RESEARCH ARTICLE
A JH
FOXO3-mTOR metabolic cooperation in the regulation of
erythroid cell maturation and homeostasis
Xin Zhang,1 Genıs Camprecios,1 Pauline Rimmele,1 Raymond Liang,1,2 Safak Yalcin,1 Sathish Kumar Mungamuri,1
Jeffrey Barminko,1,3 Valentina D’Escamard,1 Margaret H. Baron,1,2,3,4,5,6 Carlo Brugnara,7 Dmitri Papatsenko,1,8
Stefano Rivella,9 and Saghi Ghaffari1,2,3,5,8*
Ineffective erythropoiesis is observed in many erythroid disorders including b-thalassemia and anemia of
chronic disease in which increased production of erythroblasts that fail to mature exacerbate the underlying
anemias. As loss of the transcription factor FOXO3 results in erythroblast abnormalities similar to the ones
observed in ineffective erythropoiesis, we investigated the underlying mechanisms of the defective Foxo32/2
erythroblast cell cycle and maturation. Here we show that loss of Foxo3 results in overactivation of the
JAK2/AKT/mTOR signaling pathway in primary bone marrow erythroblasts partly mediated by redox
modulation. We further show that hyperactivation of mTOR signaling interferes with cell cycle progression in
Foxo3 mutant erythroblasts. Importantly, inhibition of mTOR signaling, in vivo or in vitro enhances
significantly Foxo3 mutant erythroid cell maturation. Similarly, in vivo inhibition of mTOR remarkably improves
erythroid cell maturation and anemia in a model of b-thalassemia. Finally we show that FOXO3 and mTOR
are likely part of a larger metabolic network in erythroblasts as together they control the expression of an
array of metabolic genes some of which are implicated in erythroid disorders. These combined findings
indicate that a metabolism-mediated regulatory network centered by FOXO3 and mTOR control the
balanced production and maturation of erythroid cells. They also highlight physiological interactions
between these proteins in regulating erythroblast energy. Our results indicate that alteration in the function
of this network might be implicated in the pathogenesis of ineffective erythropoiesis.
C 2014 Wiley Periodicals, Inc.
Am. J. Hematol. 89:954–963, 2014. V
䊏 Introduction
Erythroid cell maturation requires an exquisite coordination of cell proliferation and differentiation whose imbalance contributes to the pathogenesis of erythroid disorders. This coordination is compromised in many diseases of erythroid cells characterized by anemia and ineffective erythropoiesis including b-thalassemia, malaria, and anemia of chronic disease [1,2]. The importance of redox imbalance in altered erythropoiesis is
evident as, depending on its degree of severity, results in subtle to highly defective RBC production [3–7]. The redox state is tightly coupled to cellular metabolism that in mature RBC is strictly limited to glycolysis [8,9]. Although RBC glycolysis has been extensively studied, less is known
about the potential function of metabolic (glycolytic) pathways during erythroblast maturation [8–11].
The redox balance is sustained during erythroid cell maturation, at least in part, by FOXO3 transcriptional regulation of several anti-oxidant
enzymes [5,12,13]. FOXO3 belongs to the FOXO Forkhead family of winged helix transcription factors. In addition to their antioxidant response
FOXO factors exert many fundamental biological functions including the regulation of cell cycle, apoptosis, DNA repair, and metabolism [14,15]. The
transcriptional activity of FOXOs is negatively regulated by growth factor and cytokine receptor signaling via several protein kinases including AKT
[15]. Conversely, FOXO factors are phosphorylated on distinct residues and activated in response to stress stimuli via a distinct set of protein kinases.
Additional Supporting Information may be found in the online version of this article.
1
Department of Developmental and Regenerative Biology, Icahn School of Medicine at Mount Sinai, New York, New York, 10029; 2Developmental and Stem Cell Biology Multidisciplinary Training Area, Icahn School of Medicine at Mount Sinai, New York, New York, 10029; 3Division of Hematology and Medical Oncology, Department of Medicine, Icahn School of Medicine at Mount Sinai, New York, New York, 10029; 4Departments of Pediatrics Hematology-Oncology and Cell and
Developmental Biology, Weill Cornell Medical College, New York, New York, 10021; 5Tisch Cancer Institute, Icahn School of Medicine at Mount Sinai, New York, New
York, 10029; 6Department of Oncological Sciences, Icahn School of Medicine at Mount Sinai, New York, New York, 10029; 7Department of Lab Medicine, Children’s
Hospital, Boston, Massachusetts, 02115; 8Black Family Stem Cell Institute, Icahn School of Medicine at Mount Sinai, New York, New York, 10029; 9Departments of
Pediatrics Hematology-Oncology and Cell and Developmental Biology, Weill Cornell Medical College, New York, New York, 10021
Conflict of interest: Nothing to report.
X. Z. and G. C. contributed equally to this work.
Xin Zhang is currently at the First Affiliated Hospital of Shantou University Medical College, Shantou, Guangdong, 515041, China
*Correspondence to: Saghi Ghaffari; Department of Developmental and Regenerative Biology, Icahn School of Medicine at Mount Sinai, New York, NY
10029. E-mail: [email protected]
Contract grant sponsor: Spanish Ministry of Education (to G.C.); Contract grant number: RO1 HL094283 (to J.B.).
Contract grant sponsor: National Institutes of Health; Contract grant numbers: RO1 DK077174 to SG; RO1 HL116365 (SG, Co-PI).
Contract grant sponsors: a Roche foundation award, a Black Family Stem Cell Institute award, a Myeloproliferative Neoplasm (MPN) Foundation award (to
S.G.).
Received for publication: 7 June 2014; Accepted: 11 June 2014
Am. J. Hematol. 89:954–963, 2014.
Published online: 25 June 2014 in Wiley Online Library (wileyonlinelibrary.com).
DOI: 10.1002/ajh.23786
C 2014 Wiley Periodicals, Inc.
V
954
American Journal of Hematology, Vol. 89, No. 10, October 2014
doi:10.1002/ajh.23786
RESEARCH ARTICLE
In addition to phosphorylation, FOXO are regulated by a number of
post-translational modifications. FOXO factors in mammals are composed of four highly related members (FOXO1, FOXO3, FOXO4, and
FOXO6). Among these FOXO3 is the major active FOXO during erythroid cell maturation [5]. In response to erythropoietin (Epo), FOXO3
is phosphorylated by AKT protein kinase [16–20] that by promoting its
cytosolic localization represses FOXO3’s transcriptional activity.
FOXO3’s expression, nuclear localization and transcriptional activity
increases with erythroblast maturation [5]. At the steady state, Foxo3 is
required for erythroid cell formation [5] as loss of Foxo3 results in
impaired antioxidant response, cell cycle alterations associated with
delayed maturation of erythroblast precursors, as well as oxidative
stress-mediated reduction of RBC lifespan [5]. These abnormalities lead
to decreased RBC production. These combined abnormalities are highly
reminiscent of ineffective erythropoiesis in which FOXO3 may be a
participant [21,22]. Nonetheless, the precise mechanism of cell cycle
and maturation defects of Foxo3 mutant erythroblasts remains unclear.
While the phenotype of Foxo3-deficient erythroid cells is relatively
mild, these mice succumb to sudden death when exposed to exogenous
oxidative challenge likely due to severe anemia [5], suggesting that
Foxo3 has a key function in stress erythropoiesis. Recent work in our
laboratory and others’ indicate that in addition to the transcriptional
control of antioxidant enzymes, Foxo3 is implicated in an array of metabolic functions raising the possibility that Foxo3’s control of antioxidant response may be part of a broader metabolic program [23–29]
(and Camprecios and Ghaffari, manuscript in preparation).
AKT provides an important signal for erythroid cell generation and
maturation [16–18,30–33]. In addition to Foxo3, AKT regulates mammalian target of rapamycin (mTOR) kinase. mTOR signaling is one of
the major regulators of cellular metabolism. mTOR is key to cellular
growth (size) and proliferation, is highly sensitive to oxygen and
nutrients including amino acids and glucose and has a central function
in protein synthesis [34,35]. mTOR protein kinase exists in two distinct
core complexes, mTOR complexes I and II (mTORC1 and mTORC2,
respectively) which differ in their regulation and functions as well as in
their sensitivity to rapamycin. While the physiological function of
mTOR signaling during erythroid cell maturation remains unknown
[36,37], conflicting results as to whether use of the mTOR inhibitors in
patients [35] is associated with anemias have been reported [38,39].
Here we investigated mechanisms underlying alterations of Foxo3deficient erythroid cell cycling. Strikingly, we found that loss of Foxo3
results in overactivation of the JAK2/AKT/mTOR signaling pathway in
erythroblasts partly mediated by redox modulation. Activation of
mTOR leads to alterations of cycling and differentiation of immature
erythroblasts suggesting that activation of a feedback loop upstream of
FOXO3 compromises erythroid cell maturation. We further show using
in vitro and in vivo approaches that inhibition of mTOR signaling partially alleviates the abnormal maturation of Foxo3-deficient erythroblasts leading to increased red blood cells (RBC) in the peripheral
blood. Notably, we show that FOXO3 and mTOR together may be
part of a metabolic network during erythroid cell maturation. These
findings suggest that hyperactivation of mTOR signaling resulting from
loss of FOXO3 function contributes to the blockade of Foxo32/2
erythroblast maturation. In addition, they provide a platform to further
delve into redox and metabolic regulation of erythropoiesis.
FOXO3/mTOR in Erythroid Maturation
supplemented with 2 U/ml Epo, IL-6 (10 ng/ml) and SCF (100 ng/ml) (modified
from [32,33,42]). Cells were then starved in vitro for 2 hrs in IMDM supplemented
with 0.1% FCS and further stimulated with Epo (10 U/ml). In some experiments, cells
were differentiated in the presence of 100 mM NAC. Fetal liver cultures were performed using a modified protocol of [43]. Briefly, lineage negative cells were isolated
from E14.5 fetal livers and plated at <2 3 106 cells/ml with erythroid expansion
medium consisting of Stem Span SFEM (StemCell Technologies) supplemented with
2 U/ml human recombinant Epo (Amgen), 100 ng/ml SCF (PreproTech), 40 ng/ml
insulin-like growth factor-1 (PreproTech), 1026 M dexamethasone (D2915; Sigma),
0.4% cholesterol mix (Gibco), and 1% penicillin/streptomycin (Gibco). After 48 hr
cells were washed with PBS and plated at a concentration <2 3 106 with either ramapycin (20 nM; Enzo Life Sciences) or vehicle control with erythroid differentiation
medium consisting of IMDM supplemented with 2 U/ml Epo, 100 ng/ml SCF, 10%
Serum replacement (Invitrogen), 5% platelet-derived serum, glutamine and 10%
protein-free hybridoma media. After another 24 hrs, cells were collected and erythroid maturation analyzed by flow cytometry.
Retroviral production and transduction of cells. Retroviral constructs and supernatant production were performed as previously described [32,33].
Colony-forming assays. For BFU-E and CFU-E analyses, 1 3 104 and 3 3 103 total
bone marrow cells were plated respectively in triplicates as previously described [32].
Flow cytometry. Bone marrow and fetal liver single cell suspensions were prepared and maintained in IMDM 1 15% FBS, washed twice, preincubated with 10%
rat serum and stained with CD71-FITC, CD44-APC, and TER119-PE or -FITC
antibodies (BD Biosciences). Gating to distinguish erythroid populations according
to their stage of maturation was performed as in [44]. Freshly isolated bone marrow cells stained with CD44-APC and TER119-FITC, were fixed with fix/permeabilization buffer (BD Biosciences) and incubated with 1:100 dilution of antipSer473 AKT and pSer235/236 S6 antibodies (Cell Signaling Technology, Cat #9271
and #4858, respectively) followed by incubation with 1:1000 dilution of PEconjugated secondary antibody (BD Biosciences) to measure intracellular AKT and
S6 phosphorylation. Samples were washed and protein phosphorylation was analyzed by flow cytometry. Data was analyzed by FlowJo software (Treestar).
Cell proliferation assay. Mice were injected intraperitoneally with 1 mg of BrdU.
After 1 hr, bone marrow cells were isolated and stained with CD44-APC and
TER119-PE antibodies (BD-Pharmingen, CA), then fixed and stained with antiBrdU-FITC antibody (BD Biosciences) and 7-AAD for flow cytometric analysis of
cell proliferation following the manufacturer’s protocol. Similar results were
obtained when BrdU was injected 30 min before harvesting cells.
N-Acetyl-L-cysteine (NAC) treatment. Mice were injected intraperitoneally with
100 mg/kg body weight of N-acetyl-L-cysteine (NAC; Sigma, MO) in phosphate
buffered saline solution (pH7.4) daily for 2 weeks.
Western blot analysis. Cells were starved in 0.1% serum for 2 hrs and then stimulated with Epo (10 U/ml). Lysates were prepared in 13 RIPA lysis buffer (20 mM
Sodium phosphate, 300 mM sodium chloride, 4 mM EDTA) containing 2% sodium
deoxycholate, 2% NP-40, 0.2% SDS, 400 lM sodium orthovanadate, 0.2% bmercaptoethanol, 2 mM PMSF, and 100 mM sodium fluoride. The buffer was also
mixed with protease cocktail inhibitors (Roche; Cat No: 11-697-498-001). The total
protein was estimated using Bio-Rad Bradford’s Reagent (Cat #500-0006) following
manufacturer’s instructions. Retrovirally transduced GFP-positive NIH3T3 cells were
FACS sorted and cell lysates were prepared in Laemmli sample buffer (Bio-Rad),
resolved by SDS polyacrylamide gel electrophoresis (PAGE) and transferred on to
PVDF membranes. Given the size and to enhance resolution, mTOR protein was run
separately. The following primary antibodies were used for western blotting: antipSer473 Akt (#4051), anti-Akt (#9272), anti-pThr389 p70 S6 Kinase (#9205), antip70 S6 kinase (#9202), anti-pSer2448 mTOR (#2971), anti-mTOR (#2972), antipTyr1007/1008 Jak2 (#3776), anti-TSC1 (# 6935), anti-p4EBP1 (Thr37/46) (# 2855),
4EBP1 (# 9644), from Cell Signaling Technology, and antiglutamine synthase BD (#
610517); all used at 1:1000 dilutions. Anti-Tubulin: Santa Cruz Biotechnology (#sc8035) anti-Actin (#sc-1616) from Santa Cruz. Anti-JAK2 (#06–255; Upstate Biotechnology; 1:500). Horseradish peroxidase (HRP)-conjugated secondary antibodies were
used at 1:5000 (Santa Cruz).
Rapamycin treatment. Mice received intraperitoneal administration of 4 mg/kg
body weight of rapamycin (Enzo Life Sciences, NY) in PBS 1 5% Tween 80 1 5%
PEG400 1 4% Ethanol during five consecutive days/week for 2 weeks.
Hematological studies. Blood samples were obtained from the cava vein right
after sacrificing the mice and collected in EDTA or Heparin. Complete blood
counts (CBC) were measured with an Advia 120 analyzer.
Plasma Epo measurement. Plasma Epo concentrations were determined by the
QuantikineV ELISA kit for mouse erythropoietin from R&D Biosystems (Minneapolis, MN) according to the manufacturer’s instructions.
RNA isolation and QRT–PCR. Was performed as previously described [41,45,46].
Fluidigm—96.96 dynamic array IFC. For fluidigm dynamic array performance,
specific target amplification (STA) was performed according to the manufacturerÇs
protocol (PN 100-3488 B1). Briefly, cDNA was pre-amplified using the TaqManV
PreAmp Master Mix (Applied Biosystems) for the 96 genes of interest. The amplification parameters were as follows: 95 C for 100 , followed by 12 cycles at 95 C for
1500 and 60 C for 40 . After STA, we performed exonuclease I treatment as recommended by the manufacturer. Briefly, Exonuclease I and Exonuclease I buffer (New
R
䊏 Methods
Mice. Foxo31/2 mice (129xFBV/n) [40] were backcrossed >10 generations onto
C57Bl6 [41] and 10–12 week old C57Bl6 mice were used in all experiments. Protocols were approved by the Institutional Animal Care and Use Committee of Mount
Sinai School of Medicine.
Cells. Bone marrow lineage negative cells were separated from mature cells using
the EasySepTM mouse hematopoietic progenitor enrichment kit (StemCell Technologies) and differentiated on fibronectin-coated plates for 18 h with IMDM 1 15% FBS
doi:10.1002/ajh.23786
R
American Journal of Hematology, Vol. 89, No. 10, October 2014
955
Zhang et al.
RESEARCH ARTICLE
England Biolabs) were added to the STA samples, and samples were then incubated
for 300 at 37 C, followed by the enzyme inactivation at 80 C for 150 . Finally, to
load the dynamic array IFC, samples were prepared with the SsoFast EvaGreen
Supermix with Low ROX (Bio-Rad) and 203 DNA Binding Dye Sample Loading
Reagent (Fluidigm). On the other hand, primers were diluted with Assay Loading
Reagent (Fluidigm) and DNA Suspension Buffer (Teknova). After priming the
96 3 96 chip in the IFC Controller MX, samples and primers were loaded into their
respective inlets. The chip was then loaded by the IFC Controller MX (BioMarkTM
HD System). The chip was run following the GE 96 3 96 PCR 1 Melt v2.pcl protocol in the Biomark using the Data Collection Software (Fluidigm). Results were
obtained with the Fluidigm Real-Time PCR Analysis software (Fluidigm) and further analyzed by the 22DDCt method. b actin was used as a loading control. Results
shown as fold-change relative to Gate I wild type controls. Primer specific sequences are listed in Supporting Information Table II.
Statistical analysis. Fluidigm data was normalized using standard 22DDCt method
using actin readings as an internal standard in each series of PCR reactions. In experiments other than Fluidigm analysis of gene expression, the unpaired one- and twotailed Student’s t test was used. A P < 0.05 was considered to be significant.
䊏 Results
FOXO3-mTOR control erythroblast cell cycling
To examine mechanisms underlying abnormalities of Foxo3 mutant
erythroblast cycling and maturation we analyzed Epo-activated signaling pathways in primary mouse Foxo32/2 erythroid precursors. Primary lineage negative (lacking mature erythroid cells) wild type (WT)
and Foxo32/2 bone marrow cells cultured under an erythroid differentiation condition for 18 hrs (18 h) [33,42] produced 60% TER1191
cells (Supporting Information Fig. 1). These cells were serum starved
for 2 h and stimulated with Epo before lysates were analyzed by
immunoblotting (Fig. 1). As anticipated JAK2 protein tyrosine kinase
that is necessary for Epo receptor (EpoR) signaling and its downstream
effector AKT, were rapidly phosphorylated in response to Epo [16,18]
(Fig. 1). In addition, mTOR protein kinase and its downstream target
ribosomal S6 protein kinase 1 (S6K1) were phosphorylated in immature mouse erythroblasts. Epo stimulation of primary Foxo3-null erythroblast precursors also resulted in increased JAK2 phosphorylation as
in wild-type erythroblasts (Fig. 1). This was associated with enhanced
phosphorylation of signaling proteins AKT, mTOR, and mTOR target
S6K1 in Foxo3-null erythroblast precursors at levels markedly higher
than the control. Up to 1 hr after Epo stimulation the AKT/mTOR/
S6K1 remained highly phosphorylated in Foxo3-deficient erythroid precursors with a distinct kinetic from that observed in wild type cells. To
further assess whether mTOR signaling was activated in Foxo3 mutant
erythroblasts, we compared the phosphorylation of another downstream target of mTOR, the eukaryotic initiation factor 4E (eIF4E)binding protein 1 (4EBP1). mTORC1 phosphorylation and inhibition
of 4EBP1 has a more direct impact on mRNA translation [34]. Phosphorylation of 4EBP1 in response to Epo stimulation was increased in
primary freshly isolated Foxo3 mutant erythroblasts (TER 1191) (Fig.
1B, top panel) serum starved for 2 hrs as compared to control cells
(quantification, Fig. 1B, bottom panel). Detection of pAKT, pS6K1 and
p4EBP1 but not pJAK2 at time 0 even in the absence of Epo stimulation in Foxo3 mutant erythroblasts (Fig. 1) might indicate constitutive
activation of mTOR signaling independent of JAK2. These results suggested that loss of FOXO3 may alter the signaling response to Epo in
erythroblasts leading to prolonged activation of AKT/mTOR signaling
(Fig. 1, see schematic on top).
As FOXO3 is a key regulator of oxidative stress, and ROS modulate
protein phosphorylation [14,47] we evaluated whether ROS are implicated in the alteration of EpoR signaling in Foxo3 mutant erythroblasts.
In vitro treatment with ROS scavengers N-acetyl cysteine (NAC) for 2
h had a noticeable effect on reducing phosphorylation of JAK2, AKT,
mTOR, and S6K protein in primary Foxo3 mutant bone marrow erythroblasts (Fig. 2, compare lanes 8, 9 to 11, 12). These results suggest that
ROS contribute to the enhanced phosphorylation of these proteins in
Foxo3 mutant erythroblasts in response to Epo. NAC treatment also
956
American Journal of Hematology, Vol. 89, No. 10, October 2014
Figure 1. Jak2-AKT–mTOR signaling pathway is overactivated in cultured
Foxo32/2 bone marrow erythroid cells. (A) Schematic of EpoR-mediated
activation of JAK2-AKT-mTOR on the left. Western blot analysis of phosphorylation of signaling proteins. WT and Foxo32/2 lineage-negative bone
marrow cells were isolated and cultured under erythroid condition for 18
hrs, serum- and cytokine starved for 2 hrs and stimulated with Epo (10 U/
ml) for the indicated time points in vitro before preparing the whole cell
extract. One representative of two experiments is shown. (B) Schematic of
activation of feedback loop JAK2-AKT-mTOR in the absence of FOXO3 on
the left. TER 1191 WT and Foxo32/2 erythroblasts were freshly isolated
from mice serum- and cytokine starved for 2 hrs and stimulated with Epo
(10 U/ml) for the indicated time points in vitro before preparing the whole
cell extract to analyze of p4EBP1 (quantification of bands in the bottom
panel). [Color figure can be viewed in the online issue, which is available
at wileyonlinelibrary.com.]
reduced the levels of JAK2/AKT/mTOR/S6K1 phosphorylation in wildtype primary erythroid cells (Fig. 2, compare lanes 2, 3 to 5, 6).
Although under these conditions, NAC reduced notably phosphorylated AKT, mTOR and S6K in both wild type and Foxo3 mutant erythroblasts, the effect of NAC on pJAK2 was less pronounced (Fig. 2). The
phosphorylated form of ribosomal protein S6 (pS6) is a target of S6K1
and a reliable indicator of mTORC1 activity [35]. In agreement with
ROS effects on mTORC1 activity, in vivo treatment with NAC reduced
the frequency of pS6 expressing cells and the levels of pS6 (Supporting
Information Fig. 2 and data not shown). These results are consistent
with the notion that redox state modulates mTOR signaling [48,49],
and suggest that increased ROS mediate at least partially overactivation
of mTOR signaling in immature Foxo32/2 erythroblasts in vitro (Fig. 1,
Supporting Information Fig. 2).
We next investigated the function of mTOR signaling in normal
erythropoiesis. We used RNA interference to inhibit the expression of
S6K1, a direct mTORC1 target [34,35,50] in primary erythroid progenitors (Fig. 3A). Inhibition of S6K1 in bone marrow cultures using
two distinct shRNA sequences resulted in significant reduction of
BFU-E and CFU-E-derived erythroid cell colony formation (Fig. 3B).
The degree of inhibition of erythroid progenitor cell-colony formation
was consistent with the relative expression of S6K1 in response to
RNA interference targeting in BFU-Es but not in CFU-Es (Fig. 3B)
suggesting that while mTORC1 activation is required for both BFUES and CFU-Es, BFU-Es more than CFU-Es are highly sensitive to
doi:10.1002/ajh.23786
RESEARCH ARTICLE
levels of mTOR signaling. These results may reflect distinct sensitivity
of these progenitors to EpoR signaling [51]. These combined results
suggest that mTORC1 signaling is required for in vitro generation of
erythroid progenitor cell-derived colonies.
As mTOR signaling is central to cell growth and proliferation, we
asked whether inhibition of mTOR signaling by rapamycin that is a
specific inhibitor of mTORC1 [35] has any impact on erythroblast
precursor cell cycling. To address this, we used the thymidine analog
5-bromo-2-deoxyuridine (BrdU) that is incorporated in dividing cells
in vivo. Mice were treated in vivo with rapamycin for 2 weeks and
injected with BrdU 30–60 min before harvesting the bone marrow.
Erythroblast cell cycle distribution was analyzed at distinct stages of
Figure 2. Overactivation of Jak2-AKT–mTOR signaling pathway in cultured
Foxo32/2 bone marrow erythroid cells is partly mediated by ROS. Western
blot analysis of phosphorylation of signaling proteins. WT and Foxo32/2
bone marrow lineage-negative cells were isolated and cultured under erythroid condition for 18 hrs in the presence or absence of NAC (100 lM),
serum- and cytokine starved for 2 hrs and stimulated with Epo (10 U/ml)
for the indicated time points in vitro before preparing the whole cell
extract.
FOXO3/mTOR in Erythroid Maturation
maturation by flow cytometry examination of BrdU incorporation
and the DNA marker 7-aminoactinomycin D (7-AAD). Maturing
erythroblasts (proerythroblasts, basophilic erythroblasts, and polychromatophilic erythroblasts) expressing similar levels of transferrin
receptor CD71 [44] (and data not shown) were distinguished according to their size (forward scatter, FSC) and surface expression of
TER119 and CD44 [44] (Fig. 4A, FACS plot). The analysis yielded
important insights into primary bone marrow erythroblast cycling. A
significant fraction (up to 75%) of wild type bone marrow erythroblasts (Gates I to III proerythroblasts, basophilic, polychromatophilic)
were in the S phase (Fig. 4B). Rapamycin treatment strongly blocked
erythroblast cell cycle progression at the G1/S transition phase in
immature erythroblasts (Gates I and II, proerythroblasts and basophilic erythroblasts respectively) and reduced the fraction of proerythroblasts (Gate I) in the S phase. In addition, rapamycin treatment
strongly reduced the fraction of basophilic erythroblasts (Gate II) that
were in the G2/M transition without significant effects on cells at
later stages of maturation (Gates III and IV). In contrast to its effect
on proerythroblasts, rapamycin treatment resulted in an increase,
rather than a decrease in the fraction of basophilic erythroblasts
(Gate II) that were in the S phase (Fig. 4B). These results indicate
that mTOR signaling is required for normal cell cycle progression of
immature (Gates I and II) erythroblasts.
Immature Foxo3 mutant erythroblasts exhibit cell cycle defects associated with a failure to fully mature [5]. As a consequence, the rate of
Foxo3 mutant erythroblast maturation is decreased. We evaluated
whether the overactivation of mTOR signaling in Foxo3 mutant erythroblasts contributes to their cell cycle alteration and defective maturation. BrdU analysis of Foxo3 mutant erythroblasts confirmed Foxo3
mutant erythroblast cell cycle alterations [5] (Fig. 4A,B). Relative to
controls, a reduced fraction of early Foxo3 mutant erythroblasts (Gates
I) was in the S phase of cell cycle as previously observed [5]. In addition
to a slight increase in the fraction of immature Foxo3 mutant precursors (Gates I and II) in G0/G1 and a slight increase in the S phase (Gate
II), a noticeable fraction of these cells was blocked at the G2/M transition (Fig. 4B). Rapamycin treatment improved significantly the G2 to
M transition of Foxo3 mutant erythroblasts (Fig. 4B, Gate I) suggesting
that the G2/M block is mediated in part by the overactivation of mTOR
signaling in Foxo3 mutant erythroblasts. As observed in control cells,
Figure 3. RNA interference inhibition of mTOR signaling reduces bone marrow erythroid colony-formation capacity. A. Inhibition efficiency of short hairpin
RNAs targeting S6 kinase was tested by transducing NIH3T3 cells with the indicated short hairpin RNAs and analyzing S6K1 expression by Western blot.
B. Number of bone marrow BFU-E- and CFU-E-derived colonies formed by lineage negative cells transduced with short hairpin RNAs targeting S6 kinase or
scrambled control are shown. Results shown are mean 6 SEM of triplicates; *P < 0.05; **P < 0.01; ***P < 0.001; Student’s t test. [Color figure can be viewed in
the online issue, which is available at wileyonlinelibrary.com.]
doi:10.1002/ajh.23786
American Journal of Hematology, Vol. 89, No. 10, October 2014
957
RESEARCH ARTICLE
Zhang et al.
the in vivo treatment with rapamycin enhanced the fraction of Foxo3
mutant cells (Gate I) in G0/G1. The effects of rapamycin on wild type
and Foxo3 mutant cell cycle were highly similar at later stages of erythroblast maturation (Fig. 4B, lower panels).
As anticipated, the effect of rapamycin in early precursors (Gates I
and II) mediated the inhibition of mTORC1 signaling as shown by
reduced phosphorylation of S6 at distinct stages of erythroblast maturation (Fig. 4C). In contrast, rapamycin treatment did not signifi-
Rapamycin treatment increases RBC production in
Foxo32/2 mice in vivo
Figure 4. In vivo rapamycin treatment inhibits cell cycle progression in
bone marrow immature erythroblasts. A. Flow cytometry strategy to distinguish four different bone marrow erythroid populations with increasing
degree of maturation (Gates I to IV) according to their TER119 and CD44
cell surface expression and forward scatter (FSC) properties after fixation
and permeabilization. Cell cycle analysis (B) and phosphorylated S6 (C)
were analyzed by flow cytometry in each erythroid precursor population
from WT and Foxo32/2 mice treated with rapamycin (4 mg/kg*day) or control vehicle for 2 weeks. In B each graph represents a distinct population
and numbers within each graph represent percentage of cells in G0/G1, S,
and G2/M phases of cell cycle for each population. One representative of
two independent experiments is shown (n 5 3 mice in each group,
mean 6 SEM; *P < 0.05; **P < 0.01; ***P < 0.001 between vehicle- and
rapamycin-treated; #P < 0.05; ##P < 0.01; ###P < 0.001 between WT and
Foxo32/2, Student’s t test). [Color figure can be viewed in the online issue,
which is available at wileyonlinelibrary.com.]
In vivo treatment with rapamycin increased significantly RBC
numbers and hemoglobin concentration in Foxo32/2 peripheral
blood without modulating significantly the total number of erythroid
(TER1191) cells (Table I; Fig. 5A,B). Importantly, rapamycin treatment tipped the balanced production of erythroid cells towards terminal maturation (Fig. 5C,D). A picture emerging from these findings is
that by increasing the fraction of cells in G1, rapamycin may block
cell cycle progression in immature erythroblasts (proerythroblasts,
Gate I), reduce the fraction of cells in the S phase, and induce cell
cycle exit in immature Foxo3 mutant erythroblasts. In agreement
with this interpretation, the ratio of mature to immature erythroblasts
in the bone marrow also increased in response to rapamycin treatment (Fig. 5C,D).
The effect of rapamycin on maturation may be due to its intrinsic
effect in erythroblasts, or extrinsic due to its effect on bone marrow
microenvironment [52,53]. To distinguish between these alternatives,
we isolated E14.5 fetal livers that are site of definitive erythropoiesis
and followed the ex vivo differentiation of erythroblasts in the presence
or absence of rapamycin. This approach enabled us to monitor precisely stages of erythroblast maturation in response to treatment. Addition of rapamycin to erythroblasts that were all at the proerythroblast
stage reduced the frequency of immature (in Gate I) in favor of mature
erythroblasts (in Gate III) similar to the effect seen in in vivo experiments (Supporting Information Fig. 3). These findings further support
the notion that overactivation of mTOR signaling in immature Foxo3
mutant erythroblasts reduces their rate of maturation.
Elevated erythropoietin (Epo) is associated with the state of ineffective erythropoiesis and is likely to take part in abnormally enhanced
signaling in immature erythroblasts [54,55]. We argued that if the
abnormal erythropoiesis in Foxo3 mutant mice is similar to ineffective
erythropoiesis, then circulating Epo levels should be increased in
Foxo32/2 mice. In addition, rapamycin treatment should result in
reduced Epo levels in Foxo3 mutant peripheral blood. Consistent with
an ineffective erythropoiesis phenotype, circulating Epo levels were significantly increased in Foxo3 mutant peripheral blood (Supporting
Information Fig. 4). In vivo treatment with rapamycin reduced the
increased levels of circulating Epo in Foxo3 mutant peripheral blood
(Supporting Information Fig. 4) but did not significantly modulate the
levels of Epo in wild-type mice. These results further supported the
concept that the abnormal Foxo32/2 erythropoiesis has overlapping
features with ineffective erythropoiesis. They also raised the potential
cantly reduce the frequency of pAKT-expressing erythroblasts or
levels of pAKT (data not shown) suggesting that the AKT activator
mTORC2 [35] might not be involved in these effects. Together these
results suggested the effects of rapamycin in early erythroblasts (Gates
I and II) are likely mediated by mTORC1.
TABLE I. Peripheral Blood Erythrocyte Parameters
WT Vehicle
6
RBC, 310
HGB (g/l)
HCT
MCV (fL)
MCH (pg)
MCHC (g/l)
Retic (%)
10.4 6 0.3
15.0 6 0.3
54.6 6 1.6
52.8 6 1.2
14.6 6 0.2
27.4 6 0.5
3.6 6 0.3
n59
WT Rapamycin
10.9 6 0.2
15.4 6 0.3a
56.5 6 1.9
52.5 6 1.6
14.9 6 0.2
28.1 6 0.5
3.8 6 0.5
n58
Foxo3-/- Vehicle
b
9.0 6 0.1
14.8 6 0.1
52.8 6 0.9
58.5 6 0.8b
16.4 6 0.2b
28.0 6 0.3
6.6 6 0.5b
n 5 10
Foxo3-/- Rapamycin
a
9.8 6 0.4
16.0 6 0.7a
55.7 6 2.4
56.8 6 0.9
16.2 6 0.4
28.6 6 0.3
7.4 6 0.4
n59
Th3/1 Vehicle
b
7.5 6 0.2
7.9 6 0.3b
32.2 6 1.3b
42.9 6 0.8b
10.3 6 0.3b
25.0 6 0.8b
25.8 6 1.9b
n59
Th3/1 Rapamycin
8.4 6 0.2a
9.2 6 0.3a
36.4 6 1.3a
43.2 6 0.7
10.5 6 0.2
25.2 6 0.7
28.2 6 2.2
n 5 10
Mice treated with rapamycin or vehicle control for two weeks.
P < 0.05 between vehicle- and rapamycin-treated
b
P < 0.05 between WT and Foxo32/2 or Th3/1.
a
958
American Journal of Hematology, Vol. 89, No. 10, October 2014
doi:10.1002/ajh.23786
RESEARCH ARTICLE
FOXO3/mTOR in Erythroid Maturation
Figure 6. Alteration of transcripts of metabolic enzymes in Foxo32/2
Figure 5. In vivo rapamycin treatment increases erythroid cell maturation.
Flow cytometric analysis of bone marrow erythroid cell distribution in wild
type and Foxo32/2 mice treated with rapamycin (4 mg/kg*day) or control
vehicle for 2 weeks. A. Schematic of flow cytometry analysis of five distinct erythroid populations according to their TER119 and CD44 surface
expression and FSC properties. B. Percentage of TER1191 cells within the
bone marrow. C. Distribution of TER1191 cells in each of the five gates
shown in A. D. Ratio of mature cells (Gates IV and V combined) relative to
the total TER1191 erythroid population. *P < 0.05 between vehicle- and
rapamycin-treated; #P < 0.05 between WT and Foxo32/2 (n 5 12, mean6 SEM; Student’s t test). Veh: vehicle, Rapa: rapamycin. [Color figure can
be viewed in the online issue, which is available at wileyonlinelibrary.com.]
that rapamycin might have a beneficial effect on ineffective
erythropoiesis.
Together, these results (Figs. 4 and 5, Supporting Information Fig.
3, Table I) suggest that rapamycin enhances erythroid maturation by
inducing cell cycle exit of immature erythroblasts resulting in
increased RBC production. These findings reflect the amplitude of
dynamic changes that are detected specifically in immature erythroblasts (proerythroblasts) as compared to cells at later stages of maturation as previously reported [56,57].
FOXO3-mTOR regulate erythroblast metabolic gene
expression in vivo
ROS elimination is maintained in part by sustained generation of
reduced glutathione as a result of glucose metabolism and a functional
pentose phosphate pathway (Fig. 6A). Given the function of mTOR
protein kinase in cellular metabolism [34,35] and its sensitivity to redox
modulation [48], in particular in erythroblast precursors (Fig. 2, Supporting Information Fig. 2) we explored the possibility that the impact
on erythroid cell production might be mediated by mTOR’s influence
on erythroblast metabolism. Using Fluidigm microfluidics technology that
enables monitoring the expression of 96 genes in 96 samples all at once
(equivalent of 9216 real-time PCR), we interrogated the expression of an
array of metabolic genes (Fig. 6). Wild type and Foxo3 mutant mice were
treated for 2 weeks with rapamycin or control vehicle and RNA from
doi:10.1002/ajh.23786
immature erythroblasts. A. Schematic of glycolytic pathway and its interaction with tricarboxylic acid (TCA) cycle and pentose phosphate pathway,
pyruvate kinase (PK). B. QRT-PCR expression analysis by Fluidigm microfluidics technology of metabolic genes in bone marrow Gates I–IV erythroblasts from WT and Foxo32/2 mice treated with rapamycin (4 mg/kg*day)
or control vehicle for 2 weeks. Quantification of target genes is relative to
b actin. Results are mean 6 SEM of three cDNAs, each generated from
one mouse. *P < 0.05 between vehicle- and rapamycin-treated; #P < 0.05
between WT and Foxo32/2. [Color figure can be viewed in the online
issue, which is available at wileyonlinelibrary.com.]
erythroblasts at distinct stages of maturation was isolated and subjected to
Fluidigm QRT-PCR analysis (Fig. 6). These experiments led to several
interesting observations. First, in agreement with earlier observations
[10,11], the transcript of the majority of metabolic genes surveyed in wildtype erythroblasts was at least as highly expressed in early immature erythroblasts (proerythroblasts, Gate I) as it was in maturing erythroblasts
(orthochromatic erythroblast, Gate IV). Strikingly, we found the expression of many metabolic genes specifically implicated in glucose metabolism was highly altered (mostly reduced) in Foxo3 mutant erythroblasts at
different stages of maturation, specifically in immature erythroblasts (Fig.
6B). Expression of several of these genes in Foxo3 mutant erythroblasts
was sensitive to rapamycin treatment (Fig. 6B). Specifically, in vivo treatment with rapamycin normalized and/or significantly improved the transcript level of several genes including pyruvate kinase M2 (Pkm2), aldolase
A encoding fructose-bisphosphate aldolase (AldoA), Enolase 1 a (EnoA)
(Eno1/EnoA), glyceraldehyde-3-phosphate dehydrogenase (Gapdh) and
Lactate Dehydrogenase (Ldh) B (LdhB) in Foxo3 mutant immature erythroblasts (Gates I and II cells, Fig. 6B). These results indicate that altered
expression of many metabolic specifically glycolytic transcripts in Foxo3
mutant immature erythroblasts is likely due to either decreased transcription (due to lack of FOXO3), or enhanced activation of mTOR signaling
(or likely both) in Foxo3 mutant erythroblasts.
In addition to the redox state, some of the direct targets of FOXO3
[58,59] like Rictor, Tsc1, and glutamine synthetase (GS) [23] are implicated in the activation of mTOR. Rictor is part of the mTORC2 complex [60,61]. Rictor transcripts were slightly but significantly reduced in
Foxo3 mutant erythroblast (Fig. 7A) supporting a potential decreased
assembly of mTORC2 in favor of mTORC1 complex in erythroblasts
American Journal of Hematology, Vol. 89, No. 10, October 2014
959
RESEARCH ARTICLE
Zhang et al.
Figure 6. Continued
[60,62]. While transcripts for Tsc1 that is an upstream negative regulator of mTOR signaling [63] were highly reduced in Foxo3 mutant
erythroblasts (Fig. 7A), TSC1 protein was not significantly altered (Fig.
7B, quantification in the right panel) suggesting that reduction of TSC1
transcript expression was unlikely to mediate activation of mTOR in
Foxo3 mutant erythroblasts. On the other hand the transcript for glutamine synthetase (GS) that is a negative regulator of mTOR [23] was
highly and significantly reduced in Foxo3 mutant erythroblasts at all
stages of maturation, including in early precursors (Gates I and II) (Fig.
6B). GS protein expression was also significantly reduced in freshly isolated TER 1191 Foxo3 mutant erythroblasts (Fig. 7C). These findings
raise the possibility that reduced expression of GS might be implicated
in mTORC1 activation in Foxo3 mutant erythroblasts.
Increased proliferation of immature erythroblasts that are unable to
fully mature, as seen in Foxo3 mutant erythroblasts (Figs. 4 and 5, Supporting Information Figs. 3,5) characterizes ineffective erythropoiesis
observed in many erythroid disorders including b-thalassemia [52,55].
In b-thalassemia, the generation of erythroid cells is hampered by a
number of intricate mechanisms triggered partly by redox imbalance
960
American Journal of Hematology, Vol. 89, No. 10, October 2014
that together lead to an exacerbated erythropoiesis that fails to produce
sufficient numbers of mature functional RBC [54,64]. Given that rapamycin treatment improved the ineffective erythropoiesis of Foxo3
mutant mice, we evaluated the potential effect of rapamycin on bthalassemic erythropoiesis. Strikingly, a 2-week in vivo treatment with
rapamycin increased the number of RBC (*P < 0.05, Table I), hematocrit (36.4 6 1.3% in rapamycin treated mice versus 32.2 6 1.3% in mice
treated with control vehicle) and specifically hemoglobin by over 1 g/dl
in the peripheral blood of a model of b-thalassemia (intermedia
Hbbth3/1, Th3/1) [55]. These findings suggest that activation of mTOR
signaling in erythroblasts may contribute to the b-thalassemic phenotype. The effects of rapamycin on bone marrow and spleen bthalassemic erythroblast cell cycling was at best modest (data not
shown) despite its effect on Foxo3 mutant erythroblast cycling (Fig. 4)
suggesting additional mechanisms are involved. Nonetheless, consistent
with the notion that Foxo3 mutant and b-thalassemic erythropoiesis
may share common features, there was a remarkable similarity between
alteration of metabolic gene expression in b-thalassemic and Foxo3
mutant erythroblasts (Fig. 6B). In particular, expression of Pkm2, Gpi1,
doi:10.1002/ajh.23786
RESEARCH ARTICLE
EnoA-Eno1, G6pd1, Pgk1, LdhA, LdhB, Idh1, and Idh2 was similarly
altered in b-thalassemic and Foxo3 mutant erythroblasts. While rapamycin treatment increased expression of b major in b-thalassemic erythroid cells (Fig. 8), the effect of rapamycin on metabolic genes was
restricted to increasing the transcript expression of Pdk1 and Idh1 in
early and late stages of b-thalassemic erythroblast maturation respectively (Fig. 6B).
Figure 7. Glutamine synthetase protein is decreased in Foxo3 mutant
erythroblasts. A. QRT-PCR expression analysis in freshly isolated erythroblasts at distinct stages of maturation. B. Western blot and band quantification (right panel) of TSC1 in freshly isolated bone marrow TER 1191
erythroblasts from two distinct mice for each genotype. C. Western blot
analysis of glutamine syntethase in freshly isolated bone marrow TER 1191
erythroblasts from four distinct mice for each genotype. Band quantification in the right. Results shown are mean 6 SEM of triplicates; #P < 0.05;
***P < 0.001; Student’s t test. [Color figure can be viewed in the online
issue, which is available at wileyonlinelibrary.com.]
FOXO3/mTOR in Erythroid Maturation
䊏 Discussion
Defective proliferation of Foxo3 mutant erythroblasts that fail to
mature [5] is reminiscent of proliferation and maturation defects leading to ineffective erythropoiesis [54]. Here we showed that defects of
Foxo3 mutant erythroblast maturation are mediated in part by hyperactivation of mTOR signaling In immature erythroblasts. These defects
lead specifically to a G2/M block in Foxo3 mutant erythroblast cycling.
Our results suggest that physiological cooperation of mTOR signaling
with FOXO3 is key to the control of cell cycle progression of immature
erythroblasts and their rate of maturation in vivo (see Model, Fig. 9).
We have previously shown that FOXO3 is required for erythroblast cell
cycling [5]. The defects in the G2/M progression of FOXO3 mutant
erythroblasts isolated from C57Bl6 mice shown here are similar to what
we observed in Foxo3 mutant hematopoietic stem cells [41]. However,
we failed to detect the G2/M block in a heterogeneous population of
erythroblasts isolated from mice on a mixed genetic background
(FVB3129). Although our data does not implicate directly redox modulations in mTOR activation in primary immature erythroblasts in
vivo, in contrast to our observations in primitive primary Foxo3 mutant
myeloid progenitors [45] overall our findings support a model in which
a feedback loop that is in part mediated by a redox switch amplified by
loss of FOXO3 activates mTOR signaling in primary erythroblasts.
Together these results indicate that the outcome of FOXO3/mTOR
interactions may be cell context dependent [45]. Activation of mTOR
in primitive primary Foxo3 mutant myeloid progenitors [45] result in
increased cell cycle whereas in Foxo3 mutant immature erythroblasts
mTOR activation leads to relative cell cycle delay (Fig. 4).
Erythroblasts have a remarkable capacity for proliferation and as
such are likely to be especially sensitive to metabolic perturbations
[43]. FOXO3 may have a function in regulating key metabolic genes
(Fig. 6) in erythroblasts in agreement with recent studies in neural
stem and progenitor cells [65]. Defective expression of metabolic
genes may contribute to the accumulation of ROS in Foxo3 mutant
erythrocytes resulting in hemolytic anemia and ROS-mediated reduction of RBC lifespan observed in Foxo32/2 mice [5]. It would be
interesting to investigate whether in addition to oxidative stress (Fig.
2) [5], reduction of glutamine synthetase expression (Figs. 6B,7C)
contributes to the overactivation of mTOR signaling of Foxo3 mutant
erythroblasts in vitro and/or in vivo [23].
Figure 8. In vivo rapamycin treatment improves b globin expression in b-thalassemia erythroblasts. WT and Th3/1 mice treated with rapamycin (4 mg/
kg*day) or control vehicle for 2 weeks. QRT-PCR expression analysis of globin genes in WT and Th3/1 bone marrow erythroblasts from Gates I to IV. QRTPCR expression analysis was performed using Fluidigm microfluidics technology and quantification of target genes is relative to b actin. Results are mean6 SEM of three cDNAs, each generated from one mouse. Veh: vehicle, Rapa: Rapamycin. [Color figure can be viewed in the online issue, which is available
at wileyonlinelibrary.com.]
doi:10.1002/ajh.23786
American Journal of Hematology, Vol. 89, No. 10, October 2014
961
RESEARCH ARTICLE
Zhang et al.
Figure 9. Proposed model for FOXO3/ROS/mTOR regulation of erythroblast maturation. During erythroid differentiation FOXO3 becomes nuclear and active, keeping
ROS levels under control through transcriptional regulation of antioxidant enzymes and certain metabolic enzymes. The absence of FOXO3 leads to increased ROS levels that further activate mTOR, which in turn influences glycolytic enzymes and erythroblast cell cycling leading to decrease in erythroblast maturation. Red arrows
indicate the main findings described in this article. [Color figure can be viewed in the online issue, which is available at wileyonlinelibrary.com.]
One of the most unexpected findings was the increased production
of mature RBC in response to rapamycin treatment in b-thalassemic
mice. b-thalassemia arises as a consequence of mutations of b globin
gene and precipitation of excess unmatched a globin, resulting in an
increase in the pool of free iron that triggers enhanced redox reactions
and damage to RBC membrane. These changes ultimately lead to excessive compensatory proliferation of erythroid precursors that fail to
mature [2,55,66,67]. Consistent with the therapeutic effect of JAK2 protein tyrosine kinase inhibitors [55], our results implicate mTOR signaling in the ineffective b-thalassemic erythropoiesis. It is noteworthy that
rapamycin induces g-globin mRNA and fetal hemoglobin (HbF) production in cultured human erythroid progenitors from b-thalassemic
patients [68]. It will be important to investigate mechanisms whereby
rapamycin ameliorates b-thalassemic anemia and explore potential
effects on protein translation, iron flux [69] and immune response [35].
One of the noticeable findings in these studies was ROS-mediated
phosphorylation of JAK2 in primary mouse erythroblasts. JAK2 protein tyrosine kinase, an essential component of EpoR signaling is
known to generate ROS upon stimulation and be redox modulated, a
property that is relevant to myeloproliferation [70]. However whether
oxidation activates or inhibits JAK2 has been debated [71–74]. While
the mechanism of JAK2 hyperphosphorylation in Foxo3 mutant
erythroblasts is unclear, reduced expression of Lnk (SH2B3), a negative regulator of JAK2 phosphorylation, in Foxo3 mutant erythroblasts might be implicated (Supporting Information Fig. 6) [45].
䊏 References
1. McDevitt MA, Xie J, Gordeuk V, et al. The anemia of malaria infection: Role of inflammatory
cytokines. Curr Hematol Rep 2004;3:97–106.
2. Rivella S. Ineffective erythropoiesis and thalassemias. Curr Opin Hematol 2009;16:187–194.
3. Neumann CA, Krause DS, Carman CV, et al.
Essential role for the peroxiredoxin Prdx1 in
erythrocyte antioxidant defence and tumour
suppression. Nature 2003;424:561–565.
4. Friedman JS, Rebel VI, Derby R, et al. Absence
of mitochondrial superoxide dismutase results
in a murine hemolytic anemia responsive to
therapy with a catalytic antioxidant. J Exp Med
2001;193:925–934.
5. Marinkovic D, Zhang X, Yalcin S, et al. Foxo3
is required for the regulation of oxidative stress
in erythropoiesis. J Clin Invest 2007;117:2133–
2144.
962
Collectively, these studies support the notion that activation of
mTOR signaling as a result of loss of FOXO3 function [45] might be
implicated in the pathogenesis of ineffective erythropoiesis as seen in
Foxo3 mutant mice. Future studies should elucidate whether and how
metabolic abnormalities associated with overactivation of mTOR signaling contribute to erythroblast cell cycle defects.
䊏 Authors’ Contributions
XZ, designed experiments, performed experiments, analyzed data
and participated in writing the paper. GC, designed experiments, performed experiments, analyzed data and participated in writing the
paper. SY, PR, SKM, JB, RL designed experiments, performed experiments, analyzed data. VDE, performed experiments. MHB facilitated
the set up of the in vitro assay. CB, analyzed data and participated in
discussions of the data and editing. SR, participated in discussions of
the data and editing. DP, analyzed data. SG, designed experiments,
analyzed data and wrote the paper.
䊏 Acknowledgments
We thank Brigitte Izac for technical help, Dr. Jane Little (Case
Western Reserve University School of Medicine) for critical reading
of the manuscript and the Flow Cytometry Shared Research Facility
at Icahn School of Medicine at Mount Sinai School.
6. Yu D, dos Santos CO, Zhao G, et al. miR-451
protects against erythroid oxidant stress by
repressing 14-3-3zeta. Genes Dev 2010;24:1620–
1633.
7. Kong Y, Zhou S, Kihm AJ, et al. Loss of alphahemoglobin-stabilizing protein impairs erythropoiesis and exacerbates beta-thalassemia. J Clin
Invest 2004;114:1457–1466.
8. Valentine WN. Metabolism of human erythrocytes. Studies in health and disease. Arch Intern
Med 1975;135:1307–1313.
9. Bossi D, Giardina B. Red cell physiology. Mol
Aspects Med 1996;17:117–128.
10. Nijhof W, Wierenga PK, Staal GE, et al.
Changes in activities and isozyme patterns of
glycolytic enzymes during erythroid differentiation in vitro. Blood 1984;64:607–613.
11. Kim HD, Koury MJ, Lee SJ, et al. Metabolic
adaptation during erythropoietin-mediated ter-
American Journal of Hematology, Vol. 89, No. 10, October 2014
12.
13.
14.
15.
16.
minal differentiation of mouse erythroid cells.
Blood 1991;77:387–392.
Nemoto S, Finkel T. Redox regulation of forkhead proteins through a p66shc-dependent signaling pathway. Science 2002;295:2450–2452.
Kops GJ, Dansen TB, Polderman PE, et al.
Forkhead transcription factor FOXO3a protects
quiescent cells from oxidative stress. Nature
2002;419:316–321.
Zhang X, Rielland M, Yalcin S, et al. Regulation
and function of FoxO transcription factors in
normal and cancer stem cells: What have we
learned? Curr Drug Targets 2011;12:1267–1283.
van den Berg MC, Burgering BM. Integrating
opposing signals toward forkhead box O. Antioxid Redox Signal 2011;14:607–621.
Kashii Y, Uchida M, Kirito K, et al. A member
of Forkhead family transcription factor,
FKHRL1, is one of the downstream molecules
of phosphatidylinositol 3-kinase-Akt activation
doi:10.1002/ajh.23786
RESEARCH ARTICLE
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
31.
32.
33.
34.
pathway in erythropoietin signal transduction.
Blood 2000;96:941–949.
Uddin S, Kottegoda S, Stigger D, et al. Activation of the Akt/FKHRL1 pathway mediates the
antiapoptotic effects of erythropoietin in primary human erythroid progenitors. Biochem
Biophys Res Commun 2000;275:16–19.
Ghaffari S, Jagani Z, Kitidis C, et al. Cytokines
and BCR-ABL mediate suppression of TRAILinduced apoptosis through inhibition of forkhead FOXO3a transcription factor. Proc Natl
Acad Sci USA 2003;100:6523–6528.
Bakker WJ, Blazquez-Domingo M, Kolbus A,
et al. FoxO3a regulates erythroid differentiation
and induces BTG1, an activator of protein arginine methyl transferase 1. J Cell Biol 2004;164:
175–184.
Bakker WJ, van Dijk TB, Parren-van
Amelsvoort M, et al. Differential regulation of
Foxo3a target genes in erythropoiesis. Mol Cell
Biol 2007;27:3839–3854.
Franco SS, De Falco L, Ghaffari S, et al. Resveratrol accelerates erythroid maturation by activation of FoxO3 and ameliorates anemia in
beta-thalassemic mice. Haematologica 2014;99:
267–275.
Pourfarzad F, von Lindern M, Azarkeivan A,
et al. Hydroxyurea responsiveness in betathalassemic patients is determined by the stress
response adaptation of erythroid progenitors
and their differentiation propensity. Haematologica 2013;98:696–704.
van der Vos KE, Eliasson P, Proikas-Cezanne T,
et al. Modulation of glutamine metabolism by
the PI(3)K-PKB-FOXO network regulates
autophagy. Nat Cell Biol 2012;14:829–837.
Rimmele P, d’Esamard V, Fatih K, et al. Metabolic cross talk between Foxo3 and mTOR is
essential for hematopoietic stem cell function.
Blood 2012;120:a856.
Warr MR, Binnewies M, Flach J, et al. FOXO3A
directs a protective autophagy program in haematopoietic stem cells. Nature 2013.
Sandri M, Sandri C, Gilbert A, et al. Foxo transcription factors induce the atrophy-related
ubiquitin ligase atrogin-1 and cause skeletal
muscle atrophy. Cell 2004;117:399–412.
Yeo H, Lyssiotis CA, Zhang Y, et al. FoxO3
coordinates metabolic pathways to maintain
redox balance in neural stem cells. EMBO J
2013;32:2589–2602.
Camprecios G, Barminko J, Bernitz J, et al.
Steady state differences in metabolic properties
of bone marrow versus spleen erythroid cells.
Blood 2013;122:943a.
Rimmele P, Bigarella C, Liang R, et al. Aginglike phenotype and defective lineage specification in SIRT1-deficient hematopoietic stem and
progenitor cells. Stem Cell Rep 2014 (in press)
3:1–16.
Kadri Z, Maouche-Chretien L, Rooke HM, et al.
Phosphatidylinositol 3-kinase/Akt induced by
erythropoietin renders the erythroid differentiation factor GATA-1 competent for TIMP-1 gene
transactivation. Mol Cell Biol 2005;25:7412–
7422.
Hammerman PS, Fox CJ, Birnbaum MJ, et al.
Pim and Akt oncogenes are independent regulators of hematopoietic cell growth and survival.
Blood 2005;105:4477–4483.
Ghaffari S, Kitidis C, Zhao W, et al. AKT induces erythroid-cell maturation of JAK2-deficient
fetal liver progenitor cells and is required for
Epo regulation of erythroid-cell differentiation.
Blood 2006;107:1888–1891.
Zhao W, Kitidis C, Fleming MD, et al. Erythropoietin stimulates phosphorylation and activation of GATA-1 via the PI3-kinase/AKT
signaling pathway. Blood 2006;107:907–915.
Howell JJ, Manning BD. mTOR couples cellular
nutrient sensing to organismal metabolic homeostasis. Trends Endocrinol Metab 2011;22:94–
102.
doi:10.1002/ajh.23786
FOXO3/mTOR in Erythroid Maturation
35. Zoncu R, Efeyan A, Sabatini DM. mTOR: From
growth signal integration to cancer, diabetes and
ageing. Nat Rev Mol Cell Biol 2011;12:21–35.
36. Grech G, Blazquez-Domingo M, Kolbus A, et al.
Igbp1 is part of a positive feedback loop in stem
cell factor-dependent, selective mRNA translation initiation inhibiting erythroid differentiation. Blood 2008;112:2750–2760.
37. Geslain R, Uddin S, Liu H, et al. Distinct functions of erythropoietin and stem cell factor are
linked to activation of mTOR kinase signaling
pathway in human erythroid progenitors. Cytokine 2013;61:329–335.
38. Diekmann F, Rovira J, Diaz-Ricart M, et al. mTOR
inhibition and erythropoiesis: Microcytosis or anaemia? Nephrol Dial Transplant 2012;27:537–541.
39. Fishbane S, Cohen DJ, Coyne DW, et al. Posttransplant anemia: The role of sirolimus. Kidney
Int 2009;76:376–382.
40. Castrillon DH, Miao L, Kollipara R, et al. Suppression of ovarian follicle activation in mice by
the transcription factor Foxo3a. Science 2003;
301:215–218.
41. Yalcin S, Zhang X, Luciano JP, et al. Foxo3 is
essential for the regulation of ataxia telangiectasia mutated and oxidative stress-mediated
homeostasis of hematopoietic stem cells. J Biol
Chem 2008;283:25692–25705.
42. Ghaffari S, Kitidis C, Fleming MD, et al. Erythropoiesis in the absence of janus-kinase 2: BCRABL induces red cell formation in JAK2(2/2)
hematopoietic progenitors. Blood 2001;98:2948–
2957.
43. England SJ, McGrath KE, Frame JM, et al.
Immature erythroblasts with extensive ex vivo
self-renewal capacity emerge from the early
mammalian fetus. Blood 2011;117:2708–2717.
44. Chen K, Liu J, Heck S, et al. Resolving the distinct stages in erythroid differentiation based on
dynamic changes in membrane protein expression during erythropoiesis. Proc Natl Acad Sci
USA 2009;106:17413–17418.
45. Yalcin S, Marinkovic D, Mungamuri SK, et al.
ROS-mediated amplification of AKT/mTOR signalling pathway leads to myeloproliferative syndrome in Foxo3(2/2) mice. EMBO J 2010;29:
4118–4131.
46. Zhang X, Yalcin S, Lee DF, et al. FOXO1 is an
essential regulator of pluripotency in human
embryonic stem cells. Nat Cell Biol 2011;13:
1092–1099.
47. Liang R, Ghaffari S. Stem cells, redox signaling,
and stem cell aging. Antioxid Redox Signal
2014;20:1902–1916.
48. Sarbassov DD, Sabatini DM. Redox regulation
of the nutrient-sensitive raptor-mTOR pathway
and complex. J Biol Chem 2005;280:39505–
39509.
49. Yoshida S, Hong S, Suzuki T, et al. Redox regulates mammalian target of rapamycin complex 1
(mTORC1) activity by modulating the TSC1/
TSC2-Rheb GTPase pathway. J Biol Chem 2011;
286:32651–32660.
50. Ma XM, Blenis J. Molecular mechanisms of
mTOR-mediated translational control. Nat Rev
Mol Cell Biol 2009;10:307–318.
51. Gregory CJ. Erythropoietin sensitivity as a differentiation marker in the hemopoietic system:
Studies of three erythropoietic colony responses
in culture. J Cell Physiol 1976;89:289–301.
52. Ramos P, Casu C, Gardenghi S, et al. Macrophages support pathological erythropoiesis in
polycythemia vera and beta-thalassemia. Nat
Med 2013;19:437–445.
53. Hashimoto D, Chow A, Noizat C, et al. Tissueresident macrophages self-maintain locally
throughout adult life with minimal contribution
from circulating monocytes. Immunity 2013;38:
792–804.
54. Rivella S. The role of ineffective erythropoiesis
in
non-transfusion-dependent
thalassemia.
Blood Rev 2012;26 Suppl 1:S12–S15.
55. Libani IV, Guy EC, Melchiori L, et al. Decreased
differentiation of erythroid cells exacerbates
56.
57.
58.
59.
60.
61.
62.
63.
64.
65.
66.
67.
68.
69.
70.
71.
72.
73.
74.
ineffective erythropoiesis in beta-thalassemia.
Blood 2008;112:875–885.
Peslak SA, Wenger J, Bemis JC, et al. EPOmediated expansion of late-stage erythroid progenitors in the bone marrow initiates recovery
from sublethal radiation stress. Blood 2012;120:
2501–2511.
Dev A, Fang J, Sathyanarayana P, et al. During
EPO or anemia challenge, erythroid progenitor
cells transit through a selectively expandable
proerythroblast pool. Blood 2010;116:5334–5346.
Chen CC, Jeon SM, Bhaskar PT, et al. FoxOs
inhibit mTORC1 and activate Akt by inducing
the expression of Sestrin3 and Rictor. Dev Cell
2010;18:592–604.
Khatri S, Yepiskoposyan H, Gallo CA, et al.
FOXO3a regulates glycolysis via transcriptional
control of tumor suppressor TSC1. J Biol Chem
2010;285:15960–15965.
Sarbassov DD, Ali SM, Kim DH, et al. Rictor, a
novel binding partner of mTOR, defines a
rapamycin-insensitive and raptor-independent
pathway that regulates the cytoskeleton. Curr
Biol 2004;14:1296–1302.
Guertin DA, Stevens DM, Thoreen CC, et al.
Ablation in mice of the mTORC components
raptor, rictor, or mLST8 reveals that mTORC2
is required for signaling to Akt-FOXO and
PKCalpha, but not S6K1. Dev Cell 2006;11:859–
871.
Sarbassov DD, Guertin DA, Ali SM, et al. Phosphorylation and regulation of Akt/PKB by the
rictor-mTOR complex. Science 2005;307:1098–
1101.
Efeyan A, Zoncu R, Sabatini DM. Amino acids
and mTORC1: From lysosomes to disease.
Trends Mol Med 2012;18:524–533.
Thein SL. Genetic modifiers of beta-thalassemia.
Haematologica 2005;90:649–660.
Yeo H, Lyssiotis CA, Zhang Y, et al. FoxO3 coordinates metabolic pathways to maintain redox
balance in neural stem cells. EMBO J 2013;32:
2589–2602.
Shinar E, Rachmilewitz EA. Oxidative denaturation of red blood cells in thalassemia. Semin
Hematol 1990;27:70–82.
Scott MD, Eaton JW. Thalassaemic erythrocytes:
Cellular suicide arising from iron and
glutathione-dependent oxidation reactions? Br J
Haematol 1995;91:811–819.
Fibach E, Bianchi N, Borgatti M, et al. Effects of
rapamycin on accumulation of alpha-, beta- and
gamma-globin mRNAs in erythroid precursor
cells from beta-thalassaemia patients. Eur J Haematol 2006;77:437–441.
Bayeva M, Khechaduri A, Puig S, et al. mTOR
regulates cellular iron homeostasis through tristetraprolin. Cell Metab 2012;16:645–657.
Marty C, Lacout C, Droin N, et al. A role for
reactive oxygen species in JAK2 V617F myeloproliferative neoplasm progression. Leukemia
2013;27:2187–2195.
Duhe RJ, Evans GA, Erwin RA, et al. Nitric
oxide and thiol redox regulation of Janus kinase
activity. Proc Natl Acad Sci USA 1998;95:126–
131.
Maziere C, Conte MA, Maziere JC. Activation
of JAK2 by the oxidative stress generated with
oxidized low-density lipoprotein. Free Radic
Biol Med 2001;31:1334–1340.
Nieborowska-Skorska M, Kopinski PK, Ray R,
et al. Rac2-MRC-cIII-generated ROS cause
genomic instability in chronic myeloid leukemia
stem cells and primitive progenitors. Blood
2012;119:4253–4263.
Walz C, Crowley BJ, Hudon HE, et al. Activated JAK2 with the V617F point mutation
promotes G1/S-phase transition. J Biol Chem
2006;281:18177–18183.
American Journal of Hematology, Vol. 89, No. 10, October 2014
963