Recombinant Human Erythropoietin (rhEPO) in Clinical

Oncology Scientific and Clinical Aspects of Anemia in

Cancer, 1st Edition

Visit the link below to download the full version of this book:

https://medidownload.com/product/recombinant-human-erythropoietin-rhepo-in-clini
cal-oncology-scientific-and-clinical-aspects-of-anemia-in-cancer-1st-edition/

Click Download Now

 Prof. Dr. med. M. R. Nowrousian
Innere Klinik und Poliklinik (Tumorforschung), Westdeutsches Thmorzentrum
Essen, Universitatsklinikum, Essen, Deutschland

This work is subject to copyright.

All rights are reserved, whether the whole or part of the material is concerned,
specifically those of translation, reprinting, re-use of illustrations, broadcasting,
reproduction by photocopying machines or similar means, and storage in
data banks.

Product Liability: The publisher can give no guarantee for all the information
contained in this book. This does also refer to information about drug dosage and
application thereof. In every individual case the respective user must check its
accuracy by consulting other pharmaceutical literature. The use of registered
names, trademarks, etc. in this publication does not imply, even in the absence of a
specific statement, that such names are exempt from the relevant protective laws
and regulations and therefore free for general use.

© 2002 Springer-VerlaglWien
Softcover reprint of the hardcover 1st edition 2002

Typesetting: SNP Best-set Typesetter Ltd., Hong Kong

Printed on acid-free and chlorine-free bleached paper

SPIN: 10797756

With 66 (partly coloured) Figures

ISBN-13: 978-3-7091-7660-3 e-ISBN-13: 978-3-7091-7658-0

DOl: 10.1007/978-3-7091-7658-0
 Preface

During the last decade, considerable insight has been gained into the
pathogenic mechanisms of anemia in cancer and cancer treatment and its
important role in the life of cancer patients and the course of their disease.
Anemia is a frequent complication that not only presents a negative prog-
nostic factor for the outcome of treatment in a variety of malignant diseases,
but is also associated with enormous impacts on physical well-being and
quality of life (QoL) of patients. Anemia may also be involved in the develop-
ment of tumor resistance against radiotherapy and chemotherapy. Recent
studies indicate a close relationship between anemia and tumor hypoxia,
and show that the latter is a factor that significantly determines the outcome
of radiotherapy. Furthermore, there is evidence suggesting that hypoxia
stimulates angiogenesis within the tumor and contributes to a selection of a
more malignant phenotype of tumor cells and a reduced sensitivity of these
cells to irradiation and chemotherapy. These findings and the consequences
that arise from anemia for metabolic and organ functions as well as QoL
identify anemia as a much more serious problem for cancer patients than
previously considered.
Treatment of anemia has been traditionally red blood cell (RBC)
transfusion, which, however, only transiently increases the hemoglobin level
and is of minimal effect on QoL. It is also associated with a number of side
effects and risks, such as febrile and allergic reactions, alloimmunization,
transmission of infections, iron overload and suppression of cellular im-
munity that could be particularly harmful for cancer patients. In these
patients, in addition, symptoms of anemia are frequently attributed, to the
malignant disease or its treatment, and many patients remain untreated until
severe clinical symptoms occur or the hemoglobin level decreases below
8-10 g/dl.
The introduction of recombinant human erythropoietin (rhEPO) has
dramatically improved the treatment of anemia in patients with end-stage
renal disease and is also a major advantage for the treatment of anemia in
patients with malignant diseases. Using rhEPO, it is possible to achieve sus-
tained physiological, and much more effective, levels of hemoglobin than
with RBC transfusion. The use of rhEPO, in addition, has considerably
vi Preface

physical well-being and to ameliorate tumor oxygenation with the aim of

improving the outcome of cancer treatment.
This book aims to present a comprehensive and up-to-date review of
scientific and clinical aspects of anemia in cancer patients and its treatment
with rhEPO. It was a great pleasure for me, and highly appreciated, that out-
standing authors, all experts on their topics, agreed to contribute to this book
and to present the state of knowledge on anemia in cancer and the current
as well as future potential of the use of rhEPO in clinical oncology, both in
patients with solid tumors and patients with hematological malignancies. My
sincere gratitude is also extended to Mrs. Ch. Wartchow for her help in proof-
reading and Mrs. I. Demirer for her excellent organizational assistance in
preparing this book.

Essen, March 2002 M. R. Nowrousian

 Contents

I. Biology of erythropoietin 1
C. Lacombe, P. Mayeux

II. Classification and characterization of anemia in cancer 23

M. R. Nowrousian

III. Pathophysiology of cancer-related anemia 39

M. R. Nowrousian

IV. Prevalence, pathophysiology, predictive factors, and

prognostic significance of anemia in cancer chemotherapy 63
M. R. Nowrousian

V. Incidence and impact of anemia in radiation oncology 101

1 Dunst, M. Molls

VI. Relationship between anemia and tumor hypoxia 117

1 Dunst, M. Molls

VII. Tumor hypoxia and therapeutic resistance 127

P. Vaupel, M. Hackel

VIII. Impact of anemia on organ functions ·147

M. R. Nowrousian

IX. Relationship between anemia, fatigue, and quality of life in

cancer patients 173
Y. Brandberg

X. Red blood cell transfusion, risks and limitations 185

F. Mercuriali, G. Inghilleri
viii Contents

XII. Recombinant human erythropoietin (rhEPO) in anemia

associated with multiple myeloma and non-Hodgkin
lymphoma 223
A. Osterborg

XIII. Use of recombinant human erythropoietin in the treatment

of myelodysplastic syndromes 235
M. Cazzola

XIV. rhEPO in anemia associated with solid tumors and

chemotherapy 241
M. R. Nowrousian

XV. Predictive factors for response of anemia to recombinant

human erythropoietin 263
Y. Beguin

XVI. rhEPO in hematopoietic stem cell mobilization,

transplantation, and in-vitro expansion 287
S. Klaesson

XVII. Clinical trials using rhEPO in radiation oncology 301

M. Henke

XVIII. rhEPO in pediatric oncology 313

C. Cappelli, G. Ragni, A. Clerico

XIX. rHuEPO in surgical oncology 325

F. Mercuriali, G. Inghilleri

XX. Erythropoiesis, iron metabolism and iron supplementation

during erythropoietin therapy 347
L. T. Goodnough

XXI. Optimal level of hemoglobin in cancer patients 369

M. R. Nowrousian

XXII. Protection of metabolic and exercise capacity following

treatment with recombinant erythropoietin 391
K. Lundholm, P. Daneryd
 Contents ix

XXIV. Effect of rhEPO on survival in anaemic cancer patients

receiving chemotherapy 425
T. J Littlewood

xxv. Cost-effectiveness of rHuEPO in oncology 435

P.-¥. Cremieux, Ellison Dial, M. Gustafson, B. Sarokhan,
M. B. Slavin

XXVI. Current status and future developments of rhEPO in

clinical oncology 447
M. R. Nowrousian

Subject Index 493

 Chapter I

Biology of erythropoietin

C. Lacombe1,2 and P. Mayeux!

lInstitut National de la Sante et de la Recherche Medicale, ICGM,

Universite Rene Descartes,
2 Service d'Hematologie, AP-Hp, Hopital Cochin, Paris, France

Introduction

Patients undergoing chemotherapy for cancer are at risk of developing

anemia, and recombinant human erythropoietin (Epo) is an alternative to
replace transfusions of allogenic red blood cells in this setting. This article
will review the regulation of the Epo gene, the structure of the Epo receptor
(EpoR) and the Epo-induced intracellular signaling events. Finally, we will
describe other compounds and mechanisms which mimic Epo action, thereby
also leading to intracellular signalling albeit with a decreased efficiency when
compared to Epo.
The role of Epo, a 34 kDa glycoprotein hormone is to control red blood
cell production through the promotion of survival and proliferation of the
erythroid progenitors in the bone marrow. Epo is the hematopoietic growth
factor which is acting specifically on the late erythroid progenitors, so-called
CFU-E (colony forming unit-erythroid). These cells correspond to the last
amplification compartment of the erythroid lineage and give rise to the ery-
throblasts in the bone marrow. Because the main function of red cells is to
transport oxygen from the lungs to the peripheral tissues, the regulation of
Epo production is an important feature of the control of tissue oxygenation.
Accordingly, Epo is the only hematopoietic growth factor whose production
is regulated by hypoxia.
Epo acts through a specific receptor, EpoR, belonging to the family of
the hematopoietic growth factor receptors. Activation of the EpoR by its
ligand leads to the tyrosine phosphorylation of numerous proteins into the
target cell; among these proteins, some migrate to the nucleus, where they
stimulate the transcription of specific target genes.
2 C. Lacombe and P. Mayeux

1 Role of Epo in erythropoiesis

Cultures of hematopoietic progenitors in semi-solid media have shown that

the main targets of Epo are the late erythroid progenitors CFU-E (Gregory
and Eaves 1978). Epo and EpoR gene disruptions in mice confirmed that
Epo stimulation was absolutely necessary for survival and proliferation of
CFU-E (Lin et al. 1996; Wu et al. 1995). Moreover, these experiments showed
that Epo stimulation was not required for the commitment of the progeni-
tors in the erythroid lineage. Indeed, both BFU-E and CFU-E were produced
to normal levels in Epo or EpoR null mice (Lin et al. 1996; Wu et al. 1995),
demonstrating that Epo is not involved in the determination of the erythroid
lineage.
Epo is able to sustain the proliferation of several hematopoietic cell lines
either naturally expressing the EpoR, such as HCD57 (Spivak et al. 1991)
or UT7 (Komatsu et al. 1991) or after ectopic expression of the EpoR
(D'Andrea et al. 1991; Quelle and Wojchowski 1991). These cells as well as
primary erythroid cells undergo apoptosis after Epo deprivation. HCD-57
cells infected with a retroviral vector encoding either Bel-2 or Bel-XL remain
viable in the absence of Epo. However, Epo is still required for proliferation
(Silva et al. 1996). These results demonstrate that the overall action of Epo
is to protect from apoptosis and to induce the proliferation of the CFU-E
progenitors.

2 Tissue-specific Epo gene expression

The elonage of the Epo gene (Jacobs et al. 1985; Lin et al. 1985) allowed to
gain insights into the molecular biology of Epo. In the fetal stage, the liver is
the major site of Epo synthesis (Zanjani et al. 1981), however, the Epo gene
appears to be also strongly expressed in the mammalian mesonephric kidney
early in gestation. In adult mammals, the renal synthesis of Epo was first
demonstrated by Jacobson et al. (1957). Studies on mice have shOWN that Epo
gene transcription was stimulated by hypoxia or cobalt treatment (Beru
et al. 1986), and there was a clear correlation between induction of anemia
and increase of Epo mRNA content in the kidney (Bondurant and Koury
1986). It was further shown, by in situ hybridization experiments, that Epo
mRNA was produced by interstitial cells of the kidney cortex (Koury et al.
1988; Lacombe et al. 1988). This specialized population of interstitial cells
was shown to be labelled by immunohistochemical staining with antibodies
to 5' ectonucleotidase (Bachmann et al. 1993) and thereby to belong to a
fibroblast-like cell population of the renal interstitium.
Similar results were obtained in transgenic mice containing the SV40
large tumor antigen (SV40T-antigen) placed behind the Epo gene regulatory
sequences; immunohistochemical detection of T-antigen was found in the
 Biology of erythropoietin 3

same fibroblast-like renal interstitial cells (Maxwell et al. 1993a). Unfortu-

nately, the use of an oncogene like SV40 T-antigen did not induce any for-
mation of tumor in the kidney, nor the establishment of transformed cell lines
from this interstitial cell population in the kidney. In addition, in renal ade-
nocarcinomas associated with polycythemia, the tumoral cells themselves,
which derived from the epithelial tubular cells produced Epo (Da Silva et al.
1990). It is possible that a cellular cooperation in the kidney cortex is required
for Epo production. Interestingly enough, Epo mRNA could be obtained
from isolated perfused rat kidneys and never from anatomically disrupted
renal preparations (Ratcliffe et al. 1990).
The liver accounts for 20% of the Epo production. Hepatocytes sur-
rounding central veins were responsible for most of the Epo production in
the liver (Koury et al. 1991), whereas other Epo-producing cells were shown
to belong to the Ito cells which share many similarities with the fibroblast-
like interstitial cells of the kidney (Maxwell et al. 1994).
In addition to these two main sites of secretion, low levels of Epo mRNA
have been detected in lung, testes and spleen when animals were subjected
to hypoxia (Fandrey and Bunn 1993; Tan et al. 1991). The function of Epo as
a growth factor to protect cells from apoptosis extends beyond the
hematopoietic lineage, since Epo was reported to stimulate proliferation of
myoblasts and interfere with their terminal differentiation in myotubes,
thereby suggesting a potential role in muscle development (Ogilvie et al.
2000). Epo is also produced in the brain by astrocytes (Masuda et al. 1994),
accordingly Epo receptors have been detected in mouse brain (Digicaylioglu
et al. 1995) and in cell lines with neuronal properties. These data suggest that
Epo could playa neurotrophic role in the brain and that the hypoxic induc-
tion of brain Epo could protect neurons from ischemia-induced cell death
(Morishita ct al. 1997). In addition, it was recently shown that Epo was able
to cross the blood-brain barrier to protect against experimental brain injury
(Brines et al. 2000).

3 Regulation of Epo production

Epo production is regulated by hypoxia that leads to an increase of the level

of gene transcription (Schuster et al. 1989). There are no preformed stores
of Epo. To gain insights into tissue-specific Epo gene expression, Semenza
et al. (1989) developed constructs of human Epo gene containing various
lengths of cis regulatory regions for production of transgenic mice. The
pattern of human Epo gene expression in these transgenic mice led the
authors to describe different DNA sequences located in cis of the Epo gene
and required for tissue-specificity and hypoxia-inducible gene expression. Se-
quences required for expression in the kidney (KIE) have been localized to
a region located 9.5 to 14kb 5' to the human Epo gene (Semenza et al. 1991a).
4 C. Lacombe and P. Mayeux

A negative regulatory element (NRE) which represses Epo gene expression

in non-Epo-producing cells is located in a region 0.4 to 6kb to the Epo tran-
scription start site (Semenza et al. 1990). A 50 bp hypoxia-inducible enhancer
has been defined approximately 120bp 3' to the polyadenylation site, and is
responsible for hypoxia-inducible Epo gene expression (Beck et al. 1991;
Semenza et al. 1991b). Mice transgenic for a construct containing the Epo
gene and this 3' enhancer harboured hypoxia-inducible Epo gene expression
in the liver.
The 3' enhancer contains three different segments (Semenza and Wang
1992). A conserved sequence located near the 5' end of the enhancer is the
binding site for a transcription factor designated HIF-1 (hypoxia-induced
factor 1) (Beck et al. 1993; Wang and Semenza 1993a). The middle segment
is less conserved between species, but seems to playa role in the inducibil-
ity of both the human and the murine Epo enhancers (Pugh et al. 1994). The
third part corresponds to 3' DNA sequences which are binding sites for the
hepatocyte nuclear factor 4 (HNF-4). Proteins that bind to this enhancer
interact synergistically to stimulate Epo gene transcription, and HNF-4 can
augment transcriptional activation mediated by the Epo enhancer in hypoxic
cells (Galson et al. 1995). Furthermore, the C-terminal portion of HIF-1
specifically binds to P300 and overexpression of P300 enhances hypoxic
induction (Arany et al. 1996). Thus, it is likely that hypoxia induces the for-
mation of a large complex of proteins directly or indirectly bound to the
enhancer, which in turn transduce a signal to the Epo promoter, thereby per-
mitting gene transcription (Ebert and Bunn 1999; Huang et al. 1997) (Fig. 1).
The identification of HIF-1 as a DNA transcriptional complex has been
a critical step to understand the Epo gene enhancer function. Affinity purifi-
cation showed that HIF-1 was composed of two subunits (Semenza and Wang
1992; Wang and Semenza 1995). Molecular cloning of HIF-1 by Semenza and
colleagues (Wang et al. 1995) showed that the DNA binding complex was
made of two basic-loop-helix PAS proteins called HIF-1ex and HIF-1p. HIF-
1P had previously been identified as the aryl hydrocarbon nuclear receptor
translocator (ARNT), a molecule involved in the xenobiotic response. In con-
trast, HIF-1ex was a new member of this family of PAS proteins. The mecha-

...
,
~_ . . __ m . . . . __ m m _ . . __ m __ . . ___ . . [ P300 lCEXiD
Promoter HIF-1 HNF-4

Fig. 1. Cis elements and transactivating factors involved in Epo gene activation. The
5 exons of the Epo gene are represented, coding portions are solid areas. KIE: kidney
inducible elements; NRE: negative regulatory elements. (From Ebert and Bunn 1999)
 Biology of erythropoietin 5

nism of regulation by hypoxia was first studied in hepatoma cells Hep3B or

HepG2 which produced Epo. It was further shown that identical responses
could be obtained in a large array of non-Epo-producing cells and that the
system of gene regulation by oxygen was widespread from mammalian to
insect cells (Maxwell et al. 1993b; Wang and Semenza 1993b). Many genes
have now been identified as targets of HIF-1 function; these include, in addi-
tion to Epo, vascular endothelial growth factor (VEGF), several glycolytic
enzymes, glucose-transporter 1, inducible nitric oxide synthase, heme oxy-
genase and transferrin (Wenger and Gassmann 1997). These recent data
strengthen the idea that cellular response to hypoxia is an important physi-
ological process and that a similar mechanism for oxygen sensing and signal
transduction must be shared by many tissues and cells (Bunn and Poyton
1996).
In hypoxic conditions, the levels of the mRNAs encoding either HIF-1a
or HIF-1~ were not altered, suggesting that the activity ofthe HIF-1a-ARNT
complex was regulated by a posttranscriptional mechanism (Kallio et al.
1997). The major mechanism of regulation of HIF-1a involves the ubiquitin-
proteasome system: HIF-1a is constitutively degraded in normoxia, while it
accumulates rapidly following exposure to hypoxia (Salceda and Caro 1997).
The von Hippel-Lindau (VHL) tumor-suppressor protein (pVHL) has been
linked to the regulation of the transcription factor HIF-1 (Maxwell et al.
1999). Wild-type pVHL is a component of an E3 ubiquitin-ligase complex
that transfers ubiquitin onto substrates to be degraded, and one recent report
identified the a-subunits of HIF-1 as ubiquitination targets for VHL (Ohh
et al. 2000). The VHL gene is inactivated in 80% of sporadic clear-cell renal
carcinomas. These tumors lacking functional pVHL fail to degrade HIF-
la, which stimulates the transcription of a series of hypoxia-responsive genes,
among which VEGF plays an important role in tumor angiogenesis.
However, oxygen-sensing mechanisms are still not completely understood.
In the current model of oxygen sensing, a heme protein, likely a cytochrome
b-like protein senses oxygen tension and regulates production of oxygen free
radicals (Ebert and Bunn 1999; Goldberg et al. 1988).
Very recently, it was shown that the p38a MAP kinase protein played an
important role in developmental erythropoiesis through regulation of Epo
expression: in mice where the p38a locus has been disrupted, there was a
dramatic decrease in Epo gene expression. It seems that p38a affected Epo
gene expression at the posttranscriptionallevel, most likely through mRNA
stabilisation (Tamura et al. 2000).

4 Structure of the Epo receptor

Epo acts on its target cells through specific membrane receptors. They
are mainly expressed at the colony-forming unit erythroid (CFU-E) stage,
6 C. Lacombe and P. Mayeux

receptor expression then decreases with erythroid maturation (Mayeux et al.

1987). The number of EpoR at the cell surface of normal or transformed
erythroid cells is low: around one thousand per cell (reviewed in D'Andrea
and Zon 1990). EpoR are present on the surface of erythroid cells (Broudy
et al. 1991), on megakaryocytes (Fraser et al. 1989), on endothelial cells
(Anagnostou et al. 1994), and some neuronal cells (Masuda et al. 1993). The
EpoR cDNA has been cloned by D'Andrea et al. (1989) and was shown to
encode a single membrane-spanning protein of 507 amino-acids which does
not possess catalytic activity in its intracellular region. This receptor belongs
to the cytokine receptor family and shares structural homologies with recep-
tors for interleukin (IL)-2 to 7, IL-9, IL-1l-13, IL-15, GM (granulocyte-
macrophage )-CSF, leukemia inhibitory factor (LIF) , oncostatin M, throm-
bopoietin (Tpo). Most of these receptors form multimeric complexes; several
chains have been cloned for the receptors for IL-2 to IL-7, for the GM-CSF
and the LIF receptors. The 66kDa chain cloned in the EpoR is respon-
sible for intracellular signalling, since the transfection of this protein in
hematopoietic cells such as Ba/F3, 32D or DA3, which do not possess EpoR
at their cell surface, allows their proliferation in response to Epo alone
(Gobert et al. 1995b). However, chemical cross-linking experiments with
125iodine-Iabelled Epo have detected at the surface of erythroid progenitors
two additional proteins of 85 and 100kDa, respectively, which are not rec-
ognized by anti p66 antibodies (Mayeux et al. 1991). These proteins proba-
bly are accessory proteins belonging to the receptor complex but not able to
bind to the ligand. Their cloning will be required to better understand their
specific role.
The fixation of Epo on its cognate receptor leads to dimerization of
the p66 EpoR as shown by cristallization of the complex Epo/EpoR (Syed
et al. 1998). One EpoR molecule binds to the ligand with a high affinity
(Kd = 1 nM), the second receptor molecule binds to the complex with a lower
affinity (Kd = 111M). In fact, it was demonstrated by crystallographic studies
that the EpoR existed as unliganded dimers on the cell surface and that Epo
triggered a switch between a self-associated inactive conformation and an
active, ligand-bound conformation (Livnah et al. 1999; Remy et al. 1999).
Thus, Epo is a bivalent molecule for the fixation of the EpoR. This cytokine
is composed of four amphipathic a-helical bundles (Boissel et al. 1993). The
mapping of the active sites of Epo has been achieved; these two sites are com-
posed of amino acids, which are spatially close but can be more distant in the
primary sequence of Epo. The first high-affinity site is located in the vicinity
of the A-B loop and the A and D helices. This site is characterized by a central
hydrophobic binding pocket, flanked at opposite ends by hydrophilic inter-
actions (Syed et al. 1998). The EpoR is bound to Epo through several amino
acids among which the residue Phe93 is essential as already determined by
mutations in the EpoR molecule (Middleton et al. 1996). The second site
in the Epo molecule corresponds to residues in the helices A and C (Elliott
 Biology of erythropoietin 7

et al. 1997; Matthews et al. 1996; Wen et al. 1994). The Epo/EpoR interac-
tions are not numerous at this second site; however, Phe 93 still plays an
important role (Middleton et al. 1999), as well as two Epo residues Arg 14
and Arg 103 whose mutations reduce the site affinity and inhibit the Epo
binding to the second EpoR molecule for Arg 103 (Matthews et al. 1996; Wen
et al. 1994).

5 Epo-induced intracellular signalling

As we mentioned above, the conformational change of the EpoR dimer after

binding to one molecule of Epo leads to its activation and subsequent down-
stream intracellular signalling. Many groups showed that the Epo-induced
activation led to the rapid tyrosine phosphorylation of a number of proteins,
even though the EpoR does not possess endogenous tyrosine kinase activity.
The two Jak2 tyrosine kinase molecules, which are each pre-associated to the
EpoR, are positioned in sufficient proximity for their reciprocal transphos-
phorylation and activation (Remy et al. 1999; Witthuhn et al. 1993). Activated
Jak2 proteins in turn phosphorylate the EpoR tyrosine residues (Dusanter-
Fourt et al. 1992, 1994). These phosphorylated tyrosines become secondary
binding sites for signalling proteins containing SH2 (SRC homology 2)
domains. Thus, a complex of signalling proteins is generated around the
dimerized and activated receptor.

Signalling pathways activated by Epa

The PI 3-kinaselAkt pathway

PI 3-kinase is associated to EpoR in response to Epo stimulation (Mayeux

1993). It was first reported that one SH2 motif of the PI 3-kinase p85 subunit
was bound to the last tyrosine residue of the EpoR (Damen et al. 1995).
Other mechanims of PI 3-kinase activation have since been described: two
adaptor proteins, IRS2 (Verdier 1997) and GAB1 (Lecoq-Lafon et al. 1999)
are phosphorylated following Epo stimulation and associate with PI 3-
kinase; Ly294002, a specific inhibitor of PI 3-kinase inhibits Epo-induced
cell proliferation, thereby suggesting that the PI 3-kinase pathway plays
an important role in the mode of action of Epo. PI 3, 4, 5 trisphosphate,
a metabolite of the PI 3-kinase pathway, activates the serinelthreonine
kinase AKT which is known to playa major role in the inhibition of cellular
apoptosis (Franke et al. 1997) and in cell proliferation. In summary, the
activation of PI 3-kinase in response to Epo stimulation is an important
event, leading to the inhibition of apoptosis of erythroid progenitors and to
their proliferation.
8 C. Lacombe and P. Mayeux

D
Vav
Myc
B II + - - Jak2 as. ociation
and activation
IR -2 B2

y l.l3 VLD YI + - -
I tat 5,
GAB 112

y 401 TILD Y2+-- tat 5, GAB 1/_

SHP-2, I . • S3

y 429 LYLV Y3 +--/ tat 5 (+1-)

y 431 LVV Y4 ~HP- I
tat 5 (+1-)
y -J.I3 S GG Y5
y 460 SHPY Y6
y 464 E SL Y7+-- anonical Grb2
binding site

y H9 V Y8+--PI -kinase

Fig. 2. Identified binding and activation sites for signaling proteins in the intracellu-
lar domain of the EpoR. Bl and B2 represent Boxl and Box2. The eight tyrosine
residues of the EpoR are indicated together with the peptidic sequences following
each tyrosine residue

The Ras/MAP kinase pathway

Ras, Raf and MAP kinase proteins are all activated by Epo (Gobert et al.
1995a). The adaptor proteins SHe and Grb2 are associated to the EpoR,
together with the tyrosine phosphatase SHP-2, which is also able to bind
Grb2. The Ras/MAP kinase pathway could be activated by Epo via several
different mechanisms; however, these mechanisms have not been definitively
identified. This pathway is also involved in Epo-induced cell proliferation
(Damen and Krystal 1996).

The STAT pathway

The STAT proteins (Signal Transducer and Activator of Transcription) are

transcription factors activated in response to several cytokines (Ihle 1995).
 Biology of erythropoietin 9

socs-. I
SO S-3? - -

SHP-.-----

1
Internalization
Degradation

Fig. 3. Down-regulation mechanisms of EpoR intracellular signaling. Only the

down-regulation mechanisms directly involving the EpoR are presented in this figure.
Deactivation mechanisms specifically targeting each downstream signaling pathway
such as STATS, PI3-K and Ras-MAP kinase have also been reported and are not
presented

Epo activates the two isoforms of STAT5, STAT5a and STAT5b (Gouilleux
et al. 1995; Pallard et al. 1995). The STAT5 proteins bind to the Tyr 343 and
401 of the EpoR, they become phosphorylated and activated and they
translocate into the nucleus (Gobert et al. 1996). The role of STAT5 tran-
scription factors during Epo stimulation has been a matter of debate: some
reports established a correlation between STAT5 activation and cell pro-
liferation (Chretien et al. 1996), whereas others attributed a role for STAT5
in erythroid differentiation (Wakao et al. 1997). A double knock out for
Stat5a and Stat5b genes did not lead to any major defect of erythropoiesis
(Teglund et al. 1998). It was further shown that Stat5 was essential for the
high erythropoietic rate during fetal development because it bound to the
promoter of the Bcl-X gene and played a crucial role in EpoR antiapoptotic
signaling (Socolovsky et al. 1999).

Pathways leading to signalling arrest

The signal of activation, which results from stimulation by a cytokine, needs

to be terminated by additional pathways leading to signal interruption.
10 C. Lacombe and P. Mayeux

Several proteins have been recently isolated that playa negative role in Epo-
induced signal transduction.

The tyrosine phosphatase SHP-1

SHP-1 protein is involved in Epo signalling, this protein binds both to the
third Tyr of the EpoR and to Jak2 tyrosine kinase. This association leads to
Jak2 dephosphorylation and thus to the signal arrest (Klingmtiller et al. 1995).
Several cases of familial erythrocytosis have been reported, due to a trunca-
tion of the cytoplasmic domain of the EpoR, which lacked the binding site
for SHP-1 protein and thereby became hypersensitive to Epo stimulation in
vivo (De La Chapelle et al. 1993; Kralovics et al. 1997). The erythroid pro-
genitors derived from mice knocked out for the Shp-1 gene are also hyper-
sensitive to Epo (Van Zant and Shultz 1989).

The Cis and SOCS3 proteins

The Cis protein (for cytokine-inducible SRC homology 2-containing protein)

is one of the known targets of STA5 factor (Yoshimura et al. 1995). Cis is an
inhibitor of Epo-induced cell proliferation. We recently showed that Cis was
associated to the Tyr 401 of the EpoR and was ubiquitinated. The ubiquiti-
nation of Cis suggested that this protein could play an active role in the reg-
ulation of the Epo/EpoR complex by the proteasome (Verdier et al. 1998).
Indeed, we showed that the proteasome controlled the down-regulation of
EpoR in Epo-stimulated cells by inhibiting the cell surface replacement of
internalized EpoR (Verdier et al. 2000). A second member of the Cis family
called SOCS3 (for Suppressor Of Cytokine Signaling) has been described as
essential in the regulation of erythropoiesis. SOCS3 gene disruption results
in an embryonic lethality associated with marked erythrocytosis, and in con-
trast, enforced expression of SOCS3 in vivo specifically suppresses fetal liver
erythropoiesis, thereby showing that SOCS3 is critical in negatively regulat-
ing fetal liver erythropoiesis (Marine et al. 1999). Like Cis, SOCS3 binds to
Tyr 401 of the EpoR. The mechanism of action of this family of signaling
inhibitor protein is not completely elucidated. SOCS proteins recognize acti-
vated JAK kinases and cytokine receptors, and it is likely that the SOCS pro-
teins act as adaptor molecules that target activated signaling proteins to the
protein degradation pathway via the proteasome.

6 Epo mimetics

In vitro or in vivo experimental models aiming at dimerizing the EpoR have

shown that this dimerization was followed by the activation of the EpoR even