Healthy individuals at sea level continuously generate red blood cells in a process
known as “basal erythropoiesis” that is essential to life.
Erythropoiesis increases by up to 10-fold its basal rate in response to hypoxic
stress, as may occur at high altitude, or in response to anemia or hemorrhage.
Erythropoietic rate is regulated by the hormone Erythropoietin (Epo), whose
concentration in blood spans a remarkable, three orders of magnitude range, from
≈0.01 U/ml in the basal state to 10 U/ml in extreme stress. Epo exerts its
effects by binding to its receptor, EpoR, a transmembrane homodimer of the cytokine
receptor superfamily expressed by erythroid progenitors 
. Epo or EpoR-null mice die at
mid-gestation as a result of complete absence of mature red cells 
, and EpoR signaling
is essential for both basal and stress erythropoiesis 
. Binding and activation of
the EpoR results in activation of the cytoplasmic tyrosine kinase Jak2, and in
phosphorylation of EpoR cytoplasmic-domain tyrosines that act as docking sites for
signaling intermediates including Stat5 
A key challenge lies in understanding how EpoR signaling might differ between stress
and basal conditions. This challenge is of particular relevance to clinical
practice, where Epo is widely used and where erythropoietic mimetics are under
intense development to maximize benefit while reducing risk 
. Here we addressed this
question by studying Stat5, which, as suggested by mouse genetic models, is a key
mediator of both basal and stress erythropoiesis. Thus, Stat5-null mice die
perinatally due to anemia, while mice hypomorphic for Stat5 survive, but are
deficient in their response to erythropoietic stress 
Stat5 functions are due to two highly homologous proteins, Stat5a and Stat5b, of the
Signal Transducers and Activators of Transcription (STAT) family. STAT proteins are
latent cytoplasmic transcription factors that become activated by phosphorylation of
a C-terminal tyrosine in response to a variety of extracellular signals 
. Stat5 is a
key mediator of cell survival in erythroblasts and other hematopoietic progenitors.
In addition, it is frequently constitutively active in myeloproliferative disease
and in hematological malignancies 
Here we asked whether the dynamic behavior of the Stat5 activation signal, namely,
the way it varies with Epo concentration and with time, differs between stress and
basal erythropoiesis. Previously, distinct dynamic forms of ERK or Ras signaling
have been shown to specify distinct cellular responses 
. The dynamic form of a signal,
however, is often masked when measured in large populations of cells whose responses
are inherently variable. Analysis of a signal's dynamic properties therefore
requires measurement in single cells, with relatively few such studies to date.
To address this, we analyzed Stat5 signaling using flow-cytometry, in primary murine
erythroid progenitors, either in vivo or shortly following harvest (<24 h). We
combined two recent flow-cytometric assays, identifying
differentiation-stage-specific erythroblasts in tissue using cell-surface markers
and measuring their Stat5
phosphorylation signal (p-Stat5) using intracellular flow-cytometry 
. We determined
the time course and full dose-response curves of the p-Stat5 response to the entire
basal and stress Epo concentration range, in freshly harvested fetal liver
erythroblasts at five distinct stages of differentiation.
We found that Stat5 signals through two modalities, binary (digital) and graded
(analog). We characterized these modalities using wild-type mice and an EpoR mutant
mouse that we found to be restricted to the binary Stat5 signaling modality. We show
that later erythroblasts generate a low intensity but decisive, binary
“on” or “off” p-Stat5 signal that is both necessary and
sufficient for mediating Stat5 functions in basal erythropoiesis. By contrast, in
early erythroblasts Stat5 signaling is graded, reaching much higher signal
intensities that are necessary for the stress response, including the upregulation
of the transferrin receptor (CD71), a novel EpoR and Stat5 stress target.
The orderly transition in the modality of Stat5 signaling from early to later
erythroblasts is due to decreasing Stat5 protein levels with erythroid maturation.
Stat5 protein levels determine both maximal p-Stat5 signal intensity and the
steepness of the Stat5 signaling response. This contrasts with EpoR expression,
which does not appear to impose a limit on the maximal p-Stat5 response.
Our work shows that Stat5 signaling dynamics conveys information specifying the
required functional outcome in erythroblasts. The unique combination of a steep,
binary response to low Epo in the basal state, with a higher intensity graded
signaling modality during stress, allows Stat5 to transduce Epo stimuli with high
fidelity over its entire physiological and stress range.