Identification of a Multipotent Progenitor Population in the Spleen That Is Regulated by NR4A1

This article is featured in In This Issue, p.883

Hematopoiesis occurs primarily in the bone marrow of adult vertebrates and produces RBCs, WBCs, and platelets. Hematopoietic stem cells (HSC) generate all mature lineages while maintaining undifferentiated progenitor pools by undergoing self-renewal. HSC are also found elsewhere in the body, including the liver, spleen, and circulating in the blood. However, extramedullary hematopoiesis is largely auxiliary unless bone marrow function is compromised by myelofibrosis or stressed by acute infection, myocardial infarction, or blood loss (1–4). In response to these hematopoietic challenges, the spleen drives emergency hematopoiesis by promoting hematopoietic stem and progenitor cell proliferation and differentiation. Whether splenic hematopoiesis mirrors bone marrow hematopoiesis or generates distinct cell populations remains an open question.

Bone marrow and splenic HSC occupy perisinusoidal niches and rely on CXCL12 and SCF (5, 6), suggesting that their mechanisms of self-renewal are similar; however, some differences have been noted. Splenic long-term HSC (LT-HSC) appear to actively cycle and, unlike most bone marrow LT-HSC, express CCR2, a marker of proliferating myeloid-biased stem cells (7). The bone marrow contains 6-fold fewer CCR2⁺ LT-HSC, consistent with their more balanced lymphoid to myeloid output. The spleen also lacks osteoblasts, bone marrow niche cells critical for maintaining lymphoid progenitors (8, 9). In response to stress, stem and progenitor cell differentiation in the spleen favors myelopoiesis (2, 7, 10, 11). Whether hematopoiesis in a healthy unstressed spleen exhibits myeloid bias and whether the spleen supports unique myeloid developmental pathways are not known.

Many factors promoting LT-HSC maintenance in the niche have been identified and characterized, yet the signals affecting stem cell differentiation are not as well defined. Studies indicate that LT-HSC do not contribute to the bulk of the blood system in homeostasis, and show, instead, that more restricted progenitors derived from LT-HSC and short-term HSC (ST-HSC) actually maintain blood populations throughout life (12, 13). These multipotent progenitors (MPP) maintain a limited capacity for self-renewal compared with LT-HSC and ST-HSC but are capable of differentiating into all cell types of the hematopoietic system. Investigators have recently characterized three distinct subpopulations of MPP in the bone marrow (MPP2, MPP3, and MPP4, also known as lymphoid-primed MPP), which can contribute to more lineage-restricted common myeloid progenitors and common lymphoid progenitors (14). Although MPP subsets are developmentally biased (MPP2 cells toward erythrocyte/megakaryocyte lineages, MPP3 cells toward monocyte/granulocyte lineages, and MPP4 cells toward lymphocytes), their fate is not fixed and can change depending on the environmental signals that they receive. The identity of these external signals and how they are interpreted by developing blood cells are not well understood.

The orphan nuclear receptor NR4A1 (Nur77) is an attractive candidate for a transcriptional regulator that integrates external environmental signals and biases differentiation programs in multipotent hematopoietic progenitors. NR4A1, along with family members NR4A2 (Nurr1) and NR4A3 (NOR1), are immediate early-response transcription factors involved in metabolism, growth factor signaling, apoptosis, inflammation, and the development of innate and adaptive immune cells (15–21). Within this family, NR4A1 has nonredundant functions at several distinct stages of hematopoiesis, affecting multipotent and committed cells. NR4A1 plays a role in lymphocyte development and TCR-driven thymocyte selection (22–24). During the later stages of myelopoiesis, NR4A1 is required for the generation of nonclassical patrolling monocytes that function during homeostasis, tissue repair, and immune surveillance (21, 25–27). In addition to these roles in terminal immune cell development, NR4A1 functions early in hematopoiesis. NR4A1 expression biases LT-HSC toward myeloid cell fates and is induced in a PKA-dependent manner by PGE₂ (28). The role that NR4A1 plays subsequent to the HSC stage, when MPP proliferate and commit to individual blood cell lineages, is unknown.

In this study, we identify a novel specialized subset of multipotent myeloid progenitors that is overrepresented in the spleen compared with the bone marrow. We term these MPP of the spleen (MPPS). MPPS express high levels of NR4A1, which restricts their megakaryocyte and erythroid potential, instead biasing their development into other myeloid cell types, including monocytes. Thus, MPPS constitute a novel population of splenic MPP whose differentiation into the myeloid lineage is regulated by expression of NR4A1.

All mice used in this study were male and were originally purchased from The Jackson Laboratory (Bar Harbor, ME), and shared the C57BL/6J background. Nr4a1-knockout mice (B6; 129S2-Nr4a1^tm1Jmi/J) and Nr4a1^GFP+/− mice [C57BL/6-Tg(Nr4a1-EGFP/cre)820hog/J] were bred at Charles River Laboratories (Wilmington, MA) and transferred when needed. Transgenic Nr4a1^GFP+/− mice express GFP under the control of a randomly inserted Nr4a1 promoter. Only heterozygous Nr4a1^GFP+/− mice were used in this study and were identified by the presence of the transgene using genotyping primers (5′-CGGGTCAGAAAGAATGGTGT-3′ and 5′-CAGTTTCAGTCCCCATCCTC-3′). C57BL/6-Tg(UBC-GFP)30Sha/J (UBC-GFP) mice and B6.SJL-Ptprc^aPepc^b/BoyJ (CD45.1) mice were used as recipients in transplant experiments. All experiments, animal handling, and husbandry followed the guidelines required by the Columbia University Institutional Animal Care and Use Committee.

For analysis of progenitor populations, bone marrow (femur, tibia, and hip) and spleens were treated with ACK lysis buffer (Lonza) and washed in PBS containing 0.1% BSA (Thermo Fisher Scientific). Bone marrow and splenic progenitors were separated from lineage⁺ cells using a Lineage Cell Depletion Kit (Miltenyi Biotec). Lineage⁻ populations were then stained with the LIVE/DEAD Fixable Aqua Dead Cell Stain Kit (Life Technologies/Thermo Fisher Scientific), in PBS at a dilution of 1:1000, washed in 0.1% BSA PBS, and stained with cell surface Abs against lineage Ags (Mac1, Gr1, Ter119, B220, CD3, NK1.1) to detect residual lineage⁺ cells in combination with the following Abs: anti-Sca1, anti-cKit, anti-CD41, anti-CD16/32, anti-CD150, anti-CD105, anti-CD71, and anti-CD24. See Supplemental Table I for a complete list of Abs used in this study. Cells were analyzed using a MACSQuant (Miltenyi Biotec), Fortessa, or LSR II (Becton Dickinson) and sorted with a FACSAria (Becton Dickinson).

Blood samples were collected in EDTA-coated capillary tubes (Sarstedt) and measured immediately for complete blood counts, which included hematocrit levels and RBC counts, using a Hemavet HV950 Multi-species Hematology System (Drew Scientific). For analysis by flow cytometry, samples were spun down in a microcentrifuge at 100 × g for 10 min. The upper aqueous layer was removed, washed with 0.1% BSA PBS containing EDTA, and stained with anti-Ter119 and anti-CD41 Abs to identify erythrocytes and platelets, respectively. The remaining blood sample was treated with ACK lysis buffer, immediately washed in 0.1% BSA PBS, and stained with Abs against lineage markers (Mac1, Ly6C, Ly6G, CD115, CD3, CD4, CD8, and CD19).

Mouse organs (bone marrow, spleen, thymus, and lymph nodes) were dissected and dissociated without enzymes. Cells were washed in 0.1% BSA PBS containing EDTA and immediately stained for lineage markers or treated first with ACK to remove RBCs.

Progenitor cells sorted from the spleens of mice were plated at a density of 250–500 cells per well in triplicate onto six-well plates containing MethoCult M3234 (STEMCELL Technologies) with 50 ng/ml SCF, 20 ng/ml IL-3, 20 ng/ml IL-6, 10 U/ml EPO, 10 ng/ml TPO, and 10 ng/ml Flt-3 ligand (PeproTech). After 10–12 d in culture, colonies were characterized by their visual appearance using a bright-field inverted microscope (CK2; Olympus). Colonies from individual wells were pooled and stained with Aqua Dead and Abs against myeloid cell surface markers (anti-CD41, anti-Ter119, anti-Gr1, and anti-Mac1).

Five splenic Lineage⁻Sca1⁻cKit⁺ (L⁻S⁻K⁺)CD41⁻CD16/32⁻CD71^lowCD24^high cells that were NR4A1^GFP+ or NR4A1^GFP− were sorted directly into 96-well plates containing MethoCult M3234 with supplements (SCF, IL-3, IL-6, EPO, TPO, and Flt3 ligand). Five cells were plated to increase the probability of each well containing only one colony, calculated based on prior experiments in which one in every six or seven MPPS generated a colony. Five days after sorting, individual wells were analyzed for the presence of small colonies. Any well containing more than one colony was excluded from the analysis. After 10–12 d in culture, each colony was dissociated and stained.

Colony assays were performed as described previously (29–31). For erythroid CFU (CFU-E) analysis, nucleated bone marrow (1 × 10⁵ per milliliter) and splenocytes (2 × 10⁶ per milliliter) were plated onto MethoCult M3234 (STEMCELL Technologies) containing 3 U/ml EPO (PeproTech) and placed into a humidified chamber under normoxia (21% O₂). The number of small rose-colored colonies was counted 2 d after plating using an inverted microscope.

Recipient UBC-GFP mice were injected i.v. with 5,000–10,000 MPPS along with 3 × 10⁵ helper bone marrow cells from a separate UBC-GFP mouse 2 d after mice were given two equal doses of 500 cGy lethal irradiation using a [¹³⁷Cs]-based gamma-ray irradiator. UBC-GFP mice were used as recipients because of their ubiquitous GFP expression in all cell types, including erythrocytes and platelets. Donor cells were identified by their lack of GFP expression. The percentages of donor cells represent the proportion of GFP⁻ cells within each cell population.

Following lethal irradiation, as described above, individual CD45.1 recipients were injected i.v. with 2500 LT-HSC, 5000 MPP2 cells, or 5000 MPP3 cells without competitive/helper cells. All donor cells were CD45.2⁺. Eighteen days posttransfer, the spleens of recipient mice were analyzed for donor-derived MPPS. Similarly, in separate experiments, UBC-GFP mice were injected with 50 GFP⁻ LT-HSC (Lineage⁻Sca1⁺cKit⁺[LSK]CD150⁺CD48⁻) and 2 × 10⁵ UBC-GFP⁺ whole bone marrow cells and analyzed 10 wk later for GFP⁻ MPPS.

Unpaired Student t tests were performed to obtain p values using Prism 7 (GraphPad) software. The Welch correction was applied when the calculated F-test provided significantly different variances among groups.

To investigate the spleen’s myelopoietic reservoir, we first compared myeloid progenitor subsets present in the healthy spleen and bone marrow. As expected, the spleen includes L⁻S⁻K⁺ progenitors and LSK hematopoietic stem and progenitor cells (Fig. 1A, Supplemental Fig. 1). The spleen also contains all previously identified myeloid progenitor subpopulations. Specifically, we distinguished granulocyte-macrophage progenitors (G-MP) and megakaryocyte progenitors (MkP) by their expression of CD16/32 and CD41, respectively (32, 33). We observed several subtle differences among previously identified progenitor subsets, all of which were less abundant in the spleen compared with the bone marrow. Within the L⁻S⁻K⁺ population, CD41⁺ lineage-restricted MkP were present in similar proportions in both organs, and G-MP were represented at significantly lower frequencies in the spleen (Fig. 1B). When we stained L⁻S⁻K⁺CD41⁻CD16/32⁻ cells for CD71 and CD24 expression, which classically identifies erythroid burst-forming unit (BFU-E) and CFU-E progenitor populations (34), we observed a distinct previously uncharacterized population of cells in the spleen that was marked by low expression of CD71 and high expression of CD24 (CD71^lowCD24^high) (Fig. 1B). Within the L⁻S⁻K⁺ progenitor pool, the spleen contained ∼10-fold more of these cells compared with the bone marrow (Fig. 1B).

To measure the developmental potential of this cell subset, we cultured sorted splenic L⁻S⁻K⁺CD41⁻CD16/32⁻CD71^lowCD24^high cells in methylcellulose in the presence of multiple hematopoietic growth factors. These cells generated monocyte colonies (CFU-M), granulocyte colonies (CFU-G), colonies that contained both granulocyte and monocyte cells (CFU-GM), erythroid colonies (BFU-E), and megakaryocyte colonies (CFU-Mk), demonstrating their clonogenic capacity and multipotent myeloid potential (Fig. 2A). Flow cytometric analysis of pooled methylcellulose colonies confirmed the presence of cells expressing lineage-restricted markers Mac1, Gr1, Ter119, and CD41 (Fig. 2B). Therefore, we designate the novel L⁻S⁻K⁺CD41⁻CD16/32⁻CD71^lowCD24^high population MPPS.

To assess the differentiation potential of MPPS in vivo, we adoptively transferred 1 × 10⁴ wild-type (WT; nonfluorescent) sorted MPPS along with 3 × 10⁵ GFP⁺ helper bone marrow cells into lethally irradiated UBC-GFP mice (Fig. 3A). Twelve days later, blood cells from recipient mice were analyzed for lineage-specific cell surface markers, as well as GFP expression (lack of GFP expression indicates donor-derived cells). To control for the possibility that some GFP⁺ cells could lose GFP expression and confound the ability to distinguish donor (GFP⁻) cells from recipient and helper (GFP⁺) cells, some UBC-GFP recipients received GFP⁺ helper bone marrow cells only. Only trace amounts (<0.01%) of GFP⁻ cells were observed in these control recipients (Fig. 3B). Donor (GFP⁻) cells were detected in the blood, bone marrow, and spleens of recipient mice but not in their thymus or lymph nodes (Fig. 3C). Donor-derived cells contributed to erythroid populations, making up 2–5% of total Ter119⁺ cells in the peripheral blood, bone marrow, and spleen. Donor platelets were also present in the peripheral blood and spleen but not in the bone marrow. MPPS-derived cells contributed to neutrophil and monocyte populations, including Ly6C^high monocytes, in the spleen but not in the bone marrow (Fig. 3C).

Because lymphocytes are often the last lineage to repopulate from HSC after lethal irradiation (14, 35, 36), we also looked for the presence of all lineages, including lymphocytes, 4 wk posttransplant. No lymphocytes were detected, and at that point we only found donor-derived erythrocytes in the bone marrow, spleen, and peripheral blood, consistent with their 40-d half-life (Fig. 3C). Thus, although MPPS do not develop into lymphoid cells, they have the potential to differentiate in vivo into several myeloid lineage subsets, including RBCs, platelets, monocytes, and neutrophils, most notably within the spleen. Of note, <1% of MPPS were donor derived after 4 wk, suggesting that this population has limited self-renewal capacity (Fig. 3D).

The multipotent bone marrow progenitor populations MPP2 and MPP3 have been shown to give rise to mixed populations of nonlymphoid cells, with MPP2 differentiation biased toward erythrocytes and megakaryocytes, and MPP3 biased toward monocytes and granulocytes (14). To determine which progenitor population(s) give rise to MPPS, we adoptively transferred 5000 CD45.2⁺ MPP2 or MPP3 (Supplemental Fig. 2). Both MPP2 and MPP3 subpopulations are known to maintain a myeloid bias, especially during regenerative conditions (14). Recipients that received 2500 LT-HSC served as a positive control. Eighteen days after injecting purified subpopulations into lethally irradiated CD45.1 congenic recipients, we examined tissues for the presence of donor-derived CD45.2⁺ cells (Fig. 4A, 4B). We detected donor-derived MPPS in all recipient mice that received LT-HSC, with as many as 90% of splenic MPPS developing from LT-HSC donor cells (Fig. 4A, 4B). Similarly, in mice that received MPP2, ∼70% of MPPS were donor derived. In contrast, transplanted MPP3 contributed <2% of MPPS. Together, these data suggest that MPP2 and LT-HSC, not MPP3, generate MPPS in the recovery phase after myeloablation. LT-HSC continue to give rise to donor-derived MPPS, even 10 wk after transplantation, indicating that this splenic cell population continues to develop from stem cells well beyond the immediate recovery phase (Fig. 4C).

To begin to explore how the developmental fate of MPPS is controlled, we explored molecules known to regulate myeloid development. We have previously shown that NR4A1 is preferentially expressed in a subset of myeloid-biased LT-HSC (28). NR4A1 is also required for terminal monocyte maturation (21). Furthermore, the spleen serves as a reservoir for monocytes involved in diverse physiologic and pathophysiologic responses that develop locally from progenitor populations (10, 11, 37, 38). Given NR4A1’s role in dictating hematopoietic cell differentiation and the localization of monocyte progenitors to the spleen, we thought it important to examine NR4A1’s expression in splenic stem cells and progenitors, including MPPS. Using NR4A1^GFP reporter mice (39), we compared NR4A1 expression in splenic and bone marrow stem cells and progenitors.

We first noted that splenic LSK cells express significantly higher levels of NR4A1 than their bone marrow counterparts (Fig. 5A). Notably, we found that NR4A1 expression in splenic progenitors was highest within MPPS; ∼25% of MPPS expressed high levels of NR4A1 compared with nontransgenic WT controls, whereas the other progenitor subpopulations within the immature L⁻S⁻K⁺ progenitor pool expressed very little, if any, NR4A1 (Fig. 5B). In addition, NR4A1 expression within MPPS was highest in the spleen and was minimal within the few CD71^lowCD24^high bone marrow progenitors that we detected. Thus, NR4A1 expression identifies a major fraction of splenic MPPS.

To determine whether the expression of NR4A1 identifies distinct functional subsets in vitro, we purified and cultured NR4A1^GFP+ and NR4A1^GFP− MPPS from NR4A1^GFP mice in methylcellulose. After 12 d, colonies from NR4A1^GFP+ MPPS appeared mostly monocytic (CFU-M), granulocytic (CFU-G), or both (CFU-GM), whereas NR4A1^GFP− cells produced significantly fewer granulocyte/monocyte colonies (Fig. 5C). In contrast, NR4A1^GFP+ MPPS generated significantly fewer erythroid (BFU-E) and megakaryocyte colonies than NR4A1^GFP− MPPS.

To directly measure the potential of individual cells within the MPPS population, we sorted single NR4A1^GFP+ or NR4A1^GFP− MPPS into 96-well plates, cultured them in the presence of known hematopoietic growth factors, and examined their appearance and phenotype (Fig. 5D). About 35% of colonies derived from single NR4A1^GFP+ cells were multipotent and generated cells that stained positive for CD41, Ter119, and Mac1 or Gr1. Consistent with the data shown in Fig. 5C, most (∼85%) colonies from single NR4A1^GFP+ cells were Mac1⁺ or Gr1⁺; NR4A1^GFP+ cells generated significantly fewer colonies that were Ter119⁺ or CD41⁺ (Fig. 5D). In contrast, colonies derived from single NR4A1^GFP− cells were predominantly unipotent and contained very few Mac1⁺ or Gr1⁺ cells; instead, they stained positive for Ter119 or CD41, markers of the erythroid/megakaryocyte lineages. Thus, NR4A1 expression specifically identifies MPPS that are biased toward monocyte and granulocyte development. Moreover, these observations raise the possibility that NR4A1 plays a role in restricting the ability of MPPS to differentiate toward erythroid/megakaryocyte lineages.

To assess the function of NR4A1 within the MPPS population in vivo, we adoptively transferred WT or Nr4a1-deficient (Nr4a1^−/−) MPPS. Because our in vitro data suggested that NR4A1 expression specifies MPPS ability to differentiate toward monocyte/granulocyte lineages and away from erythrocytes/megakaryocytes (Fig. 5D, 5E), we measured the frequency of donor erythrocytes in the peripheral blood to assess NR4A1 function within MPPS. Consistent with our in vitro observations, Nr4a1^−/− MPPS generated 5–10-fold more Ter119⁺ erythrocytes in the peripheral blood than WT MPPS (Fig. 6A): 9 d after the cells were injected, 10% of total erythrocytes in the peripheral blood were derived from Nr4a1^−/− MPPS, whereas only ∼1% of erythrocytes were generated from WT MPPS. This trend was still evident at day 12, when >5% of erythrocytes were derived from Nr4a1^−/− MPPS donor splenocytes versus ∼1% from WT MPPS.

MPPS-derived platelet populations were also affected by NR4A1 expression. At day 12, Nr4a1^−/− MPPS contributed 1.7% of the total CD41⁺ platelet population in the peripheral blood compared with 0.4% from WT MPPS (Fig. 6B). These data show that, in the absence of NR4A1, MPPS are biased toward erythroid and megakaryocyte lineage commitment, suggesting that NR4A1 plays an unexpected role in restricting the erythroid and megakaryocyte differentiation potential in MPPS.

These data raised the possibility that the production of erythroid progenitors would be enhanced, perhaps in a spleen-specific fashion, in unmanipulated Nr4a1-deficient mice. In fact, spleens of Nr4a1^−/− mice contained a significantly higher frequency and total number of CD71⁺CD24⁺ CFU-E progenitors compared with WT mice (Fig. 7A). We confirmed this increase in CFU-E in Nr4a1-deficient mice using a distinct staining scheme that identifies CFU-E via CD105 and CD150 expression (Fig. 7B) (40). In addition, splenocytes from Nr4a1^−/− mice plated onto methylcellulose produced twice as many CFU-E–derived colonies as did WT splenocytes (Fig. 7C). Importantly, in flow cytometric and methylcellulose assays, we did not observe any differences in CFU-E populations arising from bone marrow progenitors. We detected similar numbers of RBCs and RBC precursors in the bone marrow, spleen, and peripheral blood of Nr4a1^−/− mice (Supplemental Fig. 3), although this was not surprising given that steady-state erythrocyte production predominantly occurs in the bone marrow (41, 42). Thus, these data indicate that NR4A1 plays an unanticipated role in regulating myeloid lineage fate, specifically in the spleen, through its effect on MPPS.

In this study, we identify, to our knowledge, a novel population of hematopoietic progenitors that is enriched in the spleen. We designate these MPPS and show that a proportion expresses high levels of NR4A1. MPPS are multipotent and contribute to erythroid, megakaryocyte, granulocyte, and monocyte lineages. Most importantly, we show that NR4A1 expression influences the cell fate choice adopted by MPPS. Specifically, NR4A1 inhibits MPPS development to erythroid and megakaryocyte lineages.

Although the role of the bone marrow in adult hematopoiesis is well described, the contribution of the spleen is less well understood. The spleen is known to maintain a population of Ly6C^high and Ly6C^low monocytes that play a role in atherosclerosis, myocardial infarction, cancer, and acute infection (3, 10, 11, 37). These splenic reservoir monocytes originate from bone marrow G-MP that mobilize to the spleen where they expand and differentiate (10). However, G-MP are maintained at much lower frequencies in the spleen compared with the bone marrow (Fig. 1B). In this article, we show that the spleen harbors a distinct population of cells, MPPS, which is capable of generating Ly6C^high monocytes. G-MP are not well represented within the splenic progenitor pool; therefore, we propose that MPPS serve as an attractive alternative or, at the very least, an additional local source of splenic reservoir monocytes.

Not only does the spleen harbor a higher proportion of MPPS than the bone marrow, these MPPS and their stem cell precursors express higher levels of NR4A1 in the spleen than in the bone marrow. These observations suggest that the spleen cultivates a unique hematopoietic microenvironment, a niche that also favors myeloid development. Although recent data identified typical niche factors important for the maintenance of splenic HSC (6), our data suggest that examining the spleen for cells and molecules that influence NR4A1 expression might also provide important insights into the spleen’s myeloid-favoring behavior. PGE₂, a locally active lipid compound that elicits inflammatory and immunosuppressive responses, is an especially interesting candidate (43). PGE₂ rapidly induces NR4A1 expression in a variety of cell types, including LT-HSC (28, 44–47). PGE₂ promotes granulopoiesis in response to bacterial infection and HSC proliferation and engraftment after bone marrow transplant, perhaps through its activation of NR4A1 (48–50). PGE₂ is expressed by tissue macrophages, which are important cellular components of HSC in the bone marrow and spleen (51–53). A more dominant PGE₂-secreting macrophage compartment in the spleen might trigger the increased expression of NR4A1 by MPPS and LSK cells. Alternatively, activation of NR4A1 in bone marrow cells could promote their mobilization and accumulation in the spleen. The unique ability of the spleen to maintain NR4A1-expressing cells that preferentially differentiate into monocytes and granulocytes may explain why we observe these lineages in the spleen but not the bone marrow after adoptive transfer of MPPS. Together, our observations suggest that the spleen is uniquely poised to contribute to myelopoiesis and may be an important source of monocytes.

Previous studies of NR4A1, NR4A2, and NR4A3 indicate that they are critical regulators of myeloid progenitor differentiation, as well as that these proteins often share redundant functions. Mice deficient for both Nr4a1 and Nr4a3 die of acute myeloid leukemia (AML) shortly after birth, and leukemic blasts from human patients with AML have reduced levels of NR4A1 and NR4A3, albeit through unknown mechanisms (54). In mice, genetic deletion of Nr4a1 alone is not sufficient to induce AML or myelodysplasia (42), but NR4A1 is required for the development of a specialized subset of noninflammatory patrolling monocytes (21). Previously, our laboratory demonstrated that NR4A1 expression identifies myeloid-biased LT-HSC (28), but whether NR4A1 independently regulates cell fate within the myeloid progenitor compartment was previously unknown. Our present data demonstrate that NR4A1 plays a distinct nonredundant role in myeloid progenitor development, restricting erythropoiesis and megakaryopoiesis within the MPPS population, a novel subset of multilineage myeloid progenitors.

The multipotent nature of MPPS precursors highlights the developmental plasticity that has also been attributed to other MPP, which are believed to alter their developmental pathway depending on their environmental context (14, 55). Under homeostatic conditions, different subsets of MPP show lineage bias toward megakaryocytes and erythrocytes, monocytes and granulocytes, or lymphocytes. However, in the early stages of regeneration following myeloablation, lymphoid-biased MPP activate a myeloid transcriptional program and, therefore, presumably contribute to the myeloid compartment. This indicates that their lineage commitment is not fixed but instead is influenced by extrinsic regulatory factors. Similarly, we show that activation of NR4A1 in MPPS, which limits their erythroid and megakaryocyte potential, is higher in the spleen than the bone marrow, thereby suggesting niche-specific influences on MPPS cell fate. Thus, for certain populations of hematopoietic progenitors, lineage choice is neither fixed nor completely stochastic; rather, it is mediated by external cues.

Our study suggests that the spleen makes a unique contribution to steady-state hematopoiesis. Further dissection of the regulatory factors and niche components will provide insights into myeloid-biased hematopoiesis in the spleen. This is important for understanding extramedullary hematopoiesis critical for responding to hematopoietic challenges, and it may also uncover basic mechanisms of myelopoiesis that can be exploited or hijacked by blood cancers.

We thank C. Constant and Dr. S.V. Bardina for technical expertise and assistance.

The authors have no financial conflicts of interest.