Introduction

Monocytes are integral components of the mononuclear phagocyte system (MPS) that develop from monocyte precursors in the bone marrow1. Several functionally and phenotypically distinct subsets of blood monocytes have been defined on the basis of expression of surface markers2-5. In mice, Ly6Chi monocytes (also called inflammatory or classical monocytes) characterized by Ly6ChiCCR2hiCX3CR1lo exhibit a short half-life, and are recruited to tissue and differentiate into macrophages and dendritic cells6. Ly6Chi monocytes are analogous to CD14+ human monocytes based on gene expression profiling7. A second subtype of monocytes is called Ly6Clo monocytes (also termed patrolling monocytes or non-classical monocytes), which express high level of CX3CR1 and low level of CCR2, equivalent to CD14loCD16+ human monocytes. Ly6Clo monocytes are regarded as blood-resident macrophages that patrol blood vessels and scavenge microparticles attached to the endothelium under physiological conditions, and exhibit a long half-life in the steady-state8-10. Ly6Clo monocytes may exert a protective effect by suppressing atherosclerosis in mice, while high levels of nonclassical monocytes have been associated with more advanced vascular dysfunction and oxidative stress in patients with coronary artery disease11-13. In the inflammatory settings, Ly6Clo monocytes often show anti-inflammatory properties, yet also exhibit a proinflammatory role under certain circumstances5,9,14-18. In addition, Ly6Clo monocytes have demonstrated protective functions in tumorigenesis, such as engulfing tumor materials to prevent cancer metastasis, activating and recruiting NK cells to the lungs19,20. While the functions of Ly6Clo monocytes are complex and sometimes contradictory depending on the animal models and experimental approaches, it is clear that they play a crucial role in both health and disease.

The generation and maintenance of Ly6Clo monocytes are regulated by several transcription factors, including Nr4a1 and CEBPβ21,22. The colony-stimulating factor 1 receptor (CSF1R) signaling pathway is important for the generation and survival of Ly6Clo monocytes, and neutralizing antibodies against CSF1R can reduce their numbers23. Interestingly, recent studies have suggested that different subsets of monocytes may be supported by distinct cellular sources of CSF-1 within bone marrow niches, in which targeted deletion of Csf1 from sinusoidal endothelial cells selectively reduces Ly6Clo monocytes but not Ly6Chi monocytes24. LAIR1 has been shown to be activated by stromal protein Colec12, and LAIR1 deficiency leads to aberrant proliferation and apoptosis in bone marrow non-classical monocytes25. These findings further underscore the complexity of the mechanisms that regulate Ly6Clo monocyte, and suggest that multiple factors are involved in this process.

Recombination signal-binding protein for immunoglobulin kappa J region (RBP-J; also named CSL in humans) is commonly known as the master nuclear mediator of canonical Notch signaling26. In the absence of Notch intracellular domain, RBP-J associates with co-repressor proteins to repress transcription of downstream target genes. In the immune system, the best studied functions of Notch signaling are its roles in regulating the development of lymphocytes, such as T cells and marginal zone B cells26. Notch signaling also has been reported to regulate the differentiation and function of myeloid cells including granulocyte/monocyte progenitors, osteoclasts and dendritic cells 27-30. In osteoclasts, RBP-J represses osteoclastogenesis, particularly under inflammatory conditions30. In dendritic cells, Notch-RBP-J signaling controls the maintenance of CD8- DC28. In addition, Notch2-RBP-J pathway is essential for development of CD8-ESAM+ DC in the spleen and CD103+CD11b+ DC in the lamina propria of the intestine29. However, the role of Notch-RBP-J signaling pathway in regulating monocyte subsets remains elusive. Here, we demonstrated that RBP-J controlled homeostasis of blood Ly6Clo monocytes in a cell intrinsic manner. RBP-J deficiency led to a drastic increase in blood Ly6Clo monocytes, lung Ly6Clo monocytes and CD16.2+ IM. Our results suggest that RBP-J plays a critical role in regulating the population and characteristics of Ly6Clo monocytes in the blood and lung.

Results

RBP-J is essential for maintenance of blood Ly6Clo monocytes

To investigate the role of RBP-J in monocyte subsets, we utilized mice with RBP-J specific deletion in the myeloid cells (Rbpjfl/flLyz2cre/cre mice). Efficient deletion of Rbpj in blood monocytes was confirmed by quantitative real-time PCR (qPCR) (Figure 1 - figure supplement 1A). Flow cytometric analysis revealed that RBP-J deficient mice had a significant increase in the proportion of Ly6Clo monocytes, but not Ly6Chi monocytes, in blood, compared to age-matched control mice with the genotype Rbpj+/+Lyz2cre/cre (Figure 1A). In contrast to circulating monocytes, RBP-J deficient mice exhibited minimal alterations in the percentages of monocyte subsets in bone marrow (BM) and spleen (Figure 1B; Figure 1 - figure supplement 1C). Additionally, RBP-J deficiency in myeloid cells did not appear to have an effect on neutrophils (Figure 1 - figure supplement 1B). Therefore, among circulating myeloid populations, RBP-J selectively controlled the subset of Ly6Clo monocytes.

RBP-J deficient mice display more Ly6Clo blood monocytes.

(A) Blood Ly6Chi and Ly6Clo monocytes in Rbpj+/+Lyz2cre/cre control and Rbpjfl/flLyz2cre/cre mice were determined by flow cytometry analyses (FACS). Representative FACS plots (left) and cumulative data of cell ratio (right) are shown.

(B and C) Representative FACS plots and cumulative data quantitating percentages of BM monocyte subsets (B) and myeloid progenitor cells (C) (CD45+CD11b+Ly6G- CD115+ Ly6Chi monocyte; CD45+CD11b+Ly6G-CD115+Ly6Clo monocyte; MDP, CD45+Lin-CD117+CD115+CD135+ Ly6C-; cMoP, CD45+Lin-CD11b-CD117+CD115+ CD135-Ly6C+). Lin: CD3, B220, Ter119, and Gr-1, CD11b.

(D) Experimental outline for panel (E).

(E) Cumulative data quantitating percentages of sinusoidal monocytes (CD45+) within total BM Ly6Clo monocytes.

Data are pooled from at least 2 independent experiments; n ≥ 4 in each group. Data are shown as mean ± SEM; n.s., not significant; **, P < 0.01 (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse.

BM progenitors that give rise to circulating monocytes are monocyte-dendritic cell progenitors (MDPs) and common monocyte progenitors (cMoPs)21,31-33. Next, we analyzed BM progenitors, and found that the percentages of MDPs and cMoPs were equivalent in control and RBP-J deficient mice (Figure 1C). These results in conjunction with the observations of normal BM monocyte populations suggest that alterations of peripheral blood Ly6Clo monocytes may not originate from BM. Next, we examined whether RBP-J may influence the egress of Ly6Clo monocytes from BM, and measured BM exit rate of Ly6Clo monocytes using in vivo labeling of sinusoidal cells as previously described (Figure 1D)34. The results showed that RBP-J deficiency did not affect egress of Ly6Clo monocytes from BM (Figure 1E). Taken together, these data revealed RBP-J as a critical regulator controlling homeostasis of peripheral blood Ly6Clo monocytes.

RBP-J deficiency does not affect Ly6Chi monocyte conversion or Ly6Clo monocyte survival and proliferation

Next, we aimed to determine whether the increase of blood Ly6Clo monocytes in RBP-J deficient mice was due to decreased cell death or enhanced proliferation. We first stained monocytes with Annexin V and 7-AAD to identify apoptotic cells and observed comparable percentages of apoptotic blood Ly6Clo monocytes in control and RBP-J deficient mice (Figure 2A). Given the crucial role of Nr4a1 in the survival of Ly6Clo monocytes21, we detected the expression of Nr4a1, which was similar in Ly6Clo monocytes from control and RBP-J deficient mice (Figure 2 - figure supplement 1A). We then assessed the proliferative capacity by analyzing Ki-67 expression as well as in vivo EdU incorporation. Ki-67 levels in blood monocytes displayed no differences between control and RBP-J deficient mice (Figure 2B). The percentage of EdU+ monocytes did not significantly differ between control and RBP-J deficient mice at the indicated time points, implying that the turnover of Ly6Clo monocytes was normal in RBP-J deficient mice (Figure 2C). Fluorescent latex beads as particulate tracers could be phagocytosed by monocytes after intravenous injection, and stably label Ly6Clo monocytes35. Thus, we intravenously injected fluorescent latex beads into control and RBP-J deficient mice to track circulating monocytes (Figure 2D). By day 7 post injection, only Ly6Clo monocytes were latex+ as previously reported35, whereas control and RBP-J deficient mice groups presented a similar frequency of latex+ monocytes (Figure 2E). Together, these data indicated that RBP-J did not influence monocyte survival and proliferation.

Monocyte subsets in RBP-J deficient mice display normal cell death.

(A) Representative FACS plots of monocyte subsets in blood stained with 7-AAD and Annexin V.

(B) FACS analysis of Ki-67 expression in Rbpj+/+Lyz2cre/cre control and Rbpjfl/flLyz2cre/creblood monocyte subsets. Black lines represent control mice and red lines represent RBP-J deficient mice.

(C) Analysis of time course of EdU incorporation of monocyte subsets in BM and blood after a single 1mg EdU pulsing. The percentages of EdU+ cells among the indicated monocyte subsets are shown.

(D) Experimental outline for panel (E).

(E) Analysis of time course of latex beads incorporation of monocyte subsets in blood after Latex beads injection. The percentages of latex+ cells among the indicated monocyte subsets are shown.

(F) Cartoon depicting the adoptive transfer. BM GFP+Ly6Chi monocytes were sorted from Lyz2cre/creCx3cr1gfp/+or Rbpjfl/flLyz2cre/creCx3cr1gfp/+mice and transferred into Ccr2RFP/RFP recipient mice. Sixty hours after transfer, cell fate was analyzed.

(G) Representative FACS plots are shown in the left panel, and the frequencies of GFP+ Ly6Clo monocytes within total Ly6Clo monocytes are shown in the right panel.

Data are pooled from 2 independent experiments (G); n ≥ 2 in each group (C, E and G). Data are shown as mean ± SEM; n.s., not significant; (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse (E and G).

Previous studies have shown that Ly6Clo monocytes are observed in recipient mice following the adoptive transfer of Ly6Chi monocytes, indicating that Ly6Chi monocytes can convert into Ly6Clo monocytes10,31. We next wished to evaluate whether conversion of Ly6Chi monocytes was regulated by RBP-J by isolating bone marrow GFP+Ly6Chi monocytes from Lyz2cre/creCx3cr1gfp/+control or Rbpjfl/flLyz2cre/creCx3cr1gfp/+mice and adoptively transferring them into Ccr2RFP/RFP recipients (Figure 2F). Sixty hours after transfer, a subset of cells from donors were converted into Ly6Clo monocytes (Figure 2 - figure supplement 1B), and equal percentages of Ly6Clo monocytes were derived from control and Rbpjfl/flLyz2cre/creCx3cr1gfp/+donors (Figure 2G). Thus, the conversion of Ly6Chi monocytes into Ly6Clo monocytes was not affected by RBP-J deficiency.

RBP-J regulates blood Ly6Clo monocytes in a cell intrinsic manner

We next wondered whether the increase of Ly6Clo monocytes in RBP-J deficient mice was bone marrow-derived and cell intrinsic. We performed bone marrow transplantation by engrafting lethally irradiated mice with a 1:4 mixture of Rbpj+/+Lyz2cre/cre control and Rbpjfl/flLyz2cre/cre BM cells (CD45.2) and Cx3cr1gfp/+ BM cells (CD45.1) (Figure 3A). Eight weeks after transplantation, we analyzed the frequencies of CD45.2+ donor cells in the bone marrow and blood of recipient mice and found that the frequencies of CD45.2+ cells within total cells were similar to the mixture ratio of bone marrow cells for Ly6Chi monocytes and neutrophils in both BM and blood (Figure 3B; Figure 3 - figure supplement 1A-B). Specifically, more blood Ly6Clo monocytes were derived from RBP-J deficient donors than control donors (Figure 3B), reflecting that the contribution of RBP-J deficient cells in blood Ly6Clo cells was significantly higher than that of control cells. These results implied that RBP-J deficient BM cells were highly efficient in generating blood Ly6Clo monocytes.

The role of RBP-J in blood Ly6Clo monocytes is cell intrinsic.

(A) Cartoon depicting the bone marrow transplantation. BM recipient 6-week-old male C57BL/6 mice (CD45.2) were lethally irradiated. BM cells from donor mice (Rbpj+/+Lyz2cre/cre or Rbpjfl/flLyz2cre/cre and CD45.1, Cx3cr1gfp/+) were collected and transferred into recipient mice. Mice were used after 8 weeks of bone marrow reconstitution.

(B) Representative FACS plots (left) and cumulative data (right) quantitating the frequency of Rbpj+/+Lyz2cre/cre and Rbpjfl/flLyz2cre/cre donor cells among Ly6Chi and Ly6Clo monocytes in the blood of recipient mice.

(C) Cartoon depicting the generation of Rbpj+/+Lyz2cre/cre control or Rbpjfl/flLyz2cre/cre and Cx3cr1gfp/+ parabiotic pairs.

(D) Representative FACS plots (left) and cumulative data (right) quantitating percentages of monocyte subsets derived from control or RBP-J deficient mice in control or RBP-J deficient mice.

(E) Representative FACS plots (left) and cumulative data (right) quantitating percentages of monocyte subsets derived from control or RBP-J deficient mice in Cx3cr1gfp/+ mice.

Data are pooled from at least 2 independent experiments; n ≥ 6 in each group. Data are shown as mean ± SEM; n.s., not significant; ***, P < 0.001; ****, P < 0.0001 (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse. SSC-A, side scatter area.

Given that RBP-J deficient mice had more Ly6Clo monocytes in blood than did control mice, we performed parabiosis experiments, a surgical union of two organisms, which allowed parabiotic mice to share their blood circulation. The contribution of circulating cells from one animal in another can be estimated by measuring the percentage of blood cells that originated from each animal36. We joined a CD45.1, Cx3cr1gfp/+ mouse with an age and sex-matched Rbpj+/+Lyz2cre/cre control or Rbpjfl/flLyz2cre/cre mouse (CD45.2) (Figure 3C). Four weeks after the procedure, about 50% of B cells and T cells in parabiotic mice displayed efficient exchange of their circulation (Figure 3 - figure supplement 1C). As expected, RBP-J deficient mice exhibited higher percentages of Ly6Clo monocytes than control animals (Figure 3D). Intriguingly, in the parabiotic Cx3cr1gfp/+mice, RBP-J deficient cells still constituted significantly higher proportion of circulating Ly6Clo, but not Ly6Chi monocytes, than RBP-J sufficient counterparts as indicated by the percentages of the GFP-negative population (Figure 3E). These results implicated that the enhanced ability of Ly6Clo monocytes to circulate in the peripheral blood as a result of RBP-J deficiency was cell intrinsic.

RBP-J regulates phenotypical marker genes in blood Ly6Clo monocytes

To further study the consequences of RBP-J loss of function in blood Ly6Clo monocytes, we performed gene expression profiling by RNA-seq, which revealed that Ly6Clo monocytes in RBP-J deficient mice exhibited low expression of Itgax but high expression of Ccr2, in comparison to those in control mice (Figure 4A). These differential expression patterns of phenotypic marker genes were also confirmed by qPCR and flow cytometry analysis (Figure 4B). However, Ly6Chi blood monocytes and BM monocytes displayed normal levels of CD11c and CCR2 (Figure 4B; Figure 4 - figure supplement 1A). Moreover, principal component analysis showed distinct expression pattern of monocytes between control and RBP-J deficient mice (Figure 4C). To determine whether the regulation of phenotypic markers by RBP-J in blood Ly6Clo monocytes was cell-intrinsic, we performed BM transplantation as described above (Figure 3A). The findings indicated that Ly6Clo monocytes derived from RBP-J deficient BM cells expressed high levels of CCR2 but low levels of CD11c, in comparison to those derived from control BM cells, whereas no significant changes were observed in blood Ly6Chi and BM monocytes, as expected (Figure 4D; Figure 4 - figure supplement 1B). These results collectively suggested that RBP-J regulated the expression of CCR2/CD11c in a cell-intrinsic manner.

Phenotypic markers in Ly6Clo monocytes are changed in RBP-J deficient mice.

(A) Heatmap of RNA-seq dataset showing the top 20 downregulated and upregulated genes in blood Ly6Clo monocytes from Rbpjfl/flLyz2cre/cre versus Rbpj+/+Lyz2cre/cre control mice. Blue and red font indicates downregulated and upregulated genes in RBP-J deficient Ly6Clo monocytes respectively.

(B) Representative FACS plots, cumulative MFI, and qPCR analysis of CCR2/CD11c expression in control and RBP-J deficient blood monocyte subsets. Shaded curves represent isotype control, black lines represent control mice and red lines represent RBP-J deficient mice.

(C) PCA of indicated cell types.

(D) Representative FACS plots and cumulative MFI of CCR2/CD11c expression in blood monocyte subsets derived from control or RBP-J deficient mice. Black lines represent control mice, and red lines represent RBP-J deficient mice.

Data are pooled from 2 independent experiments (B and D); n ≥ 6 in each group. Data are shown as mean ± SEM; n.s., not significant; **, P < 0.01; ****, P < 0.0001 (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse.

RBP-J-mediated control of blood Ly6Clo monocytes is CCR2 dependent

Given that RBP-J deficiency led to enhanced CCR2 expression in Ly6Clo monocytes and that CCR2 is essential for monocyte functionality under various inflammatory and non-inflammatory conditions37, we wished to genetically investigate the role of CCR2 in the RBP-J deficient background. We generated RBP-J/CCR2 double deficient (DKO) mice with the genotype Rbpjfl/flLyz2cre/creCcr2RFP/RFPwith Lyz2cre/creCcr2RFP/+ and Rbpjfl/flLyz2cre/creCcr2RFP/+mice serving as controls. As expected, the expression of CCR2 in Ly6Clo monocytes was reduced in DKO mice compared to RBP-J deficient mice (Figure 5A), confirming the successful deletion of the Ccr2 gene. DKO mice showed a lower percentage of both Ly6Chi and Ly6Clo monocytes than RBP-J deficient mice (Figure 5B-C). Notably, the percentage of Ly6Clo monocytes in DKO mice was comparable to that observed in the control mice (Figure 5B-C), implicating that deletion of CCR2 corrected RBP-J deficiency-associated phenotype of increased Ly6Clo monocytes. Whereas, both RBP-J deficient and DKO mice exhibited lower expression levels of CD11c in their Ly6Clo monocyte than control mice, suggesting that the regulation of CD11c expression was independent of CCR2 (Figure 5D). These results suggested that RBP-J regulated Ly6Clo monocytes, at least in part, through CCR2.

Blood Ly6Clo monocytes are decreased in DKO mice.

(A) Representative FACS plots and cumulative MFI of CCR2 expression in Lyz2cre/creCcr2RFP/+ control, Rbpjfl/flLyz2cre/creCcr2RFP/+ and Rbpjfl/flLyz2cre/creCcr2RFP/RFP (DKO) blood Ly6Chi and Ly6Clo monocytes are shown.

(B and C) Blood monocyte subsets in control, RBP-J deficient and DKO mice were determined by FACS. Representative FACS plots (B) and cumulative data of cell ratio

(C) are shown.

(D) Representative FACS plots and cumulative MFI of CD11c expression in control, RBP-J deficient and DKO blood Ly6Chi and Ly6Clo monocytes are shown.

Data are pooled from at least 2 independent experiments; n ≥ 5 in each group. Data are shown as mean ± SEM; n.s., not significant; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse.

Lung Ly6Clo monocytes are accumulated in RBP-J deficient mice

In mice, alveolar macrophages (AM) are maintained by local self-renewal, and arise from fetal liver derived precursors, whereas interstitial macrophages (IM) are probably originated from monocytes 38-40. Joey et al identified 3 subpopulations of IM according to expression of CD206 and CD16.2 and Ly6Clo monocytes are proposed to give rise to CD64+CD16.2+ IM41. We next compared the populations of lung monocytes and IM between control and RBP-J deficient mice. The conditional deletion of RBP-J resulted in a significant increase in the absolute and relative numbers of Ly6Clo monocytes and CD16.2+ IM, while the numbers of Ly6Chi monocytes, CD206- IM, and CD206+ IM remained unchanged (Figure 6A-B). To confirm these findings, lung sections from Lyz2cre/creCx3cr1gfp/+ control and Rbpjfl/flLyz2cre/creCx3cr1gfp/+ mice were stained with an anti-GFP antibody, which revealed a significant increase in the number of GFP+ cells in RBP-J deficient mice (Figure 6C). In addition, we evaluated the proliferative capacity of Ly6Clo monocytes and CD16.2+ IM, but did not observe enhanced EdU incorporation in Ly6Clo monocytes and CD16.2+ IM from RBP-J deficient mice (Figure 6 - figure supplement 1A-B). These findings suggested that the augmented population of these cells might not be resulted from increased in situ proliferation. Previous reports have indicated that LPS can increase the population of IM in a CCR2-dependent manner40. We thus challenged mice with LPS, and analyzed monocytes and IM at day 0 and day 4. LPS exposure induced an increase in numbers of Ly6Clo monocytes, CD16.2+ IM and CD206- IM compared to baseline both in control and RBP-J deficient mice (Figure 6D-E). There was a trend toward higher Ly6Chi monocyte after LPS treatment, although this observation was not statistically significant. Of note, RBP-J deficient mice exhibited robustly elevated numbers of Ly6Clo monocytes and CD16.2+ IM compared with control mice after LPS treatment (Figure 6D-E), suggesting that RBP-J deficient Ly6Clo monocytes were recruited to inflamed tissue in large numbers and differentiated into CD16.2+ IM.

RBP-J deficient mice exhibit more lung Ly6Clo monocytes and CD16.2+ IM.

(A and B) Indicated populations in the lungs of Rbpj+/+Lyz2cre/cre and Rbpjfl/flLyz2cre/cre mice were determined by FACS. Representative FACS plots (A) and cumulative data of cell ratio and absolute numbers (B) are shown.

(C) Immunofluorescence staining for GFP+ cells in the lung from Lyz2cre/creCx3cr1gfp/+and Rbpjfl/flLyz2cre/creCx3cr1gfp/+mice (CX3CR1 [green]; DAPI [blue]). Scale bars represent 50 µm.

(D and E) Rbpj+/+Lyz2cre/cre and Rbpjfl/flLyz2cre/cre mice were instilled intranasally with PBS or PBS containing LPS, and lungs were harvested at the indicated time points. Representative FACS plots (D) and cumulative data of cell ratio and absolute numbers

(E) are shown.

Data are pooled from at least 2 independent experiments; n ≥ 4 in each group. Data are shown as mean ± SEM; n.s., not significant; *, P < 0.05; **, P < 0.01; ****, P < 0.0001 (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse.

To investigate whether the elevated levels of lung monocytes and CD16.2+ IM in RBP-J deficient mice were derived from blood Ly6Clo monocytes, we examined monocytes and IM in Lyz2cre/creCcr2RFP/+control, Lyz2cre/creCcr2RFP/RFP, Rbpjfl/flLyz2cre/creCcr2RFP/+, and Rbpjfl/flLyz2cre/creCcr2RFP/RFP(DKO) mice. The deletion of CCR2 led to a decrease in Ly6Chi monocytes, and the Ly6Clo monocytes in DKO mice were reduced to a level similar to that of control mice (Figure 7A, B). While the DKO mice had more CD16.2+ IM compared to control mice, these cells were severely reduced compared to RBP-J deficient mice (Figure 7A, C). The above data supported the notion that increased lung Ly6Clo monocytes and CD16.2+ IM in RBP-J deficient mice were derived from Ly6Clo blood monocytes. In summary, these results suggested that in RBP-J deficient mice, recruitment of blood Ly6Clo monocytes to the lung was markedly facilitated by increased cell number as well as heightened expression of CCR2, leading to the increase of lung Ly6Clo monocytes and CD16.2+ IM.

DKO mice lack lung Ly6Clo monocytes and CD16.2+ IM.

(A) Representative FACS plots of lung monocyte and IM subsets in Lyz2cre/creCcr2RFP/+, Rbpjfl/flLyz2cre/creCcr2RFP/+, Lyz2cre/creCcr2RFP/RFP and Rbpjfl/flLyz2cre/creCcr2RFP/RFP mice.

(B and C) Cumulative data of cell ratio and absolute numbers of monocyte (B) and IM (C) subsets.

Data are pooled from 2 independent experiments; n ≥ 4 in each group. Data are shown as mean ± SEM; n.s., not significant; *, P < 0.05; **, P < 0.01; ***, P < 0.001; ****, P < 0.0001 (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse.

Discussion

In this study, we identified RBP-J as a pivotal factor in regulating blood Ly6Clo monocyte cell fate. Our results showed that mice with conditional deletion of RBP-J in myeloid cells exhibited a robust increase in blood Ly6Clo monocytes, and subsequently accumulated lung Ly6Clo monocytes and CD16.2+ IM under steady-state conditions. Bone marrow transplantation experiments in which RBP-J deficient cells were transplanted into recipient mice as well as the parabiosis experiment showed similar increase in blood Ly6Clo monocytes, demonstrating that RBP-J was an intrinsic factor required for the regulation of Ly6Clo monocytes. Further analysis revealed that RBP-J regulated the expression of CCR2 and CD11c on the surface of Ly6Clo monocytes. Moreover, the phenotype of elevated blood Ly6Clo monocytes and their progeny was further ameliorated by deleting Ccr2 in an RBP-J deficient background.

Notch-RBP-J signaling have been shown to play a role in regulating functional polarization and activation of macrophages, and regulated formation of Kupffer cells and macrophage differentiation from Ly6Chi monocytes in ischemia42-48. At steady-state, interaction of DLL1 with Notch 2 regulates conversion of Ly6Chi monocytes into Ly6Clo monocytes in special niches of the BM and spleen49. Under inflammatory conditions, Notch2 and TLR7 pathways independently and synergistically promote conversion of Ly6Chi monocytes into Ly6Clo monocytes50. However, in our study, the percentage of blood Ly6Clo monocytes increased in RBP-J deficient mice. These results somewhat differ from what was observed in mice with conditional deletion of Notch249. Actually, 4 members of the Notch family have been identified in mammals26. Thus, Notch receptors seem to be non-redundant in regulating monocyte cell fate, and distinct Notch receptors are likely to act in a concerted manner to coordinate the monocyte differentiation program.

Colony-stimulating factor 1 receptor (CSF1R) signal, CX3CR1, C/EBPβ and Nr4a1 have been suggested to be involved in the generation and survival of Ly6Clo monocytes. The lifespan of Ly6Clo monocytes is acutely shortened after blockade of CSF1R, and the numbers of these cells are reduced, and the percentage of dead cells increased in mice with endothelial cell-specific depletion of Csf123,24. CX3CR1 and C/EBPβ-knockout mice show accelerated death of Ly6Clo monocyte, display the decreased level of Ly6Clo monocytes22,51. Nr4a1 knockout mice had significantly fewer monocytes in blood, bone marrow, and spleen, due to differentiation deficiency in MDP and accelerated apoptosis of Ly6Clo monocytes in bone marrow21. RBP-J deficient mice had normal MDP and Ly6Clo monocytes, expressed normal level of Nr4a1 and Ki-67, exhibited normal half-life and turnover rate, which suggested that RBP-J may not regulate Ly6Clo monocytes through these factors. Further studies are required to elucidate the precise molecular mechanisms of RBP-J in Ly6Clo monocytes under steady state.

At the steady state, Ly6Clo monocytes patrol endothelium of blood vessel. After R848 induced endothelial injury, Ly6Clo monocytes recruit neutrophils to mediate focal endothelial necrosis and fuel vascular inflammation, and subsequently, the Ly6Clo monocytes remove cellular debris9,52. In response to Listeria monocytogenes infection, Ly6Clo monocytes can extravasate but do not give rise to macrophages or DCs8. Additional evidence shows that Ly6Clo monocytes do not merely act as luminal blood macrophages. In lung, exposure to unmethylated CpG DNA expands CD16.2+ IM, which originate from Ly6Clo monocytes and spontaneously produce IL-10, thereby preventing allergic inflammation41. A recent study shows that Ly6Clo monocytes give rise to CD9+ macrophages, which provide an intracellular replication niche for Salmonella Typhimurium in spleen, and Ly6Clo-depleted mice are more resistant to Salmonella Typhimurium infection, suggesting that Ly6Clo monocytes exert certain functions in systemic infection53. Our data provide evidence for RBP-J in the function of blood Ly6Clo monocytes, as RBP-J deficient Ly6Clo monocytes exhibited enhanced competition in blood circulation, as well as gave rise to increased numbers of lung CD16.2+ IM. In summary, we showed that RBP-J acted as a fundamental regulator of the maintenance of blood Ly6Clo monocytes and their descendent, at least in part through regulation of CCR2. These results provide insights into understanding the mechanisms that regulate monocyte homeostasis and function.

Acknowledgements

We thank the core facility at Institute for Immunology, Tsinghua University for valuable technical assistance.

Funding

This research was supported by National Natural Science Foundation of China grants (31821003, 31991174, 32030037 and 82150105 to X. Hu), a Ministry of Science and Technology of China grant (2020YFA0509100 to X. Hu), and funds from Tsinghua-Peking Center for Life Sciences and Institute for Immunology at Tsinghua University (to X. Hu).

Author contributions

T. Kou designed and performed experiments, analyzed and interpreted data, and wrote the manuscript. L. Kang designed and performed experiments, analyzed and interpreted data. B. Zhang provided advice on RNA-seq analyses. J. Li performed experiments. B. Zhao provided advice on experiments. W. Zeng provided advice on experiments. X. Hu conceptualized the project, supervised experiments, interpreted data, and wrote the manuscript.

Declaration of interests

The authors declare no competing interests.

Data availability

Sequencing data sets are deposited in the Genome Expression Omnibus with assigned accession numbers as follows: RNA-seq (GEO accession no. GSE208772).

Materials and methods

Mice

Cx3cr1gfp/gfpmice (JAX stock 005582; Jung et al., 2000) and Ccr2RFP/RFP mice (JAX stock 017586) were purchased from the Jackson Laboratory. Mice with a myeloid-specific deletion of the Rbpj were generated by crossing Rbpjfl/fl mice to Lyz2-Cre mice as described previously (Hu et al., 2008). Rbpjfl/flLyz2cre/cre were crossed to Cx3cr1gfp/gfpmice to obtain Lyz2cre/creCx3cr1gfp/+and Rbpjfl/flLyz2cre/creCx3cr1gfp/+ mice. Cx3cr1gfp/+ mice were obtained by crossing Cx3cr1gfp/gfpwith C57/BL6 CD45.1+ mice. Rbpjfl/flLyz2cre/cre mice were crossed with Ccr2RFP/RFP mice to obtain Lyz2cre/creCcr2RFP/+, Lyz2cre/creCcr2RFP/RFP, Rbpjfl/flLyz2cre/creCcr2RFP/+ and Rbpjfl/flLyz2cre/creCcr2RFP/RFP mice. All mice were maintained under specific pathogen-free conditions. All animal experimental protocols were approved by the Institutional Animal Care and Use Committees of Tsinghua University. Gender- and age-matched mice were used at 7–12 weeks old for experiments.

Quantitative RT-PCR

Blood monocytes were sorted by flow cytometry. The total RNA was extracted using total RNA purification kit (GeneMarkbio) and reversely transcribed to cDNA by M-MLV Reverse Transcriptase (Takara). qPCR was performed on a real-time PCR system (StepOnePlus; Applied Biosystems) using FastSYBR mixture (CWBIO). Gapdh messenger RNA was used as internal control to normalize the expression of target genes. Primer sequences are provided in supplemental Table 1.

Primers sequences for regular qPCR used in this study

RNA-seq Analysis

Ly6Chi and Ly6Clo blood monocytes were isolated from Rbpj+/+Lyz2cre/cre and Rbpjfl/flLyz2cre/cre mice, and total RNA was extracted using total RNA purification kit (GeneMarkbio). RNA was converted into RNA-seq libraries, which were sequenced with the pair-end option using an Illumina-HiSeq2500 platform at Beijing Genomics Institute (BGI), China. The significantly down-regulated genes were identified with P value <0.05 and (FPKM + 1) fold changes ≤0.2, and significantly up-regulated genes were identified with P value <0.05 and (FPKM + 1) fold changes ≥5.7. The RNA-seq data are deposited in Gene Expression Omnibus under accession number GSE208772

Annexin V staining

Annexin V and 7-amino-actinomycin D (7-AAD) were used for identification of apoptotic monocytes by flow cytometry. Blood cells were stained with annexin V (eBioscience) and 7-AAD (Biolegend) according to the manufacturers’ protocols.

EdU pulsing and latex beads labeling

Mice were injected intravenously with a single 1mg EdU (Thermo Scientific). BM and blood cells were collected and stained with Fluorescence-conjugated mAb against CD45, CD11b, Ly6G, CD115 and Ly6C. Cells were then fixed, permeabilized and stained with reaction cocktail using EdU Assay Kit. Labeled cells were analyzed by flow cytometry.

Mice were injected intravenously with a single 10 µL Latex beads (Polysciences) in 250 µl PBS. Blood cells were harvested at indicated time. Cells were then stained for CD45, CD11b, Ly6G, CD115, Ly6C, and analyzed by flow cytometry.

In vivo labeling of sinusoidal leukocytes

Mice were injected intravenously with 1 μg of PE-conjugated anti-CD45 antibodies. Two minutes after antibody injection, mice were sacrificed, and bone marrow were harvested. BM cells were then stained and analyzed by flow cytometry.

Generation of BM chimera

C57/BL6 CD45.2+ mice were lethally irradiated in two doses of 5.5Gy 2 h apart. 0.8 × 105 BM cells from Rbpj+/+Lyz2cre/cre mice (CD45.2+) or Rbpjfl/flLyz2cre/cre mice (CD45.2+) were mixed with 3.2× 105 BM cells from Cx3cr1gfp/+ mice (CD45.1+), and injected intravenously into recipient mice. Mice were used for experiments 8 weeks after irradiation.

Adoptive transfers

BM GFP+Ly6Chi monocytes were sorted from Lyz2cre/creCx3cr1gfp/+or Rbpjfl/flLyz2cre/creCx3cr1gfp/+ mice, and transferred to Ccr2RFP/RFP mice. Blood cells were collected 60 hours later for flow cytometry analyses.

Parabiosis

Cx3cr1gfp/+CD45.1+ mice were surgically joined with age-matched female Rbpjfl/flLyz2cre/cre or Rbpj+/+Lyz2cre/cre CD45.2+ mice at the age of 6 weeks. Bloods were obtained via cardiac puncture, and cell populations were analyzed by flow cytometry at 4 weeks after the surgery.

Intranasal instillations of LPS

Rbpj+/+Lyz2cre/cre or Rbpjfl/flLyz2cre/cre mice were anesthetized with isoflurane and intranasally instilled with 10 μg LPS in 25 μl of PBS. Lungs were harvested 4 days later for flow cytometry analyses.

Immunofluorescence histology

Lungs from Rbpjfl/flLyz2cre/creCx3cr1gfp/+ and Lyz2cre/creCx3cr1gfp/+ mice were fixed in 1% paraformaldehyde and incubated in 30% sucrose separately overnight at 4℃. The samples were then incubated in the mixture of 30% sucrose and OTC compound (Sakura Finetek) overnight at 4℃. The tissues were embedded and frozen in OCT compound and then cut at 10-µm thickness. Tissue sections were dried for 10 min at 50℃ and then fixed in 1% paraformaldehyde at room temperature for 10 min, permeabilized in PBS/0.5% Triton X-100/0.3 M glycine for 10 min and blocked in PBS/5% goat serum for 1 h at room temperature. Sections were stained with rabbit anti-mouse GFP antibodies (1:1000; EASYBIO) overnight at 4℃, and washed with PBS/0.1% Tween-20 for 30 min three times at room temperature. Sections were then incubated with AF488-conjugated goat anti-rabbit antibodies (1:200; proteintech) for 2 h at room temperature, and washed with PBS/0.1% Tween-20 for 30 min three times at room temperature. Sections were stained with DAPI (Solarbio) for 7 min, washed in PBS for 8 min two times at room temperature and mounted with SlowFade Diamond Antifade Mountant (Life Technologies).

Cell isolation and Flow cytometry

Peripheral blood was sampled by eyeball extirpating, spleens were mashed through a 70-μm strainer, and bone marrow cells were collected from femurs. Lungs were perfused with 5ml of HBSS (MACGENE) through the right ventricle, excised, and digested in HBSS containing 5% FBS, 1 mg/ml collagenase type I (Sigma-Aldrich) and 0.05 mg/ml DNase I (Sigma-Aldrich) for 1 h at 37 °C. The digested tissues were homogenized by shaking, passed through a 70-µm cell strainer to create a single cell suspension. The suspension was enriched in mononuclear cells and harvested from the 1.080:1.038 g ml-1 interface using a density gradient (Percoll from GE Healthcare). BM, Spleen, PB and lung cells were stained with fluorescently conjugated antibodies. The absolute number of cells was counted by using CountBright™ Absolute Counting Beads (Invitrogen).

Antibodies against CD45 (30-F11), Ly6C (HK1.4), CD64 (X54-5/7.1), CD206 (C068C2), CD16.2 (9E9), CD45.1 (A20), F4/80 (BM8), Biotin, CCR2 (SA203G11) and CD3ε (145-2C11) were purchased from BioLegend. Antibodies against Ly6G (1A8-Ly6g), CD11b (M1/70), CD115(AFS98), CD117(2B8), CD135 (A2F10), CD11c (N418), CD206 (MR6F3), CD45.2 (104), Nur77 (12.14), Ki-67 (SolA15) and lineage maker were purchased from eBioscience. Fluorescence-conjugated mAb against Ly6C (AL-21) were purchased from BD Biosciences. Isotype-matched antibodies (eBioscience) were used for control staining. All antibodies were used in 1:400 dilutions, and surface antigens were stained on ice for 30 min.

For intracellular staining, cells were stained with antibodies to surface antigens, fixed and permeabilized with the Cytofix/Cytoperm Fixation/Permeabilization Solution Kit (BD Biosciences), and stained with antibodies or isotype control diluted in permeabilization buffer separately for 30 min at room temperature.

Cells were analyzed on FACSFortessa or FACSAria III flow cytometer (BD Biosciences) using FlowJo software.

Figure Legends

Control and RBP-J deficient mice show similar neutrophils.

(A) qPCR analysis of Rbpj expression in sorted monocyte subsets from control and RBP-J deficient mice.

(B) Blood and BM CD45+CD11b+Ly6G+ neutrophils were determined by FACS. Cumulative data of cell ratio are shown.

(C) Spleen monocyte subsets were determined by FACS. Representative FACS plots (left) and cumulative data of cell ratio and absolute numbers (right) are shown.

Data are pooled from at least 2 independent experiments; n ≥ 2 in each group. Data are shown as mean ± SEM; n.s., not significant; *, P < 0.05 (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse.

The conversion of Ly6Chi monocytes is identical in control and RBP-J deficient mice.

(A) Representative FACS plots, cumulative MFI and qPCR analysis of Nr4a1 expression in control and RBP-J deficient blood monocyte subsets. Shaded curves represent isotype control, black lines represent control mice, and red lines represent RBP-J deficient mice.

(B) Representative FACS plots of GFP+ Ly6Chi monocytes in recipient.

Data are pooled from 2 independent experiments (A); n ≥ 4 in each group. Data are shown as mean ± SEM; n.s., not significant (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse.

Cell-intrinsic requirement of RBP-J for Ly6Clo maintenance.

(A) Representative FACS plots (left) and cumulative data (right) quantitating percentages of blood neutrophils in recipient mice.

(B) BM neutrophils and Ly6Chi monocytes in recipient mice were determined by FACS.

(C) CD4+ T cells and CD19+ B cells were analyzed by FACS in parabiotic mice.

Data are pooled from at least 2 independent experiments (A and C); n ≥ 4 in each group. Data are shown as mean ± SEM; n.s., not significant (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse.

Normal expression of CCR2 and CD11c in BM monocytes.

(A) Cumulative MFI of CCR2 expression in BM monocyte subsets.

(B) FACS plots of CCR2 and CD11c in BM Ly6Chi and Ly6Clo from control and RBP-J deficient mice.

Data are pooled from 3 independent experiments (A); n ≥ 7 in each group. Data are shown as mean ± SEM; n.s., not significant (two-tailed Student’s unpaired t test). Each symbol represents an individual mouse.

RBP-J is not required for turnover of lung Ly6Clo monocytes and CD16.2+ IM.

(A and B) Incorporation of EdU was assessed 24 hours after injection. BM monocyte subsets were used as controls. FACS plots (A) and representative data (B) are shown. Each symbol represents an individual mouse.