80^lowI-A/I-E^highCD73^–/low), which were also present in the peritoneal cavity, were not observed under the cell culture conditions used. Analysis of cell surface marker expression by t-SNE, a dimensionality reduction technique, also suggested that CPDP-derived macrophages are more similar to tissue-resident macrophages than postnatal monocyte-derived macrophages or other monocyte-macrophage lineages present in adult tissues (Fig. 3C) [21]. Notably, the phenotypes of macrophages developed from CPDPs overlapped well with those from MRPs by the t-SNE analysis (Fig. 3C).

Fig. 3.

CPDP differentiation to macrophage lineage cells. CPDPs and MRPs isolated from E9.5 YS were induced to differentiate into macrophages in culture. After 5 days (A) or 2 weeks (B, C), cells were analyzed by flow cytometry. A small panel shown in A is an unstained control from CPDP cell culture. Staining pattern of cells from peritoneal cavity (PerC) is shown for comparison in top panels in B. In right histograms in B, green, pink, blue, or orange lines indicate tissue-resident, monocyte-derived, CPDP-derived, or MRP-derived macrophages, respectively. Filled gray lines in these histograms indicate unstained controls from PerC-derived cells (top) or from CPDP-derived cells (middle). Macrophages developed from CPDPs were compared to those from MRPs or to monocyte-macrophage lineage cells present in adult tissues by the t-SNE algorithm (C). Spleen (Spl) red pulp macrophages (F4/80⁺CD11c^lowI-A/I-E⁺), spleen conventional dendritic cells (cDC; CD11c^highCD11b^–I-A/I-E⁺), and alveolar macrophages (CD11c^highCD170^high) are included in this analysis. Cell surface markers used in these analyses: CD11b, CD11c, CD73, CD80, CD170, CX3CR1, F4/80, FcRγII/III, I-A/I-E, and Ly6C. All data shown are representative of two independent experiments, each carried out with pooled embryos [n = 6 or 17]

The functional competency of the macrophages developed from CPDPs was examined and compared to that of MRP-derived and adult bone marrow (BM)-derived macrophages. For this purpose, a phagocytosis assay was performed using pHrodo-conjugated Escherichia coli. CPDPs showed comparable phagocytosis capability as MRP- and BM-derived macrophages based on the percentage of cells that phagocytosed bacterial particles (Fig. 4A, B) and fluorescence intensities derived from phagocytosed bacterial particles (Fig. 4C).

Fig. 4.

Functional phagocytic capability and osteoclast differentiation potency of CPDPs. A–C CPDP-derived, MRP-derived, or adult BM-derived macrophages were analyzed for phagocytosis capability for pHrodo E. coli. Fluorescence derived from phagocytosed particles were detected by flow cytometry. Cytochalasin D (cytoD), which inhibits phagocytosis, was added where indicated (A, B). Percentage of pHrodo^bright cells (A, B) and mean fluorescence intensity (MFI; C) of the cells pre-gated on F4/80⁺ cells are shown. Data are shown as mean ± SD of triplicate wells; n.s. denotes not significant. Data shown are representative of two independent experiments, each carried out with pooled embryos [n = 8 or 17]. D, E CPDP-derived, MRP-derived, or adult BM-derived macrophages were induced to differentiate into osteoclasts. Cultured cells were stained for TRAP activity, and TRAP⁺ cells were counted. Cell cultures were initiated with 5000 cells/well. Data are shown as mean ± SD of triplicate wells; n.s. denotes not significant (D). Representative image of osteoclasts obtained are shown in E. Scale Bar: 100 μm. All data shown are representative of three independent experiments, each carried out with pooled embryos [n = 8, 10 or 17]

YS hematopoietic cells are also thought to contribute to osteoclast lineage cells [22, 23]. To examine whether CPDPs generate osteoclasts, CPDP-derived macrophages were induced to differentiate into osteoclasts in the presence of M-CSF and RANKL and were compared to MRP-derived and adult BM-derived macrophages for its capability. As shown in Fig. 4D, CPDPs generated TRAP⁺ osteoclasts that were comparable to MRP-derived macrophages. The CPDPs matured into TRAP⁺ cells with a multinucleated, flattened phenotype, which is a feature of mature osteoclasts (Fig. 4E). Macrophages derived from CPDPs and MRPs generated osteoclasts more efficiently than BM-derived macrophages (Fig. 4D). However, the data did not necessarily show the superiority of YS-derived cells over adult BM cells in generating osteoclasts, because BM progenitors for osteoclasts are mainly found in KIT⁺ hematopoietic cells, rather than in the CSF1R⁺ cell fraction [17].

PU.1 promotes, and GATA1 inhibits formation of CSF1R⁺ progenitors in the YS

Generation of CSF1R⁺ progenitors from CPDPs prompted the examination of whether reciprocal repression between GATA1 and PU.1 was present when CSF1R⁺ progenitors appeared in the YS [24–27]. This is due to it previously being shown that PU.1 and GATA1 promoted generation of CD45⁺ lineages and primitive erythroid cells, respectively, at the expense of the generation of the other lineage [16]. First, the expression of PU.1 and GATA1 were compared between E8.5 CPDP and E9.5 CSF1R⁺ cells. As shown in Fig. 5A, PU.1 was upregulated approximately 25-fold in E9.5 CSF1R⁺ cells compared to E8.5 CPDPs. In contrast, GATA1 expression in E9.5 CSF1R⁺ cells was 300-fold lower than that in E8.5 CPDPs. These results imply that regulation of these two transcription factors is critical for the generation of CSF1R⁺ macrophage progenitors. Therefore, CPDPs were isolated from the E9.5 YS and then infected with retrovirus vectors containing one of the following: empty control, PU.1, or GATA1. The cells were analyzed after 2 days of cell culture. Transduction of PU.1 induced the enrichment of CSF1R⁺ cells (Fig. 5B, C). In contrast, transduction of GATA1 strongly inhibited the generation of CSF1R⁺ cells (Fig. 5B, C). These results suggested that expression of PU.1 promotes the generation of CSF1R⁺ progenitors at the expense of primitive erythroid cells and that continued expression of GATA1 in the YS progenitors hinders the development of CSF1R⁺ progenitors. Reciprocal antagonism between PU.1 and GATA1 might function when CSF1R⁺ macrophage progenitors and primitive erythroid cells appear in the early YS.

Fig. 5.

PU.1 promotes, and GATA1 inhibits CSF1R⁺ cell formation. A E8.5 YS CPDPs and E9.5 YS CD45⁺CSF1R⁺ cells were analyzed for the gene expression of the Pu.1 and Gata1 genes by real-time quantitative PCR. Gene expression levels are shown after internal normalization with Hprt and Polr2a relative to E8.5 CPDPs (value = 1). Values are expressed as mean ± SD of triplicate wells. A representative result from two independent cell experiments is shown. B, C CPDPs isolated from E9.5 YS were infected with retrovirus containing either Pu.1-Ires-Egfp, Gata1-Ires-Egfp, or Ires-Egfp (control). Cells were analyzed after having been cultured for two days. Representative flow cytometry plots are shown in B. Percentage of CSF1R⁺ cells within EGFP⁺ cells after transduction of the indicated genes is shown in C. Data are shown as mean ± SD of four separate experiments, each performed with pooled embryos (n = 7, 7, 13, or 22). D A model for generation of macrophage progenitors between E8 and E9 in YS.

Discussion

In this study, CSF1R⁺ macrophage progenitors were found as early as E9 in the YS. Then it was shown that CPDPs, the earliest committed hematopoietic progenitors, were the earliest source of CSF1R⁺ macrophage progenitors. The YS progenitors for tissue-resident macrophages are thought to be EMPs. The term EMP is used for cells that give rise to definitive erythrocytes and the granulocyte/macrophage lineage, but not to primitive erythrocytes and lymphoid lineages [10, 11]. The properties of CPDPs are obviously different from those of EMPs because they can generate primitive erythroid cells and lymphocytes in addition to myeloid cell lineages and definitive erythroid cells. On the other hand, EMP cells are the closest to CD45⁺KIT^highAA4.1^– MRPs (these cells are also positive for CD41), as these cells cannot give rise to primitive erythroid cells and lymphocytes [13]. The results obtained in this study showed that CPDPs, rather than EMPs, were more likely to be the first-wave source of embryonic macrophages. This was also supported by the evidence that CPDPs were present in E8 YSs, while EMP-like MRPs were only observed in E9 YSs [12, 15]. Furthermore, on E9, the first CSF1R⁺ macrophage progenitors were already present.

The E9 YS contains CD45⁺KIT^highAA4.1⁺ multipotent lympho-myeloid progenitors which lack the potential to generate primitive erythroid cells, in addition to CPDPs and the above-mentioned MRPs [12, 13]. All these hematopoietic progenitors could generate CSF1R⁺ macrophage progenitors directly. Therefore, multiple hematopoietic progenitors had the potential to give rise to CSF1R⁺ macrophage progenitors in E9 YSs. This is in contrast to the observation that precursors for CSF1R⁺ cells were limited to CPDPs in the E8 YS, just one day earlier. This timeline is summarized in Fig. 5D.

Mechanistically, enforced expression of PU.1 in CPDPs promoted the generation of CSF1R⁺ macrophage progenitors, while the expression of GATA1 strongly inhibited this process. Previously, it was shown that CPDPs express modest levels of erythroid transcription factors, including GATA1, SCL, and KLF1 [16]. Moreover, the downregulation of the primitive erythroid program by PU.1 is critical for the generation of multipotent lympho-myeloid progenitors [16]. The results of this study suggest that the same mechanism may be involved in the generation of CSF1R⁺ macrophage progenitors in the early YS. In the experiments shown in Fig. 5B, non-transduced (EGFP^–) cells did not efficiently differentiate into CSF1R⁺ cells compared to gene-transduced (EGFP⁺) cells even when empty control vectors were used. These findings suggest that efficient differentiation of CPDPs into CSF1R⁺ cells is accompanied by cell proliferation, because retrovirus vectors, which only infect proliferating cells, were used in the experiments. However, this hypothesis requires further examination.

CPDPs generated functional macrophages of tissue-resident phenotypes and osteoclasts in vitro as efficiently as MRPs. Although the initial progenitors for tissue-resident macrophages and osteoclasts are thought to be the EMPs, the data shown in this study suggest that CPDPs are the earliest tissue-resident macrophage and osteoclast progenitors, before EMP-like MRPs appear in the yolk sac. The abundance of osteoclast precursors in the yolk sac as early as E8, reported in a classic study, also supports this view, because EMP-like MRPs are absent at this stage [28].

Hematopoietic cells colonize the hepatic primordia at E10-11 [29, 30]. At this stage, the vitelline vein from the YS drains directly into the fetal liver [30, 31]. Therefore, it is likely that the earliest CSF1R⁺ macrophage progenitors generated in the YS spread through the fetal liver and establish the first-waves of macrophage and osteoclast production in the embryonic body.

Materials and methods

Mice

C57BL/6JJcl mice were purchased from Crea Japan, Inc. All animal experiments were performed at the Institute of Laboratory Animals at Mie University, according to the Mie University guidelines for laboratory animals.

Cell preparation from embryos

The E8.5–9.5 YSs obtained from C57BL/6J females, time-mated with C57BL/6J males, were incubated with 1 mg/mL collagenase (Wako) in 2% FBS/HBSS for 30 min at 37 °C. A single-cell suspension was used for the experiments. This was prepared by pipetting after incubation.

Antibodies

The following antibodies were used in this study: AA4.1 (BioLegend), CD11b (M1/70, BioLegend), CD11c (N418, BioLegend), CD41 (MWReg30, BioLegend), CD45 (30-F11, BioLegend), CD73 (TY/11.8 eBioscience), CD80 (16-10A1, BioLegend), CD170 (1RNM44N, eBioscience), CX3CR1 (SA011F11, BioLegend), CSF1R (AFS98, BioLegend), c-KIT (2B8, BioLegend), I-A/I-E (M5/114.15.2, BioLegend), F4/80 (BM8.1, TONBO), FcRγII/III (93, BioLegend), Ly6C (HK1.4, BioLegend), and TER119 (BioLegend).

Cell sorting and analysis

Cell sorting experiments were performed using FACSAria or FACSAria Fusion (BD). Cells were analyzed using FACSAria, FACSAria Fusion, or FACSCanto II (BD). Subsequent analysis was performed using FlowJo software (Tree Star).

Cell culture

For the induction of CSF1R⁺ cells, cells were cultured in MEM Alpha Medium supplemented with 20% FBS, as well as 10 ng/mL of rmIL-3 (Gibco), and rmGM-CSF (Invitrogen). These were then analyzed 1 or 2 days later. To obtain mature macrophages, cells were cultured in MEM Alpha Medium supplemented with 10% FBS and 20 ng/mL rmM-CSF (R&D). For the first 3 days, media was also supplemented with 10 ng/mL rmGM-CSF (Invitrogen). Cells were analyzed 5 days or 2 weeks later. To induce osteoclastic differentiation, the cells were cultured in the presence of 50 ng/mL rmM-CSF and 20 ng/mL RANKL (R&D) for 10 days. After cell culture, cells were subjected to TRAP staining [32]. Phagocytosis assays were performed using pHrodo Red E. coli. bioparticles conjugate (Molecular Probes) according to the manufacturer’s instructions. Cytochalasin D (10 μM, Wako) was added to the culture 5 h before and during the assay.

Real-time quantitative PCR

Total RNA was isolated using TRIzol reagent (Invitrogen). Reverse transcription reactions were performed using ReverTra Ace qPCR RT Master Mix with gDNA remover (Toyobo). First-strand synthesized cDNA was used for real-time quantitative PCR using the Fast SYBR and StepOnePlus systems (Applied Biosystems). Primer sequences used in this study are listed in Supplementary Table 1.

Retroviral vector production and cell transduction

Retroviral vectors used in this study were described in ref no [16]. Briefly, open reading frames obtained from cDNA libraries of C57BL/6J bone marrow were cloned into pMSCV-IRES-EGFP vectors. RetroNectin-treated dishes (Takara Bio) were coated with virus supernatants produced in a packaging cell line. Virus-coated dishes were used for transduction of genes into YS cells. Cells were cultured in the same dishes in MEM Alpha Medium supplemented with 20% FBS, as well as 10 ng/mL rmIL-3 (Gibco), rmGM-CSF (Invitrogen), and rmEpo (R&D Systems). These were then analyzed two days later.

Statistical analysis

Statistical significance was assessed by Student's t-test.

Supplementary Information

Below is the link to the electronic supplementary material.

Abbreviations

YS

Yolk sac

E

Embryonic day

EMP

Erythroid-myeloid progenitor

CPDP

Common primitive-definitive precursor

MRP

Myeloid-restricted progenitor

BM

Bone marrow

Author contributions

TY conceived the experiments; TY and CI designed and performed the experiments; TY and CI analyzed the data; TY, CI, and MHK contributed to conceptualization; HY contributed analysis tools; and TY wrote the paper with input from CI and MHK.

Funding

This work was supported by the JSPS KAKENHI (no. 17K09950) to TY and the Takeda Science Foundation to TY.

Availability of data and materials

Data sharing not applicable to this article as no datasets were generated or analyzed during the current study.

Declarations

Conflict of interest

The authors declare no commercial or financial conflict of interest.

Ethics approval and consent to participate

All animal experiments are approved by the institutional review board of Mie University (permission nos. 20–22 and 2021–20). This article does not contain any studies with human participants.

Consent for publication

All authors consent for the publication of this study.