Volume 37, Issue 7 p. 924-936

Regenerative Medicine

Free Access

Notch Signaling in Nestin‐Expressing Cells in the Bone Marrow Maintains Erythropoiesis via Macrophage Integrity

Tatsuhiro Sakamoto

orcid.org/0000-0001-6852-0721

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Department of Transfusion Medicine, University of Tsukuba Hospital, Tsukuba, Japan

Contributed equally.Search for more papers by this author

Naoshi Obara

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Contributed equally.Search for more papers by this author

Hidekazu Nishikii

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Takayasu Kato

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Luan Cao‐Sy

orcid.org/0000-0001-8330-1471

Graduate School of Comprehensive Human Sciences, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Ryosuke Fujimura

School of Medical Sciences, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Hideo Yagita

Department of Immunology, Juntendo University School of Medicine, Tokyo, Japan

Search for more papers by this author

Mamiko Sakata‐Yanagimoto

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Satoru Takahashi

Department of Anatomy and Embryology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Shigeru Chiba

Corresponding Author

schiba-tky@umin.net

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Life Science Center for Survival Dynamics, Tsukuba Advanced Research Alliance, University of Tsukuba, Tsukuba, Japan

Correspondence: Shigeru Chiba, M.D., Ph.D., Department of Hematology, Faculty of Medicine, University of Tsukuba, 1‐1‐1 Tennodai, Tsukuba, Ibaraki 305‐8575, Japan. Telephone: (81‐29) 853‐3103; e‐mail: schiba-tky@umin.net Search for more papers by this author

Tatsuhiro Sakamoto

orcid.org/0000-0001-6852-0721

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Department of Transfusion Medicine, University of Tsukuba Hospital, Tsukuba, Japan

Contributed equally.Search for more papers by this author

Naoshi Obara

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Contributed equally.Search for more papers by this author

Hidekazu Nishikii

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Takayasu Kato

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Luan Cao‐Sy

orcid.org/0000-0001-8330-1471

Graduate School of Comprehensive Human Sciences, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Ryosuke Fujimura

School of Medical Sciences, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Hideo Yagita

Department of Immunology, Juntendo University School of Medicine, Tokyo, Japan

Search for more papers by this author

Mamiko Sakata‐Yanagimoto

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Satoru Takahashi

Department of Anatomy and Embryology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Search for more papers by this author

Shigeru Chiba

Corresponding Author

schiba-tky@umin.net

Department of Hematology, Faculty of Medicine, University of Tsukuba, Tsukuba, Japan

Life Science Center for Survival Dynamics, Tsukuba Advanced Research Alliance, University of Tsukuba, Tsukuba, Japan

First published: 01 April 2019

https://doi.org/10.1002/stem.3011

Abstract

Notch signaling plays pivotal roles in both hematopoietic stem/progenitor and their niche cells. Myeloproliferative phenotypes are induced by disruption of Notch signaling in nonhematopoietic bone marrow (BM) cells. Nestin‐expressing cells in the BM reportedly represent a component of the hematopoietic stem cell niche. We established mice in which rare Nestin‐expressing cells in the BM were marked by green fluorescent protein, and Notch signaling was conditionally disrupted in these cells specifically. We observed impairment of erythropoiesis in the BM accompanying splenomegaly with BM hematopoietic programs in other lineages undisturbed. Transplantation experiments revealed that the microenvironmental rather than the hematopoietic cells were attributable to these phenotypes. We further found that the erythroid‐island‐forming ability of BM central macrophages was compromised along with the transcriptional upregulation of interleukin‐6. Various Inflammatory conditions hamper BM erythropoiesis, which often accompanies extramedullary hematopoiesis. The mouse model demonstrated here may be of relevance to this common pathophysiologic condition. Stem Cells 2019;37:924–936

Significance Statement

Notch signaling plays a crucial role in both hematopoietic stem cells (HSCs) and their niche cells. Nestin‐expressing mesenchymal stromal cells (MSCs) in a bone marrow (BM) are reported as a component of HSC niche. Results of the study showed that Notch signaling in Nestin‐expressing MSCs regulates erythroid differentiation in the BM. Disruption of Notch signaling in Nestin‐expressing MSCs induced the upregulation of interleukin‐6 (IL‐6) in central macrophages, and IL‐6 can cause the impairment of erythroid‐island‐forming capacity in central macrophages. The results suggest that the erythropoiesis in the BM is regulated by interaction between the central macrophages and Nestin‐expressing MSCs.

Introduction

Notch signaling acts between neighboring cells and is highly conserved across mammals 1. In canonical Notch signaling, binding of one of several ligands (Delta‐like 1 [Dll1], Dll2, Dll4, Jagged 1, and Jagged 2) to one of four receptors (Notch1, Notch2, Notch3, and Notch4) expressed on adjacent cells triggers the signaling cascade. The intracellular domain of a Notch protein generated by two cleavage steps translocates to the nucleus, which forms a transcriptional activation complex directly associating with recombination signal binding protein for immunoglobulin kappa J region (RBPJ). Several target genes then influence a variety of developmental and differentiation programs in a context‐dependent manner 1-4.

Notch signaling is also involved in regulating hematopoiesis 1. Hematopoietic stem cell (HSC)/hematopoietic progenitor cell‐endogenous Notch signaling has been proven to contribute to development and differentiation of hematopoietic cells in many contexts 5-8. In addition, Notch signaling in BM microenvironmental cells has been revealed to be essential for the maintenance of hematopoietic homeostasis 9-13. For example, a myeloproliferative phenotype was caused by disruption of the Mindbomb‐1 gene (Mib‐1), which works for Dll1 endocytosis and positively regulates Notch signaling, in Mx‐Cre‐expressing host environmental cells. The phenotype could not be rescued by wild‐type syngeneic HSC/hematopoietic progenitor cell transplantation 11. A similar myeloproliferative phenotype was also observed when Rbpj was disrupted by Mx‐Cre in the host cells 12. In these mice, a wide range of interferon‐responsive cells lose Notch signaling, and thus, the responsible niche cells in which Notch signaling plays a crucial role for regulation of blood production are unknown.

Recent studies have shown substantial kinds of distinct niche cells in the BM for HSC regulation, such as osteoblasts 14, endothelial cells 15, nonmyelinating Schwann cells 16, chemokine C‐X‐C motif ligand 12‐abundant reticular cells 17, and Nestin‐expressing mesenchymal stromal cells (MSCs) 18-20. Disrupting Rbpj by Tie2‐Cre resulted in the same myeloproliferative phenotype as that seen in mice with the Rbpj disruption by the Mx‐Cre system. This indicates that Notch signaling in vascular endothelial cells, among other environmental cells, may play a dominant role in suppressing myeloproliferation 11, 12. Nevertheless, there is still no information about the roles of Notch signaling in the BM Nestin‐expressing cells.

Most of the niche cells that have been characterized are those for HSCs, whereas studies on niche cells supporting specific lineage hematopoiesis have been rare, except macrophages that create erythroid islands. Some other examples of the latter include the intercellular adhesion molecule‐1 (ICAM‐1)‐positive pericytes, which have been proposed as the osteogenic progenitors for the B‐lymphopoietic niche 21, 22, and the podoplanin‐expressing perivascular cells for the megakaryopoietic niche, respectively 23. Erythroid islands have been extensively studied, leading to the conclusion that central macrophages were involved in the proliferation of relatively mature erythroblasts and their terminal differentiation processes, such as enucleation 24-26.

All these backgrounds prompted us to delete Rbpj specifically in Nestin‐expressing cells. We confirmed that very rare nonhematopoietic cells were marked by green fluorescent protein (GFP) induced by Nestin promoter‐driven Cre‐ER^T2. GFP marking coincided substantially with endogenous Nestin expression. Disruption of Notch signaling in Nestin‐expressing cells disturbed the BM erythropoiesis. We here show that this phenotype is attributable to the compromised erythropoietic activity of macrophages in the BM.

Materials and Methods

Mice

Rbpj^flox/flox mice were obtained from T. Honjo 27. Nestin‐CreER^T2 mice were obtained from R. Kageyama 28. Z/EG mice, in which enhanced GFP is expressed in the presence of Cre recombinase, were obtained from the Jackson Laboratories 29. C57BL/6J‐Ly5.2 (Ly5.2) and C57BL/6J‐Ly5.1 (Ly5.1) mice were purchased from Japan SLC (Hamamatsu, Japan). Nestin‐CreER^T2 expression was induced by intraperitoneal injection with tamoxifen. The animal experiments were performed following the protocols approved by the University of Tsukuba Animal Care and Use Committee.

Flow Cytometry and Cell Sorting

To analyze the GFP⁺ cells in the BM, cells were obtained from Nestin‐CreER^T2;Z/EGRbpj^flox/flox mice and from littermates Nestin‐CreER^T2;Z/EG;Rbpj^flox/wt mice after tamoxifen injection, by flushing out of the femurs and tibias and by dissecting various bones.

To analyze the hematopoietic cells, BM and spleen cells were obtained from Nestin‐CreER^T2;Rbpj^flox/flox (Rbpj cKO) mice and littermate Rbpj^flox/flox (littermate control) mice after tamoxifen injection by flushing out of the bones and spleens.

Immunofluorescence Imaging

Tibias were obtained from mice, and thin frozen sections (5 μm) of the specimen were prepared using the Kawamoto's film method 30. The immunofluorescence data were obtained and analyzed using the Leica TCS SP5 confocal microscope and the Leica LAS AF (Leica Microsystems, Wetzlar, Germany) software, respectively.

In Vitro Colony‐Forming Assays

For colony‐forming unit‐fibroblast (CFU‐F) assay, 50 of CD45⁻Ter119⁻CD31⁻GFP⁺ cells, CD45⁻Ter119⁻CD31⁻GFP⁻ cells, and CD45⁻Ter119⁻CD31⁻ platelet‐derived growth factor receptor α‐chain (PDGFRα)⁺Sca‐1⁺ (PαS) cells in the BM were sorted into flat‐bottom 96‐well plate. The cells were cultured in Dulbecco's modified Eagle's medium with GlutaMAX (Thermo Fisher Scientific, Waltham, MA, USA) and 20 ng/ml basic fibroblast growth factor (FUJIFILM Wako Pure Chemical, Osaka, Japan). The medium was changed once a week. CFU‐Fs were scored 2 weeks after plating. The colony was defined when cell number exceeded 10.

For colony‐forming unit‐erythroid (CFU‐E), burst‐forming unit‐erythroid (BFU‐E), and colony‐forming unit‐granulocyte/macrophage (CFU‐GM) assays, BM and spleen cells were obtained from Rbpj cKO mice and littermate control mice and cultured in Methocult M3231 medium (Stem Cell Technologies, Vancouver, Canada) with mixtures of cytokines.

Transplantation Assay

Total BM cells (1 × 10⁶) from Ly5.1 wild‐type mice were transplanted into lethally irradiated, tamoxifen‐treated Rbpj cKO or littermate control recipients. Total BM cells (1 × 10⁶) from tamoxifen‐treated Rbpj cKO mice or littermate control mice (Ly5.2) were transplanted into lethally irradiated (9.5 Gy) Ly5.1 wild‐type recipients. Eight weeks after transplantation, BM cells and spleen cells were collected and analyzed by flow cytometry.

Erythroid Island Reconstitution Assay

For the erythroid island reconstitution assay, a previously published protocol 31, 32 was used with some modifications. Macrophages were prepared from the BM of the BM of Rbpj cKO mice or littermate control mice and incubated with Ter119⁺ erythroid cells from the BM of tamoxifen‐untreated C57BL/6J mice. As a counterpart experiment, macrophages from the BM of tamoxifen‐untreated C57BL/6J mice were incubated with Ter119⁺ erythroid cells from the BM of Rbpj cKO mice or littermate control mice. Additionally, macrophages from the spleen of Rbpj cKO mice or littermate control mice were incubated with Ter119⁺ erythroid cells from the BM of tamoxifen‐untreated C57BL/6J mice.

To assess the effect of interleukin‐6 (IL‐6) blockade, the adherent macrophages were incubated for 4 hours with 5 μg/ml of rat anti‐mouse IL‐6 receptor (anti‐IL‐6R) antibody (MR16‐1, Chugai Pharmaceutical Co., Ltd, Tokyo, Japan) that blocks binding of IL‐6 to IL‐6 receptor or control rat IgG (MP Biomedical, Santa Ana, CA, USA).

Quantitative Reverse Transcription Polymerase Chain Reaction Experiments

Total RNA from purified GFP⁺ cells and macrophages derived from the BM were isolated. cDNA was prepared using Superscript III (Thermo Fisher Scientific) with random hexamers (Thermo Fisher Scientific). In single cell quantitative polymerase chain reaction (qPCR) assay, single cells were sorted into 96‐well PCR plates. Then, reverse transcription and specific target amplification were carried out. All qPCR assays were performed with the ABI7500 Fast real‐time system (Thermo Fisher Scientific) and analyzed by comparative CT method (ddCT method). All the experiments were conducted in technical duplicate and normalized to the mRNA level of β‐actin (Actb) or glyceraldehyde‐3‐phosphate dehydrogenase (Gapdh).

In Vivo IL‐6 Blockade

Rbpj cKO mice were injected with rat anti‐mouse anti‐IL‐6R antibody or control rat IgG once a week for 4 weeks after 4‐week administration of tamoxifen. The dose of anti‐IL‐6R antibody or control rat IgG was 2 milligram per body in the first administration and 0.5 milligram per body in subsequent administrations.

Statistical Analysis

All data are expressed as the means ± SDs of at least three independent experiments, unless otherwise mentioned in the figure legends. Probability values less than .05 were considered significant. The details of all methods, including those for antibodies, primers, and probes, are described in Supporting Information.

Results

GFP Marks BM Stromal Cells Endogenously Expressing Nestin

To understand the role of Notch signaling in Nestin‐expressing cells, Rbpj was conditionally disrupted in Nestin‐expressing cells from Nestin‐CreER^T2 transgenic mice mated to Rbpj^flox/flox mice and Z/EG transgenic mice to visualize Nestin‐expressing cells. A schematic representation of the construction of the gene of the Rbpj^flox/flox allele and Z/EG is shown in Supporting Information Figure S1A.

To identify CreER^T2 expression in the BM, Nestin‐CreER^T2;Z/EG;Rbpj^flox/wt (heterozygous) mice were intraperitoneally injected with tamoxifen. GFP expression was sparsely detected in a small population of whole BM cells (Fig. 1A–1C). Some GFP‐expressing cells were located near VE‐cadherin⁺ or CD31⁺ endothelial cells (Fig. 1A[ii]; Supporting Information Fig. S1B[i]), and the others were not. These results corresponded to the findings of a previous paper, in which Nestin‐GFP‐marked cells were identified in both perivascular region and nonperivascular BM cavity 19. Most, albeit not 100%, of the GFP⁺ cells overlapped with fluorescein antibody‐labeled endogenous Nestin‐expressing cells in the BM (Fig. 1B). On flow cytometry, GFP expression was found in 0.52% (±0.19%) of the CD45⁻Ter119⁻CD31⁻ cells, which represent nonhematopoietic and nonendothelial stromal cells (Fig. 1C). The background GFP signal was very few in the BM of the same genotype mice without tamoxifen injection (Fig. 1C). The Nestin mRNA levels were markedly higher in the CD45⁻Ter119⁻CD31⁻GFP⁺ BM cells than in the whole BM cells (Fig. 1D).

**Figure 1**
Open in figure viewer PowerPoint

GFP marks BM stromal cells endogenously expressing Nestin. **(A):** Distribution of GFP⁺ cells in the BM. (i) Low‐magnification image. Scale bar, 200 μm. (ii) The imaging for VE‐cadherin, CD31, and GFP. VE‐cad, VE‐cadherin. Scale Bar, 15 μm. Arrowhead, perivascular GFP⁺ cells. *, the cavity of vessels. **(B):** GFP⁺ cells and antibody‐labeled Nestin‐expressing cells in the BM. Arrowhead, copresence of GFP⁺ and endogeneous Nestin. Scale bar, 50 μm. **(C):** Identification of GFP⁺ cells in the BM by flow cytometry. Reanalysis represents the data of sorted GFP⁺ cells. A representative analysis of BM cells from tamoxifen‐treated (Tamoxifen [+]) and ‐untreated (Tamoxifen [−]) *Nestin‐CreER*^T2;Z/EG;Rbpj^flox/wt mice. **(D):** *Nestin* mRNA levels in whole BM cells and CD45⁻Ter119⁻CD31⁻GFP⁺ cells (n = 3, each). **(E):** CFU‐F assay of the BM cells. (i) The frequency of colony‐growing wells in sorted 50 of CD45⁻Ter119⁻CD31⁻GFP⁺ cells, CD45⁻Ter119⁻CD31⁻GFP⁻ cells, and CD45⁻Ter119⁻CD31⁻PDGFRα⁺Sca‐1⁺ cells (n = 4, each). (ii) The morphology of colonies in CD45⁻Ter119⁻CD31⁻GFP⁺ cells derived from homozygous mice and heterozygous mice. Representative data. Scale bar, 100 μm. **(F):** Cell surface antigen expression of BM stromal cells (representative data). (i) Whole CD45⁻Ter119⁻ cells from tamoxifen‐treated *Nestin‐CreER*^T2;Z/EG;Rbpj^flox/wt mice. (ii) CD45⁻Ter119⁻GFP⁺ cells. **(G):** Expression of *Rbpj* and *Hes1* in sorted CD45⁻Ter119⁻CD31⁻GFP⁺ cells in the BM (n = 3–5, each). mRNA levels were normalized to the *Gapdh* mRNA level. *, p < .05. Abbreviations: BM, bone marrow; DAPI, 4′,6‐diamidino‐2‐phenylindole; f/f, tamoxifen‐treated *Nestin‐CreER*^T2;Z/EG;Rbpj^flox/flox mice; f/w, tamoxifen‐treated *Nestin‐CreER*^T2;Z/EG;Rbpj^flox/wt mice; GFP, green fluorescent protein; PDGFRα, Platelet‐derived growth factor receptor α‐chain; SSC, side scatter; 7‐AAD, 7‐amino‐actinomycin D.

A previous study demonstrated that the Nestin‐expressing cells in the BM were MSCs 18. In addition, another study showed that PαS cells were also MSCs 33. Therefore, we performed a CFU‐F assay for the sorted CD45⁻Ter119⁻CD31⁻GFP⁺, CD45⁻Ter119⁻CD31⁻GFP⁻, and PαS cells in the BM. These cells were derived from the Nestin‐CreER^T2;Z/EG;Rbpj^flox/flox (homozygous) and the heterozygous mice after tamoxifen injection (Fig. 1E). CD45⁻Ter119⁻CD31⁻GFP⁺ cells from both genotype mice formed colonies at comparable plating efficiencies with the corresponding PαS cells, whereas CD45⁻Ter119⁻CD31⁻GFP⁻ cells did not (Fig. 1E[i]), in agreement with the previous findings 18, 33. The deletion of Rbpj did not affect the plating efficiency on both CD45⁻Ter119⁻CD31⁻GFP⁺ cells and PαS cells. The morphology of colonies derived from CD45⁻Ter119⁻CD31⁻GFP⁺ cells from both genotype is shown in Figure 1E(ii). Some colonies contained more than 500 cells, whereas others contained less than 100.

Next, we analyzed GFP⁺ cells in the BM from Nestin‐CreER^T2;Z/EG;Rbpj^flox/flox (homozygous) mice after tamoxifen injection in comparison with those from the heterozygous mice by flow cytometric analysis (Fig. 1F; Supporting Information Fig. S1C, S1D). The CD45⁻Ter119⁻CD31⁻GFP⁺ cell fraction size was small (∼0.5%) and unchanged in the tamoxifen‐injected homozygous mice as compared with the heterozygous mice (Supporting Information Fig. S1C). To examine the expression of cell surface markers including endothelial markers, we analyzed CD45⁻Ter119⁻GFP⁺ cells from the BM of homozygous and heterozygous mice. Among the mesenchymal cell markers, integrin αV (CD51) was expressed in all of CD45⁻Ter119⁻GFP⁺ BM cells (Fig. 1F, left panel; Supporting Information Fig. S1D[i]). Other mesenchymal cell markers such as CD105, CD90, and PDGFRα were variably expressed in ∼40%–60% of the CD45⁻Ter119⁻GFP⁺ cells (Supporting Information Fig. S1D[i]). Sca‐1, a MSC marker, was expressed in only ∼20%–30% of the CD45⁻Ter119⁻GFP⁺ BM cells (Fig. 1F, left panel; Supporting Information Fig. S1D[i]). Among the Notch receptor proteins, Notch2 was expressed in 60%–80% of the CD45⁻Ter119⁻GFP⁺ BM cells, whereas Notch1, Notch3, and Notch4 were variably expressed in their minor fractions, in agreement with the findings of a previous study (Supporting Information Fig. S1D[ii]) 34. Separated from the major population, less than 10% of CD45⁻Ter119⁻GFP⁺ BM cells were demonstrated as CD31^hiICAM‐1⁺ cells, indicating that GFP also marked a fraction of endothelial cells (Fig. 1F, middle panel). There might be a distinct population of endothelial cells that coexpresses Nestin, although these cells were not detected by immunofluorescent staining (Fig. 1A[ii]; Supporting Information Fig. S1B).

Although PDGFRα⁺CD51⁺ cells reportedly represent MSCs 33, GFP partially marked PαS and PDGFRα⁺CD51⁺ MSCs (Fig. 1F, middle and right panels). There appeared no phenotypic differences between CD45⁻Ter119⁻GFP⁺ BM cells from Rbpj heterozygous mice and those from homozygous mice.

These results imply that GFP grossly marks endogenous Nestin‐expressing BM stromal cells. Furthermore, these GFP⁺ stromal cells contain MSCs on the basis of the results of CFU‐F assay. These GFP⁺ MSCs constitute multiple distinct phenotypic populations including endothelial cells and MSCs. Moreover, impairment of Notch signaling in these cells does not affect their numbers.

Rbpj Is Deleted Specifically in GFP⁺ BM Stromal Cells

Next, we analyzed Rbpj deletion in fractionated BM cells from homozygous mice after tamoxifen injection, with reference DNA extracted from the brains of the same mice before and after tamoxifen injection. The Rbpj deletion band was detected in brain‐derived DNA even before tamoxifen injection, which was markedly increased after the injection. In DNA extracted from variable fractions of BM cells, the Rbpj deletion band was detected only in CD45⁻Ter119⁻CD31⁻GFP⁺ cells but not in whole BM cells or in any other fractions (Supporting Information Fig. S1E). We intended to analyze mRNA expression of Rbpj and 15 genes known to be the targets of the Notch‐Rbpj transcriptional regulator in sorted CD45⁻Ter119⁻CD31⁻GFP⁺ fraction by real‐time PCR (Fig. 1G; Supporting Information Table S2 and Fig. S1F). The decrease of Rbpj was confirmed in a statistical manner, although the expression did not disappear (Fig. 1G).

To explore the cause of incomplete Rbpj deletion, we performed single‐cell qPCR assay with a specific target amplification method. At the single‐cell level, Rbpj mRNA amounts were remarkably variable in the CD45⁻Ter119⁻CD31⁻GFP⁺ fraction from the heterozygous mice, with some cells showing below the detection limit (Supporting Information Fig. S2A). Nevertheless, we confirmed that the number of single cells showing Rbpj mRNA expression below the limit was significantly greater in homozygous than heterozygous mice. A statistical analysis also showed that overall expression levels of Rbpj mRNA were lower in the homozygous mice at the single‐cell level (Supporting Information Fig. S2B). However, we also confirmed that some single cells still remained expressing Rbpj mRNA. Thus, remaining expression most likely reflects incompleteness of the deletion of the Rbpj gene in the CD45⁻Ter119⁻CD31⁻GFP⁺ cell compartment.

Among the Notch signaling targets, we could obtain results for only 6 out of 15 genes: Hes1, Myc, Notch2, Notch3, Dtx3, and Dtx4 in the bulk analysis. We observed a decrease in Hes1 mRNA in the bulk analysis and Dtx2, Dtx3, and Dtx4 mRNA in the single cell analysis (Fig. 1G; Supporting Information Fig. S2B). Consequently, Rbpj is specifically deleted, albeit partially, in Nestin‐expressing stromal cells.

Rbpj Conditional Knockout in Nestin‐Expressing Cells Impairs BM Erythropoiesis and Triggers Extramedullary Hematopoiesis in the Spleen

To investigate the role of Notch signaling in Nestin‐expressing cells in the BM, we first compared the general appearance of the BM from Nestin‐CreER^T2;Rbpj^flox/flox mice after tamoxifen injection (Rbpj cKO mice). The numbers of nucleated cells and red blood cells (RBCs) in the BM was almost the same in Rbpj cKO and littermate control mice (Fig. 2A). Histologically, no apparent difference was observed in H&E staining (Fig. 2B).

**Figure 2**
Open in figure viewer PowerPoint

*Rbpj* deletion in Nestin‐expressing cells induced impaired erythroid differentiation. **(A):** Number of nucleated cells and RBCs in the BM (n = 5, each). (i) Number of nucleated cells. (ii) Number of RBCs. **(B):** H&E staining of the BM. **(C):** Each erythroid fraction in the BM measured by CD71 and Ter119 expression. (i) Representative flow cytometric analysis. (ii) Number of each erythroid fraction in the BM measured by flow cytometry (n = 5, each). **(D):** Each erythroid fraction in the BM measured by CD44 expression and FSC. (i) Representative flow cytometric analysis. (ii) Number of each erythroid fraction in the BM measured by flow cytometry (n = 5, each). I, proerythroblasts; II, polychromatic erythroblasts; III, orthochromatic erythroblasts; IV, reticulocytes; V, RBCs (n = 5, each). **(E):** Immunofluorescence staining of erythroid cells in the BM. Scale bar, 50 μm. **(F):** Number of whole erythroid (1 + 2 + 3 + 4 of Fig. 2B) cells (n = 5, each). **(G):** The complete blood cell counts. *, p < .05. Abbreviations: BM, bone marrow; cKO, *Rbpj* cKO mice; Ctrl, *Rbpj*^flox/flox littermate control mice; DAPI, 4′,6‐diamidino‐2‐phenylindole; FSC, forward scatter; Hb, hemoglobin level; Plt, platelet count; RBC, red blood cell; WBC, white blood cell.

The decrease of CD71⁺Ter119⁺ immature erythroid cells (II) was observed by flow cytometry using CD71 and Ter119 expression (Fig. 2C) 35. For further analysis of erythroid differentiation, we performed flow cytometric analysis using CD44 antibody 36. The decrease in erythroid cells was observed in the orthochromatic erythroblast (Fr‐III) and reticulocyte (Fr‐IV) fractions (Fig. 2D). Furthermore, there was a tendency of decrease in polychromatic erythroblast (Fr‐II) fraction (Fig. 2D). The decrease of CD71⁺Ter119⁺ erythroid cells was confirmed by immunofluorescent staining (Fig. 2E; imaging for isotype control is shown in Supporting Information Fig. S3A). The total erythroid cells tended to be decreased in Rbpj cKO mice (Fig. 2F). Additionally, the numbers of HSC/hematopoietic progenitor cell and other lineage cells (granulocytes, CD41⁺ megakaryocytes, Gr1⁻Mac1⁺F4/80⁺ macrophages, CD3⁺ T‐cells, and B220⁺ B‐cells) in the BM did not differ between the Rbpj cKO mice and the control mice (Supporting Information Fig. S3B, S4). Moreover, the complete blood cell counts were not different between control and Rbpj cKO mice (Fig. 2G). These results suggest that Notch signaling in Nestin‐expressing BM MSCs is essential for erythropoiesis in the BM. Nevertheless, Rbpj cKO mice did not show anemia or any abnormalities in blood cell counts in the other lineages.

The Rbpj cKO mice showed mild splenomegaly as compared with the littermate control mice (Fig. 3A, 3B). Histologically, the margin between the red pulp and the white pulp of the spleen was blurred in the Rbpj cKO mice (Fig. 3C), and the red pulp seemed to be extended in the Rbpj cKO spleen (Supporting Information Fig. S5A). To confirm the expansion of red pulp, we examined the ratio of erythroid cells (Ter119⁺ cells) to lymphoid cells (CD3⁺ or B220⁺ cells). The erythroid/lymphoid ratio was significantly increased in the Rbpj cKO spleen (Supporting Information Fig. S5B). The numbers of nucleated cells and RBCs were increased in the Rbpj cKO spleen reflecting splenomegaly (Fig. 3D).

**Figure 3**
Open in figure viewer PowerPoint

*Rbpj* deletion in Nestin‐expressing cells induced compensatory erythropoiesis. **(A):** Appearance of the spleens. Red bar, 10 mm. **(B):** The spleen weights. Circles and squares represent each mouse. **(C):** H&E staining of the spleen. The scale bar represents 100 μm. **(D):** Number of nucleated cells and RBCs in the spleen (n = 6, each). **(Di):** Number of nucleated cells. **(Dii):** Number of RBCs. **(E):** Each erythroid fraction in the spleen. **(Ei):** Representative data of flow cytometric analysis. **(Eii):** Number of each erythroid fraction in the spleen (n = 5, each). **(F):** Number of whole erythroid cells (1 + 2 + 3 + 4 of Fig. 2E[i]) in the spleen (n = 5, each). *, p < .05. Abbreviations: cKO, *Rbpj* cKO mice; Ctrl, *Rbpj*^flox/flox littermate control mice; RBC, red blood cell.

The numbers of CD71⁺Ter119⁺ immature erythroid cells and total erythroid cells were also increased in the spleen of cKO mice (Fig. 3E, 3F). Additionally, there was a tendency of increase in a variety of cell populations including mature and immature cells in the Rbpj cKO mice (Supporting Information Fig. S4, S5C). Significant differences were observed in some fractions, such as Linage⁻cKit⁺Sca‐1⁺, megakaryocyte erythroid progenitors, macrophages, and granulocytes but not in others including preCFU‐E and CFU‐E (Supporting Information Fig. S5C).

To exclude the effect of Cre toxicity on these phenotypes, we analyzed the BM and spleen from tamoxifen‐injected Nestin‐CreER^T2;Rbpj^wt/wt mice in comparison with Rbpj^flox/flox mice. There were no significant differences in the number of each erythroid fraction and in the weight of the spleen between Nestin‐CreER^T2;Rbpj^wt/wt mice and Rbpj^flox/flox mice (Supporting Information Fig. S6A).

Taken together, these results suggest that extramedullary erythropoiesis compensates for impaired BM erythropoiesis in Rbpj cKO mice.

Impairment of BM Erythropoiesis and Splenomegaly Are Caused by Microenvironmental Factors Not by Hematopoietic Progenitor‐Endogenous Factors

To validate the hypothesis of the microenvironmental origin of the impaired erythropoiesis in Rbpj cKO mice, a methylcellulose‐based colony assay was performed for BM mononuclear cells. Despite the morphological abnormalities in the BM, the numbers of the BFU‐E, CFU‐E, and CFU‐GM cells were unchanged when compared with the control BM mononuclear cells (Supporting Information Fig. S6B[i]), indicating that the endogenous capacity of hematopoietic progenitors was unaffected. Similarly, the colony assay for spleen cells did not show any changes in progenitor cell frequencies (Supporting Information Fig. S6B[ii]).

To further confirm the origin of the phenotype in Rbpj cKO mice, BM transplantation experiments were performed. First, Rbpj cKO and littermate control mice were transplanted with the whole BM cells from wild‐type Ly5.1 mice after a lethal dose of irradiation (Fig. 4A). Almost 100% donor chimerism was demonstrated at 8 weeks after the transplant in both the recipient and the control mice with similar cellularities (Fig. 4B; Supporting Information Fig. S7A, S7B). The flow cytometric analysis at 8 weeks post‐transplant demonstrated a marked decrease in the CD71⁺Ter119⁺ immature erythroid cells (Fig. 4C, 4D), recapitulating the results of the original Rbpj cKO mice (Fig. 2C). Splenomegaly was also observed in Rbpj cKO recipients (Fig. 4E, 4F), similarly to that seen in the original Rbpj cKO mice (Fig. 3A, 3B).

**Figure 4**
Open in figure viewer PowerPoint

Impairment of BM erythropoiesis and splenomegaly was caused by microenvironmental factors. **(A):** Experimental design to test the microenvironment of cKO mice. **(B–F):** Phenotypes of tamoxifen‐treated *Nestin‐CreER*^T2;Rbpj^flox/flox recipients and tamoxifen‐treated *Rbpj*^flox/flox littermate recipients. **(B):** Frequency of donor‐derived cells in BM and Sp mononuclear cells. Circles and squares represent each mouse. **(C):** Representative flow cytometric data of erythroid‐lineage cells in the BM. **(D):** Number of each erythroid fraction in the BM (n = 5, each). **(E):** Appearace of Sps. Red bar, 10 mm. **(F):** Sp weights. Circles and squares represent each mouse. **(G):** Experimental design to test hematopoietic cells in *Rbpj* cKO mice. **(H, I):** Phenotypes of Ly5.1 recipients transplanted with whole BM cells from *Rbpj* cKO mice and from tamoxifen‐treated *Rbpj*^flox/flox littermates. **(H):** Representative flow cytometric data of erythroid‐lineage cells in the BM. **(I):** Number of each erythroid fraction in the BM (n = 5, each). *, p < .05. Abbreviations: BM, bone marrow; cKO, *Rbpj* cKO mice; cKO, *Rbpj* cKO recipients; cKO, recipients transplanted with whole BM cells from *Rbpj* cKO mice; Ctrl, *Rbpj*^flox/flox littermate control mice; Ctrl, *Rbpj*^flox/flox littermate control recipients; Ctrl, recipients transplanted with whole BM cells from *Rbpj*^flox/flox littermate control mice; Ly5.1, C57BL/6J‐Ly5.1; Sp, spleen.

Next, wild‐type Ly5.1 mice were used as recipients of transplants from Rbpj cKO mice and littermate control mice (Fig. 4G). Chimerism at 8 weeks post‐transplant was almost 100%, with both grafts showing indistinguishable cellularities (Supporting Information Fig. S7C, S7D). Erythropoiesis judged by flow cytometric analysis did not differ between the grafts from the Rbpj cKO and the control mice (Fig. 4H, 4I), and no splenomegaly was evident (Supporting Information Fig. S7E, S7F).

These results supported the hypothesis that the impaired erythropoiesis with splenomegaly in Rbpj cKO mice is caused by microenvironmental factors and not by hematopoietic progenitor endogenous factors.

The Erythroid‐Island‐Forming Capacity in BM Macrophages of Rbpj cKO Mice Is Damaged

Given the impairment of BM erythropoiesis caused by microenvironmental factors, we focused on the erythroid‐island‐forming capacity of BM macrophages in Rbpj cKO mice. The erythroid islands are recognized as a niche for late‐phase erythropoiesis and reportedly consists of central macrophages and surrounding erythroid cells.

An erythroid island reconstitution assay was performed using BM macrophages prepared as adherent cells from Rbpj cKO and littermate control mice and using Ter119⁺ erythroid cells prepared from the BM of wild‐type mice. The number of F4/80⁺ macrophages remaining on the dish was indistinguishable between Rbpj cKO and control mice. However, the number of reconstituted erythroid islands using macrophages from Rbpj cKO mice was significantly reduced (Fig. 5A, 5B). When each erythroid island was observed at a higher magnification, no major differences were detected in the size of the erythroid island or in the number of erythroid cells constituting each island (Supporting Information Fig. S8A). No differences were found in erythroid island formation when BM macrophages from wild‐type mice were incubated with Ter119⁺ erythroid cells prepared from Rbpj cKO and littermate control mice (Supporting Information Fig. S8B). Additionally, no differences were found in erythroid island formation when the spleen macrophages from Rbpj cKO mice and littermate control mice were incubated with Ter119⁺ erythroid cells prepared from the BM of wild‐type mice (Supporting Information Fig. S8C).

**Figure 5**
Open in figure viewer PowerPoint

The erythroid‐island‐forming capacity in bone marrow (BM) macrophages (Mo) of *Rbpj* cKO mice was damaged. **(A):** Immunofluorescence staining of reconstituted erythroid islands with BM Mo from littermate Ctrl or cKO and BM Ery from WT. Arrow: Sole Mo. Arrowheads, erythroid islands. Scale bar, 200 μm. **(B):** Number of reconstituted erythroid islands with BM Ery from WT. BM Mo were obtained from the Ctrl and cKO (n = 4, each). **(C):** Relative mRNA expression levels of *Il6* in whole BM or each Mo fraction from the Ctrl or cKO (n = 7–15, each). mRNA levels were normalized to *β‐actin* (*Actb*) mRNA expression. Circles represent each mouse. **(D):** Relative mRNA expression levels of *Hepcidin* in whole BM or each Mo fraction from the BM of Ctrl or cKO (n = 7–15, each). mRNA levels were normalized to *Actb* mRNA expression. p values were obtained by Mann‐Whitney U test. Circles represent each mouse. **(E):** BM Ery from cKO and Ctrl injected with control rat IgG and anti‐IL‐6R Ab. (i) Experimental design. (ii) Representative data of flow cytometric analysis. (iii) Number of each erythroid fraction in the BM (n = 3–5, each). p value is calculated with analysis of variance. *, p < .05. Abbreviations: anti‐IL‐6R Ab, anti‐IL‐6 receptor antibody; cKO, *Rbpj* cKO mice; Ctrl, *Rbpj*^flox/flox littermate control mice; DAPI, 4′,6‐diamidino‐2‐phenylindole; Ery, erythroid cells; LPF, low power field; Mo, macrophages; wBM, whole BM cells; WT, wild‐type C57BL/6J mice.

These results suggested that a fraction of macrophages residing in the BM, but not in the spleen, which otherwise possess the erythroid island‐forming capacity, lost such capacity in Rbpj cKO mice. It was also clear that the erythroid cells themselves kept the erythroid‐island‐forming capacity in Rbpj cKO mice.

IL‐6 Expression Is Increased in BM Central Macrophages of Rbpj cKO Mice and the Blockade of IL‐6 Reverts the Impaired BM Erythropoiesis In Vivo

The frequencies of Gr1⁻Mac1⁺F4/80⁺ whole macrophages, Gr1⁻Mac1⁺F4/80⁺Vcam1⁺ central macrophages, and Gr1⁻Mac1⁺F4/80⁺Vcam1⁻ macrophages did not differ between the Rbpj cKO and the control mice (Supporting Information Fig. S8D, S8E). Because some inflammatory cytokines are known to cause anemia, it was hypothesized that certain cytokines secreted from the central macrophages could be changed in the Rbpj cKO mice. Therefore, we measured the mRNA levels of several inflammatory cytokines such as IL‐6, IL‐1β, IL‐12β, and tumor necrosis factor‐α in the central macrophages, as well as in whole BM cells and Gr1⁻F4/80⁺Vcam1⁻ macrophages (Fig. 5C; Supporting Information Fig. S8F). Among the cells examined, a difference was found only in Il6 mRNA in the central macrophages (Fig. 5C). In addition, the IL‐6 protein levels in serum and BM fluid were not significantly different between control mice and Rbpj cKO mice (Supporting Information Fig. S8G)

The production of hepcidin is induced by IL‐6, and the hyperproduction of hepcidin causes inflammation‐induced anemia 37, 38. Agreeing with this, Hepcidin (Hamp1) mRNA expression increased in the central macrophages of cKO BM (Fig. 5D). In comparison, Hamp1 mRNA expression was not increased both in the whole BM cells and in the liver, which is the main source of hepcidin (Fig. 5D; Supporting Information Fig. S9A). Therefore, the upregulation of IL‐6 and hepcidin appeared specific to the BM central macrophages of Rbpj cKO mice.

To further investigate whether IL‐6 overproduction causes impairment of erythroid island formation in Rbpj cKO mice, anti‐IL‐6R antibody and control IgG were added to the erythroid island reconstitution assay. However, anti‐IL‐6R antibody did not rescue the impairment of erythroid island formation reconstitution with macrophages from the Rbpj cKO mice (Supporting Information Fig. S9B). This failure could be caused by the in vitro condition, and thus, we examined the effect of anti‐IL‐6 antibody an in vivo condition. Therefore, Rbpj cKO mice were injected with anti‐IL‐6R antibody or control IgG (Fig. 5E[i]). The administration of anti‐IL‐6R antibody restored the decrease of CD71⁺Ter119^dim‐high erythroid cells (II and III) in the BM of Rbpj cKO mice (Fig. 5E[ii], [iii]).

Consequently, the defect of Notch signaling in Nestin‐expressing stromal cells seems to disturb BM central macrophages, in which IL‐6 production is increased. Furthermore, IL‐6 blockade rescues the impaired BM erythropoiesis of Rbpj cKO mice in vivo conditions. These findings may explain that the phenotype of impaired erythropoiesis in the Rbpj cKO mice is induced by IL‐6 hyperproduction of central macrophages, possibly through hepcidin overproduction.

Discussion

In this study, we showed faithful marking of Nestin‐expressing BM stromal cells with GFP. Rbpj deletion specifically in these Nestin‐expressing cells was also demonstrated. The Rbpj cKO mice demonstrated impaired erythropoiesis in the BM, with compensatory extramedullary erythropoiesis in the spleen. The impaired erythropoiesis in the BM could be attributed to dysfunction of BM central macrophages through hyperproduction of IL‐6.

Some studies of BM MSCs showed discrepant results from each other depending on the mouse strains or experimental systems 19, 20, 39-41, blurring the whole picture of MSCs 42, 43. Among them, Nestin‐expressing cells have been found as the hematopoiesis‐supporting MSCs in BM 18. Analysis of previous Nestin‐GFP transgenic mice revealed two types of cells in the BM with different GFP expression levels: one existing in areas surrounding arterioles and the other existing in areas close to sinusoids 19, 20. In this study, we showed a high coincidence of GFP⁺ cells and endogenous Nestin‐expressing cells in the BM from multiple points of view. The Nestin‐expressing cells partially marked PαS and PDGFRα⁺CD51⁺ MSCs (Fig. 1). Previously, the activities of the Nestin promoter were showed to have diverse cell specificities within the central nervous system, depending on the mouse lines 28. The similar integration‐site effect of the Nestin promoter might have been traced in the BM cells, making the differences between the previous studies and ours. Nonetheless, by confining the Rbpj disruption to a very small population, we succeeded in showing here, for the first time, an erythroid lineage‐restricted phenotype.

Erythroid islands were first described based on microscopic observations in the 1950s 44. Later, CD169⁺ macrophages were found to play a role as a niche for late‐phase erythropoiesis, because mice lacking CD169⁺ macrophages showed a reduction of erythroblasts in the BM 45. Furthermore, such CD169⁺ central macrophages were shown to contribute to the maintenance of HSCs (Fig. 6, black box) 46. Nestin‐expressing MSCs are reportedly located in proximity to CD169⁺ central macrophages. According to this report, very interestingly, the capacity of Nestin‐expressing MSCs for HSC retention was weakened by the depletion of these macrophages 46. This finding implied a close communication between these cells, although the action of Nestin‐expressing cells on macrophages was not the focus of the study. Our results suggest that Nestin‐expressing cells in the BM support the ability of central macrophages to form erythroid islands through endogenous Notch signaling in Nestin‐expressing cells (Fig. 6, red box).

**Figure 6**
Open in figure viewer PowerPoint

Schematic representation of the major findings of this study in reference to the previous study 46. Black box, the findings of the previous study. Red box, the findings of this study. Abbreviations: HSC, hematopoietic stem cell; IL‐6, interleukin‐6; Nes(+) MSC, nestin‐expressing mesenchymal stromal cells.

Impairment of BM erythropoiesis accompanied by splenomegaly is a general finding in systemic chronic inflammations, such as collagen diseases, Castleman disease, and infectious endocarditis 47-49. In these conditions, extramedullary erythropoiesis is likely to compensate for the impaired erythropoiesis in the BM. Thus, one possibility is that the Rbpj cKO mice shown here model such pathophysiology. It is also known that systemic inflammation can induce anemia via hepcidin upregulation induced by IL‐6 upregulation, although the source of IL‐6 production is undefined 37, 47. In our cKO mice, elevation of Il6 and Hamp1 expression was discovered in the central macrophages but not in the other cell populations examined in the BM. In another line of reports, disruption of Notch signaling in the diverse environmental cells induced systemic inflammatory statuses 45, 50, although the hematopoietic phenotypes were largely different from the one shown here. Recently, it is reported, in a mouse model, that the aging of BM microenvironment induced ineffective hematopoiesis mimicking deletion (5q) myelodysplastic syndrome via the inflammation 51. In that study, however, specific niche cells accounting for the phenotype remains unclear. The context of inflammation is extremely variable. The current study demonstrates that limited inflammation at a very local microenvironment can indeed induce a specific type of disease status.

Rbpj was also knocked out in Nestin‐expressing cells outside the BM milieu in our cKO mice. Nestin is expressed in central and peripheral nervous system cells, including those of the brain and spinal cord, as well as in MSCs of various tissues/organs 52-54. Our current study does not exclude the possibility that those extra‐BM Nestin‐expressing cells affected erythropoiesis in a remote manner. Indeed, many nonhematopoietic organs can affect BM hematopoiesis; the most important erythropoietic regulator, erythropoietin, is produced mainly by the kidney 55, 56. Given that nerve cells could regulate hematopoiesis 16, 57, 58, the remote effect is always an open possibility.

Conclusion

We propose that, in the BM, Nestin‐expressing cells functionally support the ability of central macrophages to form erythroid islands and support erythropoiesis. It would be interesting to examine whether the current model represents human pathophysiology, even if only in part, given that Notch signaling is a universally used machinery to maintain homeostasis.

Acknowledgments

This work was supported by JSPS KAKENHI Grants JP18K16107 to T.S., JP25461408 and JP17K09898 to N.O., and JP15K15359 and JP23659482 to S.C. H.N. is supported by Gilead Sciences International Research Scholars Program in Hematology/Oncology. We thank Dr. Tasuku Honjo, Department of Immunology and Genomic Medicine, Kyoto University, for providing Rbpj floxed mice and Dr. Ryoichiro Kageyama, Institute for Frontier Life and Medical Sciences and Institute for Integrated Cell‐Material Sciences, Kyoto University, for Nestin‐CreER^T2 mice. We thank F. Miyamasu, Medical English Communications Center, University of Tsukuba, for the editorial assistance. We thank Dr. Y. Mabuchi, Department of Biochemistry and Biophysics, Graduate School of Health Care Sciences, Tokyo Medical and Dental University, for giving useful advice for CFU‐F assay. Anti‐IL‐6 receptor antibody was kindly provided by Chugai Pharmaceutical Co., Ltd. (Tokyo, Japan). This work is submitted in partial fulfillment of the requirement for the Ph.D of T.S.

Author Contribution

T.S., N.O.: conception/design, performed experiments, data analysis and interpretation, manuscript writing; H.N. conception/design, data analysis and interpretation, manuscript writing; T.K., C.S.L., R.F.: performed the experiments; H.Y.: provided antibodies, data interpretation, manuscript writing; M.S.‐Y.: data interpretation, manuscript writing; S.T.: data interpretation, manuscript writing; S.C.: data interpretation, provided overall research supervision, manuscript writing.

Disclosure of Potential Conflicts of Interest

The authors indicated no potential conflicts of interest.

Supporting Information

Filename	Description
stem3011-sup-0001-Supinfo.docxWord 2007 document , 99.2 KB	Appendix S1: Supporting Information
stem3011-sup-0002-FigureS1.tiffTIFF image, 27.9 MB	Figure S1 Schematic representation of Cre recombinase‐induced excision, profiling of GFP⁺ cells and phenotypes of GFP⁺ *cells induced by Rbpj* cKO.** A. Schematic representation of Cre recombinase‐induced excision. [i] Construction of the Rbpj floxed allele. The number refers to the exon number. [ii] Construction of the Z/EG allele. Arrow, the start point of transcription. β‐gal, β‐galactosidase. EGFP, enhanced green fluorescent protein. B. Immunofluorescent staining of the BM. [i] The imaging for VE‐cadherin, CD31 and GFP. VE‐cad, VE‐cadherin. VE‐cad, VE‐cadherin. DAPI, 2‐(4‐amidinophenyl)‐1H‐indole‐6‐carboxamidine. Scale bar, 20 μm. ★, the cavity of vessel. [ii] Isotype control. Related image to Figure 1B[ii] and supplemental Figure 1B[i]. Scale bar, 50 μm. (iii) Isotype control. Related image to Figure 1C. Scale bar, 50 μm. C. Frequency of GFP⁺ cells in CD45⁻Ter119⁻CD31⁻ cells in the BM (n = 4, each). D. Cell surface antigen expression in CD45⁻Ter119⁻GFP⁺ cells in the BM analyzed by flow cytometry (n = 3, each). f/w, tamoxifen‐treated Nestin‐CreER^T2; Z/EG; Rbpj^flox/wt mice. f/f, tamoxifen‐treated Nestin‐CreER^T2; Z/EG; Rbpj^flox/flox mice [i] Stromal cell marker expression in CD45⁻Ter119⁻GFP⁺ cells. (ii) Expression of Notch family proteins in CD45⁻Ter119⁻GFP⁺ cells. E. Representative PCR results for the Rbpj gene. Genomic DNA were obtained from sorted samples or tissues. The data was obtained by nested PCR. F. Expression of Notch‐related genes in sorted CD45⁻Ter119⁻CD31⁻GFP⁺ cells in the BM (n = 3‐5, each). mRNA levels were normalized to the Gapdh mRNA level.
stem3011-sup-0003-FigureS2.tiffTIFF image, 27.9 MB	Figure S2 Single cell qPCR of GFP⁺ cells. A. Expression of Notch‐related genes in GFP⁺ cells in the BM with single cell qPCR (Representative data). A technical difficulty in this qPCR method hampered demonstrating Nestin mRNA expression. mRNA levels were normalized to the Gapdh mRNA level. B. Rbpj mRNA expression level of single CD45⁻Ter119⁻CD31⁻GFP⁺ BM cells (pooled data from 3 independent experiments). mRNA levels were normalized to the Gapdh mRNA level. Bars indicate means. Circles indicate the data of each single cells.
stem3011-sup-0004-FigureS3.tiffTIFF image, 27.9 MB	*Figure S3 Imaging and hematopoietic stem or progenitor cells, each lineage cells of the BM and from Rbpj* cKO mice.** A. Isotype control staining. Related image to Figure 2F. Scale bar, 100 μm. B. Hematopoietic stem or progenitor cells, each lineage cells in the BM of Rbpj cKO mice (n = 5, each).
stem3011-sup-0005-FigureS4.tiffTIFF image, 27.9 MB	Figure S4 Immunofluorescent staining of CD41⁺megakaryocytes in the BM and spleen (Sp). Scale bar, 100 μm.
stem3011-sup-0006-FigureS5.tiffTIFF image, 27.9 MB	*Figure S5 Histological findings and hematopoietic stem or progenitor cells, each lineage cells in the spleen of Rbpj* cKO mice.** A. Immunofluorescent staining of the spleen for CD71 and Ter119. Scale bar, 100 μm. B. The ratio of erythroid cells (Ter119⁺ cells) to lymphoid cells (CD3⁺ or B220⁺ cells) in the spleen (n = 10, each, pooled from 2 independent experiments) C. Number of hematopoietic stem/progenitor fractions and mature cell fractions in the spleen (n = 5, each).
stem3011-sup-0007-FigureS6.tiffTIFF image, 27.9 MB	Figure S6 Phenotype of control mice and result of colony‐forming assay of control mice and cKO mice. A. Phenotypes of the BM erythroid cells and the spleen from Nestin‐CreER^T2; Rbpj^wt/wt mice (Cre + WT) and Rbpj^flox/flox mice (Cre‐ Rbpj f/f) (n = 5, each). [i] Number of each erythroid fraction of the BM. (ii) Spleen weight. B. Number of CFU‐E, BFU‐E, and CFU‐GM colonies. [i] BM (n = 6, each). (ii) Spleen (n = 6, each).
stem3011-sup-0008-FigureS7.tiffTIFF image, 27.9 MB	Figure S7 Cellularity, donor cell chimerism, and spleen in the recipients. A. Number of mononuclear cells (MNCs) in the bone marrow (BM) of the littermate control recipients and cKO recipients transplanted with BM cells from Ly5.1 mice (n = 5, each). Ctrl, Rbpj^flox/flox littermate control recipients. cKO, Rbpj cKO recipients. B. Detection of donor‐derived cells in the BM by flow cytometry. Ctrl, Rbpj^flox/flox littermate control recipients. cKO, Rbpj cKO recipients. C‐F. Phenotype of Ly5.1 recipients transplanted with whole BM cells from Rbpj cKO mice, and tamoxifen‐treated Rbpj^flox/flox littermates. Ctrl, Ly 5.1 recipients transplanted with whole BM cells from Rbpj^flox/flox littermate control mice. cKO, recipients transplanted with whole BM cells from Rbpj cKO mice. C. Number of nucleated cells in the BM (n = 5, each). D. Frequency of donor‐derived cells in the BM and spleens. Circles or squares represent each mouse. Sp, spleen. E. Appearance of the spleen. Red bar, 10 mm. F. Spleen weights. Circles or squares represent each mouse.
stem3011-sup-0009-FigureS8.tiffTIFF image, 27.9 MB	Figure S8 Erythroid island reconstitution and bone marrow macrophages. A. Immunofluorescence staining of reconstituted erythroid islands with BM macrophages (Mo) from littermate control or Rbpj cKO mice and BM erythroid cells (Ery) from wild‐type C57BL/6J mice. Scale bar, 20 μm. WT, C57BL/6J mice. Ctrl, Rbpj^flox/flox littermate control mice. cKO, Rbpj cKO mice. B. Number of reconstituted erythroid islands with BM macrophages from wild‐type C57BL/6J mice. Erythroid cells were obtained from the BM of littermate control mice and Rbpj cKO mice (n = 3, each). C. Number of reconstituted erythroid islands with erythroid cells from the BM of wild‐type C57BL/6J mice. Macrophages were obtained from the spleens of littermate control mice or Rbpj cKO mice (n = 3, each). D. Gating strategy of macrophages in the BM. Red Box, Gr1⁻Mac1⁺F4/80⁺Vcam1⁺ central macrophages. E. Frequency of each macrophage fraction in the BM measured by flow cytometry (n = 4, each). Ctrl, Rbpj^flox/flox littermate control mice. cKO, Rbpj cKO mice. F. Relative mRNA expression levels of inflammatory cytokines in whole BM or each macrophage fraction from the control mice or Rbpj cKO mice (n = 8‐12, each). mRNA levels were normalized to Actb mRNA expression. G. IL‐6 levels of the serum and the BM fluid (n = 5‐6, each).
stem3011-sup-0010-FigureS9.tiffTIFF image, 27.9 MB	*Figure S9 Hepcidin* mRNA expression of the liver and the effect of anti‐IL‐6R antibody on the erythroid‐island reconstitution capacity in vitro.** A. Hamp1 mRNA Expression of liver from Rbpj cKO mice and littermate control mice (n = 4, each). mRNA levels were normalized to the Actb mRNA level B. Number of erythroid islands reconstituted with control and Rbpj cKO macrophages treated with anti‐IL‐6R antibody (anti‐IL‐6R Ab) or control IgG (IgG). (n = 5, each)
stem3011-sup-0011-TableS1.xlsxExcel 2007 spreadsheet , 34.4 KB	Table S1 List of antibodies used in this study.
stem3011-sup-0012-TableS2.xlsxExcel 2007 spreadsheet , 30.8 KB	Table S2 List of primers and Taqman probes used in this study.

Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.

References

1Chiba S. Notch signaling in stem cell systems. Stem Cells 2006; 24: 2437– 2447.
Wiley Online Library CAS PubMed Web of Science®Google Scholar
2Iso T, Hamamori Y, Kedes L. Notch signaling in vascular development. Arterioscler Thromb Vasc Biol 2003; 23: 543– 553.
Crossref CAS PubMed Web of Science®Google Scholar
3Radtke F, Fasnacht N, Macdonald HR. Notch signaling in the immune system. Immunity 2010; 32: 14– 27.
Crossref CAS PubMed Web of Science®Google Scholar
4Apelqvist A, Li H, Sommer L et al. Notch signalling controls pancreatic cell differentiation. Nature 1999; 400: 877– 881.
Crossref CAS PubMed Web of Science®Google Scholar
5Hozumi K, Mailhos C, Negishi N et al. Delta‐like 4 is indispensable in thymic environment specific for T cell development. J Exp Med 2008; 205: 2507– 2513.
Crossref CAS PubMed Web of Science®Google Scholar
6Hozumi K, Negishi N, Suzuki D et al. Delta‐like 1 is necessary for the generation of marginal zone B cells but not T cells in vivo. Nat Immunol 2004; 5: 638– 644.
Crossref CAS PubMed Web of Science®Google Scholar
7Kumano K, Chiba S, Kunisato A et al. Notch1 but not Notch2 is essential for generating hematopoietic stem cells from endothelial cells. Immunity 2003; 18: 699– 711.
Crossref CAS PubMed Web of Science®Google Scholar
8Robert‐Moreno A, Espinosa L, de la Pompa JL et al. RBPjkappa‐dependent Notch function regulates Gata2 and is essential for the formation of intra‐embryonic hematopoietic cells. Development 2005; 132: 1117– 1126.
Crossref CAS PubMed Web of Science®Google Scholar
9Yao D, Huang Y, Huang X et al. Protein O‐fucosyltransferase 1 (Pofut1) regulates lymphoid and myeloid homeostasis through modulation of Notch receptor ligand interactions. Blood 2011; 117: 5652– 5662.
Crossref CAS PubMed Web of Science®Google Scholar
10Yoda M, Kimura T, Tohmonda T et al. Dual functions of cell‐autonomous and non‐cell‐autonomous ADAM10 activity in granulopoiesis. Blood 2011; 118: 6939– 6942.
Crossref CAS PubMed Web of Science®Google Scholar
11Kim YW, Koo BK, Jeong HW et al. Defective Notch activation in microenvironment leads to myeloproliferative disease. Blood 2008; 112: 4628– 4638.
Crossref CAS PubMed Web of Science®Google Scholar
12Wang L, Zhang H, Rodriguez S et al. Notch‐dependent repression of miR‐155 in the bone marrow niche regulates hematopoiesis in an NF‐kappaB‐dependent manner. Cell Stem Cell 2014; 15: 51– 65.
Crossref CAS PubMed Web of Science®Google Scholar
13Klinakis A, Lobry C, Abdel‐Wahab O et al. A novel tumour‐suppressor function for the Notch pathway in myeloid leukaemia. Nature 2011; 473: 230– 233.
Crossref CAS PubMed Web of Science®Google Scholar
14Calvi LM, Adams GB, Weibrecht KW et al. Osteoblastic cells regulate the haematopoietic stem cell niche. Nature 2003; 425: 841– 846.
Crossref CAS PubMed Web of Science®Google Scholar
15Kiel MJ, Yilmaz OH, Iwashita T et al. SLAM family receptors distinguish hematopoietic stem and progenitor cells and reveal endothelial niches for stem cells. Cell 2005; 121: 1109– 1121.
Crossref CAS PubMed Web of Science®Google Scholar
16Yamazaki S, Ema H, Karlsson G et al. Nonmyelinating Schwann cells maintain hematopoietic stem cell hibernation in the bone marrow niche. Cell 2011; 147: 1146– 1158.
Crossref CAS PubMed Web of Science®Google Scholar
17Sugiyama T, Kohara H, Noda M et al. Maintenance of the hematopoietic stem cell pool by CXCL12‐CXCR4 chemokine signaling in bone marrow stromal cell niches. Immunity 2006; 25: 977– 988.
Crossref CAS PubMed Web of Science®Google Scholar
18Mendez‐Ferrer S, Michurina TV, Ferraro F et al. Mesenchymal and haematopoietic stem cells form a unique bone marrow niche. Nature 2010; 466: 829– 834.
Crossref CAS PubMed Web of Science®Google Scholar
19Kunisaki Y, Bruns I, Scheiermann C et al. Arteriolar niches maintain haematopoietic stem cell quiescence. Nature 2013; 502: 637– 643.
Crossref CAS PubMed Web of Science®Google Scholar
20Asada N, Kunisaki Y, Pierce H et al. Differential cytokine contributions of perivascular haematopoietic stem cell niches. Nat Cell Biol 2017; 19: 214– 223.
Crossref CAS PubMed Web of Science®Google Scholar
21Pierini A, Nishikii H, Baker J et al. Foxp3(+) regulatory T cells maintain the bone marrow microenvironment for B cell lymphopoiesis. Nat Commun 2017; 8: 15068.
Crossref CAS PubMed Web of Science®Google Scholar
22Wu JY, Purton LE, Rodda SJ et al. Osteoblastic regulation of B lymphopoiesis is mediated by Gs{alpha}‐dependent signaling pathways. Proc Natl Acad Sci USA 2008; 105: 16976– 16981.
Crossref CAS PubMed Web of Science®Google Scholar
23Tamura S, Suzuki‐Inoue K, Tsukiji N et al. Podoplanin‐positive periarteriolar stromal cells promote megakaryocyte growth and proplatelet formation in mice by CLEC‐2. Blood 2016; 127: 1701– 1710.
Crossref CAS PubMed Web of Science®Google Scholar
24Chasis JA, Mohandas N. Erythroblastic islands: Niches for erythropoiesis. Blood 2008; 112: 470– 478.
Crossref CAS PubMed Web of Science®Google Scholar
25Korolnek T, Hamza I. Macrophages and iron trafficking at the birth and death of red cells. Blood 2015; 125: 2893– 2897.
Crossref CAS PubMed Web of Science®Google Scholar
26Klei TR, Meinderts SM, van den Berg TK et al. From the Cradle to the Grave: The role of macrophages in erythropoiesis and erythrophagocytosis. Front Immunol 2017; 8: 73.
Crossref PubMed Web of Science®Google Scholar
27Han H, Tanigaki K, Yamamoto N et al. Inducible gene knockout of transcription factor recombination signal binding protein‐J reveals its essential role in T versus B lineage decision. Int Immunol 2002; 14: 637– 645.
Crossref CAS PubMed Web of Science®Google Scholar
28Imayoshi I, Ohtsuka T, Metzger D et al. Temporal regulation of Cre recombinase activity in neural stem cells. Genesis 2006; 44: 233– 238.
Wiley Online Library CAS PubMed Web of Science®Google Scholar
29Novak A, Guo C, Yang W et al. Z/EG, a double reporter mouse line that expresses enhanced green fluorescent protein upon Cre‐mediated excision. Genesis 2000; 28: 147– 155.
Wiley Online Library CAS PubMed Web of Science®Google Scholar
30Kawamoto T, Kawamoto K. Preparation of thin frozen sections from nonfixed and undecalcified hard tissues using Kawamot's film method (2012). Methods Mol Biol 2014; 1130: 149– 164.
Crossref CAS PubMed Web of Science®Google Scholar
31Fraser ST, Midwinter RG, Coupland LA et al. Heme oxygenase‐1 deficiency alters erythroblastic island formation, steady‐state erythropoiesis and red blood cell lifespan in mice. Haematologica 2015; 100: 601– 610.
Crossref CAS PubMed Web of Science®Google Scholar
32Kusakabe M, Hasegawa K, Hamada M et al. c‐Maf plays a crucial role for the definitive erythropoiesis that accompanies erythroblastic Island formation in the fetal liver. Blood 2011; 118: 1374– 1385.
Crossref CAS PubMed Web of Science®Google Scholar
33Morikawa S, Mabuchi Y, Kubota Y et al. Prospective identification, isolation, and systemic transplantation of multipotent mesenchymal stem cells in murine bone marrow. J Exp Med 2009; 206: 2483– 2496.
Crossref CAS PubMed Web of Science®Google Scholar
34Sato Y, Mabuchi Y, Miyamoto K et al. Notch2 signaling regulates the proliferation of murine bone marrow‐derived mesenchymal stem/stromal cells via c‐Myc Expression. PloS One 2016; 11: e0165946.
Crossref PubMed Web of Science®Google Scholar
35Socolovsky M, Nam H, Fleming MD et al. Ineffective erythropoiesis in Stat5a(−/−)5b(−/−) mice due to decreased survival of early erythroblasts. Blood 2001; 98: 3261– 3273.
Crossref CAS PubMed Web of Science®Google Scholar
36Chen K, Liu J, Heck S et al. Resolving the distinct stages in erythroid differentiation based on dynamic changes in membrane protein expression during erythropoiesis. Proc Natl Acad Sci USA 2009; 106: 17413– 17418.
Crossref CAS PubMed Web of Science®Google Scholar
37Nemeth E, Rivera S, Gabayan V et al. IL‐6 mediates hypoferremia of inflammation by inducing the synthesis of the iron regulatory hormone hepcidin. J Clin Invest 2004; 113: 1271– 1276.
Crossref CAS PubMed Web of Science®Google Scholar
38Wang CY, Babitt JL. Hepcidin regulation in the anemia of inflammation. Curr Opin Hematol 2016; 23: 189– 197.
Crossref CAS PubMed Web of Science®Google Scholar
39Greenbaum A, Hsu YM, Day RB et al. CXCL12 in early mesenchymal progenitors is required for haematopoietic stem‐cell maintenance. Nature 2013; 495: 227– 230.
Crossref CAS PubMed Web of Science®Google Scholar
40Ding L, Saunders TL, Enikolopov G et al. Endothelial and perivascular cells maintain haematopoietic stem cells. Nature 2012; 481: 457– 462.
Crossref CAS PubMed Web of Science®Google Scholar
41Sung JH, Yang HM, Park JB et al. Isolation and characterization of mouse mesenchymal stem cells. Transplant Proc 2008; 40: 2649– 2654.
Crossref CAS PubMed Web of Science®Google Scholar
42Joseph C, Quach JM, Walkley CR et al. Deciphering hematopoietic stem cells in their niches: A critical appraisal of genetic models, lineage tracing, and imaging strategies. Cell Stem Cell 2013; 13: 520– 533.
Crossref CAS PubMed Web of Science®Google Scholar
43Chen KG, Johnson KR, McKay RDG et al. Concise Review: Conceptualizing Paralogous Stem‐Cell Niches and Unfolding Bone Marrow Progenitor Cell Identities. Stem Cells 2018; 36: 11– 21.
Wiley Online Library PubMed Web of Science®Google Scholar
44Bessis M. Erythroblastic island, functional unity of bone marrow. Rev Hematol 1958; 13: 8– 11.
CAS PubMed Google Scholar
45Chow A, Huggins M, Ahmed J et al. CD169(+) macrophages provide a niche promoting erythropoiesis under homeostasis and stress. Nat Med 2013; 19: 429– 436.
Crossref CAS PubMed Web of Science®Google Scholar
46Chow A, Lucas D, Hidalgo A et al. Bone marrow CD169+ macrophages promote the retention of hematopoietic stem and progenitor cells in the mesenchymal stem cell niche. J Exp Med 2011; 208: 261– 271.
Crossref CAS PubMed Web of Science®Google Scholar
47Fraenkel PG. Understanding anemia of chronic disease. Hematology/the Education Program of the American Society of Hematology American Society of Hematology Education. Dent Prog 2015; 2015: 14– 18.
Google Scholar
48Chapman J, Azevedo AM. Splenomegaly. Treasure Island, FL: StatPearls, 2017.
Google Scholar
49Motyckova G, Steensma DP. Why does my patient have lymphadenopathy or splenomegaly? Hematol Oncol Clin North Am 2012; 26: 395– 408.
Crossref PubMed Web of Science®Google Scholar
50Dumortier A, Durham AD, Di Piazza M et al. Atopic dermatitis‐like disease and associated lethal myeloproliferative disorder arise from loss of Notch signaling in the murine skin. PloS One 2010; 5: e9258.
Crossref CAS PubMed Web of Science®Google Scholar
51Mei Y, Zhao B, Basiorka AA et al. Age‐related inflammatory bone marrow microenvironment induces ineffective erythropoiesis mimicking del(5q) MDS. Leukemia 2018; 32: 1023– 1033.
Crossref CAS PubMed Web of Science®Google Scholar
52Matsumura S, Takagi K, Okuda‐Ashitaka E et al. Characterization of nestin expression in the spinal cord of GFP transgenic mice after peripheral nerve injury. Neuroscience 2010; 170: 942– 953.
Crossref CAS PubMed Web of Science®Google Scholar
53Vaittinen S, Lukka R, Sahlgren C et al. The expression of intermediate filament protein nestin as related to vimentin and desmin in regenerating skeletal muscle. J Neuropathol Exp Neurol 2001; 60: 588– 597.
Crossref CAS PubMed Web of Science®Google Scholar
54Jeremias Tda S, Machado RG, Visoni SB et al. Dermal substitutes support the growth of human skin‐derived mesenchymal stromal cells: Potential tool for skin regeneration. PloS One 2014; 9: e89542.
Crossref PubMed Web of Science®Google Scholar
55Souma T, Suzuki N, Yamamoto M. Renal erythropoietin‐producing cells in health and disease. Front Physiol 2015; 6: 167.
Crossref PubMed Web of Science®Google Scholar
56Obara N, Suzuki N, Kim K et al. Repression via the GATA box is essential for tissue‐specific erythropoietin gene expression. Blood 2008; 111: 5223– 5232.
Crossref CAS PubMed Web of Science®Google Scholar
57Katayama Y, Battista M, Kao WM et al. Signals from the sympathetic nervous system regulate hematopoietic stem cell egress from bone marrow. Cell 2006; 124: 407– 421.
Crossref CAS PubMed Web of Science®Google Scholar
58Mendez‐Ferrer S, Lucas D, Battista M et al. Haematopoietic stem cell release is regulated by circadian oscillations. Nature 2008; 452: 442– 447.
Crossref CAS PubMed Web of Science®Google Scholar

Volume37, Issue7

July 2019

Pages 924-936

This article also appears in:

Key Hematopoietic Papers from the STEM CELLS Journals

Journal list menu

Tools

Follow journal

Notch Signaling in Nestin‐Expressing Cells in the Bone Marrow Maintains Erythropoiesis via Macrophage Integrity