This paper is dedicated to the memory of Sandy C Marks Jr.
Characterization of a Novel Bipotent Hematopoietic Progenitor Population in Normal and Osteopetrotic Mice†
Version of Record online: 22 MAR 2004
Copyright © 2004 ASBMR
Journal of Bone and Mineral Research
Volume 19, Issue 7, pages 1137–1143, July 2004
How to Cite
Blin-Wakkach, C., Wakkach, A., Rochet, N. and Carle, G. F. (2004), Characterization of a Novel Bipotent Hematopoietic Progenitor Population in Normal and Osteopetrotic Mice. J Bone Miner Res, 19: 1137–1143. doi: 10.1359/JBMR.040318
The authors have no conflict of interest
- Issue online: 2 DEC 2009
- Version of Record online: 22 MAR 2004
- Manuscript Accepted: 22 MAR 2004
- Manuscript Revised: 4 FEB 2004
- Manuscript Received: 17 DEC 2003
- bone marrow;
- B-cell differentiation
Several reports indicate that osteoclasts and B-lymphocytes share a common progenitor. This study focuses on the characterization of this bipotent progenitor from the bone marrow of the osteopetrotic oc/oc mouse, where the bipotent progenitor population is amplified, and of normal mice.
Introduction: Osteoclasts have a myelomonocytic origin, but they can also arise in vitro from pro-B-cells, suggesting that a subset of normal pro-B-cells is uncommitted and may reorient into the myeloid lineage representing a B-lymphoid/osteoclastic progenitor. The aim of this study was to characterize this progenitor population.
Materials and Methods: The osteopetrotic oc/oc mouse was used as a choice model because it displays an increased number of both osteoclasts and pro-B-cells in the bone marrow. Our results have been confirmed in normal littermates. Bone marrow cells from these animals were analyzed by flow cytometry. After sorting, the cells were cultured under different conditions to assess their differentiation capacity.
Results: Pro-B-cells from oc/oc and normal mice include an unusual biphenotypic population expressing markers from the B-lymphoid (CD19, CD43, CD5) and the myeloid (F4/80) lineages. This population also expresses progenitor markers (CD34 and Flt3) and is uncommitted. After sorting from the oc/oc bone marrow, this population is able to differentiate in vitro into osteoclast-like cells in the presence of RANKL and macrophage colony-stimulating factor (M-CSF), into dendritic-like cells in the presence of granulocyte/macrophage colony-stimulating factor (GM-CSF), interleukin (IL)-4, and TNFα, and into immature B-cells when seeded onto ST2 cells in the presence of IL-7.
Conclusion: Our results show the existence of a novel bipotent biphenotypic hematopoietic progenitor population present in the bone marrow that has retained the capacity to differentiate into myeloid and B-lymphoid cells.
BONE MORPHOGENESIS AND remodeling are tightly controlled by a balance between osteoblasts, the bone forming cells arising from mesenchymal stem cells, and osteoclasts (OCLs), the bone resorbing cells that differentiated from the myeloid lineage. OCLs are multinucleated giant cells formed by the fusion of myelomonocytic precursors. This myelomonocytic origin has been well documented and is similar to that of macrophages and dendritic cells.(1) However, several studies have shown that OCLs could also be obtained in vitro from B-lymphoid progenitors.(2–4) B-lymphocytes differentiate from common lymphoid progenitors, and this differentiation pathway includes various steps, based on the phenotypic analysis of the cells: pro-B-, pre-B-, immature, and mature B-cells. The myelomonocytic and the lymphoid lineage cells are thought to be distantly related, but the in vitro differentiation of OCLs from B-lymphoid progenitors strongly suggests the existence of a bipotent progenitor population. However, such a bipotent osteoclastic/B-lymphoid population probably represents a very rare population, making its characterization difficult. Given the previously reported data,(2,3) this population represents a subset of B220+ pro-B-cells that have retained the potential for differentiating into diverse hematopoietic lineages, including OCLs. Therefore, analysis of a mouse model displaying both a pro-B-cell amplification and an increased osteoclastogenesis would be very useful for the characterization of such a bipotent progenitor population.
Osteopetrosis is a disease characterized by impaired bone resorption leading to a severe reduction of the medullary cavity and extramedullary hematopoiesis. Several genetically modified or spontaneous mutant osteopetrotic mice have been described.(5) However, many of these mutants present an arrest or a decrease in the OCL differentiation pathway and therefore are not adapted for the analysis of OCL progenitors.(6) Among the osteopetrotic models, the oc/oc mouse is characterized by an absence of bone resorption caused by inactive but fully differentiated OCLs.(7) We have shown that the oc mutation consists of a 1.6-kb deletion in the Tcirg1 gene encoding the 116-kDa subunit of the vacuolar proton pump (V-ATPase).(8) The V-ATPase is a protein complex that mediates H+ transport into the resorption lacunae where a low pH is necessary to dissolve mineralized bone material and for optimal function of proteases that degrade the organic bone matrix.(9) In the OCLs of the oc/oc mouse, the V-ATPase is absent at the cell surface in contact with the bone matrix,(10) and consequently, the degradation of the bone matrix is impaired. Recently, we have shown that myelopoiesis and osteoclastogenesis are increased in oc/oc mice (C Blin-Wakkach, A Wakkach, PM Sexton, N Rochet, GF Carle, unpublished data, 2004). Furthermore, we have also shown that B-lymphopoiesis is blocked at the pro-B to pre-B transition, leading to an increased pro-B-cell population (C Blin-Wakkach, A Wakkach, PM Sexton, N Rochet, GF Carle, unpublished data, 2004). Thus, we hypothesized that the oc/oc mouse could be a good model to isolate and characterize bipotent progenitors. To test this hypothesis, we have further analyzed the pro-B-cells present in the bone marrow of oc/oc mice and investigated whether they have a bipotent differentiation potential.
MATERIALS AND METHODS
Two pairs of (C57BL/6J × C3HheB/FeJ) F1 oc/+ mice were initially obtained from the Jackson Laboratory (Bar Harbor, ME, USA) and maintained in our central animal facility in accordance with the general guidelines of the Direction des Services Veterinaires. Mice were genotyped using a PCR test using three primers, 5′-CCCTTCTCTGCCTTTCACC-3′, 5′-CTGCTTACAATTTGGGGAGG-3′, and 5′-CAAGTGGG-GACACACATCG-3′, allowing the detection of the 1.6-kb deletion present in the mutant allele.(8)
Cell preparation and flow cytometry analysis
Bone marrow cells from 17-day-old normal mice were collected by flushing femoral shafts, and those from 17-day-old oc/oc mice were obtained by crushing the femora into small pieces and vigorous pipetting. Single cell suspensions were incubated with Fc block CD16/32 (clone 2.4G2) and stained with FITC-, phycoerythrin-, or Cy-chrome antibodies reactive to CD11b (M1/70), CD43 (S7), B220 (RA3-6B2), CD19 (1D3), CD5 (53-7.3), CD34 (RAM34), CD62L (MEL14), CD8 (53-6.7), CD23 (B3B4), CD40 (3/23), Gr1 (RB6-8C5), and F4/80 (all purchased from Becton-Dickinson). All staining steps were performed at 4°C in PBS with 0.1% bovine serum albumin (BSA) and 0.02 mM NaN3. After three washes, the labeled cells were analyzed on a FACScan (Becton Dickinson).
The B220+CD11b+ population from 17-day-old oc/oc mice was sorted onto a cell sorter (FACS-vantage; Becton Dickinson) after labeling with FITC-anti B220 (RA3-6B2) and phycoerythrin-anti CD11b (M1/70) antibodies. The B220+ population from normal mice was enriched by sorting with the B220 microbead system (Miltenyi-Biotech).
Cell morphology studies
After a cytospin procedure, the morphology of sorted B220+CD11b+ cells was assessed by light microscopy examination after May Grünwald-Giemsa staining.
Total RNA was extracted by adsorption onto silica membranes according to the manufacturer's procedure (Macherey-Nagel). PCR analysis was performed on 1 μl of the RT reaction products in 25 μl with 10 pmol of the following primers for 30 cycles: Pax5, 5′-AGCAGCC-CCCCAATCAG-3′ and 5′-TGCGTCACGGAGCCTGTA-3′ (GenBank accession NM_008782, position 463-479 and 512-529); Flt3, 5′-TTCGGAACAGACATCAGATGCT-3′ and 5′-AAGGGTTCCCCCACTTTCAG-3′ (GenBank accession NM_010229, position 706-727 and 817-836); Scl 5′-AGGCCCTCCCCATATGAGAT-3′ and 5′-GCTCAGCAA-ATGCCCCATT-3′ (GenBank accession U01530, position 3197-3216 and 3352-3370); PU.1 5′-CGTGCAA-AATGGAAGGGTTT-3′ and 5′-GCTCTGAATCGTA-AGTAACCAAGTCA-3′ (GenBank accession X17463, position 851-870 and 894-919); c-Fms 5′-AACAA-GTTCTACAAACTGGTG-3′ and 5′-AAGCCT-GTAGTCTAAGCATCT-3′ (GenBank accession BC043054, position 2654-2674 and 3386-3406); and Gapdh 5′-ACCACAGTCCATGCCATCAC-3′ and 5′-TCCACCAC-CCTGTTGCTGTA-3′ (GenBank accession NM_008084.1, position 566-585 and 998-1017). In all analyses, RNA in absence of RT served as a control for genomic DNA contamination.
Immunoglobulin gene rearrangements
DJH rearrangements in the Ig locus were detected according to the protocol described by Schlissel et al.(11) DNA was extracted in a solution containing 0.5% SDS, 0.1 M NaCl, 5 mM EDTA, and 50 mM Tris-HCl (pH 7.5) in the presence of proteinase K at 50°C overnight, followed by a phenol/chloroform extraction. The PCR amplification used a mixture of degenerated oligonucleotides homologous to all the members of the Dfl16 and Dsp2D gene families (5′-GGAATTCGMTTTTTGTSAAGGGATCTACTACTGTG-3′) and a primer located downstream of the JH4 gene (5′-TCCCTCAAATGAGCCTCCAAAGTCC-3′). DNA amplification was performed on 50 ng of genomic DNA for 30 cycles of 1 minute at 94°C, 1 minute at 60°C, and 1.75 minutes at 72°C as described.(11)
In vitro differentiation
For OCL differentiation, sorted B220+CD11b+ cells were plated in a collagen-coated 24-well plate at 1.5 × 105 cells/well in αMEM medium supplemented with 5% FCS (Hyclone, Perbio), 25 ng/ml mouse recombinant macrophage colony-stimulating factor (M-CSF), and 30 ng/ml mouse RANKL (R&D Systems). After a 4-week culture, cells were fixed in a 2% glutaraldehyde solution and stained for TRACP activity with the leukocyte acid phosphatase kit (Sigma). OCL differentiation was evaluated by the presence of multinucleated TRACP+ cells.
Dendritic-like cells were obtained after seeding 5 × 104 sorted B220+CD11b+ cells in a 24-well collagen-coated plate in αMEM medium supplemented with interleukin (IL)-4 (20 ng/ml), granulocyte/macrophage colony-stimulating factor (GM-CSF; 2.5 ng/ml), and TNFα (10 ng/ml). After 10 days, cells were analyzed after May Grünwald-Giemsa staining for their morphology. The capacity of the sorted B220+CD11b+ cell population to differentiate into B-cells was tested by seeding overconfluent ST2 stromal cells in the presence of 10 ng/ml IL-7. After 10 days, cells were analyzed for the expression of B-lymphoid markers.
Amplification of an uncommitted bilineage population in the bone marrow of oc/oc mice
We have recently characterized the hematopoietic populations present in the oc/oc bone marrow. We have shown that the percentage of CD11b+ cells and pre-OCLs is increased compared with normal littermates (NLMs; C Blin-Wakkach, A Wakkach, PM Sexton, N Rochet, GF Carle, unpublished data, 2004). We have also described a block in B-lymphopoiesis at the pro-B- to pre-B-cell transition, leading to an increased percentage of B220+CD43+ pro-B-cells, in the bone marrow of oc/oc mice (Fig. 1A; C Blin-Wakkach, A Wakkach, PM Sexton, N Rochet, GF Carle, unpublished data, 2004).
Given the increased percentage of CD11b+ myeloid population, the B220+CD43+ pro-B population of oc/oc mice was further analyzed for the expression of CD11b. A triple flow cytometry analysis revealed that 85% of the B220+CD43+ pro-B-cells expressed CD11b (Fig. 1B). Bone marrow cells of normal and oc/oc mice were thus analyzed for the expression of B220 and CD11b. Interestingly, we identified an unusual B220+CD11b+ cell population that represents 0.6% of the hematopoietic cells in normal mice and that is amplified up to 15% in oc/oc mice (Fig. 2A). Phenotypic triple flow cytometry analysis of the oc/oc B220+CD11b+ cell population showed a coexpression of markers from both lymphoid and myeloid lineages such as the lymphoid markers CD43, CD19, and CD5 (50% of CD5+ cells) and the macrophage marker F4/80 (Fig. 2B). However, the B220+CD11b+ cell population seemed to be negative for the CD8 antigen and for the granulocyte antigen Gr1 (Fig. 2B). This biphenotypic B220+CD11b+ population also expressed the progenitor marker CD34 and the homing receptor CD62L (Fig. 2B).
The B220+CD11b+ population was sorted from the bone marrow of oc/oc mice (Fig. 3A). These cells displayed the morphological characteristics of undifferentiated and immature cells (Fig. 3B). These results were further supported by semiquantitative RT-PCR analysis of the FACS-sorted B220+CD11b+ population that revealed that it expressed Flt3, whose expression is restricted to early progenitors in normal bone marrow,(12)PU.1, a regulator of myelopoiesis and B-lymphopoiesis,(13) and c-fms, the M-CSF receptor (Fig. 3C). However, it expressed Pax5 and SCL (stem cell leukemia) mRNA at a very low level, two transcription factors that are expressed by lymphoid and myeloid precursors in a mutually exclusive way (Fig. 3C).(14,15) Altogether, these findings strongly suggest that the B220+CD11b+ cell population is uncommitted. Finally, analysis of the Ig gene locus by PCR amplification revealed that the sorted B220+CD11b+ population displayed the characteristic DJH rearrangements, as observed for genuine pro-B-cells (Fig. 3D).(16)
In the bone marrow from NLMs, the B220+CD11b+ cell population represents <1% of the hematopoietic cells. To determine whether these cells are equivalent to the one observed in the oc/oc mice, we first sorted the B220+ cells from normal mice, allowing the enrichment of the B220+CD11b+ population up to 4% (Fig. 4). A triple flow cytometry on these B220+CD11b+ cells revealed that this population expressed the lymphoid markers CD43+ and CD19+ as well as the myeloid marker F4/80+ but not CD8, indicating that this population is biphenotypic (Fig. 4). These cells also expressed CD62L+ and the progenitor marker CD34+ as the population analyzed in the mutant mice (Fig. 4), showing that they correspond to a progenitor population. This B220+CD11b+ biphenotypic population is different, for the expression of the analyzed markers, from the B220+CD11b− cells, which are also CD19+ CD43low CD34− CD62L− F4/80− CD8−, and from the B220−CD11b+ cells, which are also CD19− CD43high CD34− CD62Llow F4/80+ CD8−, and therefore forms a distinct population (Fig. 4). Thus, the biphenotypic population described in the oc/oc bone marrow also exists in NLMs and could define a new hematopoietic progenitor population.
B220+CD11b+ population has multiple differentiating potential
The characterization of the B220+CD11b+ population indicates that these cells are committed neither to the lymphoid nor to the myeloid lineages. Hence, the question arose of whether these cells could give rise to both myeloid and lymphoid lineages.
Sorted B220+CD11b+ cells from oc/oc mice were thus cultured in the presence of RANKL and M-CSF for 4 weeks to induce osteoclast differentiation. At this stage, the presence of multinucleated TRACP+ cells reflected the osteoclastogenic potential of this population (Fig. 5A). These results show that these cells are capable to give rise to osteoclast-like cells and suggest that the B220+CD11b+ population may participate in the increase of osteoclastogenesis in the bone marrow of oc/oc mice.
OCL and dendritic cells share a common progenitor.(6) We therefore tested the capacity of the biphenotypic population to generate dendritic-like cells after 10 days of culture in the presence of GM-CSF, IL-4, and TNFα. Light microscope examination after May Grünwald-Giemsa (MGG) staining showed the characteristic morphology of dendritic cells with multiple cytoplasmic projections for a large majority of the cells (Fig. 5B).
Finally, to determine whether the B220+CD11b+ cell population has a B-lymphoid potential, the cells were seeded on ST2 stromal cells in the presence of IL-7 and cultured for 10 days. Over this period of time, the cells displayed a marked proliferation (Fig. 5C). They switched from a B220+CD43+ pro-B phenotype to a B220lowCD43− pre-B and immature B-cell phenotype (Fig. 5D), while losing the CD11b marker at the same time (data not shown). These B220lowCD43− cells were CD19+, and ∼50% expressed the co-stimulatory molecule CD40 (Fig. 5E). These CD19+CD40+ cells express CD23, a marker of B-cell activation,(17) whereas the CD19+CD40− cells do not (Fig. 5E). All these results strongly suggest that the B220+CD11b+ cell population has retained the capacity to differentiate into pre-B- and immature B-lymphocytes.
Altogether, the above data show that the biphenotypic B220+CD11b+ population has a bipotent B-lymphoid/myeloid capacity.
Recent studies have challenged the current concept of pro-B-cell commitment. Several lines of evidence indicate that a subset of normal pro-B-cells is uncommitted and may reorient into the myeloid lineage. The Pax5−/− mice display a block in B-cell development at the pro-B stage.(2) Pro-B-cells purified from these mice have a multilineage potential, and on in vitro stimulation with appropriate cytokines, can differentiate into macrophages, OCLs, dendritic cells, granulocytes, and natural killer (NK) cells. The in vivo reimplantation of these Pax5−/− pro-B-cells can give rise to myeloid cells and T-lymphocytes.(18) Furthermore, OCLs have been obtained in vitro from a subset of normal pro-B-cells.(3,4) All these data support the existence of a normal bipotent progenitor population able to differentiate into OCLs as well as into B-lymphocytes. However, as for all progenitor cells, this bipotent population is probably very rare, and the only phenotypic markers reported to date for these cells are the B220 and CD43 antigens.(18)
Based on these observations, we have used the osteopetrotic oc/oc mouse as a model for the characterization of this progenitor population. Indeed, in the oc/oc bone marrow, the pro-B-cell population is 5-fold increased compared with NLMs, and the number of OCLs is also elevated (C Blin-Wakkach, A Wakkach, PM Sexton, N Rochet, GF Carle, unpublished data, 2004). Analysis of this pro-B population revealed the presence in the bone marrow of an unusual B220+CD11b+ population expressing B-lymphoid markers such as CD19 and CD5, as well as the macrophage F4/80 protein. This biphenotypic population is distinct from common lymphoid progenitors (CLPs) and from monocytes, because CLPs are B220−CD11b and monocytes are CD43+B220−. However, we show that this population displays DJH rearrangements characteristic of pro-B-cells and therefore represent a subset of pro-B-cells.(16) The very weak expression of Pax-5 and SCL mRNA, specifically expressed by lymphoid and myeloid precursors, respectively,(14,15) indicates that the B220+CD11b+ cell population is uncommitted. Furthermore, the expression of CD34 and Flt3, involved in the expansion and mobilization of early hematopoietic progenitors,(12) their immature morphology, and their ability to give rise in vitro to immature B-cells, to dendritic cells and OCLs indicate that they may represent a bipotent progenitor population. However, further analysis is required to clarify whether the bipotent potential of these cells does also exist in vivo. The analysis of the oc/oc model led us to isolate and characterize a biphenotypic population that possess myeloid and lymphoid differentiation capacity in vitro. Based on these results, we were then able to define a set of characteristic markers allowing the analysis of this population in normal mouse. Indeed, we have also detected in the bone marrow of normal mouse, a biphenotypic population B220+ CD11b+ CD43+ CD19+ CD62L+ CD34+, similar to the one described in oc/oc bone marrow and representing 0.6% of total mononuclear bone marrow cells.
The commitment of progenitors into the B-cell lineage is controlled by the Pax5 transcription factor, which is essential for the downregulation of the non-B-lineage genes such as c-fms, the receptor for the M-CSF.(18) In agreement with this data, whereas the expression of Pax5 is very low, the c-fms mRNA is expressed by the bipotent progenitors. Furthermore, the B220+CD11b+ population also expresses PU.1 mRNA, whose product regulates the expansion of macrophage and B-cell progenitors.(13) Recently, a B-progenitor population with a similar pattern of gene expression and a bipotent lymphoid/myeloid capacity has been described in human cord blood.(20) Thus, the transcription profile of Pax5, PU.1, and c-fms expression may explain the myeloid differentiation potential of the B220+CD11b+ population.
In conclusion, we have characterized a biphenotypic and potentially bipotent progenitor population present in the bone marrow of normal mouse and amplified in the oc/oc mouse. This amplification is probably a consequence of the profound alterations in the bone architecture and in hematopoiesis found in the oc/oc mouse (C Blin-Wakkach, A Wakkach, PM Sexton, N Rochet, GF Carle, unpublished data, 2004). We show that this population has retained the capacity to differentiate into B-lymphocytes and myeloid cells in vitro. We can thus hypothesize that, in vivo, this population might be involved in both osteoclastogenesis and B-lymphopoiesis. In the oc/oc bone marrow, the block in the B-lymphoid differentiation pathway could lead to an accumulation of pro-B-cells including the bipotent population. This high number of progenitors could be recruited into the myelomonocyte differentiation pathway to participate to the increased osteoclastogenesis.
The existence of bipotent progenitors may also have implications in hematopoietic disorders. In severe forms of lymphomas, cells displaying phenotypic markers for both B-cells (CD19 and CD5) and macrophages (CD11b, F4/80), and defined as “B-macrophages,” have been associated with a poor survival prognostic, while their origin remains unknown.(21) In various animal models such as the osteoporosis induced by ovariectomy,(22) the murine model of aging Kl−/−,(4) and the IL-7−/− model in which the stromal factor IL-7 essential for B-lymphopoiesis is absent,(23) the variations in OCL number are correlated with changes in B-cell number. Therefore, modifications in the differentiation or in the number of the bipotent B220+CD11b+ population may be associated with various pathological conditions including osteoporosis.
This work was supported by the Association Française contre les Myopathies, the Association pour la Recherche sur le Cancer, and the Fondation Singer-Polignac. We thank A Loubat (IFR50, Nice, France) for help in cell sorting and Z. M'Roueh for technical assistance. Animal care was performed by M Topi. We are indebted to Dr L Ollier for helpful discussion and Dr D Burke for the careful reading of this manuscript and comments.
- 12000 Bone resorption by osteoclasts. Science 289: 1504–1508.
- 21997 Essential functions of Pax5 (BSAP) in pro-B cell development: Difference between fetal and adult B lymphopoiesis and reduced V-to-DJ recombination at the IgH locus. Genes Dev 11: 476–491., , ,
- 32001 Generation of bone-resorbing osteoclasts from B220+ cells: Its role in accelerated osteoclastogenesis due to estrogen deficiency. J Bone Miner Res 16: 2215–2221., , ,
- 42001 Connection between B lymphocyte and osteoclast differentiation pathways. J Immunol 167: 2625–2631., , , , , , , , , ,
- 52001 Mouse models of abnormal skeletal development and homeostasis. Trends Genet 17: S38–S43.,
- 62003 Genetic regulation of osteoclast development and function. Nat Rev Genet 4: 638–649.,
- 71985 Morphological evidence of reduced bone resorption in the osteosclerotic (oc) mouse. Am J Anat 172: 141–153.,
- 82000 The gene encoding the mouse homologue of the human osteoclast-specific 116-kDa V-ATPase subunit bears a deletion in osteosclerotic (oc/oc) mutants. Bone 26: 207–213., , , , , , , , ,
- 92002 The vacuolar (H+)-ATPases—nature's most versatile proton pumps. Nat Rev Mol Cell Biol 3: 94–103.,
- 101997 Lack of vacuolar proton ATPase association with the cytoskeleton in osteoclasts of osteosclerotic (oc/oc) mice. FEBS Lett 401: 207–212., , , , , , ,
- 111991 Virus-transformed pre-B cells show ordered activation but not inactivation of immunoglobulin gene rearrangement and transcription. J Exp Med 173: 711–720., ,
- 122002 The roles of FLT3 in hematopoiesis and leukemia. Blood 100: 1532–1542.,
- 131999 The key role of PU. 1/SPI-1 in B cells, myeloid cells and macrophages. Immunol Today 20: 184–189., ,
- 142000 A clonogenic common myeloid progenitor that gives rise to all myeloid lineages. Nature 404: 193–197., , ,
- 152001 The fetal liver counterpart of adult common lymphoid progenitors gives rise to all lymphoid lineages, CD45+CD4+CD3− cells, as well as macrophages. J Immunol 166: 6593–6601., , , , , ,
- 161998 Murine B lymphopoiesis: Towards a unified model. Immunol Today 19: 65–68., ,
- 171994 The CD40 antigen and its ligand. Annu Rev Immunol 12: 881–922., , , , , , , ,
- 181999 Commitment to the B-lymphoid lineage depends on the transcription factor Pax5. Nature 401: 556–562., , ,
- 191997 Identification of clonogenic common lymphoid progenitors in mouse bone marrow. Cell 91: 661–672., ,
- 202003 In vitro identification of human pro-B cells that give rise to macrophages, natural killer cells, and T cells. Blood 101: 4313–4321., , , ,
- 211996 The B/macrophage cell: An elusive link between CD5+ B lymphocytes and macrophages. Immunol Today 17: 471–475.,
- 222000 Comparative effects of estrogen and raloxifene on B lymphopoiesis and bone loss induced by sex steroid deficiency in mice. J Bone Miner Res 15: 541–549., , , , ,
- 232003 Interleukin-7 is a direct inhibitor of in vitro osteoclastogenesis. Endocrinology 144: 3524–3531., , , ,