Macrophages form erythropoietic niches and regulate iron homeostasis to adapt erythropoiesis in response to infections and inflammation

Abstract

It has recently emerged that tissue resident macrophages are key regulators of several stem cell niches orchestrating tissue formation during development, as well as postnatally where they also organise the repair and regeneration of many tissues including the haemopoietic tissue. The fact that macrophages are also master regulators and effectors of innate immunity and inflammation allows them to co-ordinate haematopoietic response to infections, injuries and inflammation. After recently reviewing the roles of phagocytes and macrophages in regulating normal and pathological haematopoietic stem cell niches, we now focus on the key roles of macrophages in regulating erythropoiesis and iron homeostasis. We review herein the recent advances in understanding how macrophages at the centre of erythroblastic islands form an erythropoietic niche that controls the terminal differentiation and maturation of erythroblasts into reticulocytes, how red pulp macrophages in the spleen control iron recycling and homeostasis, how these macrophages coordinate emergency erythropoiesis in response to blood loss, infections and inflammation and how persistent infections or inflammation can lead to anaemia of inflammation via macrophages. Finally, we discuss the technical challenges associated with the molecular characterisation of erythroid island macrophages and red pulp macrophages.

Full text

Journal Pre-proof
Macrophages form erythropoietic niches and regulate iron
homeostasis to adapt erythropoiesis in response to infections and
inflammation
Jean-Pierre L´
evesque , Kim M Summers , Kavita Bisht ,
Susan M Millard , Ingrid G Winkler , Allison R Pettit
PII: S0301-472X(21)00291-5
DOI: https://doi.org/10.1016/j.exphem.2021.08.011
Reference: EXPHEM 3950
To appear in: Experimental Hematology
Received date: 11 August 2021
Revised date: 30 August 2021
Accepted date: 31 August 2021
Please cite this article as: Jean-Pierre L´
evesque , Kim M Summers , Kavita Bisht , Susan M Millard ,
Ingrid G Winkler , Allison R Pettit , Macrophages form erythropoietic niches and regulate iron home-
ostasis to adapt erythropoiesis in response to infections and inflammation, Experimental Hematology
(2021), doi: https://doi.org/10.1016/j.exphem.2021.08.011
This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition
of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of
record. This version will undergo additional copyediting, typesetting and review before it is published
in its final form, but we are providing this version to give early visibility of the article. Please note that,
during the production process, errors may be discovered which could affect the content, and all legal
disclaimers that apply to the journal pertain.
©2021 Published by Elsevier Inc. on behalf of ISEH – Society for Hematology and Stem Cells.

1
Macrophages form erythropoietic niches and regulate iron
homeostasis to adapt erythropoiesis in response to infections and
inflammation
Jean-Pierre Lévesque, Kim M Summers, Kavita Bisht, Susan M Millard, Ingrid G Winkler,
Allison R Pettit
Mater Research Institute – The University of Queensland, Woolloongabba, QLD, Australia
Key words: Erythropoiesis, erythroblastic island, macrophage, red pulp macrophage, iron
homeostasis, niche, anaemia, inflammation
Corresponding author:
Jean-Pierre Levesque, PhD
Mater Research Institute – The University of Queensland
Translational Research Institute
37 Kent Street
Woolloongabba, QLD 4102
Australia
e-mail: jp.levesque @ mater.uq.edu.au

2
Abstract
It has recently emerged that tissue resident macrophages are key regulators of several stem cell
niches orchestrating tissue formation during development, as well as postnatally where they also
organise the repair and regeneration of many tissues including the haemopoietic tissue. The fact that
macrophages are also master regulators and effectors of innate immunity and inflammation allows
them to co-ordinate haematopoietic response to infections, injuries and inflammation. After recently
reviewing the roles of phagocytes and macrophages in regulating normal and pathological
haematopoietic stem cell niches, we now focus on the key roles of macrophages in regulating
erythropoiesis and iron homeostasis. We review herein the recent advances in understanding how
macrophages at the centre of erythroblastic islands form an erythropoietic niche that controls the
terminal differentiation and maturation of erythroblasts into reticulocytes, how red pulp
macrophages in the spleen control iron recycling and homeostasis, how these macrophages
coordinate emergency erythropoiesis in response to blood loss, infections and inflammation and
how persistent infections or inflammation can lead to anaemia of inflammation via macrophages.
Finally, we discuss the technical challenges associated with the molecular characterisation of
erythroid island macrophages and red pulp macrophages.
Highlights
 Macrophages in erythroblastic islands and spleen red pulp are key players of erythropoiesis
and iron homeostasis.
 Erythroblastic island macrophages support the maturation of erythroblasts into reticulocytes.
 Erythroblastic island macrophages regulate erythropoiesis response to infections and
inflammation.
 Red pulp macrophages are essential for iron recycling to feed erythropoiesis.
 Macrophage activation by inflammatory cytokines and PAMPs plays important roles in
anaemia of inflammation.

3
Introduction
Macrophages are phagocytic mononucleated myeloid cells present at similarly high density in every
single tissue of the body [1]. Macrophages are key effectors and regulators of inflammation and
acquired and innate immunity, through their ability to detect and phagocytose pathogens and
foreign bodies, and release bioactive molecules and proteins such as cytokines and interleukins that
modulate innate and adaptive immunity [2]. In addition to these immune roles, tissue resident
macrophages have essential functions in regulating tissue development, homeostasis, repair and
regeneration [3], including the skeletal and haematopoietic tissues [4]. Indeed, it has recently
emerged that macrophages play key roles in regulating the development of the haematopoietic
system in vertebrates from fish to mammals [5-8]. In the adult bone marrow (BM) they are key
orchestrators of haematopoietic stem cell (HSC) niche function and maintenance [9-15] in addition
to driving endosteal bone formation [9, 16]. In a previous issue of Experimental Hematology, we
have reviewed how macrophages, as effectors and regulators of innate immunity and inflammation,
are perfectly placed to adapt haematopoietic output in the presence of physiological stressors,
particularly in response to infections and inflammation. We discussed how their dysregulation can
contribute to the pathogenesis and response to treatments in several haematopoietic diseases,
including acquired aplastic anaemia, lymphomas, multiple myeloma, lymphoblastic leukaemia,
myelodysplastic syndrome, juvenile myelomonocytic leukaemia and possibly acute myeloid
leukaemia [17].
Macrophage directed formation of erythropoietic niches in the BM was one of the first
reported examples of the important role these cells play in supporting adult and stem and progenitor
cell homeostasis. These specialised erythroblastic island macrophages (EBI) control many aspects
of erythroid cell maturation. Of interest, macrophages influence the entire life cycle of the
erythrocytes. Specialised red pulp macrophages in the spleen are essential to the removal of old
erythrocytes [18-21] and the homeostasis of iron, a key chemical element for erythropoiesis [22].
The turnover of erythrocytes in humans is ~2.5 million erythrocytes per second throughout adult
life [23], with this number being constantly produced in bone marrow while the same number of
exhausted erythrocytes are continuously degraded in the spleen. These two processes run in parallel
to maintain erythrocyte numbers in the circulation.

4
Erythroblastic island (EBI) macrophages and erythropoiesis
Erythropoiesis is the critical process to continuously replace old erythrocytes (Figure 1A). It
involves differentiation of HSCs into megakaryocyte-erythroid progenitors (MEPs) in the BM [24]
which further commit to the erythroid lineage to generate erythroid progenitors such as erythroid
burst-forming units (BFU-E) and colony-forming units (CFU-E). These progenitors then
differentiate into pro-erythroblasts, followed by their progressive maturation into basophilic,
polychromatic and finally orthochromatic erythroblasts [20, 24, 25]. Erythropoiesis concludes with
the enucleation of orthochromatic erythroblasts into immature reticulocytes which mature into
typical biconcave shaped erythrocytes devoid of intracellular organelles. The whole erythroblast
maturation process and enucleation occurs in an erythroid niche called the erythroblastic island
(EBI) with 5-20 erythroblasts at various degrees of maturation rosetting around a central ferritin-
rich macrophage called the EBI macrophage [21, 26-29]. Historically, EBIs with their central
macrophages were first identified in the human BM [30, 31] and subsequently in the mouse BM
[26], mouse spleen [27], and rat BM [28]. EBIs, with the central macrophage promoting
erythroblast maturation, can also be reconstituted in BM cultures in erythrogenic conditions [32].
The EBI central macrophage is rich in ferritin [26, 30, 31], the main protein storing iron cations
within cells. Recent advances with inflight imaging flow cytometry have revealed that in the mouse
BM, stressed spleen and foetal liver, EBI macrophages express at their surface the
monocyte/macrophage antigen F4/80, tissue macrophage-restricted antigen CD169 and vascular cell
adhesion molecule-1 (VCAM-1) [21, 33] (Figure 1B).
EBI macrophages provide growth factors such as insulin-like growth factor-1 (IGF-1)
supporting erythroblast proliferation and differentiation [34-36], and iron ions for haemoglobin
synthesis within maturing erythroblasts [22, 29, 37]. EBI macrophages are also crucial for
erythroblast fission into an enucleated reticulocyte and a small nucleated pyrenocyte which is
engulfed and degraded by the EBI macrophage [38]. The molecular aspects of the erythroblast
enucleation process have been recently reviewed [38, 39]. Although erythroblast enucleation can
take place in cultures of BM erythroid progenitors, particularly in presence of sheer stress [40],
physiologically enucleation takes place in EBIs and involves several cell surface proteins such as
macrophage erythroblast attacher (MAEA) [41] and VCAM-1 [40] expressed at the surface of EBI
macrophages [21, 33], and intercellular adhesion molecule-4 (ICAM-4) expressed by erythroblasts
[42]. Conditional deletion of the Maea gene in Csf1r expressing macrophages leads to reduced
medullary erythropoiesis in steady-state and under stress, with reduced frequency of EBIs [41].
However, conditional deletion of the Vcam1 gene in macrophages did not have such an effect [41]

5
although infusion of a neutralising monoclonal antibody specific for VCAM-1 delayed medullary
erythropoiesis recovery after BM transplantation [18]. Icam4-/- mice are not anaemic in steady-state
but have lower frequencies of EBIs in the BM [42]. Once enucleated, the pyrenocyte must be
engulfed and degraded by EBI macrophages. Mer tyrosine kinase (MERTK), which binds protein S
bound to phosphatidyl serine at the surface of the pyrenocyte, plays an important role in pyrenocyte
attachment to the EBI macrophage [29, 40, 41]. DNAse 2α within EBI macrophages is essential to
degrade the DNA within the pyrenocyte nucleus [43, 44]. Interestingly, although the presence of
pyrenocytes was not stated explicitly, mice and rats with depleted or absent macrophages in the
foetal liver do not accumulate obvious nuclear material [45-47], suggesting that during foetal
haematopoiesis, another cell type may be able to degrade the expelled nucleus.
There are several lines of evidence that EBI macrophages are key for successful completion
of erythropoiesis in vivo. EBI macrophages are known to express the CD169 antigen (also called
sialoadhesin or sialic acid-binding immunoglobulin-type lectin-1 / SIGLEC1) [19, 21, 33], which is
uniquely expressed by tissue-resident macrophages [48] including BM niche macrophages [13].
Depletion of CD169+ macrophages in mice with a diphtheria toxin receptor transgene knocked-in to
the Siglec1 gene (Siglec1DTR mice) causes the collapse of EBIs and erythropoiesis arrest in both BM
and spleen with dramatic reductions in the numbers of erythroblasts in these two tissues [18, 20].
Furthermore, treatment of mice with granulocyte colony-stimulating factor (G-CSF), which
depopulates BM macrophages [9], or bacterial lipopolysaccharides (LPS) which activates BM
macrophages [49], also causes a collapse in the numbers of EBIs in the BM with similar arrest in
medullary erythropoiesis [19, 21]. However, under these conditions in the mouse splenic stress-
induced extramedullary erythropoiesis is activated to minimise disruption to erythrocyte supply [19,
20]. Therefore, the EBI macrophage forms a unique erythropoietic niche in the BM to support the
final stages of reticulocyte formation and can recreate this niche within the mouse spleen during
stress /emergency erythropoiesis.
Eliminating the old and stiff, the role of red pulp macrophages
The spleen acts as a sieve to eliminate old erythrocytes that have lost their plasma membrane
flexibility. Specifically, splenic red pulp macrophages phagocytose old blood erythrocytes to
recycle their iron content. Red pulp macrophages are distinct from medullary and splenic EBI
macrophages. Phenotypic differences include that splenic red pulp macrophages do not express the
CD169 antigen on their surface in steady-state [14]. However, at some stage during red pulp

6
macrophage development, Siglec1 is expressed as these cells are targeted in transgenic mouse
models utilising the Siglec1 promoter. Indeed specific ablation of CD169+ macrophages in
Siglec1DTR mice not only depletes both red pulp and EBI macrophages resulting in arrest of
medullary and splenic erythropoiesis [18, 20], but also increases lifespan of blood erythrocytes [18].
Development of red pulp macrophages is controlled by the transcription factor Spi-C [50, 51] and a
subset of reticular fibroblasts uniquely present in the red pulp expressing Wilms’ Tumour1 (WT1)
and CSF1 [52]. Haem molecules released from haemoglobin degradation can differentiate
monocytes into Spi-C+ macrophages in vitro and haem injection expands Spi-C+ red pulp
macrophages in vivo [50]. Red pulp macrophages maintain iron homeostasis and stable numbers of
healthy haemoglobin-loaded erythrocytes in the blood. The balance between medullary
erythropoiesis and erythrophagocytosis in the spleen is key to maintaining steady erythrocyte levels
and their lifespan of approximately 120 days in humans and 40 days in mice.
Most iron requirements are from recycling by macrophages
Iron is vital to all forms of life because its ions in oxidation states II (Fe2+ ferrous cations) and / or
III (Fe3+ ferric cations) are essential components of the haems in O2 and CO2 transporters (e.g.
haemoglobins and myoglobins) and cytochromes, and in the catalytic centres of many non-haem
mononuclear iron containing enzymes [53] and iron-sulphur cluster containing enzymes [54] (e.g.
oxidases, dioxygenases, reductases, hydroxylases, dehydrogenases, mitochondrial enzymes of the
Krebs cycle, respiratory chain complexes, DNA metabolism enzymes, etc). Oxidation and reduction
reactions between Fe(II) and Fe(III) oxidation states enable donation or capture of electrons to
catalyse many reactions essential to life:
    
However, the absorption and handling of elemental iron presents a few challenges because the most
abundant form of iron on earth since the cyanobacteria-driven Great Oxidation Event that occurred
2.4 billion years ago [55, 56], is the more oxidised state III (Fe3+), which is almost insoluble in
aqueous solutions at physiological pH range 7-8. The more water soluble Fe2+ cations, which are
used in haem molecules, can react with peroxides produced by cell metabolism to form extremely
toxic and reactive hydroxyl radicals OH in the Fenton reaction [57, 58] described below:
     

7
Since no enzyme detoxifies OH radicals (unlike peroxides), the over two billion of years of
evolution have generated ingenious ways to capture, utilise and recycle elemental iron while
containing its inherent toxicity. This iron handling machinery is summarised in Table 1 and Figure
2. As described below, erythrophagocytosis by red pulp macrophages in the spleen is critical to iron
homeostasis.
Alimentary iron absorption is a very inefficient process in part because the most abundant
form of alimentary iron is the more oxidised Fe3+ cation which is very poorly soluble in aqueous
solutions. To compensate for this, iron recycling is very efficient. Indeed, only 1-2 mg/day
alimentary iron are absorbed by duodenal enterocytes in humans, which is insufficient for the daily
20-25 mg iron required to replace haemoglobin and erythrocytes that are phagocytosed by and
degraded within splenic red pulp macrophages (Figures 2 and 3). Alimentary iron is released during
the digestive process as Fe3+ ions and must be first reduced to into Fe2+ by cytochrome b reductase 1
(CYBRD1) at the luminal surface of enterocytes in the duodenum in order to be imported across the
enterocyte plasma membrane by divalent metal transporter-1 (DMT-1/SLC11A2). Cytoplasmic Fe2+
is then re-exported into the blood via ferroportin (FPN / SLC40A1), a Fe2+ transmembrane iron
transporter [59-61] (Figures 2 and 3A). Of note the alimentary Fe2+-containing haem molecule
released from meat digestion can also be directly imported by enterocytes via heme carrier protein-1
(HCP-1/SLC46A1) (Figure 2). Fe2+ is then released from the haem molecule by heme oxidase-1
(HMOX-1) in enterocyte cytosol [62]. Once released out of enterocytes by FPN, Fe2+ is oxidised to
Fe3+ at the basal surface of duodenal enterocytes by the transmembrane ferroxidase hephaestin
(HEPH) [63, 64] as well as by the plasma copper-dependent ferroxidase ceruloplasmin [65, 66]
(Figure 2) because iron can only bind to its blood transporter transferrin (TF) in its Fe3+ oxidised
state (Figure 3B). Each TF molecule binds two Fe3+ cations.
The other 20 mg daily iron necessary to maintain erythropoiesis in humans are from iron
recycled through splenic and hepatic erythrophagocytosis of old erythrocytes by red pulp
macrophages and Kupfer cells [67] (Figure 1F). Once old erythrocytes are phagocytosed, Fe2+ is
extracted from haem molecules of haemoglobin by HMOX-1 (Figure 2), an enzyme which is also
key to maintaining functionality of both medullary EBI macrophages and erythrophagocytic
macrophages in the spleen [68]. Once extracted from the haem molecule and released into the
cytosol, Fe2+ ions can be 1) incorporated in haem molecules and other iron-containing enzymes, 2)
stored in large polymers of ferritin within EBI and red pulp macrophages, or 3) re-exported to the
circulation or neighbouring cells by via FPN (iron recycling) (Figure 3C-I). Ferritin is essential for
intracellular iron storage such as in EBI and red pulp macrophages as well as storage of excessive

8
iron in hepatocytes [69]. Ferritin is a large polymer of 24 heavy and light subunits that forms a 75 Å
diameter nano cage within which up to 4,500 iron atoms can be stored [70] in the form of ferric
oxyhydroxide (FeOOH) associated with various amounts of phosphate [71]. As iron is present in
the cytoplasm as water soluble Fe2+ but stored in ferritin in its Fe(III) oxidation state, ferritin heavy
subunits have a ferroxidase activity enabling its oxidation to Fe3+ for storage in a reaction that
produces toxic hydroxyl radicals OH [72]. Of note, a proportion of ferritin leaks into the blood and
plasma ferritin concentration is a marker of iron overload (if > 300 ng/mL) or iron deficiency (<30
ng/ mL). The release of intracellular iron stores involves nuclear receptor coactivator-4 (NCOA4)
which cargoes ferritin-iron complexes to autophagosomes where it is degraded enabling Fe3+
release. NCOA4 is critical to iron release from ferritin storage in macrophages and enterocytes.
Indeed, Ncoa4-/- mice accumulate iron in spleen, liver and BM macrophages as well as duodenum
enterocytes resulting in typical microcytic hypochromic iron-deficiency anaemia (IDA) [73].
Although NCOA4 expressed both cell autonomously (erythroblasts) and non-autonomously
(macrophages, hepatocytes, enterocytes) is thought to be critical to iron storage mobilisation to
complete erythropoiesis [74], recent experiments with transplantation chimeras of WT and Ncoa4-/-
mice have shown that NCOA4 expressed by macrophages is critical to iron mobilisation,
particularly in the context of an iron deficient diet [75]. Once released in the cytoplasm from ferritin
storage, Fe3+ is then reduced to Fe2+ by 6-transmembrane epithelial antigen of the prostate-3
(STAEP-3) enzyme (Figure 2) [76]. Fe2+ is exported from the macrophage cytosol into the blood by
FPN where it is oxidised to ferric Fe3+ by plasma ferroxidases such as copper-dependent
ceruloplasmin and other plasma ferroxidases yet to be identified [65, 66, 77]. Fe3+ bound to TF can
be transported through the whole body. The same mechanism takes place in EBI macrophages, red
pulp macrophages and hepatocytes to import, recycle, store and re-export iron (Figure 2). All cells
needing iron (particularly erythroblasts and all dividing cells) express the transmembrane TF
receptor (TFRC / CD71). The Fe3+-TF-TFRC complex is endocytosed by erythroblasts and upon
acidification of endosomes, Fe3+ is released from the complex and reduced to Fe2+ by STAEP-3
enzyme with TFRC and iron-free apoTF recycled to the cell surface. Fe2+ is then transported across
the endosome membrane into the cytoplasm by DMT-1 (Figure 2) where it can enter in the
synthesis of key enzymes and the haem [22, 78]. As FPN is expressed by duodenal enterocytes, EBI
in the BM and spleen, red pulp macrophages in spleen and hepatocytes, FPN is at the centre of iron
biodistribution during the whole erythropoietic and iron recycling processes (Figure 3).

9
Hepcidin and macrophages are key regulators of erythropoiesis adaptation to infections and
inflammation
A key regulator of iron homeostasis is hepcidin antimicrobial peptide (HAMP) (Figure 3H). During
inflammation and infections, IL-6 expression is induced. IL-6 binding to hepatocytes induces
HAMP transcription, production and release into the blood [79]. HAMP reduces iron bioavailability
by binding to FPN on duodenal enterocytes, hepatocytes and splenic red pulp and BM EBI
macrophages which causes FPN internalisation [80] and subsequent ubiquitination by E3 ubiquitin
ligase RNF217 leading to FPN degradation [81], thereby blocking iron export from these cells.
Once internalised, FPN cannot shuttle iron outside of cells and iron bioavailability to other cells
drops. Reduced serum and cellular iron concentrations inhibit the survival and virulence of
infectious organisms hence limiting the infection [82]. The key role of HAMP in iron homeostasis
is shown by mice overexpressing a Hamp transgene, which develop a hypochromatic microcytic
anaemia similar to IDA [83], whereas mice lacking the Hamp gene have iron overload and
haemochromatosis [84] (Figure 1I).
Of note, the iron storage disease haemochromatosis is the most common recessive disease in
populations of European origin and is most frequently associated with low expression of HAMP, or
defective HAMP protein due to mutations in the HAMP gene [85]. Haemochromatosis in northern
European populations is usually caused by mutations in the HFE gene encoding high ferric protein
(HFE) [86], which is bound to TFRC at the surface of hepatocytes and is necessary to transduce the
signal emanating from the binding of iron saturated TF to activate HAMP transcription [87]. HFE
mutations leading to haemochromatosis, fail to activate HAMP transcription in hepatocytes when
TF is highly saturated with iron [86]. Due to low HAMP expression in people with
haemochromatosis, there is increased absorption of iron by duodenal enterocytes as well as
excessive export of iron from macrophages and hepatocytes via FPN, leading to much higher iron
levels in blood and tissues than in normal individuals. The excessive circulatory and tissue iron in
haemochromatosis and related conditions of iron excess is stored largely in liver hepatocytes using
the same protein machinery as in macrophages and duodenal enterocytes as summarised in Table 1
and Figure 3 but is also deposited in heart and endocrine organs. Similar mechanisms leading to
haemochromatosis involve mutations in hemojuvelin (HJV gene), which is complexed with BMP
receptors at the surface of hepatocytes to induce HAMP expression in response to bone
morphogenetic protein-6 (BMP-6) [88], transferrin receptor 2 (TFR2), and iron transporter FPN
(SLC40A1) [61] as well as mutations in HAMP itself [85, 89]. Patients consequently have low levels

10
of HAMP and high blood iron levels, highlighting the importance of this protein in sensing and
regulating circulatory iron levels.
An illustration of the key role of EBI macrophages and red pulp macrophages in coupling
erythropoiesis with iron homeostasis is given by the mechano-sensing ion channel PIEZO1.
Recently identified gain-of-function mutations in the PIEZO1 gene in humans and mice cause iron
overload and haemochromatosis with enhanced erythropoiesis in both species [90]. Conditional
expression of a clinically relevant gain-of-function PIEZO1 mutant specifically in macrophages was
sufficient to replicate this phenotype in mice with increased iron content in liver and plasma,
increased TF saturation, increased erythropoiesis, increased erythrophagocytosis by red pulp
macrophages and decreased erythrocyte lifespan [90]. Importantly specific conditional expression
of this mutant in erythroid cells had no such effect [90].
While inflammation reduces iron bioavailability by upregulating HAMP in response to IL-6,
inflammation also increases erythrophagocytosis by red pulp macrophages via interferon-γ [91] and
IL-4 [92], and consequently reduces blood erythrocyte lifespan. Several pathogen-associated
molecular patterns (PAMPs) such as zymosan and LPS reduce erythrocyte lifespan in mice by
activating red pulp macrophages [93].
As essential effectors of iron homeostasis, erythrophagocytosis and EBIs, macrophages play a
key role in the development of anaemia of inflammation (AoI), a form of anaemia caused by acute
or chronic inflammatory conditions [94]. AoI is estimated to affect 40% of the over one billion
individuals suffering anaemia worldwide [94, 95]. AoI most frequently occurs as a consequence of
sepsis [96], chronic infections [97], chronic inflammatory diseases such as inflammatory bowel
disease [95, 98, 99], haematological malignancies [94] and autoimmune disorders such as
rheumatoid arthritis [100]. It has been assumed that AoI is caused by a combination of 1) iron
deficiency caused by the induction of the IL-6 /HAMP loop in response to inflammation / infection,
and 2) enhanced erythrophagocytosis and reduced erythrocyte lifespan. However, patients with true
AoI and true IDA have opposite profiles in respect to iron homeostasis blood markers. True IDA
patients have low serum ferritin (low iron stores) and high serum soluble TFRC [101] suggesting
high but ineffective or incomplete erythropoiesis. In contrast, true AoI patients have high serum
ferritin and low soluble TFRC, suggesting they are not iron-deficient [101] but erythropoiesis is
reduced. Therefore, low iron availability caused by HAMP up-regulation may not be the main cause
of AoI, even if it is a contributing factor to the anaemia. This may explain why approximately 50%
of inflammatory bowel disease patients with anaemia do not respond to iron supplementation [98].

11
Erythroblastic island macrophages and anaemia of inflammation
We have recently reported that injection of bacterial LPS to mimic endotoxemia, as well as
injection of G-CSF, causes a rapid collapse of EBIs in the BM with a rapid arrest of medullary
erythropoiesis [19, 21] (Figure 3G). The suppressive effect of LPS on medullary erythropoiesis is
completely dependent on the LPS receptor toll-like receptor-4 (TLR-4) and its intracellular adaptor
protein MyD88 as mice germinally knocked-out for either gene have normal medullary
erythropoiesis following LPS treatment. TLR-4 is not expressed at mRNA or protein levels by
CFU-E, pro-erythroblasts or erythroblasts, and as LPS has no effect on BFU-E/CFU-E erythroid
colony output in culture [19] whereas macrophages express high levels of TLR4 [19] and are well
known effectors of the response to LPS [49] implicating that the LPS effect on erythropoiesis is
indirectly mediated by EBI macrophages. Similar collapse of medullary EBIs and arrest in
medullary erythropoiesis were identified by imaging flow cytometry in mice treated with G-CSF
[21]. Public RNA -sequencing database
(https://haemosphere.org/expression/show?geneId=ENSMUSG00000028859) as well as our own
qRT-PCR on sorted cells (Figure 4), show that erythroblasts do not express the G-CSF receptor
gene Csf3r. Therefore the suppressive effect of G-CSF on EBIs is presumably mediated through the
macrophages of the EBIs as BM macrophages do express Csf3r [11] (Figure 4). Surprisingly, unlike
LPS-mediated HSPC mobilisation which is indirectly mediated by G-CSF secreted in response to
LPS, LPS-mediated suppression of medullary erythropoiesis was G-CSF-independent as well as IL-
1-independent and TNF-independent [19]. A role of interferon-γ (IFN-γ) is possible as mice
chronically overexpressing IFN-γ are anaemic with reduced numbers of BFU-E and erythroblasts in
the BM (but with normal numbers of CFU-E) [102]. Although the effect of high IFN-γ on EBIs was
not investigated in this model, it is important to note that erythroblasts do not express IFN-γ
receptors IFNGR1 and IFNGR2 unlike BM macrophages which express high levels of both
(https://haemosphere.org/expression/show?geneId=ENSMUSG00000020009 and
https://haemosphere.org/expression/show?geneId=ENSMUSG00000022965). Thus, the suppressive
effect of IFN-γ could also be in part mediated via EBI macrophages and this will need to be further
investigated. Therefore although reduced iron bioavailability by activation of the IL-6 / HAMP
pathway [103] and shortened erythrocyte lifespan [91] may contribute in part to AoI, we propose
that EBI macrophages are key effectors of AoI by sensing innate immune stimuli such as LPS and
possibly other PAMPs such as zymosan during infections. Detection of these PAMPs cause EBI

12
macrophages to rapidly shift their erythropoietic function to an inflammatory response resulting in
the loss of medullary EBIs and a profound suppression of medullary erythropoiesis.
Another fascinating aspect of EBI macrophages in the mouse is that their response to
physiological challenges depends on their tissue of residence. In steady state, most of background
erythropoietic activity takes place within the BM with little activity in the spleen. However in the
mouse, medullary erythropoiesis is switched off with impairment of EBI macrophage function in
response to inflammatory or infectious challenges while temporary extramedullary emergency /
stress erythropoiesis rapidly ramps up [19, 20, 104] to limit the extent of anaemia [105]. This is
associated with rapid expansion of EBI and EBI macrophages in the spleen and/or liver [27].
Phlebotomy and haemolytic anaemia induced by injection of phenyl hydrazine also stimulate mouse
splenic stress erythropoiesis and increased number of splenic EBI macrophages [106].
Several pathways have been identified to explain the increased number of splenic EBI
macrophages in response to haemolytic or inflammatory stress in the mouse. Inflammation induces
Spi-C expression in red pulp macrophages [107]. Spi-C induction not only enhances
erythrophagocytosis but also activates expression of growth differentiation factor-15 (GDF-15)
[107]. GDF-15 is necessary to the expansion of splenic EBI macrophages and to stress
erythropoiesis in response to phenyl hydrazine-mediated haemolysis or to PAMPs. Stress
erythropoiesis in response to phenyl hydrazine or PAMPs is blunted in Gdf15-/- mice with much
reduced expansion of EBI macrophages and erythropoiesis in the spleen of these animals [106,
107]. GDF-15 increases the recruitment of BM-derived monocytes from the circulation into the
spleen where they differentiate into EBI macrophages [108]. GDF-15 has a dual effect in the spleen
as it also stimulates expansion of stress erythroid progenitors recruited from the BM into the spleen
by promoting expression of enzymes involved in glucose metabolism and glutaminolysis [106]. The
effect of GDF-15 on splenic EBI macrophages is mediated by BMP-4. GDF-15 together with Indian
Hedgehog (IHH) stimulates BMP-4 expression in splenic red pulp macrophages in a HIF-2α-
dependent manner [106, 109]. Depletion of CD169+ macrophages in Siglec1DTR mice prevents the
expression of BMP-4 in the spleen in response to myeloablation [18]. This BMP-4-dependent
induction of splenic stress erythropoiesis is also operative in response to PAMPs such as zymosan
and LPS [93]. BMP-4 expressed by splenic macrophages then stimulates proliferation of stress
erythroid progenitors (SEPs), which are derived from short-term CD34+ HSCs in the spleen under
signalling from BMP-4 and IHH, and initiation of stress erythropoiesis [109, 110]. This mechanism
is also effective in hepatic stress haematopoiesis [111]. Interestingly BMP-4 does not stimulate
erythroid progenitors in the BM.

13
An additional mechanism to stimulate extramedullary stress erythropoiesis may involve
adrenal glucocorticoids which are the centre piece of the stress response. In vitro, glucocorticoid
agonists induce differentiation of monocytes into macrophages that can form EBIs and support
erythroblast enucleation in culture [112, 113]. However, whether this mechanism is operative in
vivo in response to acute inflammation remains to be determined.
Intriguingly, rats have a different response to phenyl hydrazine-induced haemolytic stress
compared to mice. Unlike mice, emergency erythropoiesis does not take place in the spleen and as a
consequence, anaemia is more prolonged in rats than in mice with no induction of BMP-4
expression in the rat spleen in response to phenyl hydrazine [114]. Even more surprising was the
observation that the BM is the main site of emergency erythropoiesis in response to phenyl
hydrazine in rats with expansion of KIT+ SEPs and up-regulation of BMP-4 in the rat BM rather
than the spleen [114]. Intriguingly, glycophorin A+ KIT+ SEPs were also abundant in human BM
within 30 days post HSC transplantation, suggesting that the BM is also the site of emergency
erythropoiesis in humans [114]. A hypothesis to explain this difference between rats and mice is
that the rat BM is more fatty (similar to human BM) than mouse BM and this allows for the
expansion of SEPs in place of adipocytes in the rat and human BM, whereas SEPs have to move to
the spleen in mice due to the lack of vacant space in the mouse BM for emergency erythropoiesis to
take place [114]. Therefore, in respect to emergency erythropoiesis, the mouse seems to be different
from rats and humans [114].
An enduring question is why EBI macrophages collapse in the BM while they expand in the
mouse spleen in response to inflammatory cytokines (e.g. G-CSF or Flt3 ligand) or PAMPs such as
LPS or zymosan. This has not been explored extensively and side by side comparison of medullary
and splenic EBI macrophages has not been made. A factor contributing to this lack of knowledge
could be the difficulty of purifying tissue-resident macrophage populations for molecular profiling
(see Challenges section). Despite this lack of knowledge, some cell surface receptors may have
different expression levels or function between these two populations in mouse. Although splenic
and medullary EBI macrophage populations both express CD169, as CD169+ macrophage depletion
in vivo arrests both medullary and splenic erythropoiesis [10, 20], deletion of the Maea gene in
macrophages mostly alters medullary EBIs and erythropoiesis in steady-state [41], whereas
inducible deletion of the α4 integrin gene (Itga4), which forms receptors for VCAM-1 expressed by
EBI macrophages, mostly affects splenic stress erythropoiesis in response to phenyl hydrazine [115,
116].

14
Challenges to study the molecular biology of EBI and red pulp macrophages
As alluded to earlier, characterisation of the unique molecular signature underlying macrophage
functional adaptations to BM niches and their support EBI has been elusive [17, 117]. This is
reflected in the inability to achieve precise identification and consequently isolation of the multiple
macrophage subsets present in BM e.g. EBI versus HSC niche or other BM macrophages [4]. When
quantified in situ using F4/80 staining, macrophages make up about 15% of the cells in the BM [14]
and more than 30% of cells in spleen [14], and yet single cell sorting frequently recovers only a
small proportion of these. For example, the Tabula Muris data of single cell sequencing results in
mouse identified only 3.4% of FACS-sorted BM cells and 2.8% of spleen cells as macrophages, and
surprisingly, few of these had detectable expression of macrophage signature genes such as Csf1r,
Adgre1 and Siglec1 [118].
We have recently provided an explanation for anomalous results relating to macrophage
frequency and expression profiles in single cell data from mouse studies. A large majority of cell
events that would be assigned as BM or splenic macrophages in single cell suspensions using
traditional flow cytometry are actually membrane-encapsulated macrophage remnants attached to
the surface of other cell types [117]. In spleen single cell suspensions, F4/80+ macrophage remnants
are attached to Ter119+ red blood cells with biconcave morphology among other cells [117]. This is
likely due to the extraordinary reticulation of F4/80+ macrophages in the BM and spleen with
processes contacting many neighbouring nucleated cells [1, 117]. To illustrate this specifically in
the context of EBI, we analysed by imaging flow cytometry single cell suspensions from the BM of
Siglec1-Cre x Rosa26-lox-STOP-lox-ZsGreen (abbreviated Siglec1ZsGreen mice) in which the bright
coral-derived fluorescent reporter ZsGreen is expressed specifically in CD169+ macrophages. In
BM suspensions from these mice, a significant portion of Ter119+ events that fell in the single cell
gate had detectably F4/80 expression (Figure 5). Most of these Ter119+ F4/80+ double positive
events were in fact intact Ter119+ erythroblasts with one or more F4/80+ puncta attached at their
surface (Figure 5), most likely representing remnants of EBI macrophages to which these
erythroblasts were attached in the tissue [117]. The low stain area of F4/80 within the cell
population compare to the stain area of Ter119, further verifies that F4/80 distribution is not
consistent with cell surface expression in the majority of Ter119+ gated cells (Figure 5). Of note,
BM erythroblasts had F4/80+ macrophage remnants that could be ZsGreen+ or ZsGreen- suggesting
that the cytoplasmic reporter was not consistently carrier within the remnants or that both CD169+

15
and CD169- macrophage subsets have been disrupted during isolation and interact with
erythroblasts (Figure 5). Actual intact macrophages were very rare in single cell suspension of
disaggregated mouse BM and spleen [117] suggesting that traditional flow cytometry macrophage
gating strategies are predominantly capturing non-macrophage cells with attached macrophage
remnants. For instance, we and others have reported a subset of myeloid cells co-expressing the
F4/80 and CD169 macrophage markers together with the Ly6G granulocyte marker in the mouse
BM [20]. We interpreted these F4/80+ CD169+ Ly6G+ events as containing EBI macrophages due to
1) their expression of CD169 and 2) the observation that these cells were specifically depleted in the
BM of mice treated with agents that cause arrest medullary erythropoiesis such as clodronate-
loaded liposomes, CD169+ macrophage depletion, G-CSF or LPS [9, 15, 20]. However subsequent
studies of whole EBIs using imaging flow cytometry showed that the central macrophage of EBIs in
the mouse BM do not express Ly6G [19, 21, 33]. We since have confirmed that F4/80+ CD169+
Ly6G+ BM events represent Ly6G+ granulocytes with attached CD169+ macrophage remnants
[117]. As macrophage remnants contain macrophage RNA, intracellular and cell surface proteins
[117], profiling of BM macrophage populations including EBI macrophages should not rely solely
on ex vivo single cell suspension strategies, with all expression profiles validated using imaging
techniques prior to attributing to specific cell types. It is clear that ongoing precision analysis,
guided by these recent observations, will be essential in achieving in depth, accurate profiling of the
tissue resident macrophage subsets that play integral roles in regulation of erythropoiesis.
Conclusions
While EBIs with their central macrophage were discovered in mammal BM half a century ago, the
realisation that macrophages are key to couple erythropoiesis with iron metabolism in steady-state
and in response to stress is much more recent. EBI macrophages support the final stages of
erythroblast maturation to produce enucleated reticulocytes and adapt erythropoiesis output to
infections and inflammation. Red pulp macrophages in the spleen are essential to recycle iron from
haemoglobin and the haem, and to deliver it to the BM to ensure sufficient haemoglobin synthesis
in erythroblasts. This exquisite coupling of medullary erythropoiesis with iron metabolism exerted
in tandem by EBI in the BM and red pulp macrophages in the spleen is illustrated with macrophage-
specific conditional expression of loss-of-function or gain-of-function mutants of the mechano-
sensing ion channel PIEZO1, which alters both erythropoietic output and iron recycling and
metabolism [90]. Likewise, inactivation of the Ncoa4 gene in macrophages, which is necessary for

16
the mobilisation of iron intracellular stores from ferritin, also alters erythropoiesis and iron
metabolism in tandem [75].
From recent experiments summarised in this review, it is increasingly evident that the
anaemia that develops acutely in patients with sepsis [96], chronic infections such as tuberculosis
[97], or chronic inflammatory diseases such as inflammatory bowel disease [98, 99] or rheumatoid
arthritis [100] is not entirely due to increased erythrophagocytosis by red pulp macrophages and
increased iron sequestration caused by induction of HAMP secretion and consequent FPN
internalisation in hepatocytes and macrophages [119] to starve prokaryotic and eukaryotic
infectious micro-organisms that also need iron for their own metabolism. Indeed, several PAMPs
released by infectious micro-organisms (e.g. LPS from gram-negative bacteria walls or zymosan
from fungi and yeasts), as well as certain inflammatory cytokines such as G-CSF can rapidly arrest
medullary erythropoiesis by causing EBI disaggregation in the BM [19, 21, 93], and may explain
why iron supplementation is not always effective at correcting anaemia of inflammation in these
conditions [98]. This effect of PAMPs on both medullary erythropoiesis and iron homeostasis
illustrates again the coupling between these two parallel biological processes. Elucidation of the
molecular mechanisms by which inflammatory cytokines and PAMPs dislocate EBIs may reveal
alternative therapeutic approaches to restore the functionality of EBI macrophages and EBIs and
thus restore medullary erythropoiesis in the context of sepsis, chronic infections and inflammation.
However, recent imaging flow cytometry experiments [117] (Figure 4) have shown that
isolation of pure intact macrophages devoid of erythroid cell contamination from the BM or spleen
for molecular profiling and characterisation is very challenging as illustrated in this and previous
reviews [17]. Macrophage preparations used to sort red pulp and EBI macrophages to generate their
transcriptome profile should be further validated to confirm that these cell preparations contained
the intended pure cell type. Nevertheless, complementary experimental approach can de designed to
further characterise red pulp and EBI macrophages. One approach is to culture monocytes [112,
113] or induced pluripotent stem cells [120] in conditions that force the erythropoiesis-supportive
function of macrophages or promote Spi-C expression and the iron recycling function [50]. Unique
transcripts identified in in vitro generated EBI-like or red pulp-like macrophages should be
validated in vivo using adequate fluorescent reporters and techniques (such as imaging flow
cytometry) to confirm expression of the translated transcript in actual EBI or red pulp macrophages.
Alternative approaches could be adapted from proximity-based single cell RNA sequencing
analyses developed to isolate RNA from single HSC niche cells from sections of the neonatal BM
[121]. The original method involved isolation of niche cells proximal to fluorescently labelled

17
HSCs from sections of neonatal BM frozen sections to avoid fixation and decalcification of the
bone. Similar technique combining fluorescent reporter mouse strains such as Siglec1ZsGreen or
SpicGFP mice to label EBI macrophages and red pulp macrophages in the BM and spleen combined
with fluorescent Ter119 antibody to label Ter119+ erythroblasts, should enable to selectively pick
single EBI macrophages and red pulp macrophages from neonatal BM and spleen for molecular
characterisation. Finally, genetic approaches are also possible with conditional gene deletion or
induction using Spic-Cre mice to target red pulp macrophages. Although transcripts specifically
expressed in EBI macrophages have not yet been identified, the fact that CD169 is exclusively
expressed in tissue resident macrophages including EBI macrophages, Siglec1-Cre mice enable
direct gene deletion or expression in these macrophages without affecting monocytes, granulocytes
and dendritic cells or other cell types. These approaches will uncover the key proteins, their
regulation and their interactions that allow macrophages to control iron homeostasis and
erythropoiesis in the body and may reveal new therapeutic targets to correct pathologies associated
with iron homeostasis and erythropoiesis.
Acknowledgements
JPL is funded by Research Fellowship 1136130 from the National Health and Medical Research
Council of Australia, and ARP by a University of Queensland Amplify Fellowship. The authors are
also funded by the Mater Foundation. The Translational Research Institute receives funding from
the Australian Government.
Conflict of interest disclosure
The authors declare no competing financial interests.

18
References
1. Grabert K, Sehgal A, Irvine KM, Wollscheid-Lengeling E, Ozdemir DD, Stables J, et al. A
transgenic line that reports CSF1R protein expression provides a definitive marker for the
mouse mononuclear phagocyte system. J Immunol. 2020; in press.
2. Guilliams M, Mildner A, Yona S. Developmental and functional heterogeneity of monocytes.
Immunity. 2018; 49:595-613.
3. Ginhoux F, Schultze JL, Murray PJ, Ochando J, Biswas SK. New insights into the
multidimensional concept of macrophage ontogeny, activation and function. Nat Immunol.
2016; 17:34-40.
4. Kaur S, Raggatt LJ, Batoon L, Hume DA, Levesque J-P, Pettit AR. Role of bone marrow
macrophages in controlling homeostasis and repair in bone and bone marrow niches. Semin
Cell Dev Biol. 2017; 61:12-21.
5. Travnickova J, Tran Chau V, Julien E, Mateos-Langerak J, Gonzalez C, Lelièvre E, et al.
Primitive macrophages control HSPC mobilization and definitive haematopoiesis. Nat
Commun. 2015; 6:6227.
6. Li Z, Vink CS, Mariani SA, Dzierzak E. Subregional localization and characterization of
Ly6aGFP-expressing hematopoietic cells in the mouse embryonic head. Dev Biol. 2016;
416:34-41.
7. Li Z, Mariani SA, Rodriguez-Seoane C, He W, Ning X, Liu B, et al. A role for macrophages
in hematopoiesis in the embryonic head. Blood. 2019; 134:1929-40.
8. Mariani SA, Li Z, Rice S, Krieg C, Fragkogianni S, Robinson M, et al. Pro-inflammatory
aorta-associated macrophages are involved in embryonic development of hematopoietic stem
cells. Immunity. 2019; 50:1439-52.e5.
9. Winkler IG, Sims NA, Pettit AR, Barbier V, Nowlan B, Helwani F, et al. Bone marrow
macrophages maintain hematopoietic stem cell (HSC) niches and their depletion mobilizes
HSCs. Blood. 2010; 116:4815-28.
10. Chow A, Lucas D, Hidalgo A, Méndez-Ferrer S, Hashimoto D, Scheiermann C, 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; 208:261-71.
11. Christopher MJ, Rao M, Liu F, Woloszynek JR, Link DC. Expression of the G-CSF receptor
in monocytic cells is sufficient to mediate hematopoietic progenitor mobilization by G-CSF in
mice. J Exp Med. 2011; 208:251-60.
12. Hur J, Choi J-I, Lee H, Nham P, Kim T-W, Chae C-W, et al. CD82/KAI1 maintains the
dormancy of long-term hematopoietic stem cells through interaction with DARC-expressing
macrophages. Cell Stem Cell. 2016; 18:508-21.
13. Kaur S, Raggatt LJ, Millard SM, Wu AC, Batoon L, Jacobsen RN, et al. Self-repopulating
recipient bone marrow resident macrophages promote long-term hematopoietic stem cell
engraftment. Blood. 2018; 132:735-49.
14. Kaur S, Sehgal A, Wu AC, Millard SM, Batoon L, Sandrock CJ, et al. Stable colony-
stimulating factor 1 fusion protein treatment increases hematopoietic stem cell pool and
enhances their mobilisation in mice. J Hematol Oncol. 2021; 14:3.
15. McCabe A, Zhang Y, Thai V, Jones M, Jordan MB, MacNamara KC. Macrophage-lineage
cells negatively regulate the hematopoietic stem cell pool in response to interferon gamma at
steady state and during infection. STEM CELLS. 2015; 33:2294-305.
16. Chang MK, Raggatt L-J, Alexander KA, Kuliwaba JS, Fazzalari NL, Schroder K, et al. Osteal
tissue macrophages are intercalated throughout human and mouse bone lining tissues and
regulate osteoblast function in vitro and in vivo. J Immunol. 2008; 181:1232-44.

19
17. Levesque JP, Summers KM, Millard SM, Bisht K, Winkler IG, Pettit AR. Role of
macrophages and phagocytes in orchestrating normal and pathological haematopoietic niches.
Exp Hematol. 2021; 100:12-31.
18. Chow A, Huggins M, Ahmed J, Hashimoto D, Lucas D, Kunisaki Y, et al. CD169+
macrophages provide a niche promoting erythropoiesis under homeostasis and stress. Nat
Med. 2013; 19:429-36.
19. Bisht K, Tay J, Wellburn RN, McGirr C, Fleming W, Nowlan B, et al. Bacterial
lipopolysaccharides suppress erythroblastic islands and erythropoiesis in the bone marrow in
an extrinsic and G- CSF-, IL-1-, and TNF-independent manner. Front Immunol. 2020;
11:2548.
20. Jacobsen RN, Forristal CE, Raggatt LJ, Nowlan B, Barbier V, Kaur S, et al. Mobilization
with granulocyte colony-stimulating factor blocks medullar erythropoiesis by depleting
F4/80+VCAM1+CD169+ER-HR3+Ly6G+ erythroid island macrophages in the mouse. Exp
Hematol. 2014; 42:547-61.e4.
21. Tay J, Bisht K, McGirr C, Millard SM, Pettit AR, Winkler IG, et al. Imaging flow cytometry
reveals that granulocyte colony-stimulating factor treatment causes loss of erythroblastic
islands in the mouse bone marrow. Exp Hematol. 2020; 82:33-42.
22. Winn NC, Volk KM, Hasty AH. Regulation of tissue iron homeostasis: the macrophage
“ferrostat”. JCI Insight. 2020; 5:e132964.
23. Palis J. Primitive and definitive erythropoiesis in mammals. Front Physiol. 2014; 5:3.
24. Koury MJ. Tracking erythroid progenitor cells in times of need and times of plenty. Exp
Hematol. 2016; 44:653-63.
25. Chen K, Liu J, Heck S, Chasis JA, An X, Mohandas N. Resolving the distinct stages in
erythroid differentiation based on dynamic changes in membrane protein expression during
erythropoiesis. Proc Nat Acad Sci USA. 2009; 106:17413-8.
26. Berman I. The ultrastructure of erythroblastic islands and reticular cells in mouse bone
marrow. J Ultrastruct Res. 1967; 17:291-313.
27. Sadahira Y, Mori M, Kimoto T. Isolation and short-term culture of mouse splenic
erythroblastic islands. Cell Struct Funct. 1990; 15:59-65.
28. Mohandas N, Prenant M. Three-dimensional model of bone marrow. Blood. 1978; 51:633-43.
29. Jacobsen RN, Perkins AC, Levesque JP. Macrophages and regulation of erythropoiesis. Curr
Opin Hematol. 2015; 22:212-9.
30. Bessis M. L'îlot érythroblastique. Unité fonctionelle de la moelle osseuse (Erythroblastic
island. Functional unit of the bone marrow). Rev Hematol. 1958; 13:8-11.
31. Bessis MC, Breton-Gorius J. Iron metabolism in the bone marrow as seen by electron
microscopy: A critical review. Blood. 1962; 19:635-63.
32. Allen TD, Dexter TM. Ultrastructural aspects of erythropoietic differentiation in long-term
bone marrow culture. Differentiation. 1982; 21:86-94.
33. Seu KG, Papoin J, Fessler R, Hom J, Huang G, Mohandas N, et al. Unraveling macrophage
heterogeneity in erythroblastic islands. Front Immunol. 2017; 8:1140.
34. Miyagawa S-I, Kobayashi M, Konishi N, Sato T, Ueda K. Insulin and insulin-like growth
factor I support the proliferation of erythroid progenitor cells in bone marrow through the
sharing of receptors. Br J Haematol. 2000; 109:555-62.
35. Li W, Wang Y, Zhao H, Zhang H, Xu Y, Wang S, et al. Identification and transcriptome
analysis of erythroblastic island macrophages. Blood. 2019; 134:480-91.
36. Ramos P, Casu C, Gardenghi S, Breda L, Crielaard BJ, Guy E, et al. Macrophages support
pathological erythropoiesis in polycythemia vera and [beta]-thalassemia. Nat Med. 2013;
19:437-45.
37. Leimberg MJ, Prus E, Konijn AM, Fibach E. Macrophages function as a ferritin iron source
for cultured human erythroid precursors. J Cell Biochem. 2008; 103:1211-8.

20
38. Menon V, Ghaffari S. Erythroid enucleation: a gateway into a “bloody” world. Exp Hematol.
2021; 95:13-22.
39. Yeo JH, Lam YW, Fraser ST. Cellular dynamics of mammalian red blood cell production in
the erythroblastic island niche. Biophys Rev. 2019;
40. Toda S, Segawa K, Nagata S. MerTK-mediated engulfment of pyrenocytes by central
macrophages in erythroblastic islands. Blood. 2014; 123:3963-71.
41. Wei Q, Boulais PE, Zhang D, Pinho S, Tanaka M, Frenette PS. Maea expressed by
macrophages, but not erythroblasts, maintains postnatal murine bone marrow erythroblastic
islands. Blood. 2019; 133:1222-32.
42. Lee G, Lo A, Short SA, Mankelow TJ, Spring F, Parsons SF, et al. Targeted gene deletion
demonstrates that the cell adhesion molecule ICAM-4 is critical for erythroblastic island
formation. Blood. 2006; 108:2064-71.
43. Kawane K, Fukuyama H, Kondoh G, Takeda J, Ohsawa Y, Uchiyama Y, et al. Requirement
of DNase II for definitive erythropoiesis in the mouse fetal liver. Science. 2001; 292:1546-9.
44. Porcu S, Manchinu MF, Marongiu MF, Sogos V, Poddie D, Asunis I, et al. Klf1 affects
DNase II-alpha expression in the central macrophage of a fetal liver erythroblastic island: A
non-cell-autonomous role in definitive erythropoiesis. Mol Cell Biol. 2011; 31:4144-54.
45. Rojo R, Raper A, D. Ozdemir D, Lefevre L, Grabert K, Wollscheid-Lengeling E, et al.
Deletion of a Csf1r enhancer selectively impacts CSF1R expression and development of
tissue macrophage populations. Nat Comm. 2019; 10:3215.
46. Hoeffel G, Chen J, Lavin Y, Low D, Almeida Francisca F, See P, et al. C-Myb+ erythro-
myeloid progenitor-derived fetal monocytes give rise to adult tissue-resident macrophages.
Immunity. 2015; 42:665-78.
47. Keshvari S, Caruso M, Teakle N, Batoon L, Sehgal A, Patkar OL, et al. CSF1R-dependent
macrophages control postnatal somatic growth and organ maturation. PLOS Genet. 2021;
17:e1009605.
48. O'Neill ASG, van den Berg TK, Mullen GED. Sialoadhesin – a macrophage-restricted marker
of immunoregulation and inflammation. Immunology. 2013; 138:198-207.
49. Fujihara M, Muroi M, Tanamoto K-i, Suzuki T, Azuma H, Ikeda H. Molecular mechanisms
of macrophage activation and deactivation by lipopolysaccharide: roles of the receptor
complex. Pharmacol Ther. 2003; 100:171-94.
50. Haldar M, Kohyama M, So Alex Y-L, Kc W, Wu X, Briseño Carlos G, et al. Heme-mediated
SPI-C induction promotes monocyte differentiation into iron-recycling macrophages. Cell.
2014; 156:1223-34.
51. Kohyama M, Ise W, Edelson BT, Wilker PR, Hildner K, Mejia C, et al. Role for Spi-C in the
development of red pulp macrophages and splenic iron homeostasis. Nature. 2009; 457:318-
21.
52. Bellomo A, Mondor I, Spinelli L, Lagueyrie M, Stewart BJ, Brouilly N, et al. Reticular
fibroblasts expressing the transcription factor WT1 define a stromal niche that maintains and
replenishes splenic red pulp macrophages. Immunity. 2020; 53:127-42.e7.
53. Solomon EI, Decker A, Lehnert N. Non-heme iron enzymes: Contrasts to heme catalysis.
Proc Natl Acad Sci USA. 2003; 100:3589.
54. Rouault TA. Mammalian iron–sulphur proteins: novel insights into biogenesis and function.
Nat Rev Mol Cell Biol. 2015; 16:45-55.
55. Bekker A, Holland HD, Wang PL, Rumble D, Stein HJ, Hannah JL, et al. Dating the rise of
atmospheric oxygen. Nature. 2004; 427:117-20.
56. Schirrmeister BE, Gugger M, Donoghue PCJ. Cyanobacteria and the Great Oxidation Event:
evidence from genes and fossils. Palaeontology. 2015; 58:769-85.
57. Kehrer JP. The Haber–Weiss reaction and mechanisms of toxicity. Toxicology. 2000; 149:43-
50.

21
58. Eid R, Arab NTT, Greenwood MT. Iron mediated toxicity and programmed cell death: A
review and a re-examination of existing paradigms. Biochim Biophys Acta Mol Cell Res.
2017; 1864:399-430.
59. McKie AT, Marciani P, Rolfs A, Brennan K, Wehr K, Barrow D, et al. A novel duodenal
iron-regulated transporter, IREG1, implicated in the basolateral transfer of iron to the
circulation. Mol Cell. 2000; 5:299-309.
60. Donovan A, Brownlie A, Zhou Y, Shepard J, Pratt SJ, Moynihan J, et al. Positional cloning of
zebrafish ferroportin1 identifies a conserved vertebrate iron exporter. Nature. 2000; 403:776-
81.
61. Njajou OT, Vaessen N, Joosse M, Berghuis B, van Dongen JWF, Breuning MH, et al. A
mutation in SLC11A3 is associated with autosomal dominant hemochromatosis. Nat Genet.
2001; 28:213-4.
62. Winter WE, Bazydlo LAL, Harris NS. The molecular biology of human iron metabolism. Lab
Med. 2014; 45:92-102.
63. Vulpe CD, Kuo Y-M, Murphy TL, Cowley L, Askwith C, Libina N, et al. Hephaestin, a
ceruloplasmin homologue implicated in intestinal iron transport, is defective in the sla mouse.
Nat Genet. 1999; 21:195-9.
64. Chen H, Attieh ZK, Su T, Syed BA, Gao H, Alaeddine RM, et al. Hephaestin is a ferroxidase
that maintains partial activity in sex-linked anemia mice. Blood. 2004; 103:3933-9.
65. Roeser HP, Lee GR, Nacht S, Cartwright GE. The role of ceruloplasmin in iron metabolism. J
Clin Invest. 1970; 49:2408-17.
66. Vasilyev VB. Looking for a partner: ceruloplasmin in protein–protein interactions. BioMetals.
2019; 32:195-210.
67. Camaschella C, Nai A, Silvestri L. Iron metabolism and iron disorders revisited in the
hepcidin era. Haematologica. 2020; 105:260-72.
68. Fraser ST, Midwinter RG, Coupland LA, Kong S, Berger BS, Yeo JH, et al. Heme
oxygenase-1 deficiency alters erythroblastic island formation, steady-state erythropoiesis and
red blood cell lifespan in mice. Haematologica. 2015; 100:601-10.
69. Anderson ER, Shah YM. Iron homeostasis in the liver. Compr Physiol. 2013; 3:315-30.
70. Melino G, Stefanini S, Chiancone E, Antonini E. Stoichiometry of iron oxidation by
apoferritin. FEBS Lett. 1978; 86:136-8.
71. de Silva D, Guo JH, Aust SD. Relationship between iron and phosphate in mammalian
ferritins. Arch Biochem Biophys. 1993; 303:451-5.
72. Van Eden ME, Aust SD. The consequences of hydroxyl radical formation on the
stoichiometry and kinetics of ferrous iron oxidation by human apoferritin. Free Radic Biol
Med. 2001; 31:1007-17.
73. Bellelli R, Federico G, Matte’ A, Colecchia D, Iolascon A, Chiariello M, et al. NCOA4
deficiency impairs systemic iron homeostasis. Cell Rep. 2016; 14:411-21.
74. Santana-Codina N, Gableske S, del Rey MQ, Małachowska B, MJedrychowski MP, Biancur
DE, et al. NCOA4 maintains murine erythropoiesis via cell autonomous and non-autonomous
mechanisms. Haematologica. 2019; 104:1342-54.
75. Nai A, Lidonnici MR, Federico G, Pettinato M, Olivari V, Carrillo F, et al. NCOA4-mediated
ferritinophagy in macrophages is crucial to sustain erythropoiesis in mice. Haematologica.
2020; 106:795-805.
76. Ohgami RS, Campagna DR, Greer EL, Antiochos B, McDonald A, Chen J, et al.
Identification of a ferrireductase required for efficient transferrin-dependent iron uptake in
erythroid cells. Nat Genet. 2005; 37:1264-9.
77. Gray Lawrence W, Kidane Theodros Z, Nguyen A, Akagi S, Petrasek K, Chu Y-L, et al.
Copper proteins and ferroxidases in human plasma and that of wild-type and ceruloplasmin
knockout mice. Biochem J. 2009; 419:237-45.

22
78. Nemeth E, Ganz T. Anemia of inflammation. Hematol Oncol Clin North Am. 2014; 28:671-
81, vi.
79. Nemeth E, Rivera S, Gabayan V, Keller C, Taudorf S, Pedersen BK, et al. IL-6 mediates
hypoferremia of inflammation by inducing the synthesis of the iron regulatory hormone
hepcidin. J Clin Invest. 2004; 113:1271-6.
80. Billesbølle CB, Azumaya CM, Kretsch RC, Powers AS, Gonen S, Schneider S, et al.
Structure of hepcidin-bound ferroportin reveals iron homeostatic mechanisms. Nature. 2020;
586:807-11.
81. Jiang L, Wang J, Wang K, Wang H, Wu Q, Yang C, et al. RNF217 regulates iron homeostasis
through its E3 ubiquitin ligase activity by modulating ferroportin degradation. Blood. 2021;
138:689-705.
82. Marx JJM. Iron and infection: competition between host and microbes for a precious element.
Best Practice & Research Clinical Haematology. 2002; 15:411-26.
83. Roy CN, Mak HH, Akpan I, Losyev G, Zurakowski D, Andrews NC. Hepcidin antimicrobial
peptide transgenic mice exhibit features of the anemia of inflammation. Blood. 2007;
109:4038-44.
84. Zumerle S, Mathieu JRR, Delga S, Heinis M, Viatte L, Vaulont S, et al. Targeted disruption
of hepcidin in the liver recapitulates the hemochromatotic phenotype. Blood. 2014; 123:3646-
50.
85. Brissot P, Pietrangelo A, Adams PC, de Graaff B, McLaren CE, Loréal O.
Haemochromatosis. Nat Rev Dis Primers. 2018; 4:18016.
86. Barton JC, Edwards CQ, Acton RT. HFE gene: Structure, function, mutations, and associated
iron abnormalities. Gene. 2015; 574:179-92.
87. Vujić Spasić M, Kiss J, Herrmann T, Galy B, Martinache S, Stolte J, et al. Hfe acts in
hepatocytes to prevent hemochromatosis. Cell Metab. 2008; 7:173-8.
88. Babitt JL, Huang FW, Wrighting DM, Xia Y, Sidis Y, Samad TA, et al. Bone morphogenetic
protein signaling by hemojuvelin regulates hepcidin expression. Nat Genet. 2006; 38:531-9.
89. Anderson GJ, Bardou-Jacquet EJAoTM. Revisiting hemochromatosis: genetic vs . phenotypic
manifestations. Ann Transl Med. 2021; 9:731.
90. Ma S, Dubin AE, Zhang Y, Mousavi SAR, Wang Y, Coombs AM, et al. A role of PIEZO1 in
iron metabolism in mice and humans. Cell. 2021; 184:969-82.e13.
91. Gomes AC, Moreira AC, Silva T, Neves JV, Mesquita G, Almeida AA, et al. IFN-γ–
dependent reduction of erythrocyte life span leads to anemia during mycobacterial infection. J
Immunol. 2019; 203:2485-96.
92. Milner JD, Orekov T, Ward JM, Cheng L, Torres-Velez F, Junttila I, et al. Sustained IL-4
exposure leads to a novel pathway for hemophagocytosis, inflammation, and tissue
macrophage accumulation. Blood. 2010; 116:2476-83.
93. Millot S, Andrieu V, Letteron P, Lyoumi S, Hurtado-Nedelec M, Karim Z, et al.
Erythropoietin stimulates spleen BMP4-dependent stress erythropoiesis and partially corrects
anemia in a mouse model of generalized inflammation. Blood. 2010; 116:6072-81.
94. Weiss G, Ganz T, Goodnough LT. Anemia of inflammation. Blood. 2019; 133:40-50.
95. Gisbert JP, Gomollón F. Common misconceptions in the diagnosis and management of
anemia in inflammatory bowel disease. Am J Gastroenterol. 2008; 103:1299-307.
96. Jiang Y, Jiang F-Q, Kong F, An M-M, Jin B-B, Cao D, et al. Inflammatory anemia-associated
parameters are related to 28-day mortality in patients with sepsis admitted to the ICU: a
preliminary observational study. Ann Intensive Care. 2019; 9:67.
97. Gil-Santana L, Cruz LAB, Arriaga MB, Miranda PFC, Fukutani KF, Silveira-Mattos PS, et al.
Tuberculosis-associated anemia is linked to a distinct inflammatory profile that persists after
initiation of antitubercular therapy. Sci Rep. 2019; 9:1381.

23
98. Eriksson C, Henriksson I, Brus O, Zhulina Y, Nyhlin N, Tysk C, et al. Incidence, prevalence
and clinical outcome of anaemia in inflammatory bowel disease: a population-based cohort
study. Aliment Pharmacol Ther. 2018; 48:638-45.
99. Høivik ML, Reinisch W, Cvancarova M, Moum B, group tIs. Anaemia in inflammatory
bowel disease: a population-based 10-year follow-up. Aliment Pharmacol Ther. 2014; 39:69-
76.
100. Martí‐Carvajal AJ, Agreda‐Pérez LH, Solà I. Erythropoiesis‐stimulating agents for anemia in
rheumatoid arthritis. Cochrane Database Sys Rev. 2013; 2:CD000332.
101. Punnonen K, Irjala K, Rajamaki A. Serum transferrin receptor and its ratio to serum ferritin in
the diagnosis of iron deficiency. Blood. 1997; 89:1052-7.
102. Libregts SF, Gutiérrez L, de Bruin AM, Wensveen FM, Papadopoulos P, van Ijcken W, et al.
Chronic IFN-γ production in mice induces anemia by reducing erythrocyte life span and
inhibiting erythropoiesis through an IRF-1/PU.1 axis. Blood. 2011; 118:2578-88.
103. Gardenghi S, Renaud TM, Meloni A, Casu C, Crielaard BJ, Bystrom LM, et al. Distinct roles
for hepcidin and interleukin-6 in the recovery from anemia in mice injected with heat-killed
Brucella abortus. Blood. 2014; 123:1137-45.
104. Jacobsen RN, Nowlan B, Brunck ME, Barbier V, Winkler IG, Levesque J-P. Fms-like
tyrosine kinase 3 (Flt3) ligand depletes erythroid island macrophages and blocks medullar
erythropoiesis in the mouse. Exp Hematol. 2016; 44:207-12.e4.
105. Paulson RF, Hariharan S, Little JA. Stress erythropoiesis: definitions and models for its study.
Exp Hematol. 2020; 89:43-54.e2.
106. Hao S, Xiang J, Wu D-C, Fraser JW, Ruan B, Cai J, et al. Gdf15 regulates murine stress
erythroid progenitor proliferation and the development of the stress erythropoiesis niche.
Blood Adv. 2019; 3:2205-17.
107. Bennett LF, Liao C, Quickel MD, Yeoh BS, Vijay-Kumar M, Hankey-Giblin P, et al.
Inflammation induces stress erythropoiesis through heme-dependent activation of SPI-C. Sci
Signal. 2019; 12:eaap7336.
108. Liao C, Prabhu KS, Paulson RF. Monocyte-derived macrophages expand the murine stress
erythropoietic niche during the recovery from anemia. Blood. 2018; 132:2580-93.
109. Perry JM, Harandi OF, Porayette P, Hegde S, Kannan AK, Paulson RF. Maintenance of the
BMP4-dependent stress erythropoiesis pathway in the murine spleen requires hedgehog
signaling. Blood. 2009; 113:911-8.
110. Lenox LE, Perry JM, Paulson RF. BMP4 and Madh5 regulate the erythroid response to acute
anemia. Blood. 2005; 105:2741-8.
111. Lenox LE, Shi L, Hegde S, Paulson RF. Extramedullary erythropoiesis in the adult liver
requires BMP-4/Smad5–dependent signaling. Exp Hematol. 2009; 37:549-58.
112. Heideveld E, Hampton-O’Neil LA, Cross SJ, van Alphen FPJ, van den Biggelaar M, Toye
AM, et al. Glucocorticoids induce differentiation of monocytes towards macrophages that
share functional and phenotypical aspects with erythroblastic island macrophages.
Haematologica. 2018; 103:395-405.
113. Falchi M, Varricchio L, Martelli F, Masiello F, Federici G, Zingariello M, et al.
Dexamethasone targeted directly to macrophages induces macrophage niches that promote
erythroid expansion. Haematologica. 2015; 100:178-87.
114. Zhang J, Liu Y, Han X, Mei Y, Yang J, Zhang ZJ, et al. Rats provide a superior model of
human stress erythropoiesis. Exp Hematol. 2019; 78:21-34.e3.
115. Ulyanova T, Jiang Y, Padilla S, Nakamoto B, Papayannopoulou T. Combinatorial and distinct
roles of α5 and α4 integrins in stress erythropoiesis in mice. Blood. 2011; 117:975-85.
116. Ulyanova T, Georgolopoulos G, Papayannopoulou T. Reappraising the role of α5 integrin and
the microenvironmental support in stress erythropoiesis. Exp Hematol. 2020; 81:16-31.e4.

24
117. Millard SM, Heng O, Opperman KS, Sehgal A, Irvine KM, Kaur S, et al. Fragmentation of
macrophages during isolation confounds analysis of single cell preparations from mouse
hematopoietic tissues. bioRxiv. 2021; 2021.04.28.441876.
118. The Tabula Muris Consortium, Overall Coordination, Schaum N, Karkanias J, Neff NF, May
AP, et al. Single-cell transcriptomics of 20 mouse organs creates a Tabula Muris. Nature.
2018; 562:367-72.
119. Abreu R, Quinn F, Giri PK. Role of the hepcidin-ferroportin axis in pathogen-mediated
intracellular iron sequestration in human phagocytic cells. Blood Adv. 2018; 2:1089-100.
120. Lopez-Yrigoyen M, Yang C-T, Fidanza A, Cassetta L, Taylor AH, McCahill A, et al. Genetic
programming of macrophages generates an in vitro model for the human erythroid island
niche. Nat Comm. 2019; 10:881.
121. Silberstein L, Goncalves KA, Kharchenko PV, Turcotte R, Kfoury Y, Mercier F, et al.
Proximity-based differential single-cell analysis of the niche to identify stem/progenitor cell
regulators. Cell Stem Cell. 2016; 19:530-43.

25
Figure legends
Figure 1. Schematic of erythropoiesis hierarchy and examples of erythroblastic islands isolated
from the mouse bone marrow and visualised by imaging flow cytometry. (A) Schematic of
erythropoiesis hierarchy in the BM. (B) EBIs were prepared from the mouse BM as previously
described [19, 21, 33] and stained with fluorescent monoclonal antibodies specific for CD71,
VCAM-1, erythroid antigen Ter119, and macrophage antigen F4/80. Stain for each antigen is
shown in separate channels. Hoechst33342 stains cell nuclei. On the right is the merge of all the
different stains. Note that the central macrophages express F4/80, CD169 and VCAM-1 whereas the
rosetted Ter119+ CD71+ erythroblasts don’t.
Figure 2. Schematic of iron metabolism, and iron oxidation and reduction reactions in mammals.
The transit of alimentary iron through duodenal enterocytes to EBI macrophages, red pulp
macrophages, erythroblasts and hepatocytes via the blood are shown together with the main
receptors and enzymes involved in iron cation oxidation, reduction, transport and storage. Ferric
cations and oxidation steps are shown in red whereas ferrous cations and reduction reactions are
shown in green. Abbreviations for protein names are detailed in Table 1.
Figure 3. Schematic overview of macrophage dependent iron-homeostasis and erythropoiesis. (A)
Alimentary iron ions are adsorbed by enterocytes in the duodenum via divalent metal transporter-1
(DMT-1) and exported into the circulation (2 mg/day) through ferroportin (FPN). (B) Fe3+ ions bind
with high affinity to transferrin (TF) and are transported in the blood plasma throughout the body.
(C) Erythroblasts and EBI macrophages (EBI MΦ) in the BM import Fe3+ via TF receptors. Iron
sorted in Fe3+-ferritin complexes in EBI macrophages can be released to be then exported to
maturing erythroblasts in erythroblastic islands (EBI) to synthetise haemoglobin (Hb) under (D) the
effect of erythropoietin (EPO) produced by the kidneys to promote proliferation and survival or
erythroid progenitors in the BM. (E) Following enucleation of erythroblasts in erythroblastic
islands, erythrocytes are released in the circulation to transport O2 and CO2. (F) Old erythrocytes
are captured while passing through the spleen and phagocytosed by red pulp macrophages (RPMΦ)
that extract Fe2+ ions from haemoglobin and re-export in the circulation via ferroportin (iron
recycling 20 mg/day). (G) During infections, PAMPs such as LPS block EBI macrophage function
causing the dissociation of EBIs and arrest of erythropoiesis in the BM. (H) IL-6 produced in
response to infections, activate HAMP expression by hepatocytes. HAMP binds to ferroportin
causing its internalisation and blocking iron export into the circulation. (I) Low levels of HAMP

26
cause excessive iron transfer from duodenum resulting in excessive iron load in the blood, storage
of excess iron in hepatocytes expressing ferritin and excessive erythropoiesis leading to
haemochromatosis.
Figure 4. Expression of Csf3r mRNA encoding the G-CSF receptor in myeloid cells and erythroid
progenitors in the mouse BM. Single cell suspension from C57BL/6 mouse BM in steady-state were
prepared as previously described [19]. (A) BM cells were stained with monoclonal antibodies for
Ter119, CD11b, F4/80, Ly6G, Ly6C VCAM-1 and CD169 and Ter119- CD11b+ F4/80+ Ly6G-
CD169+ VCAM-1+ macrophages, Ter119- CD11b+ F4/80+ Ly6G- CD169- VCAM- Ly6Cbright
monocytes and Ter119- CD11b+ F4/80 Ly6G+ neutrophils and RNA extracted. (B) Erythroid
progenitors were sorted accorded to the following phenotypes: megakaryocyte-erythroid
progenitors (MEPs: Lin- Sca1- KIT+ CD16/32- CD41- CD150+ CD105-), pre-megakaryocytes (Pre-
Meg: Lin- Sca1- KIT+ CD41+), BFU-E (Lin- Sca1- KIT+ CD16/32- CD41- CD150+ CD105+) and
CFU-E (Lin- Sca1- KIT+ CD16/32- CD41- CD150- CD105+) as previously described [19].
Erythroblasts were sorted according to expression of Ter119, CD44 and CD45 and forward scatter
as previously described [19]. RNA was extracted, reverse transcribed and Csfr3 and Hprt mRNA
were quantified by qRT-PCR using primer-probe sets Mm00432735_m1 and Mm03024075_m1
respectively.
Figure 5. Imaging flow cytometry of murine bone marrow singlets demonstrates that abundant
Ter119+F4/80+ events are consistent with F4/80+ macrophage remnants adhered to intact Ter119+
erythroblasts. Single cell BM suspensions were flushed from the femurs of adult
Siglec1Cre;R26ZsGreen reporter mice as previously described [117] and stained with erythroid antigen
Ter119 and macrophage antigen F4/80. A. Ter119+ events were segregated into large and small
events. B. Representative F4/80 and ZsGreen staining profile for large Ter119+ events. C.
Representative images of large Ter119+F4/80+ events with each channel displayed independently,
followed by two merge images; #1 Ter119 and F4/80, #2 Brightfield (BF), F4/80 and ZsGreen.
Violin plots display the stain area distribution for Ter119 and F4/80 stains for the large
Ter119+F4/80+ population (n=481). D. Representative F4/80 and ZsGreen staining profile for small
Ter119+ events. E. Representative images of small Ter119+ F4/80+ events and violin plots
displaying stain area distribution for this population (n=297).

27
Table 1. Proteins regulating iron homeostasis
Protein names
Gene names
Roles
Ceruloplasmin
CP
Oxidises Fe2+ into Fe3+ in plasma
Cytochrome b reductase 1
CYBRD1
Reduces Fe3+ into Fe2+ at the luminal surface of
duodenal enterocytes
Divalent metal transporter-1
(DMT-1)
SLC11A2
Imports Fe2+ cations into enterocytes and
macrophages
Ferritin
FTL (light chain)
FTH1 (heavy chain)
Oxidises intracellular Fe2+ to Fe3+ to store in the 24-
mer protein in macrophages and hepatocytes
Ferroportin (FPN)
SLC40A1 previously
SLC11A3
Exports Fe2+ cations from enterocytes, macrophages
and hepatocytes
Heme carrier protein-1 (HCP-1)
SLC46A1
Transports intact haem molecule across enterocyte
and hepatocyte plasma membrane
Hemojuvelin
HJV
Co-receptor of BMP receptors on hepatocytes.
Necessary to induce HAMP expression in response to
BMP-6
Hepheastin
HEPH
Oxidises Fe2+ into Fe3+ at the basal surface of
enterocytes
Hepcidin antimicrobial peptide
HAMP
Binds to FPN causing its internalisation and blocking
its Fe2+ transporter function
High ferric protein (HFE)
HFE
Complexed with TFRC, necessary to enhance HAMP
transcription in response to high iron plasma
concentration in hepatocytes
Interleukin-6
IL6
Activates HAMP transcription in hepatocytes
Nuclear receptor co-activator-4
(NCOA4)
NCOA4
Cargoes ferritin to autophagosome for degradation
and Fe3+ release from intracellular ferritin storage in
macrophages and hepatocytes
Piezo-type mechanosensitive ion
channel component 1 (PIEZO1)
PIEZO1
PIEZO1 expressed in EBI and red pulp macrophages
regulates erythropoiesis, erythrophagocytosis, and
iron recycling and storage
RNF217
RNF217
Poly-ubiquinates FPN to induce its degradation
Transferrin (TF)
TF
Transports Fe3+ in the plasma
Transferrin receptor (TFRC /
CD71)
TFRC
Binds to and import Fe3+-TF complexes and releases
Fe3+ in endosomes upon acidification
Transferrin receptor 2 (TFR2)
TFR2
Complexed with TFRC and HJV at the surface of
hepatocytes, necessary to enhance HAMP
transcription in response to high iron plasma
concentration
6-transmembrane epithelial
antigen of the prostaste-3
(STEAP-3)
STEAP3
Reduces Fe3+ into Fe2+ within the endosomes of
macrophages, erythroblasts and hepatocytes

28
Figure 1

29

30
Figure 3

31
Figure 4

32
Figure 5