Adult erythropoiesis is a highly controlled sequential differentiation of hematopoietic stem cells (HSCs) into mature red blood cells (RBCs) in the bone marrow (BM). At steady state, adult humans make more than 2.5 × 10⁶ erythrocytes per second, whereas adult mice make approximately 7,000 erythrocytes per second (Palis, 2014). Erythropoiesis is characterized by the movement of erythroid lineage committed cells through three different compartments, erythroid progenitors, precursors and mature RBCs (An and Mohandas, 2011). Erythroid progenitors consist of a burst-forming unit erythroid (BFU-E) and colony-forming unit erythroid (CFU-E) (Palis, 2014) cells. These cells give rise to nucleated erythroid precursors that progress from proerythroblast, through basophilic, polychromatic and orthochromatic erythroblasts. During precursor maturation the size and RNA content of the cells decrease, hemoglobin synthesis increases, and there is progressive nuclear condensation. The orthochromatic erythroblasts expel the nucleus to form reticulocytes (An and Mohandas, 2011; Dzierzak and Philipsen, 2013; Palis, 2014). Erythroid differentiation occurs in a specific niche in the BM, the erythroblastic islands (EBIs). EBIs consist of a central macrophage surrounded by developing erythroid cells from CFU-E to enucleating erythroblasts (Bessis, 1958; Manwani and Bieker, 2008). Circulating reticulocytes and mature RBCs are the final products of the erythropoietic process.

Bone marrow is a spongy tissue that fills the cavity of most bones. BM is usually regarded as being homogeneous and sampling bone marrow from one bone, such as the femur, is thought to be representative of the whole BM “organ” (Lassailly et al., 2013). However, bones are heterogenous in their cellular embryonic origin and mode of ossification. Different bones originate from three distinct embryonic lineages. The flat bones of the skull, clavicle and cranium develop from cranial neural crest cells. The axial skeleton (the vertebral column and associated ribs) comes from paraxial mesoderm, and the appendicular skeleton (the limbs) comes from lateral plate mesoderm (Kornak and Mundlos, 2003; Provot et al., 2013). Bones have different modes of ossification; endochondral ossification is involved in the growth of long bones and intramembranous ossification occurs in cranial bones (Provot et al., 2013). In mice, the long bones and calvaria were shown to be heterogenous in bone remodeling activity, blood volume fraction, HSCs function (Lassailly et al., 2013), osteoclast activity (Everts et al., 2009), osteoblast activity (Quarto et al., 2010), hormonal and neuronal responses (Everts et al., 2009) and HSPCs hypoxic state (Lassailly et al., 2013). In humans, by 18 years of age the sites of hematopoiesis decrease in long bones and persist in the skull, sternum, ribs, vertebrae, pelvis and epiphyseal regions of femur, and humerus (Ershler and Longo, 2010). Thus, there appear to be functional differences in the cells that reside in these bones. It is not known though, if there are differences in the rate of steady state adult erythropoiesis in the different mouse bones.

The spleen is one of the main hematopoietic organs in mice during development and exhibits extramedullary hematopoiesis in physiological stress and pathological conditions. During embryonic development definitive erythropoiesis shifts from the fetal liver to the spleen, and around the time of birth, the site of hematopoiesis switches to the bone marrow from the spleen (Barminko et al., 2016). It was reported that the myelopoiesis support activity in the spleen is lost by a week after birth (Ohno et al., 1993). However, it is unclear at what age splenic erythropoietic activity is lost.

In this study, we examined steady state adult mouse BM erythroid precursors and progenitors isolated from several skeletal sites: the femur, tibia, pelvis, sternum, vertebrae, radius, humerus, frontal, and parietal bones. We compared all sites to the femur to learn whether sampling bone marrow from the femur is representative of erythropoiesis in the whole BM organ. In addition, we examined trabecular (spongy) bone and cortical (compact) bone erythroid precursors and progenitors in the femur to determine whether there are differences in erythropoiesis in the same bone. Postnatal spleen was studied to determine the age at which erythropoietic activity is lost. We provide evidence that there are steady state differences in erythroid precursors among different bones, and within the trabecular and cortical regions of the same bone. Postnatal spleen supports erythroid precursors until 5–6 weeks of age. These differences need to be taken into consideration in studies of erythropoiesis.

In this study, we examined steady state erythroid precursor and progenitors of different mouse bones, trabecular and cortical regions of the femur, and juvenile spleen. In adult mice, we observed difference in the percentage of BM Ter119⁺ cells, erythroid precursors and MPPs of different bones compared to femur. These bones are categorized into two groups, group one, bones that have BM Ter119⁺ cells, erythroid precursors and MPPs comparable to femur are tibia, pelvis, humerus, sternum. Group two, bones that have a significantly lower percentage than femur are radius, vertebrae, frontal, and parietal bones. We also report that within femur, the epiphysis and metaphysis region (trabecular bone) have lower numbers of erythroid precursors, progenitors, MPPs and myeloid progenitors than the diaphysis (cortical bone).

The differences in erythropoiesis among the different bones described in this work were predominantly observed at the level of the erythroid precursors. This might be due to increased/decreased number or intrinsic differences among the central macrophages in the BM microenvironment in the different bones. Zhang et al. (2021) recently reported that a small subset of endothelial cells of BM sinusoids expresses the colony-stimulating factor 1 (CSF1). This endothelial cell derived CSF1 regulates the differentiation of myeloid cells (Zhang et al., 2021) suggesting that local cues by different blood vessels might control myeloid differentiation. The differences we found in erythropoiesis among the different bones could also be due to these CSF1 producing subtype of sinusoid vessels affecting the differentiation of central macrophages of the EBIs. Central macrophages of EBI were initially identified by the presence of a combination of specific adhesion molecules, such as F4/80⁺, VCAM1⁺, and CD169⁺ (Seu et al., 2017; Li et al., 2020). However, selective depletion of CD169⁺ macrophages in steady state decreased the frequency of EBI in the BM but did not result in anemia (Chow et al., 2013). Interestingly, similarly to central macrophages, osteomacs or osteal macrophages also express CD169⁺, F4/80⁺, VCAM1⁺. Osteomacs support osteoblasts in the maintenance of bone homeostasis and maintain HSCs in the BM (Chang et al., 2008; Batoon et al., 2019). These findings strongly suggest that it is unlikely that all F4/80⁺ VCAM1⁺ CD169⁺ macrophages are EBI central macrophages. Recently Li et al. (2019) showed that in addition to the known adhesion molecules described above the expression of Epo-receptors (Epo-R) is also necessary to define a macrophage to be central macrophage of EBI. Interestingly, osteoclast (bone removing cells) also express Epo-R just as central macrophages. Epo stimulates monocytic lineage cells by increasing BM monocytes, macrophages, preosteoclasts, and mature osteoclasts. Osteoclasts appear to be required for the Epo induced erythropoietic response and inhibition of osteoclasts blunts the Epo response (Singbrant et al., 2011; Hiram-Bab et al., 2015). Osteoclasts exhibit bone site specific differences as osteoclastic activity is higher in the long bones compared to the calvaria (Everts et al., 1999, 2009; Zenger et al., 2010). Given the differences in the osteoclast in different bones, and the similarity of osteoclasts to central macrophages we speculate that central macrophages might also display bone site specific differences. Further studies are needed to support (or defy) this hypothesis. The expression of Epo-R in both central macrophage and osteoclast may explain why Epo can simultaneously regulate erythropoiesis and bone homeostasis.

To date, a limited number of studies have focused on differences in the numbers and functions of HSCs in different regions inside bones (Guezguez et al., 2013; Lassailly et al., 2013). Our work shows that different regions within femur show differences in erythropoiesis. We show that erythroid progenitors and precursors in the Epi and metaphysis are significantly lower than that of diaphysis regions of the femur. These differences could be due the specific type of vasculature in these regions, which might have influenced the formation of the central macrophage of the EBI. The type of stromal cell population associated with vasculature creates the specific environment where HSCs and HSPCs are maintained. Metaphysis has type H vessels which express high levels of CD31 and endomucin (EMCN) described as CD31^hi EMCN^hi. Diaphysis has type L vessels which express lower levels of CD31 and endomucin described as CD31^lo EMCN^lo (Kusumbe et al., 2014; Ramasamy et al., 2014, 2016). The CSF1 sinusoids which support myelopoiesis described by Zhang et al. (2021) are a subset of type L vessels in the diaphysis. The higher number of CFU-GM colonies seen in the diaphysis might be due to the higher number of sinusoids in the diaphysis than in the epi- and metaphysis (Zhang et al., 2021). The CSF1 sinusoids might affect the formation of the EBIs, as well as their migration toward the sinusoids as the erythroid differentiation progresses from proerythroblast to orthochromatic erythroblasts on the EBIs (Yokoyama et al., 2003). In addition, the type of vessels in a region influences the degree of hypoxia present in different regions of the bone. Diaphysis lacks a direct arterial supply and the arteries exclusively connect to type H vessels in the metaphysis which makes the diaphysis highly hypoxic. The hypoxic environment maintaining HSCs dormancy and a low metabolic state which results in a large number of HSCs in the diaphysis compared to the metaphysis (Kusumbe et al., 2014; Ramasamy et al., 2014, 2016). These studies indicate that local cues produced by distinct blood vessels might be responsible for the differentiation of specific blood cells. This fits well with our finding that erythropoiesis seem to be different within different sites of the femur, possibly due to differences in the number of HSCs, HSPCs, and EBIs in different regions. In humans, unlike in the mouse, by the age of 18 active hematopoiesis occurs only in the vertebrae, ribs, sternum, skull, pelvis, and epiphyseal regions of humerus and femur (Ershler and Longo, 2010) suggesting a species difference in the sites of erythropoiesis.

We also demonstrate that the spleen supports erythroid precursors until 5–6 weeks of age, while the residual erythroid progenitor support capacity seems to reach adult levels by 3 months of age. The role of the spleen in supporting erythropoiesis during the postnatal period has received little attention. While this paper was being revised Chen et al. (2021) published a paper on spleen erythropoiesis in developing mice which supports our results. We have shown that in the spleen a significant decrease in erythroid precursors are seen at about 4 weeks of age. This is followed by a gradual disappearance of erythroid precursors by the end of the 6th week. This event coincides with a significant down-regulation of the total number of macrophages and the RPMs during this period. This gradual loss of precursors and RPMs parallels the appearance of well-defined white pulp, red pulp and marginal zone macrophages in the spleen (A-Gonzalez, and Castrillo, 2018). This observation suggests that the compartmentalization might play a role in the loss of erythroid precursor support. In our study the erythroid progenitors, the BFU-Es, were significantly decreased from 6 to 8 weeks and 8 weeks to 3 months old spleen. This is in contrast to the report of Wolber et al. (2002) who found no significant change in the number of BFU-Es in the spleen between 8 days and 8 weeks in mice. The difference in precursor and progenitor support suggests that the niches that support progenitors vs. precursors in the spleen are different, even though both progenitor and precursors are found in the red pulp (Inra et al., 2015). In addition, this could also be due to the differences in the cytokine environment such as the levels of SCF and CXCL12 secreted by endothelial perivascular stromal cells in the red pulp. In adult mice, stress erythropoiesis increased the number of Scf-expressing endothelial cells and Scf and Cxcl12 expressing perivascular stromal cells in the red pulp. Conditional deletion of Scf-expressing endothelial cells, and Scf and Cxcl12 expressing perivascular stromal cells were shown to reduce spleen erythropoiesis without effecting BM erythropoiesis (Inra et al., 2015). Splenectomy of neonatal pups did not affect their ability to grow and thrive, or alter the progenitor numbers in the BM, but the number of erythroid progenitors was increased in the liver (Wolber et al., 2002). This is similar to stress erythropoiesis induced in adult splenectomized mice, where the BM did not expand the number of progenitors to compensate for the loss of the spleen. Instead, progenitors expanded in the liver of the splenectomized mice (Lenox et al., 2009). The stress erythropoiesis in adults recapitulates the postnatal spleen erythropoiesis suggesting that endothelial and stromal cell secretion of cytokines might also be modulated during postnatal time, which requires further study. These studies suggest that in times of need (e.g., during development of bone or when stress erythropoiesis is required) the BM needs help from organs such as the spleen or the liver.

In conclusion, our work shows that steady state erythropoiesis in the mouse femur may reflect RBC production in the tibia, pelvis, humerus and sternum, however, the differences need to be considered when analyzing other bones. Furthermore, studies of erythropoiesis in mice younger than 5–6 weeks of age should consider the contribution of the spleen to RBC production early in the life and analyzed when using mice less than 6 weeks old.

The original contributions presented in the study are included in the article/Supplementary Material, further inquiries can be directed to the corresponding authors.

The animal study was reviewed and approved by the Animal Care and Use Committee of National Institute of Dental and Craniofacial Research (NIDCR), NIH.

VM led the conception and design of the experiments, collected and interpreted data, and wrote the manuscript. IS collected colony forming assay data. EM interpreted the results, prepared Figures, and edited the manuscript. All authors contributed to the article and approved the submitted version.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.