Erythropoietin

Jerry L. Spivak*
Department of Medicine, Johns Hopkins University School of Medicine,
Traylor 924 720 Rutland Avenue, Baltimore, MD 21205, USA
* corresponding author tel: 410-955-5454, fax: 410-955-0185, e-mail: jlspivak@mail.jhu.edu
DOI: 10.1006/rwcy.2000.09007.

SUMMARY putative humoral mediator was named erythropoie-

tin, as suggested previously (Grant and Root, 1947).
This chapter reviews current knowledge about With scientific interest now rekindled, it was soon
erythropoietin from the level of its gene to the appli- established that the kidneys were a major site of
cation of its recombinant version for the correction of erythropoietin production (Jacobson et al., 1957). An
anemia with special emphasis on the molecular bio- in vivo bioassay was developed and efforts were
logy, physiology and pharmacology of the protein, its initiated to purify erythropoietin, an accomplishment
tissue-specific expression and regulation of that which proved exceedingly difficult owing to the
expression in health and disease. minute quantities of the hormone normally present in
the plasma and urine. Indeed, the initial biochemical
purification of erythropoietin required 2550 L of
urine from extremely anemic individuals (Miyake
BACKGROUND et al., 1977). Once purified, it was possible to raise
antibodies to erythropoietin and thereby develop an
Discovery immunoassay sufficiently sensitive and specific, in
contrast to the existing bioassays, to define the
A relationship between hypoxia and increased physiology of the hormone (Egrie et al., 1987). With
erythropoiesis was first recognized in 1863 and by development of recombinant DNA technology,
the end of the nineteenth century not only were the knowledge of the amino acid sequence of erythro-
kinetics of this relationship well established, but a poietin permitted cloning of its gene (Jacobs et al.,
hypothesis linking enhanced erythropoiesis with 1985; Lin et al., 1985).
marrow hypoxia had been developed (Erslev, 1953).
In 1906, Carnot and Deflandre proposed an alter-
native hypothesis that hypoxia enhanced erythropoi- Alternative names
esis not by a direct effect on the bone marrow but
rather indirectly through a humoral mechanism. This EPO. The two commercially available erythropoietin
hypothesis was attractive but could not be substan- preparations have been designated Epoietin  and
tiated by other investigators using the Carnot and Epoietin .
Deflandre protocol and lay dormant until 1950. At
this time, Kurt Reissmann, using parabiotic rats,
conclusively demonstrated that hypoxic induction of Structure
erythropoiesis was under humoral control (Reissmann,
1950). For technical reasons, Reissmann's demonstra- This is a globular protein (predicted radius 20.2 AÊ)
tion was indirect and it remained for Erslev to with an observed Stokes radius of 32 AÊ due to its
demonstrate directly and conclusively using large carbohydrate residues (Davis and Arikawa, 1987),
quantities of anemic rabbit plasma that Carnot and constituted as a four antiparallel  helical bundle with
Deflandre were indeed correct in their intuition if not one short and two long intervening loops and several
their experimental methods (Erslev, 1953). The stretches of  sheets (Elliott et al., 1997), a structure
942 Jerry L. Spivak

which is characteristic of hematopoietic growth fac- et al., 1986; Powell et al., 1986). There are no docu-
tors in general, even though they share little amino mented relevant linkages, mutations or related genes.
acid sequence homology (Bazan, 1990a). The protein
has two independent binding sites for its receptor,
which differ in their binding affinity (Philo et al., Regulatory sites and corresponding
1996; Elliott et al., 1997). transcription factors
A GATA sequence is present 50 to the cap site, but the
Main activities and putative promoter region of the erythropoietin gene
pathophysiological roles does not contain the TATA or CAAT sequences typi-
cal of a classical promoter (Shoemaker and Mitsock,
Erythropoietin is a member of the hematopoietic 1986). Rather, the erythropoietin gene contains a 50
growth factor family and is one of the few hema- nucleotide enhancer sequence in its immediate 30
topoietic growth factors which behaves like a flanking sequence, a liver inducible element mapping
hormone. The principal function of erythropoietin is between 0.4 kb 50 to the gene and 0.7 kb 30 , a negative
to couple oxygen delivery by circulating red cells to regulatory element between 0.4 and 6 kb 50 to the gene
long-term tissue oxygen needs. Produced primarily in which represses expression in most tissues, and a
the kidneys and to a small extent in the liver in adults, region between 6 and 14 kb 50 which controls renal
erythropoietin interacts in the bone marrow with expression (Semenza et al., 1991; Semenza and Wang,
specific receptors on the surface of erythroid pro- 1992). Adding to this level of complexity is the
genitor cells to initiate their entry into cell cycle if existence of multiple transcription initiation sites
dormant (Spivak et al., 1991, 1996) or to maintain which appear to be differentially utilized in a tissue-
their viability while differentiating, if they are already specific fashion.
in active cell cycle (Koury and Bondurant, 1990). In addition to hypoxia-induced transactivating
Erythropoietin achieves its effects by causing homo- factors that upregulate erythropoietin gene transcrip-
dimerization of its receptor with the resultant auto- tion, factors that can negatively regulate erythropoie-
phosphorylation of the tyrosine kinase JAK2 tin production have also been described. These
(Witthuhn et al., 1993) and phosphorylation of the include an as yet unidentified 47 kDa nuclear protein
receptor itself, as well as various substrate proteins and a ribonucleoprotein (Beru et al., 1990) and the
leading to the upregulation of a number of signaling transcription factors GATA-1, -2, and -3 (Imagawa
pathways and the activation of gene transcription et al., 1997). The von Hippel-Lindau (VHL) gene
(Damen and Krystal, 1996). In keeping with its high product may also be a negative regulator of erythro-
level of evolutionary conservation and its vital role in poietin production under normoxic conditions (Krieg
maintaining an adequate supply of red cells for et al., 1998). Finally, in contrast to the genes for other
oxygen transport, no mutant erythropoietin molecule hematopoietic growth factors such as IL-3, GM-CSF,
has ever been described and antibody production to and G-CSF, the erythropoietin gene lacks the 30
endogenous erythropoietin or its exogenous recombi- ATTA sequences that enhance mRNA degradation
nant counterpart is exceedingly rare (Casadevall et al., (Shaw and Kamen, 1986). This may be a consequence
1996). of its oxygen-dependent expression.

GENE AND GENE REGULATION Cells and tissues that express

the gene
Accession numbers
Cloning of the erythropoietin gene provided an
Gene: M11319, X02158 opportunity to define its tissue-specific expression
cDNA: X02157 formally. In the normoxic state, erythropoietin
mRNA was detected in the kidneys, liver, lung,
testes, and brain (both infra- and supratentorial) (Tan
Chromosome location et al., 1991; Digicaylioglu et al., 1995). In the setting
of hypoxia, erythropoietin mRNA was upregulated in
The human erythropoietin gene is present as a single kidney, liver, testes, and brain and become detectable
copy on chromosome 7(q11-q22) and is composed of in the spleen. Using RT-PCR analysis, it has also
five exons (582 bp) and four introns (1562 bp) (Law been possible to study the ontogeny of erythropoietin
 Erythropoietin 943

gene expression. Erythropoietin mRNA was detect- radius of 20.2 AÊ with an observed Stokes radius of
able in embryonic stem cells before their differentia- 32 AÊ, presumably due to its carbohydrate side chains,
tion together with erythropoietin receptor mRNA heterogeneity with respect to its pI (4.2±4.6) owing to
(Schmitt et al., 1991; Keller et al., 1993). Thus, variable degrees of sialation, and a net negative
erythropoietin gene expression occurs before the charge at physiologic pH also due to its sialic acid
formation of the kidneys or liver and even before the residues (Davis and Arakawa, 1987). The pI of the
expression of multipotent growth factors such as aglycone is 9.2 and the protein becomes destabilized
IL-1, IL-3, and GM-CSF (Schmitt et al., 1991). below its isoelectric point, presumably as a result of
To date, only two cell lines, HepG2 (HB-8065) and the loss of sialic acid residues. In spite of its acidic pI,
Hep3B (HB-8064), derived from human hepatomas, fluorescent quenching studies indicate the protein
have been documented to produce erythropoietin interacts better with negatively charged ions, suggest-
in vitro (Goldberg et al., 1987). Recently, data have ing that the net positive charge of the protein as
been obtained suggesting that erythropoietin is also opposed to the net negative charge of the whole
produced by erythroid burst-forming units (BFU-E) molecule is dominant in ionic interactions (Davis and
and might be upregulated by erythropoietin itself Arakawa, 1987), a behavior which is consonant with
(Stopka et al., 1998b). respect to the interaction of erythropoietin with its
receptor.
Erythropoietin is a hydrophobic protein, a char-
PROTEIN acteristic which is also important with regard to its
ligand±receptor interactions. While human erythro-
Accession numbers poietin contains four cysteine residues (at positions 7,
29, 33, and 161), which form two internal disulfide
PO1588 bonds, cysteine 33 is replaced by proline in murine
erythropoietin (Wen et al., 1993). The disulfide bridge
between cysteine 7 and 161 is, however, necessary for
Sequence biological activity since alkylation of these residues or
their replacement through site-directed mutagenesis
See Figure 1. inactivated the protein under physiologic circum-
stances. Normally this internal disulfide bond is not
exposed, since erythropoietin in solution is not inacti-
Description of protein vated by reducing agents. Internal disulfide bonds are
also a feature of GM-CSF, G-CSF, and IL-3.
The erythropoietin gene codes for a 193 amino acid Early studies indicated that iodination of more
protein, of which 27 serve as the signal peptide, while than one of erythropoietin's four tyrosine residues,
the remaining 166 constitute the native protein. occupation of its lysine residues, or alteration of their
However, based on studies employing fast atom charge also inactivated erythropoietin. With respect
bombardment mass spectrometry and peptide map- to tyrosine iodination, this might have been due to
ping, the terminal amino acid, arginine 166, is absent steric hindrance since site-directed mutagenesis at
not only from recombinant erythropoietin produced tyrosine residues 49 and 145 did not abolish biologic
by Chinese hamster ovary cells but also from human activity, while only nonaromatic substitutions for
urine erythropoietin, the only native form of erythro- tyrosine 15 or 156 affected biologic activity (Elliott
poietin purified to date (Recny et al., 1987). The site et al., 1997). Futhermore, erythropoietin can be iodi-
for this posttranslational modification has not been nated without loss of biological activity. The net
identified, nor is its biologic significance understood. charge of erythropoietin is positive and neutralization
Physicochemical studies indicate that erythropoie- of certain basic amino acids such as lysine impairs
tin has a high  helix content, a calculated Stokes receptor-ligand binding (Elliott et al., 1997). Indeed,

Figure 1 Amino acid sequence for erythropoietin.

MGVHECPAWL WLLLSLLSLP LGLPVLGAPP RLICDSRVLE RYLLEAKEAE

NITTGCAEHC SLNENITVPD TKVNFYAWKR MEVGQQAVEV WQGLALLEAV
LRGQALLVNS SQPWEPLQLH VDKAVAGLRS LTTLLRALGA QKEAISPPDA
ASAAPLRTIT ADTFRKLFRV YSNFLRGKLK LYTGEACRTG DR
944 Jerry L. Spivak

electrostatic interactions between hematopoietic crystallography at 1.90 AÊ of an erythropoietin agly-

growth factors and their receptors appear to be an cone±erythropoietin receptor extracellular ligand-
essential factor in ligand binding (Wlodawer et al., binding domain complex has recently been reported
1993, Demchuk et al., 1994). (Syed et al., 1998). The crystallized ligand±receptor
Mapping studies using site-directed mutagenesis complex confirmed studies employing mutational
and monoclonal antibodies have confirmed secondary analysis, epitope mapping and ligand±receptor cross-
and tertiary structure predictions based on computer linking, indicating that one erythropoietin molecule
modeling as well as X-ray diffraction and NMR anal- binds to two molecules of its cognate receptor. Thus,
ysis of other hematopoietic growth factors (De Vos erythropoietin was found to interact with its receptor
et al., 1992; Boissel et al., 1993; Elliott et al., 1996, at two sites with both hydrophobic and hydrophilic
1997). Indeed, in spite of their different target cells interactions. Phenylalanine 93 of the erythropoietin
and cognate receptors, hematopoietic growth factors receptor was the dominant residue with respect to
as well as growth hormone and prolactin share in hydrophobic interactions and mutations of this resi-
common a four  helical antiparallel bundle structure due resulted in loss of erythropoietin binding
with long crossover loops between the first two and (Middleton et al., 1996). As predicated by epitope
last two helices. Erythropoietin follows this pattern, mapping, the charge interactions between erythro-
containing long crossover loops between the A and B poietin and its receptor involve lysine and arginine
and C and D helices, and a short loop between B and resides of the former and aspartate and glutamate
C. There are also two short antiparallel  strands residues of the latter (Syed et al., 1998). The most
formed by the long AB and CD crossover loops (Syed important erythropoietin residues with respect to its
et al., 1998). Epitope mapping and site-directed interaction with its receptor appear to be glycine 151,
mutagenesis revealed four regions of the protein which which dictates the appropriate conformation of the D
are important for biological activity: amino acid resi- helix and the arginine residues 14 and 103.
dues 11±15 in helix A; 44±51 in the A±B connecting
loop; 100±108 in helix C and 147±151 in helix D
(Matthews et al., 1996; Elliott et al., 1997). These four Important homologies
regions map to two sites on erythropoietin based on
their effects on receptor binding and biological activ- Erythropoietin is highly active across mammalian
ity. Mutations of the amino acids comprising site 1 species, a behavior which correlates well with the high
impair both receptor binding and bioactivity, while degree of homology amongst the cloned human,
mutations in site 2 have lesser effects on receptor simian, and rodent genes (McDonald et al., 1986;
binding but greater effects on bioactivity. Shoemaker and Mitsock, 1986). Whether analyzed at
These observations support a model of ligand± the level of the gene or its protein, with one exception
receptor binding where erythropoietin initially binds erythropoietin is highly conserved amongst all mam-
to one erythropoietin receptor with high affinity malian species studied to date (Wen et al., 1993). For
(Kd=1 nM) and subsequently to a second receptor example, the human erythropoietin gene is 91%
molecule with a lower affinity (Kd ˆ 1 mM), causing identical to monkey erythropoietin, 85% identical to
receptor dimerization. Site 1, the high-affinity site, is cat and dog erythropoietin, and 80±82% identical to
composed of portions of the A, B, and D helices and pig, sheep, and rat erythropoietin. Interestingly, while
the AB loop while site 2, the low-affinity site, involves mouse and human erythropoietin stimulate guinea
portions of the A and C helices (Elliott et al., 1996, pig erythroid progenitor cells, guinea pig erythro-
1997). Physiochemical analysis using soluble erythro- poietin did not stimulate mouse or human erythro-
poietin receptors containing only its extracellular poietin progentor cells in vitro, suggesting loss of
domain (Philo et al., 1996) support this model, as epitope homology between guinea pig erythropoietin
have chemical crosslinking studies using soluble full- and other mammalian erythropoietins (Stopka et al.,
length erythropoietin receptors (Avedissian et al., 1998a). With respect to gene conservation, there are
1995) and studies using activating monoclonal recep- extensive regions of sequence homology between
tor antibodies (Elliott et al., 1996; Schneider et al., human and mouse erythropoietin genes upstream of
1997). the cap site and in the 30 UTR as well as in the first
intron, together with a high frequency of Alu se-
quences (or their murine homologs) in the 50 region
Discussion of crystal structure and within the transcription unit (Shoemaker and
Mitsock, 1986; Galson et al., 1993).
Although the native erythropoietin molecule has Erythropoietin has no significant homology to
not been crystallized nor studied by NMR, any other known protein except for thrombopoietin,
 Erythropoietin 945

the hematopoietic growth factor which regulates erythropoietin exposes the penultimate galactose
megakaryocyte and platelet production as well as residues which are recognized by hepatic galactosyl
stimulation of primitive, uncommitted hematopoi- receptors and desialated erythroporetin is rapidly
etic progenitor cells (Bartley et al., 1994). cleared from the plasma and degraded in hepatocytes.
Thrombopoietin can be divided into two domains Oxidation of the exposed galactose residues restored
on the basis of its amino acid sequence; the N- the ability of the hormone to remain in the circulation
terminal domain of 155 resides has 21% sequence (Spivak and Hogans, 1989). There is no evidence that
identity and overall a 46% sequence similarity to the sialic acid residues serve another purpose.
erythropoietin. It is this domain which binds to the To date, no role has been defined for the single
thrombopoietin receptor. O-glycosylation site of erythropoietin. Murine ery-
thropoietin has a proline at amino acid 126 rather
than serine, suggesting that O-glycosylation per se
Posttranslational modifications does not have a physiologic role (Wen et al., 1993).
Supporting this contention is the observation that
Based on its amino acid composition, the estimated erythropoietin produced in cells defective in
molecular weight of erythropoietin is 18,398, while O-glycosylation behaved identically in vivo to
based on its behavior during sedimentation equili- O-glycosylated erythropoietin (Wasley et al., 1991).
brium, its apparent molecular weight is 30,400 (Davis Studies employing enzymatic O-deglycosylation have
and Arakawa, 1987). The difference between the yielded similar results (Higuchi et al., 1992). Although
estimated and apparent molecular weights is due to one study employing site-directed mutagenesis sug-
glycosylation and, as might be expected from the gested that O-glycosylation was essential for bio-
extent of glycosylation, this is the most important synthesis and secretion of the hormone (Dube et al.,
posttranslational modification of the protein. 1988), other studies also employing site-directed
Erythropoietin contains one O-linked (serine 126) mutagenesis indicated that this result was an artifact
and three N-linked (asparagine 24, 38, and 83) gly- due to the amino acid (glycine) chosen to replace
cosylation sites and the latter are invariant amongst serine at site 126. A similar result was obtained with
mammalian erythropoietins. The N-linked carbohy- alanine but not with valine, threonine, histidine, or
drates consist primarily of tetra-antennary saccha- glutamic acid (Delorme et al., 1992). Since serine 126
rides with or without one or more N-acetyl occurs in the long loop between the C and D helices of
lactosaminyl repeats and lesser quantities of bian- erythropoietin which is not near its active sites, it is
tennary and triantennary saccharides, the latter of likely that the alanine or glycine substitutions altered
which may also contain N-acetyl lactosaminyl protein conformation and were not truly neutral
repeats. All the N-linked saccharides are sialated, as substitutions. At the same time, it is possible that the
is the O-linked saccharide. On a mole/mole basis, the O-glycosylation site serves to stabilize the conforma-
carbohydrate composition of human urinary erythro- tion of the protein, a role served by proline in murine
poietin consists of fucose (2.9), mannose (9.2), erythropoietin.
galactose (12.9), N-acetylglucosamine (16.3), N- In contrast to O-glycosylation, the highly con-
acetylgalactosamine (0.9), and N-acetylneuraminic served N-linked saccharides of erythropoietin are
acid (10.4) (Sasaki et al., 1987). With the exception of required for its secretion and in vivo biologic activity.
the type of sialic acid linkage, the glycosylation of With respect to secretion, glycosylation at either site
human recombinant erythropoietin is similar to 38 or 83 was sufficient but glycosylation at site 24 was
human urinary erythropoietin. not (Delorme et al., 1992), while complete loss of N-
The erythropoietin aglycone, produced in linked glycosylation markedly impaired secretion of
Escherichia coli or by enzymatic deglycosylation, has erythropoietin but incomplete glycosylation did not
a conformation similar to fully glycosylated erythro- (Yamaguchi et al., 1991). Thus, erythropoietin pro-
poietin as defined by circular dichroism (Narhi et al., duced in insect cells or certain mammalian cells which
1991). Thus, glycosylation is not required for proper lack the capacity for complex glycosylation secreted
folding of the protein or even refolding after dena- erythropoietin in a comparable fashion to mamma-
turation by a chaotrophic agent. However, glycosyla- lian cells which have this capacity (Wojchowski et al.,
tion does appear important for stabilizing the tertiary 1987; Wasley et al., 1991). Glycosylation of erythro-
structure of the protein under a variety of denaturing poietin was also required for biologic activity in vivo
conditions. and, in contrast to the differential effects on secretion,
The sialic acid residues of erythropoietin, in keep- each N-linked site appeared to contribute equally to
ing with their abundance, are essential for maintain- the in vivo activity of the hormone. This is presumably
ing erythropoietin in the circulation. Desialation of because their sialation prevents rapid degradation of
946 Jerry L. Spivak

circulating erythropoietin since elimination or oxida- With hypoxia, erythropoietin mRNA was not only
tion of the penultimate galactose residues also stabi- upregulated in these organs but became detectable in
lized the hormone in the circulation (Spivak and the spleen as well (Tan et al., 1992). Erythropoietin
Hogans, 1989). mRNA synthesis in an oxygen-dependent manner has
The erythropoietin aglycone produced in E. coli is also been demonstrated in murine, primate, and
said to have in vitro biologic activity but this has been human brain in a broad distribution (Digicaylioglu
incompletely quantified, while studies employing et al., 1995) and erythropoietin has been identified in
enzymatically deglycosylated erythropoietin appear the cerebrospinal fluid (Marti et al., 1997). In mice,
to be confusing in this regard because of the definition astrocytes were identified as a site of erythropoietin
of the extent of deglycosylation. Based on molecular production (Masuda et al., 1994). Since neurons
weight determinations, fully deglycosylated but express erythropoietin receptors (Morishita et al.,
otherwise intact eukaryotic erythropoietin has very 1997) erythropoietin may have a role in maintaining
little in vitro biologic activity (Takeuchi et al., 1990). neuronal viability under conditions of hypoxic stress
This is probably a consequence of a loss of stability or (Sakanaka et al., 1998).
a change in conformation of the protein since the N-
linked glycosylation sites are not near its putative
active sites. Thus, the saccharides of erythropoietin Eliciting and inhibitory stimuli,
appear to be necessary to sustain the physical half-life
of the hormone in the circulation and to maintain it in
including exogenous and
a biologically active conformation. Specific carbohy- endogenous modulators
drate structural motifs on erythropoietin may also
have roles in determining its circulatory residence Hypoxia is the only physiological stimulus for
time and rate of elimination. For example, it has been erythropoietin production (Reissmann, 1950). In the
demonstrated that erythropoietin molecules rich in kidneys, erythropoietin production is constitutive and
biantennary oligosaccharides not only have lower maximal in each cell. With hypoxia, additional renal
in vivo biologic activity, but are also preferentially interstitial fibroblasts are recruited to produce it in a
cleared by the kidney (Misaizu et al., 1995). watershed fashion (Koury et al., 1989). By contrast,
Furthermore, erythropoietin molecules containing in the liver, erythropoietin production occurs in every
three N-acetyl lactosamine repeats are preferentially hepatocyte, is most marked in hepatocytes closest to
cleared from the circulation by hepatocyte galactose central veins and with increasing hypoxia can be
receptors (Fukuda et al., 1989). Thus, there appear to upregulated in individual hepatocytes (Koury et al.,
be mechanisms operative in vivo which control the 1991). The characteristics of erythropoietin produc-
specific forms of erythropoietin that are physically tion in hepatic interstitial fibroblasts have not yet
available to the bone marrow. been investigated.
During gestation, the liver is the major site of
erythropoietin production and the switch to renal
dominance appears to occur after birth (Zanjani et al.,
CELLULAR SOURCES AND 1981; Dame et al., 1998). In adults, under both
TISSUE EXPRESSION normoxic and hypoxic conditions, the kidneys are the
major site of erythropoietin production. This is in
Cellular sources that produce part due to a differential sensitivity to hypoxia in this
organ as compared to the liver. Thus, with mild
Early studies of erythropoietin production identified hypoxia, the increase in erythropoietin mRNA is
the kidneys and the liver as major sites for this. In the much greater in the kidneys than liver and only when
kidneys, erythropoietin is produced in peritubular hypoxia becomes severe is hepatic synthesis of the
fibroblastoid cells in the inner cortex and outer hormone substantial. Indeed, with severe hypoxia, the
medulla (Koury et al., 1988; Lacombe et al., 1988). In liver can account for over 35% of erythropoietin
the liver, erythropoietin is produced by hepatocytes mRNA production (Tan et al., 1991). With renal
and interstitial fibroblasts (Koury et al., 1991). With parenchymal damage, hepatic erythropoietin produc-
cloning of the erythropoietin gene, it was finally tion becomes even more dominant. However, it is
possible to establish definitively the sites of erythro- never sufficient to sustain erythropoiesis at an ade-
poietin production using more sensitive and specific quate level when renal function is severely impaired
techniques. Under normoxic conditions, expression of or in the anephric state. This appears to be a con-
erythropoietin mRNA was identified in the lung, sequence of the relative insensitivity of hepatocytes to
testes, and brain, in addition to the liver and kidneys. hypoxia, possibly because of the dual blood supply
 Erythropoietin 947

that the liver enjoys. Alternatively, it could be a not zinc, iron, calcium, or tin (Goldberg et al., 1988).
consequence of a difference between fetal and adult Cycloheximide inhibited cobalt, nickel and manga-
hepatocytes. nese-stimulated erythropoietin production, indicating
In this regard, it is well established that in patients that protein synthesis was required for transcriptional
with chronic renal failure amelioration of anemia activation of the erythropoietin gene. Because these
could occur in the setting of acute hepatitis. Indeed, in elements are interchangeable with iron in the por-
one such anephric patient with acute viral hepatitis, phyrin ring of heme and because their effects were not
plasma erythropoietin rose and the hematocrit was additive with respect to themselves or hypoxia, it
restored to normal transiently but subsequently both appeared that a heme protein served as the oxygen
fell to their previous levels, even though chemical sensor in erythropoietin-producing cells. Support for
evidence of hepatocellular injury was still present. this conclusion was provided by the observation that
(Klassen and Spivak, 1990). This suggested that either carbon monoxide inhibited erythropoietin production
erythropoietin production occurred as part of a by hypoxic hepatoma cells but not by those exposed
recapitulation of ontogeny associated with hepatocyte to cobalt or nickel, to which carbon monoxide cannot
regeneration or, less likely, that hepatocyte injury was bind. Additionally, desferrioxamine, which inhibits
associated with a greater reduction in erythropoietin heme synthesis, also inhibited hypoxia-, cobalt- or
catabolism than production. nickel-stimulated erythropoietin synthesis (Goldberg
Chronic inflammatory or infectious disorders and et al., 1988). In this regard, it is of interest that iron
neoplasms are commonly associated with anemia. deficiency can potentiate the production of erythro-
With the development of a sensitive assay for poietin out of proportion to the degree of tissue
erythropoietin and the identification of the cells and hypoxia (Kling et al., 1996a), which is consonant with
cell lines which produce the hormone, it has been the observation that desferrioxamine induced ery-
possible to evaluate the mechanisms involved in the thropoietin gene expression (Wang and Semenza,
suppression of erythropoiesis by infection, inflamma- 1993).
tion, or neoplasia. The inflammatory cytokines, IL-1, The process involved in the induction of erythro-
TNF, and TGF inhibited erythropoietin produc- poietin production by hypoxia or cobalt appears to be
tion by Hep3B and Hep2G cells while IL-6 and IFN a universal mechanism for oxygen-regulated gene
did not (Faquin et al., 1992; Jelkmann et al., 1992). expression. To date, genes that are regulated in this
IL-1 also inhibited erythropoietin production by an fashion include vascular endothelial growth factor
isolated perfused kidney (Jelkmann et al., 1992). IL-6, (VEGF), heme oxygenase I, nitric oxide synthase, and
however, actually potentiated the effect of hypoxia on a variety of glycolytic enzymes (Ratcliffe et al., 1995).
Hep3B cell erythropoietin production. Endotoxin The oxygen-sensing heme protein has not been iden-
(Schade and Fried, 1976), IL-1, TNF, TGF, and tified to date, but a cytochrome b with NAD(P)H
IFN also inhibited the proliferation of erythroid oxidase activity is considered to be a putative candi-
progenitor cells in response to erythropoietin (Means date (Bunn and Poyton, 1996).
and Krantz, 1992), while IL-6 and its soluble receptor Identification of a cis-acting enhancer sequence in
were capable of stimulating erythroid progenitor cell the 30 UTR of the erythropoietin gene provided an
proliferation in the absence of erythropoietin (Sui opportunity to define the activators involved in
et al., 1996). oxygen-regulated gene transcription. When this cis-
Renal erythropoietin production can be stimulated acting enhancer sequence was transfected into cells,
in vivo or in vitro by exposure to cobalt or by it responded in an identical fashion in the presence
intrarenal injections of nickel but the mechanisms of cobalt or hypoxia to its endogenous counterpart
involved remained undefined because of lack of a in the intact erythropoietin gene (Wang and
model system to study the molecular regulation of Semenza, 1993). With this enhancer sequence, it was
erythropoietin production. With cloning of the possible to identify the transactivating factor
erythropoietin gene it was finally possible to identify (hypoxia-inducible factor 1, HIF-1) that was respon-
cells that produced erythropoietin in vitro in response sible for activation of erythropoietin gene transcrip-
to hypoxia. It is of interest that, to date, no renal cell tion (Wang et al., 1995a). As might be expected,
line has been identified that displays this property, renal innervation has no influence on oxygen-
but several hepatoma cell lines have been useful in mediated erythropoietin gene expression (Eckardt
defining the mechanism for hypoxia-induced erythro- et al., 1992). Among the antioxidant vitamins,
poietin production (Goldberg et al., 1987). vitamin A (Jelkmann et al., 1997) and its derivative,
Oxygen-sensitive hepatoma cell lines respond to retinoic acid (Okano et al., 1994) upregulate both
cobalt as well as hypoxia and also produce erythro- renal and hepatic erythropoietin expression in vitro
poietin after exposure to nickel and manganese but and in vivo.
948 Jerry L. Spivak

HIF-1 is a heterodimeric DNA binding complex (Schuster et al., 1987). While erythropoietin mRNA
involving two basic helix-loop-helix PAS proteins levels increase 50- to 100-fold in response to hypoxia,
(HIF-1 and HIF-1) (Wang et al., 1995a). HIF-1 is nuclear run-off studies indicated only a 5- to 10-fold
a novel, 120 kDa, nonheme iron-containing protein increase in the actual rate of transcription (Goldberg
(Srinivas et al., 1998), while HIF-1 is the 91±90 kDa et al., 1991). Resolution of this discrepancy was
product of the aryl hydrocarbon nuclear translocator provided by the observation that with hypoxia not
(ARNT) gene (Semenza et al., 1997). The HIF-1 only was erythropoietin gene transcription increased
heterodimer binds to a specific (TACGTGCT) 50 but there was also an increase in the stability of its
sequence in the 30 erythropoietin gene enhancer mRNA which was similar to that achieved under
element and to p300, a non-DNA binding transcrip- normoxic conditions with actinomycin or cyclohex-
tional activator through its  subunit. HNF-4, an imide (Goldberg et al., 1991). The hypoxia-mediated
orphan nuclear steroid receptor, that binds to two stability of erythropoietin mRNA appears to be due
random hormone-responsive elements in the 30 end of to the upregulation of an erythropoietin mRNA-
the enhancer, also binds to HIF-1, completing the binding protein (Rondon et al., 1991; McGary et al.,
transcriptional complex (Huang et al., 1997). The 1997) in association with a heat shock protein
expression of HIF-1 and  mRNAs is ubiquitous (Scandurro et al., 1997). Therefore, posttranscrip-
and their protein products are upregulated by tional events involving erythropoietin mRNA parallel
hypoxia, consistent with the universal role of this those of HIF-1 and have an important role in the
complex in modulating oxygen-regulated gene expres- regulation of erythropoietin production.
sion. Since the HIF-1 mRNAs are constitutively
expressed, it appears that regulation of HIF-1 levels
does not simply involve protein synthesis but also
posttranslational stabilization of these proteins RECEPTOR UTILIZATION
(Wenger et al., 1997). HIF-1 protein is present in
excess of HIF-1 which is constitutively synthesized Erythropoietin interacts with its target cells through a
and rapidly degraded by the ubiquitin±proteasome specific receptor that is a member of the hematopoie-
pathway (Huang et al., 1998). HIF-1 is stabilized tic growth factor receptor superfamily (Bazan, 1989,
by hypoxia or iron chelation and dimerizes with 1990b). The human erythropoietin receptor gene,
HIF-1 (Kallio et al., 1997). Sulfhydryl reduction is located on chromosome 19, consists of eight exons
also required for HIF-1 and  dimerization. and seven introns and codes for a 508 residue gly-
Importantly, dimerization is required not only for coprotein (Winkelmann et al., 1990) (507 in the
high-affinity DNA binding but also for the resistance mouse: D'Andrea et al., 1989). The promoter region
to proteolysis which permits HIF-1 accumulation. of the gene contains CACC, GATA-1 and SP-1-
Posttranslational modifications such as protein phos- binding sites and one inhibitory and two enhancer
phorylation (both serine/threonine and tyrosine) that regions 50 to the ATG intiation site. Erythroid-
appear to be required not only for DNA binding but specific expression of the gene appears to be provided
also protein synthesis confer another level of com- by elements both 50 and 30 to the promoter which
plexity on transcription factor complex behavior exert a negative effect on nonerythroid tissues and
(Wang et al., 1995b). Finally, induction of both HIF- elements within the first intron which appear to
1 activity and hypoxia-induced expression of ery- provide erythroid specificity (Youssoufian et al.,
thropoietin is blocked by inhibitors of RNA or 1993). The erythropoietin receptor is expressed not
protein synthesis. only by erythroid cells but also by embryonic stem
Under normoxic conditions there is a low level of cells (Schmitt et al., 1991), endothelial cells
constitutive erythropoietin gene expression in hepa- (Anagnostou et al., 1990), and neural cells
toma cell lines that is not suppressible by increasing (Morishita et al., 1997).
tissue oxygenation. The half-life of erythropoietin The receptor has a 24 residue signal peptide, a 224
mRNA is normally approximately 2 hours but is amino acid extracellular domain, a single 24 amino
increased 3-fold in the presence of actinomycin D or acid membrane-spanning domain, and a 236 amino
cycloheximide, suggesting that a short-lived protein acid cytoplasmic domain (D'Andrea et al., 1989).
was involved in its degradation (Goldberg et al., There are 11 cysteine residues but no disulfide
1991). With hypoxia, erythropoietin mRNA tran- bridges and only one N-linked glycosylation site.
scription was initiated within 30 minutes and Based on its amino acid composition, the molecular
increased exponentially, reached a maximal level by weight of the erythropoietin receptor is 55,000 but
6 hours and eventually leveled off or declined to after posttranslational processing apparent molecular
approximately 50% of the maximum level achieved weights as high as 78 kDa have been observed for
 Erythropoietin 949

membrane-bound receptor (Sawyer and Hankins, Bioassays used

1993). The posttranslational modifications include
glycosylation and tyrosine and serine-threonine phos- The standard in vivo assay for erythropoietin is the ex-
phorylation. In general, the form of the receptor hypoxic polycythemic mouse assay which is relatively
present in the plasma membrane has a higher insensitive (lower limit of detection ˆ 0.05 U/mL),
molecular weight than its cytoplasmic counterpart, cumbersome, time-consuming, and expensive. In vitro
presumably due to phosphorylation. The isoelectric assays for the hormone are more sensitive and specific
point of a purified baculovirus-expressed human and there are a variety of techniques from which
erythropoietin receptor is 5.6 (Avedissian and Spivak, to choose that differ with respect to their end points.
1995). The erythropoietin receptor is a member of the The mouse spleen bioassay is based on the incor-
hematopoietic growth factor receptor superfamily poration of labeled thymidine into the DNA of mouse
whose other members include the receptors for IL-2 splenic erythroblasts after exposure to erythropoietin
(chain), IL-3, IL-4, IL-5, IL-6, IL-7, GM-CSF, G- (Krystal, 1983). In vitro clonal assays using freshly
CSF, thrombopoietin, leukemia-inhibitory factor explanted marrow or peripheral blood mononuclear
(LIF), ciliary neutrotropic factor (CNTF), oncostatin cells are based on the number of erythroid colonies
M, growth hormone, and prolactin (Bazan, 1989, which form in the presence of erythropoietin. These
1990b). They share in common four positionally colony-forming assays which can employ either
conserved cysteines in their extracellular domain as human or animal cells require strict attention to the
well as a tryptophan-serine-X-tryptophan-serine components of the culture medium to ensure repro-
(WSXWS) motif or its homolog located near the ducibility (Spivak and Seiber, 1983). Erythropoietin-
transmembrane region, and each lacks kinase motifs dependent cell lines are also useful for bioassay
in their intracellular domain. purposes and are, perhaps, the simplest, most direct
method for the in vitro assay of erythropoietin. With
the development of recombinant reagents, in vitro
IN VITRO ACTIVITIES immunoassays that are specific, sensitive, and highly
reproducible have replaced the bioassay procedures
In vitro findings for clinical purposes (Egrie et al., 1987).
Erythropoietin promotes the proliferation of its target
cells and maintains their viability as they differentiate.
The effect of the hormone varies with the particular
IN VIVO BIOLOGICAL
target cell. For primitive erythroid progenitor cells ACTIVITIES OF LIGANDS IN
(BFU-E) which express a low number of erythro- ANIMAL MODELS
poietin receptors (Sawada et al., 1990), erythropoietin
acts as a mitogen, promoting their proliferation Normal physiological roles
(Spivak et al., 1991). For late erythroid progenitor
cells, which express a high number of erythropoietin The major physiologic role of erythropoietin is to
receptors and are actively cycling, erythropoietin acts couple blood oxygen transport with long-term tissue
as a survival factor (Koury and Bondurant, 1990). oxygen requirements. Erythropoietin production is
The hormone is not required for erythroid cell dif- controlled at the level of its gene. Hypoxia upregu-
ferentiation since erythroid progenitor cells over- lates erythropoietin production and hyperoxia sup-
expressing Bcl-xL can differentiate in the absence of presses erythropoietin production, although never
erythropoietin (Silva et al., 1996). In the absence of completely. Because erythropoietin is a survival factor
erythropoietin, erythroid cells become quiescent and as well as a mitogen, it is produced constitutively
enter a G0ÿG1 state (Spivak et al., 1996). and is always present in the plasma (Spivak and
Other growth factors also participate in stimulating Hogans, 1987).
the proliferation of erythroid progenitor cells. They
include insulin-like growth factor type 1 (IGF-1)
(Sawada et al., 1989), IL-3 (Goldwasser et al., 1983), Species differences
stem cell factor (Dai et al., 1991), and thrombopoietin
(Kobayashi et al., 1995). Other agents which can As mentioned previously, erythropoietin is highly
enhance the effects of erythropoietin include tyrosine conserved and, with the exception of guinea pig
kinase inhibitors such as zinc chloride and sodium erythropoietin (Stopka et al., 1998a), species differ-
vanadate and activators of protein kinase C such as ences are negligible with respect to molecular homo-
bryostatin (Spivak et al., 1992). logy: this is also true for the erythropoietin receptor.
950 Jerry L. Spivak

Knockout mouse phenotypes causes an increase in circulating red blood cells, it is

associated with increases in the number of various
Both knockout and transgenic mouse studies have types of hematopoietic progenitor cells in active cell
provided additional insight into the behavior of this cycle within the bone marrow without a change in
highly conserved ligand±receptor pair. Mice hetero- their number in the blood (Dessypris et al., 1988).
zygous for erythropoietin or its receptor appear Myeloid growth factors also appear to have over-
phenotypically normal. However, homozygosity for lapping effects with erythropoietin. For example, G-
loss of either was embryonically lethal due to failure CSF can potentiate the effects of erythropoietin in vivo
of definitive erythropoiesis (Wu et al., 1995). Impor- (Miles et al., 1991). Since most growth factors, in
tantly, however, committed erythroid progenitor cells contrast to erythropoietin, are produced and act
were present in homozygous knockout mice, indicat- locally, it is likely that a number of synergistic inter-
ing that neither erythropoietin nor its receptor are actions are occurring with respect to the growth
required for lineage-specific commitment. This should factor network in the bone marrow, and we are as yet
not be a surprising observation given the overlapping unaware of the details of this.
functions of many of the growth factors which
influence primitive hematopoietic progenitor cells
(Walker et al., 1985), and also the stochastic nature
of lineage±specific commitment (Nakahata et al., PATHOPHYSIOLOGICAL ROLES
1982). Indeed, immortalized, factor-independent mul- IN NORMAL HUMANS AND
tipotent hematopoietic progenitor cells are capable of DISEASE STATES AND
lineage-specific commitment in the absence of either
hematopoietic growth factors or serum (Fairbairn DIAGNOSTIC UTILITY
et al., 1993). Interestingly, thrombopoietin was capa-
ble of rescuing a fraction of the erythroid progenitor Normal levels and effects
cells which lacked the receptor for erythropoietin in
homozygous knockout mice (Kieran et al., 1996). Cloning of the human erythropoietin gene provided
an opportunity to develop the reagents necessary to
measure accurately the quantity of erythropoietin in
the circulation. The immunoassay for erythropoietin
Transgenic overexpression has proved to be useful because:
Transgenic mice that overexpress erythropoietin 1. There is only one form of erythropoietin in the
behave in concordance with the lineage-restricted circulation.
behavior of the hormone. Thus, these mice develop 2. Immunoreactive erythropoietin is equivalent to
erythrocytosis but not leukocytosis or thrombocytosis biologically active erythropoietin.
(Semenza et al., 1989). 3. Erythropoietin is biochemically unique.
4. There are no preformed stores of the protein.
5. Production of erythropoietin is independent of its
Pharmacological effects plasma level.
6. The plasma clearance of erythropoietin is
The major pharmacologic effect of erythropoietin is independent of its plasma level.
to increase the red blood cell mass, an effect which 7. Production of erythropoietin is controlled at the
comes at the expense of the plasma volume (Lim et al., level of its gene.
1989). Other effects attributed to erythropoietin, such 8. There is no effect of age or gender on plasma
as hypertension or improved hemostasis, are a con- erythropoietin.
sequence of the increased red blood cell mass. 9. Erythropoietin is always present in the plasma.
10. Its plasma level is constant in a given individual
(Spivak, 1993).
Interactions with cytokine network Erythroid progenitor cells metabolize erythropoie-
tin and, therefore, the extent of bone marrow
The flu-like syndrome experienced by patients receiv- erythroid activity influences the plasma clearance of
ing bolus injections of large quantities of recombinant erythropoietin (Cazzola et al., 1998). Generally,
erythropoietin in vivo may be a cytokine storm marrow metabolism of erythropoietin is influenced
triggered by the hormone. In this regard, it is of by the same environmental factors which modify
interest that, while recombinant erythropoietin only erythropoietin production and in the same direction.
 Erythropoietin 951

For example, situations such as inflammation, 1985; Haga et al., 1987). In part, this is due to the
infection, or neoplasia which downregulate erythro- wide range of normal values for plasma erythropoie-
poietin production also impair erythroid progenitor tin and also to the increased metabolism of the
cell proliferation, while the relative increase in plasma hormone by the expanded marrow as well as by the
erythropoietin due to an increase in its production suppressive effect on erythropoietin production by
caused by tissue hypoxia depends on the extent of the the increase in blood viscosity associated with the
marrow committed erythroid precursor pool. This is elevated red cell mass (Singh et al., 1993). Thus, a
because erythropoietin is metabolized by its targets `normal' plasma erythropoietin level in a patient with
cells (Cazzola et al., 1998). Thus, in anemic patients erythrocytosis does not exclude hypoxia as a cause for
the plasma erythropoietin level should be high if the the erythrocytosis.
kidneys are normal and marrow function is impaired, Other situations can elevate plasma erythropoietin,
while a low plasma erythropoietin level in an anemic although only transiently, including chemotherapy
patient suggests impaired erythropoietin production which reduces erythroid progenitor cell proliferation
alone. However, it is not possible in the latter situa- and parenchymal liver damage when hepatocyte
tion to predict whether the marrow will respond to regeneration results in upregulation of erythropoietin
recombinant erythropoietin without a clinical trial; in production. For unknown reasons, zidovudine also
the former situation, if the marrow is cellular and the causes an elevation in plasma erythropoietin (Spivak
erythropoietin level is greater than 1000 mU/mL, a et al., 1989). Initially chemotherapy can increase
response to exogenous erythropoietin is unlikely. erythropoietin levels (Birgegard et al., 1989), pre-
There is a small diurnal variation in erythropoietin sumably by reducing the number of erythroid pro-
production with the highest levels occurring in the genitors but prolonged treatment with certain
morning and this needs to be considered when chemotherapeutic drugs can also reduce erythropoie-
comparative or serial measurements are made (Wide tin production (Miller et al., 1990).
et al., 1989). A number of conditions are associated with
impaired erythropoietin production. Intrinsic renal
disease is the most common and once the serum
Role in experiments of nature and creatinine rises above 1.5 mg%, the inverse linear
disease states correlation between hemoglobin and plasma erythro-
poietin is lost. However, it is important to note that
Since tissue hypoxia is the only physiologic stimulus patients with renal failure are capable of increasing
for erythropoietin production and production of the erythropoietin production in response to hypoxia but
protein is regulated at the level of its gene, there is an the threshold for this is higher than normal and the
inverse, log-linear relationship between plasma response is not sustained (Walle et al., 1987; Chandra
erythropoietin and tissue hypoxia whether measured et al., 1988). In patients with diabetes mellitus, the
by the hemoglobin level or arterial oxygen saturation. effect of the disease on renal endocrine function is
Unfortunately, however, plasma erythropoietin can- usually out of proportion to its effect on renal exo-
not simply be used as a surrogate measure of tissue crine function and there can be impaired erythro-
oxygenation because of a central tendency for down- poietin production and anemia without clinical
regulation of erythropoietin production normally, evidence of a significant decrease in renal excretory
and because of factors which positively or negatively function. Other situations with impaired erythropoie-
regulate erythropoietin production. Thus, within the tin production include inflammatory disorders, infec-
normal range for hemoglobin or hematocrit for both tions, cancer, pregnancy, surgery, and prematurity
men and women, there is no linear correlation (Spivak, 1993), liver disease (Siciliano et al., 1995)
between plasma erythropoietin and the hemoglobin and post bone marrow transplantation (Miller et al.,
or hematocrit level (Spivak, 1993). Importantly, when 1992).
anemia develops, even if it is mild, plasma erythro-
poietin will increase but because the normal range is
wide (4±26 mU/mL), the level will not rise outside the
normal range until the hemoglobin level falls below
IN THERAPY
10.5 mg% (hematocrit ˆ 30%) (Spivak, 1993).
In states of compensated hemolytic anemia, the
Preclinical ± How does it affect
erythropoietin level generally does not rise indefi- disease models in animals?
nitely or even outside the normal range. This is also
true in patients with tissue hypoxia due to cardiac Recombinant human erythropoietin is the most
shunts or obstructive lung disease (Wedzicha et al., successful recombinant protein currently in clinical
952 Jerry L. Spivak

use in terms of safety and efficacy. It can truly be Pharmacokinetics

considered a medical miracle since it provides most of
the benefits of blood transfusion without transfusing The plasma clearance of erythropoietin is complex
blood. Its development was remarkably timely, and best explained by a two-compartment model with
occurring when the nation's blood supply was under exponential clearance (Emmanouel et al., 1984;
duress, not only with respect to the adequacy of its Steinberg et al., 1986; Spivak and Hogans, 1989).
inventory but also with respect to its safety. Indeed, The volume of distribution of the hormone is slightly
with the focus of transfusion medicine on blood larger than the plasma volume and the initial phase of
conservation rather than blood transfusion, recombi- clearance reflects the distribution of the protein
nant erythropoietin was a welcome therapeutic addi- between the intravascular and extravascular spaces.
tion. In comparison to the other hematopoietic Thereafter, there is monoexponential elimination of
growth factors, erythropoietin not only had the erythropoietin from the plasma. The half-time of
advantage of being the first to be discovered but also elimination for plasma erythropoietin in humans is
being one of the few such growth factors to behave between 6 and 10 hours and is not appreciably
like a hormone. As a consequence, its physiology was affected by its plasma level, absence of the kidneys
easier to define and erythropoietin deficiency could be (McMahon et al., 1990), or cirrhosis of the liver
directly identified by bioassay or immunoassay. (Jensen et al., 1995). Under normal circumstances,
Furthermore, since decades of study of erythropoietin less than 10% of erythropoietin is eliminated by the
physiology preceded the development of recombinant kidneys (Emmanouel et al., 1984). In humans,
erythropoietin, a solid base of information was also erythropoietin does not cross the placenta (Widness
available to delineate those situations in which the et al., 1991). Administration of recombinant erythro-
hormone would be most effective and to guide clinical poietin by subcutaneous or intraperitoneal injection
trials. For example, the ability of exogenous erythro- results in remarkably different clearance kinetics than
poietin to increase the red blood cell mass in both intravenous injection (MacDougall et al., 1989). In
small and large animal models of renal insufficiency contrast to the exponential clearance from the plasma
(Anagnostou et al., 1977; Mladenovic et al., 1985), as following intravenous injection, following subcuta-
well as the documented impairment of erythropoietin neous or intraperitoneal administration there is a
production in anemic patients with end-stage renal depot-like gradual increase in plasma erythropoietin
disease (Gallagher et al., 1960), strongly supported followed by decline. Subcutaneous injection results in
the contention that recombinant erythropoietin would higher plasma levels than intraperitoneal injection.
prove effective in correcting anemia in this group of Importantly, although a higher plasma concentration
patients. Additionally, both in vivo and in vitro of erythropoietin is achieved by intravenous admin-
models (DeGowin and Gibson, 1979; Dainiak et al., istration, from a dose-response perspective, subcuta-
1983) indicated that erythroid progenitor cells were neous administration is approximately 30% more
responsive to exogenous erythropoietin in the pres- effective (Albitar et al., 1995; Kaufman et al., 1998).
ence of cancer or chemotherapeutic agents. This may be due to its more sustained plasma resi-
In vitro studies (Shannon et al., 1987) also dence time. Furthermore, the peak plasma levels of
supported the contention that erythropoietin might erythropoietin achieved by intravenous administra-
be useful in the treatment of the anemia of prematurity tion probably far exceed the available number of
while studies of serum erythropoietin using a more erythropoietin receptors, resulting in elimination of
sensitive and specific immunoassay (Egrie et al., 1987) much of the hormone before it has any biological
further defined specific groups of anemic patients in effect.
whom serum erythropoietin was inappropriately low The metabolic fate of erythropoietin over and
and who might, therefore, be appropriate candidates above that utilized by erythroid progenitor cells is
for receiving the hormone. These included patients unknown. Following infusion of labeled recombinant
with infectious (Spivak et al., 1989) or inflammatory human erythropoietin into rats, initial accumulation
disorders (Hochberg et al., 1988) or neoplasia (Miller in the liver, spleen, kidneys, or bone marrow was
et al., 1990), premature infants (Kling et al., 1996b) insufficient to account for its elimination from the
and patients undergoing elective surgery (Clemens plasma, indicating equilibration with the extracellular
and Spivak, 1994) or bone marrow transplantation space (Spivak and Hogans, 1989). No accumulation
(Miller et al., 1992). Of course, before recombinant of acid-soluble, labeled metabolites occurred over the
erythropoietin therapy is used all correctable causes 4-hour period of observation, indicating that catabo-
for anemia must be excluded, and the patient should lism of the hormone did not contribute to its
be symptomatic from the anemia or transfusion- plasma clearance. Native erythropoietin eventually
dependent. accumulated in the bone marrow but only to a
 Erythropoietin 953

slightly greater extent than in the kidneys and spleen. therapy with a low dose (50 U/kg) of the hormone
Importantly, the plasma clearance kinetics of de- and to avoid excessive elevation of the hematocrit.
sialated, oxidized, biologically inactive erythropoietin There is no evidence from either in vitro studies or
were identical with native erythropoietin, indicating clinical trials that recombinant erythropoietin can
determinants other than its carbohydrate moieties promote the growth of tumor cells (Berdel et al.,
or its biological activity could be involved in its 1991). Erythropoietin can, however, stimulate extra-
catabolism under certain circumstances. Of course, medullary erthropoiesis and the development of
repeated infusions of recombinant erythropoietin re- splenomegaly and even splenic infarcts in patients
sult in a shortening of its plasma half-life (McMahon with myelodysplasia or myeloproliferative disorders
et al., 1990), presumably due to an increase in the has been described. In this regard, recombinant ery-
erythroid progenitor cell proliferation and metabo- thropoietin has no significant influence on platelet
lism of the hormone (Cazzola et al., 1998). production.
Iron deficiency was a significant problem in renal
dialysis patients receiving recombinant erythropoietin
and emphasized the importance of an adequate
Toxicity supply of iron to sustain a response to the hormone.
However, hemodialysis patients represent a unique
In the decade since recombinant erythropoietin was population who are at constant risk of negative iron
first introduced into clinical practice, it has had a balance due to blood loss incurred through diagnostic
remarkable safety record, particularly in patients with phlebotomies, gastrointestinal hemorrhage or in dia-
normal renal function. The flu-like syndrome observed lyzer dead space. It is well established, however, that
when erythropoietin was administered intravenously individuals with normal body iron stores can sustain a
by bolus injection is rarely encountered when it is weekly blood loss of 500 mL while taking oral iron
administered subcutaneously. Improvements in the without developing significant anemia (Coleman et al.,
excipient have also reduced the discomfort and local 1953). As a corollary, they can also increase their red
skin reactions associated with subcutaneous injection. cell mass in response to erythropoietin without
Allergic reactions are rare, as is the development of receiving iron. As a general rule, if the serum ferritin
antibodies to the recombinant protein. Hypertension is greater than 100 ng/mL in normal individuals or
or exacerbation of pre-existing hypertension oc- 200 ng/mL in patients with renal disease, body iron
curred frequently in the initial clinical trials involving stores are sufficient for a maximal response to recom-
anemic hemodialysis patients (Eschbach et al., 1989). binant erythropoietin (Rutherford et al., 1994).
Although the frequency of hypertension is much Certainly, in the absence of iron deficiency, there is
lower in patients with normal renal function, they no evidence that parenteral iron as opposed to oral
are not immune to this complication, which generally iron is a necessary adjunct to erythropoietin therapy.
occurs in the first several months of therapy. There-
fore, it is worth monitoring the blood pressure
periodically during the early phase of therapy, parti- Clinical results
cularly if there is pre-existing hypertension.
Anemia Associated with Renal Disease
Vascular thrombosis is the most serious side-effect
of recombinant erythropoietin and, while hemostasis Since erythropoietin is produced primarily in the
is improved with successful erythropoietin therapy kidneys in adults and renal disease is associated with
(Moia et al., 1987), therapy-associated thrombosis impaired erythropoietin production, anemic patients
is purely a function of the extent to which the red with end-stage renal disease were the first to parti-
cell mass is increased. This should not be surprising cipate in clinical trials of human recombinant
since there is a direct linear correlation between the erythropoietin (Winearls et al., 1986; Eschbach et al.,
hematocrit and whole-blood viscosity. Indeed, a 1987). As anticipated, recombinant erythropoietin
recent study indicated that elevating the hematocrit proved to be remarkably effective in correcting
to 42% in hemodialysis patients was associated with anemia and alleviating transfusion requirements in
an increased cardiovascular mortality (Besarab et al., 97% of patients (Eschbach et al., 1989). Importantly,
1998). While the mechanism for hypertension is still correction of anemia was associated with improve-
undefined, the reduction in the plasma volume (Lim ment in the quality of life as a consequence of
et al., 1989) that accompanies the increase in red cell improved tissue oxygenation, cardiac function,
mass is undoubtedly involved. Thus, it is always muscle strength, and cognitive function (Evans et al.,
prudent, particularly in patients with renal disease 1990). Recombinant erythropoietin therapy also
or preexisting hypertension, to initiate erythropoietin proved to be effective in anemic patients with
954 Jerry L. Spivak

predialysis renal failure (Watson et al., 1990). (Dainiak et al., 1983). Importantly, however, che-
Contrary to initial concerns, recombinant erythro- motherapeutic agents did not abrogate erythropoie-
poietin did not accelerate renal failure in predialysis tin-responsiveness (Reissmann and Udupa, 1972). To
patients, was not associated with an increased date, four randomized, double-blind, placebo-con-
incidence of seizures, dialyzer heparin requirements, trolled studies have established that recombinant
hypercalcemia, or access thrombosis, at least with human erythropoietin can alleviate the anemia asso-
respect to native fistulas. Elevation of the hematocrit ciated with cancer and its chemotherapy and reduce
was also associated with an improvement in hemo- transfusion requirements (Abels et al., 1991; Cascinu
stasis (Moia et al., 1987). Subcutaneous administra- et al., 1994; Heiss et al., 1996; Wurnig et al., 1996).
tion of erythropoietin has proved to be more effective The overall response rate irrespective of the type or
than intravenous administration (Albitar et al., 1995; malignancy or treatment regimen was approximately
Kaufman et al., 1998) and, most recently, adminis- 60%. In the absence of chemotherapy, the response
tration of the total weekly dose of erythropoietin as a rate appeared to be lower for hematologic malig-
single injection has proved to be as effective as its nancies than solid tumors. Failure to respond to
administration in divided doses (Goldberg et al., erythropoietin had significant prognostic implications
1996). In patients with end-stage renal disease, it is because such patients usually had a shorter survival
best to initiate erythropoietin therapy at a low dose (Ludwig et al., 1994). A reduction in transfusion
(30±50 U/kg s.c.) and escalate gradually if necessary. requirements was obtained in approximately 50% of
Resistance to recombinant erythropoietin in pa- patients but is rarely absolute, since it takes 4±8 weeks
tients with chronic renal failure can occur due to a for recombinant erythropoietin to elevate the hemo-
variety of causes other than inadequate dosing. They globin level (Glaspy et al., 1997). Importantly,
include iron deficiency, infection, inflammation erythropoietin therapy has been shown to improve
(surgery, connective tissue disorder), bleeding, alumi- the quality of life in responders (Demetri et al., 1998).
num toxicity, metabolic bone disease, folate deficiency, Unfortunately, no study to date has demonstrated
hemolysis, a hemoglobinopathy, or splenomegaly. that the use of erythropoietin perioperatively either
Iron deficiency is the most common cause of erythro- alone or in conjunction with autologous blood dona-
poietin resistance and may require the administration tion results in a reduction in allogeneic blood expo-
of intravenous iron if oral supplementation is not sure in cancer patients. Since there is no evidence that
effective. The extent to which anemia needs to be allogeneic blood exposure promotes tumor metas-
corrected in patients with end-stage renal disease has tases, recombinant erythropoietin cannot be recom-
been the subject of substantial debate not only from mended for perioperative use in cancer patients
the perspective of cost but also because this particular (Busch et al., 1993).
patient population has a high incidence of cardiovas- With respect to dose, it has been established that
cular disease. Furthermore, although erythropoietin 300 U/kg is not better than 150 U/kg. Approximately
can alleviate anemia and prevent iron overload, it 30,000±40,000 U/week should be an effective dose in
does not change the need for dialysis. Thus, while it most patients and this can be administered once a
appears that a target hematocrit of 33% is too low, week subcutaneously (Goldberg et al., 1998). Given
the extent to which it should be raised above 36% is the importance of cost considerations, algorithms
unclear. A recent study of anemic hemodialysis have been developed to predict which patients are
patients with ischemic heart disease or congestive heart most likely to respond. An erythropoietin level
failure indicated that elevation of the hematocrit to greater than 400 mU/mL in the absence of che-
42% was associated with an increased risk of cardiac- motherapy suggested that a response was unlikely
related morbidity and mortality (Besarab et al., 1998). (Osterborg et al., 1996). Failure to achieve an increase
in hemoglobin of at least 0.5 g after 2 weeks of
therapy (or > 1 g after 4 weeks), together with a
Anemia Associated with Cancer
serum erythropoietin level of > 100 mU/mL, or a
Anemia is a common complication of cancer and, serum ferritin level of > 400 ng/mL, were also highly
although there are multiple reasons for this, the most predictive for nonresponders (Ludwig et al., 1994).
common cause is impairment of erythropoietin pro- The reticulocyte count is, unfortunately, a sensitive
duction presumably secondary to the elaboration of indicator of responsiveness (Henry et al., 1995).
inflammatory cytokines (Means and Krantz, 1992) or Although recombinant erythropoietin is only
chemotherapy (Miller et al., 1990). At the same time, approved for use in preventing symptomatic anemia
erythroid progenitor cells from cancer patients were in patients with nonmyeloid malignancies undergoing
found to be responsive to erythropoietin in vitro, chemotherapy, it has been widely used in patients
although the response was somewhat blunted with anemia due to myelodysplasia (Mittleman and
 Erythropoietin 955

Lessin, 1994). In this situation the hormone has impact on survival in HIV-infected patients and that
proved to be most effective in patients with refractory a response to erythropoietin is associated with
anemia or refractory anemia with ringed sideroblasts. improved survival suggests that recombinant erythro-
However, responses have been obtained in patients poietin may still have a role in the management of
with all types of myelodysplasia regardless of cyto- anemic HIV-infected patients.
genetic status unless the serum erythropoietin level
was greater than 1000 mU/mL. In many patients, the
Perioperative Use
addition of G-CSF may be necessary to obtain a
response to erythropoietin (Negrin et al., 1993, 1996; Most blood use occurs with surgical procedures and,
Hellstrom-Lindberg et al., 1998). Recombinant eryth- given public concern about blood safety, the
ropoietin therapy does not seen to accelerate the perioperative use of recombinant erythropoietin has
underlying myelodyplasia but can cause splenomegaly received intense scrutiny. There is a certain irony to
and splenic infarction. this since the blood supply is safer now then ever
before due to the development of sensitive tests for
the detection of known viral pathogens. Indeed, the
Anemia Associated with HIV Infection
risk of acquiring HIV by transfusion is equivalent to
Anemia is a common complication of HIV infection the risk of a fatal hemolytic transfusion reaction
and is multifarious with respect to its etiology (Popovski, 1998). The impetus to employ recombi-
(Holland and Spivak, 1990). Serum erythropoietin nant erythropoietin perioperatively in part stemmed
levels are depressed in HIV-infected patients (Spivak from the observation that erythropoietin production
et al., 1989) and HIV is capable of suppressing ery- did not increase substantially unless the hemoglobin
thropoietin production by Hep3B cells in vitro (Wang fell below 10.5 g% (Spivak, 1993), and that auto-
et al., 1993). Although HIV does not directly infect logous blood donors bled weekly did not increase
erythroid progenitor cells (Potts et al., 1992), the erythropoietin production substantially and become
immunosuppressed state it creates permits other progressively anemic (Kickler and Spivak, 1988).
pathogens such as MAI, parvovirus B19, or cyto- Furthermore, the time available before surgery is
megalovirus to attack the marrow and inhibit sometimes insufficient to collect the quantity of blood
erythropoiesis. Anemia is a poor prognostic sign in needed to avoid allogeneic blood exposure. Thus,
HIV-infected patients (Moore et al., 1998; Sullivan it appeared that recombinant erythropoietin might
et al., 1998) and blood transfusions can activate HIV have a useful role in the perioperative period.
expression in infected individuals (Mudido et al., Unfortunately, this has proved not to be the case.
1996). Furthermore, the antiretroviral agent zidovu- Rather, it was observed that, with a more aggressive
dine suppresses erythropoiesis. blood donation schedule, nonanemic, iron-replete
These observations led to clinical trials of recom- individuals with an adequate blood volume
binant erythropoietin in anemic HIV-infected patients (> 5000 mL) could donate the quantity of blood
(Henry et al., 1992). These trials demonstrated that desired (Biesma et al., 1994). However, estimated
erythropoietin therapy could alleviate anemia and blood needs and actual blood requirements proved
reduce tranfusion requirements in patients receiving not to be equal since surgical blood loss is not pre-
zidovudine whose endogenous erythropoietin level dictable. Thus, neither erythropoietin (Canadian
was less than 500 mU/mL. In a large open-label trial, Perioperative Orthopedic Erythropoietin Study
using doses of 40,000 U s.c. 6 days per week, trans- Group, 1993) nor erythropoietin coupled with auto-
fusion requirements were reduced by 50%, and 44% logous blood donation prevented exposure to
of the patients archived a 6% increase in hematocrit allogeneic blood in either nonanemic or anemic
(Phair et al., 1993). Combined therapy with erythro- patients (Goodnough et al., 1989, 1994).
poietin and G-CSF was also effective in 70% of Given the current safety of the blood supply,
patients receiving zidovudine who were unable to autologous blood donation is not cost-effective
tolerate this drug because of marrow toxicity (Miles (Etchason et al., 1995) and obviously erythropoietin
et al., 1991). Correction of anemia by erythropoietin therapy does not improve the situation. While recom-
in HIV-infected patients was also associated with an binant erythropoietin is now approved clinically for
improvement in quality of life (Revicki et al., 1994). use in mildly anemic patients (hemoglobin <
HIV infection does not appear to impair erythro- 13.5 g% > 11.5 g%) undergoing noncardiac surgical
poietin-responsiveness in patients with end-stage renal procedures, it must be remembered that it may not
disease (Ifudu et al., 1997). Although high does of abrogate allogeneic blood exposure under these con-
zidovudine are no longer used in treating HIV infec- ditions. Whether erythropoietin together with acute
tion, the recognition that anemia has an adverse normovolemic hemodilution will be superior to
956 Jerry L. Spivak

autologous blood donation remains to be determined may be unavoidable; in part due to more conservative
but recent results are encouraging (Ness et al., 1992) transfusion and phlebotomy usage as well as improved
and may be applicable to patients with cardiac transfusion safety; and finally, in part to unavoidable
disease who cannot donate autologously (Sowade medical complications which affect the need for
et al., 1997) and also to adherents of the Watchtower transfusion independently of erythropoietin therapy.
Society who refuse blood transfusion on a religious When recombinant erythropoietin is used for the
basis. Individuals with rare blood groups, severe anemia of prematurity, its use should be restricted to
alloimmunization, a small blood volume, or severe very low birth weight infants and at a dose of not
IgA deficiency are other candidates for perioperative more than 750 U/kg week (Maier et al., 1998) in
use of recombinant erythropoietin. combination with iron supplementation.

Anemia of Prematurity Miscellaneous Uses of Recombinant Erythropoietin

After birth, there is a reduction in hemoglobin which Recombinant erythropoietin has been demonstrated
is most marked in premature infants, often resulting to ameliorate anemia in patients with rheumatoid
in the need for transfusions (Keyes et al., 1989). arthritis (Pincus et al., 1990) and inflammatory bowel
Studies of serum erythropoietin postpartum indicate disease (Schreiber et al., 1996), to accelerate erythro-
that, while these infants are capable of producing poiesis after allogenic but not autologous
erythropoietin in response to hypoxia, their response marrow transplantation (Miller et al., 1992), to
is blunted when compared to the responses of adults improve blood pressure control in patients with
to a similar degree of anemia (Brown et al., 1984). orthostatic hypotension (Hoelktke and Streeten, 1993),
Since the number of circulating erythroid progenitor and to improve erythropoiesis in patients with
cells is not reduced in premature infants and their aplastic anemia (Bessho et al., 1990) or thalassemia
sensitivity to erythropoietin is normal, it was con- major (Rachmilewitz et al., 1995). However, the
cluded that inadequate erythropoietin production was extent to which its use can be justified in these
a major factor in the anemia of prematurity (Shannon situations on a cost±benefit basis has not been estab-
et al., 1987). A recent evaluation of erythropoietin lished. Nevertheless, given patient perceptions about
pharmacokinetics challenges that assumption in part, blood safety as well as the demonstrated improve-
since it demonstrated that the plasma clearance of ment in quality of life associated with alleviation of
erythropoietin was accelerated in premature infants anemia, it is difficult to fault the appropriate use of
(Widness et al., 1996). Given the evidence of increased the hormone in `off-label' situations. Indeed, when-
circulating erythroid progenitor cells in premature ever it can be shown that recombinant erythropoietin
infants, this should not be a surprising observation alleviates the need for blood transfusions, an
(Shannon et al., 1987). Thus, although there appears argument can be made for its use.
to be a defect in erythropoietin production in infants,
it is apparently magnified by an increase in erythro- References
poietin metabolism.
Because of the documented adverse effects of blood Abels, R. I., Larholt, K. M., Krantz, K. D., and Bryand, E. C.
transfusion, erythropoietin therapy was widely (1991). In ``Blood Cell Growth Factors'' (ed M. J. Murphy),
embraced for the treatment of the anemia of Recombinant human erythropoietin (r-HuEPO) for the treat-
prematurity (Maier et al., 1994; Ohls et al., 1997; ment of the anemia of cancer pp. 122±140. AlphaMed Press,
Dayton, OH.
Kumar et al., 1998; Shannon et al., 1998). The results Albitar, S., Meulders, Q., Hammoud, H., Soutif, C., Bouvier, P.,
of these trials indicate that recombinant erythropoie- and Pollini, J. (1995). Subcutaneous versus intravenous admin-
tin therapy is safe in premature infants but that the istration of erythropoietin improves its efficiency for the treat-
dose required to enhance erythropoiesis was larger ment of anaemia in haemodialysis patients. Nephrol. Dial.
than the effective dose in adults, that erythropoietin Transplant. 6, 40±43.
Anagnostou, A., Barone, J., Kedo, A., and Fried, W. (1977).
therapy was not beneficial in infants weighing more Effect of erythropoietin therapy on the red cell volume of urae-
than 1000 g and that, despite erythropoietin therapy, mic and non-uraemic rats. Br. J. Hematol. 37, 85±91.
exposure to allogeneic blood could not be avoided Anagnostou, A., Lee, E. S., Kessimian, N., Levinson, R., and
(Strauss, 1994). Furthermore, the cost±benefit ratio Steiner, M. (1990). Erythropoietin has a mitogenic and positive
of employing erythropoietin was not favorable (Fain chemotactic effect on endothelial cells. Proc. Natl Acad. Sci.
USA 87, 5978±5982.
et al., 1995). This was in part due to the fact that it Avedissian, L. S., and Spivak, J. L. (1995). In ``Human Protein
takes several weeks for erythropoietin therapy to be Data'' (ed A. Haebeck), Erythropoietin receptor. VCH
effective and within this time interval transfusions Verlagsgesellschaft, Weinheim.
 Erythropoietin 957

Avedissian, L. S., Poola, I., and Spivak, J. L. (1995). Ligand bind- Canadian Orthopedic Perioperative Erythropoietin Study Group
ing kinetics of a soluble full-length murine erythropoietin recep- (1993). Effectivness of perioperative recombinant human
tor. Biochem. Biophys. Res. Commun. 216, 62±68. erythropoietin in elective hip replacement. Lancet 341, 1227±
Bartley, T. D., Bogenberger, J., Hunt, P., Li, Y.-S., Lu, H. S., 1232.
Martin, F., Chang, M. S., Samal, B., Nichol, J. L., Switft, S., Casadevall, N., Dupuy, E., Molho-Sabatier, P., Tobelem, G.,
Johnson, J., Hsu, R. Y., Parker, V. P., Suggs, S., Skrine, J. D., Varet, B., and Mayeux, P. (1996). Autoantibodies against
Metewether, L. A., Clogston, C., Hsu, E., Hokom, M. M., erythropoietin in a patient with pure red-cell aplasia. N. Engl.
Hornkohl, A., Choie, E., Pangelinan, M., Sun, Y., Mar, V., J. Med. 334, 630±633.
McNinch, J., Simonet, L., Jacbsen, F., Xie, C., Shutter, J., Cascinu, S., Fedeli, A., Del Ferro, E., Luzi Fedeli, S., and
Chute, H., Basu, R., Selander, L., Trollinger, D., Sieu, L., Catalano, G. (1994). Recombinant human erythropoietin treat-
Padilla, D., Trail, G., Xu, W., Del Castillo, K., Biron, J., ment in cisplatin-associated anemia: a randomized, double-
Cole, S., Hu, M. C. T., Pacific, R., Ponting, I., Saris, C., blind trial with placebo. J. Clin. Oncol. 12, 1058±1062.
Wen, D., Yung, Y. P., Lin, H., and Bosselman, R. A. (1994). Cazzola, M., Gurnone, R., Cerani, P., Centenara, E., Rovati, A.,
Identification and cloning of a megakaryocyte growth and and Beguin, Y. (1998). Red blood cell precursor mass as an
development factor that is a ligand for the cytokine receptor independent determinant of serum erythropoietin level. Blood
Mpl. Cell 77, 1117±1124. 91, 2139±2145.
Bazan, J. F. (1989). A novel family of growth factor receptors. Chandra, M., Clemons, G. K., and McVicar, M. I. (1988).
Biochem. Biophys. Res. Commun. 164, 788±796. Relation of serum erythropoietin levels to renal excretory func-
Bazan, J. F. (1990a). Haemopoietic receptors and helical cyto- tions: evidence for lowered set point for erythropoietin produc-
kines. Immunol. Today 11, 350±354. tion in chronic renal failure. J. Pediatr. 113, 1015±1021.
Bazan, J. F. (1990b). Structural design and molecular evolution of Clemens, J., and Spivak, J. L. (1994). Serum immunoreactive
a cytokine receptor superfamily. Proc. Natl Acad. Sci. USA 87, erythropoietin during the perioperative period. Surgery 115,
6934±6938. 510±515.
Berdel, W. E., Oberberg, D., Reufi, B., and Thiel, E. (1991: Coleman, D. H., Stevens, A. R., Dodge, H. T., and Finch, C. A
Studies on the role of recombinant human erythropoietin in (1953). A rate of blood regeneration after blood loss. Arch.
the growth regulation of human nonhematopoietic tumor cells Intern. Med. 92, 341±349.
in vitro. Ann. Hematol. 63, 5±8. Dai, C. H., Krantz, S. B., and Zsebo, K. M. (1991). Human burst-
Beru, N., Smith, D., and Goldwasser, E. (1990). Evidence suggest- forming units-erythroid need direct interaction with stem cell
ing negative regulation of the erythropoietin receptor by ribo- factor for further development. Blood 78, 2493±2524.
nucleoprotein. J. Biol. Chem. 265, 14100±14104. Dainiak, N., Kulkarni, V., Howard, D., and Kalamanti, M.
Besarab, A., Bolton, W. K., Browne, J. K., Egrie, J. C., (1983). Mechanisms of abnormal erythropoiesis in malignancy.
Nissenson, A. R., Okamoto, D. M., Schwab, S. J., and Cancer 51, 1101±1106.
Goodkin, D. A. (1998). The effects of normal as compared Dame, C., Fahnenstich, H., Freitag, P., Hofmann, D., Abdul-
with low hematocrit values in patients with cardiac disease Nour, T., Bartmann, P., and Fandrey, J. (1998). Erythro-
who are receiving hemodialysis and epoietin. N. Engl. J. Med. poietin mRNA expression in human fetal and neonatal tissue.
339, 584±901. Blood 92, 3218±3225.
Bessho, M., Hirashima, K., Asano, S., Ikeda, Y., Ogawa, N., Damen, J. E., and Krystal, G. (1996). Early events in erythro-
Tomonaga, M., Toyama, K., Nakahata, T., Nmura, T., poietin-induced signaling. Exp. Hematol. 24, 1455±1459.
Mizoguchi, H., Yoshida, Y., Niitsu, Y., and Kohogo, Y. D'Andrea, A. D., Lodish, H. F., and Wong, G. G. (1989).
(1997). Treatment of the anemia of aplastic anemia patients Expression cloning of the murine erythropoietin receptor. Cell
with recombinant human erythropoietin in combination with 57, 277±285.
granulocyte colony-stimulating factor: a multicenter random- Davis, J. M., and Arakawa, T. (1987). Characterization of recom-
ized controlled study. Eur. J. Hematol. 58, 265±272. binant human erythropoietin produced in Chinese hamster
Biesma, D. H., Marx, J. J. M., Karaaijenhagen, R. J., Franke, W., ovary cells. Biochemistry 26, 2633±2638.
Messinger, D., and van de Wiel, A. (1994). Lower homo- DeGowin, R. L., and Gibson, D. P. (1979). Erythropoietin and the
logous blood requirment in autologous blood donors after anemia of extramedullary tumor. J. Lab. Clin. Med. 94, 303±311.
treatment with recombinant human erythropoietin. Lancet Delorme, E., Lorenzini, T., Giffin, J., Martin, F., Jacobsen, F.,
334, 367±370. Boone, T., and Elliott, S. (1992). Role of glycosylation on the
Birgegard, G., Wide, L., and Simonsson, B. (1989). Marked eryth- secretion and biological activity of erythropoietin. Biochemistry
ropoietin increase before fall in Hb after treatment with cyto- 31, 9871±9876.
static drugs suggests mechanism other than anaemia for Demchuk, E., Mueller, T., Oschkinat, H., Sebald, W., and
stimulation. Br. J. Haematol. 72, 462±466. Wade, R. C. (1994). Receptor binding properties of four-
Boissel, J. P., Lee, W.-R., Presnell, S. R., Cohen, F., and helix-bundle growth factors deduced from electrostatic analysis.
Bunn, H. F. (1993). Erythropoietin structure-function relation- Protein Sci. 3, 920±935.
ships. J. Biol. Chem. 268, 15983±15993. Demetri, G. D., Kris, M., Wade, J., Degos, L., and Cella, D.
Brown, M. S., Garcia, J. F., Phibbs, R. H., and Dallman, P. R. (1998). Quality-of-life benefit in chemotherapy patients treated
(1984: Decreased response of plasma immunoreactive erythro- with epoetin alfa is independent of disease response. J. Clin.
poietin to ``available oxygen'' in anemia of prematurity. Oncol. 16, 3412±3425.
J. Pediatr. 105, 793±798. Dessypris, E., Graber, S. E., Krantz, S. B., and Stone, W. J.
Bunn, J. F., and Poyton, R. O. (1996). Oxygen sensing and molec- (1988). Effects of recombinant erythropoietin on the concentra-
ular adaptation to hypoxia. Physiol. Rev. 76, 839. tion and cycling status of human marrow hematopoietin pro-
Busch, O. R. C., Hop, W. C. J., Hoynck van Papendrecht, gentior cells in vivo. Blood 72, 2060±2062.
M. A. W., Marquet, R. L., and Jeekel, J. (1993). Blood trans- De Vos, A. M., Ultsch, M., and Kossiakoff, A. A. (1992). Human
fusions and prognosis in colorectal cancer. N. Engl. J. Med. 32, growth hormone and extracellular domain of its receptor: crys-
1372±1376. tal structure of the complex. Science 255, 306±312.
958 Jerry L. Spivak

Digicaylioglu, M., Bichet, S., Marti, H. H., Wenger, R. H., Galson, D. L., Tan, C. C., Ratcliffe, P. J., and Bunn, H. F. (1993).
Rivas, L. A., Bauer, C., and Gassmann, M. (1995). Comparison of the human and mouse erythropoietin genes
Localization of specific erythropoietin binding sites in defined shows extensive homology in the flanking regions. Blood 82,
areas of the mouse brain. Proc. Natl Acad. Sci. USA 92, 3717± 3321±3326.
3720. Glaspy, J., Bukowski, R., Steinberg, T. C., Taylor, C.,
Dube, S., Fisher, J. W., and Powell, J. S. S. (1988). Glycosylation Tchekmedyian, S., and Vadhan-Raj, S. (1997). Impact of ther-
at specific sites of erythropoietin is essential for biosynthesis, apy with epotein alfa on clinical outcomes in patients with non-
secretion and biological function. J. Biol. Chem. 263, 17516± myeloid malignancies during cancer chemotherapy in
17521. community oncology practice. J. Clin. Oncol. 15, 1218±1234.
Eckardt, K.-U., LeHir, M., Tan, C. C., Ratcliffe, P. J., Goldberg, M. A., Glass, G. A., Cunningham, J. H., and
Kaissling, B., and Kurtz, A. (1992). Renal innervation plays Bunn, H. F. (1987). The regulated expression of erythropoietin
no role in oxygen-dependent control of erythropoietin mRNA by two human hepatoma cell lines. Proc. Natl Acad. Sci. USA
levels. Am. J. Physiol. 263, F925±F930. 84, 7972±7976.
Egrie, J. C., Cotes, P. M., Lane, J., Gaines Das, R. E., and Goldberg, M. A., Dunning, S. P., and Bunn, H. F. (1988).
Tam, R. C. (1987). Development of radioimmunoassays for Regulation of the erythropoietin gene: evidence that the oxygen
human erythropoietin as tracer and immunogen. J. Immunol. sensor is a heme protein. Science 242, 1412±1414.
Methods 99, 235±241. Goldberg, M. A., Gaut, C. C., and Bunn, H. F. (1991).
Elliott, S., Lorenzini, T., Chang, D., Brazilay, J., Delorme, E., Erythropoietin mRNA levels are governed by both the rate of
Griffin, J., and Hesterberg, L. (1996). Fine-structure epitope gene transcription and posttranscriptional events. Blood 77,
mapping of antierythropoietin monoclonal antibodies reveals 271±277.
a model of recombinant human erythropoietin structure. Goldberg, M. A., McGutchen, J. W., Jove, M., Di Cesare, P.,
Blood 87, 2702±2713. Friedman, R. J., Pos, R., Guilfoyle, M., Frei, D., and
Elliott, S., Lorenzini, T., Chang, D., Barzilay, J., and Delorme, E. Young, D. (1996). A safety and efficacy comparison study of
(1997). Mapping of the active site of recombinant human two dosing regimens of epoetin alfa in patients undergoing
erythropoietin. Blood 89, 493±502. major orthopedic surgery. Am. J. Orthop. 25, 544±556.
Emmanouel, D. S., Goldwasser, E., and Katz, A. I. (1984). Goodnough, L. T., Rudnick, S., Price, T. H., Ballas, S. K.,
Metabolism of pure human erythropoietin in the rat. Am. Collins, M. L., Crowley, J. P., Kosman, M., Drushall, M. S.,
J. Physiol. 247, F168±F176. Lenes, B. A., Menitove, J. E., Silberstein, L. E., Smithe, K. J.,
Erslev, A. (1953). Humoral regulation of red cell production. Wallas, C. H., Abels, R., and Von Tress, M. (1989). Increased
Blood 8, 349±357. preoperative collection of autologous blood with recombinant
Eschbach, J. W., Egrie, J. C., Dowining, M. R., Browne, J. K., human erythropoietin therapy. N. Engl. J. Med. 321, 1163±
and Adamson, J. W. (1987). Correction of the anemia of end- 1168.
stage renal disease with recombinant human erythropoietin. N. Goodnough, L. T., Price, T. H., Friedman, K. D., Johnson, M.,
Engl. J. Med. 321, 73±78. Ciavarella, D., Khan, N., Sacher, R., Vogler, W. R., Wissel, M.,
Eschbach, J. W., Abdulhardi, M. H., Browne, J. K., Delano, B. C., and Abels, R. I. (1994). A phase III trial of recombinant human
Downing, M. R., Egrie, J. C., Evans, R. W., Friedman, E. A., erythropoietin therapy in nonanemic orthopedic patients sub-
Graber, S. E., and Haley, N. R. (1989). Recombinant human jected to aggressive removal of blood for autologous use: dose,
erythropoietin in anemic patients with end-stage renal disease. response, toxicity, and efficacy. Transfusion 34, 66±71.
Results of a phase III multicenter clinical trial. Ann. Intern. Grant, W. E., and Root, W. S. (1947). The relation of O2 in bone
Med. 111, 992±1000. marrow blood to posthemorrhagic erythropoiesis. Am. J.
Etchason, J., Petz, L., Keeler, E., Calhoun, L., Kleinman, S., Physiol. 150, 618±627.
Snider, C. Y., Fink, A., and Brook, R. (1995). The cost effec- Haga, P., Cotes, P. M., Till, J. A., Minty, B. D., and
tiveness of preoperative autologous blood donations. N. Engl. Shinebourne, E. A. (1987). Serum immunoreactive erythropoie-
J. Med. 332, 719±724. tin in children with cyanotic and acyanotic congenital heart
Evans, R. W., Rader, B., and Mannine, D. L. (1990). The quality disease. Blood 70, 822±826.
of life of hemodialysis recipients treated with recombinant Heiss, M. M., Tarabichi, A., Delanoff, C., Allgayer, H.,
human erythropoietin. J. Am. Med. Assoc. 263, 825±830. Jauch, K. W., Hernandez-Richter, T., Mempel, W.,
Fain, J., Hilsenrath, P., Widness, J. A., Strauss, R. G., and Beck, K. G., Schildberg, F. W., and Messmer, K. (1996).
Mutnick, A. H. (1995). A cost analysis comparing erythropoie- Perisurgical erythropoietin application in anemic patients with
tin and red cell transfusions in the treatment of anemia of pre- colorectal cancer: a double-blind randomized study. Surgery
maturity. Transfusion 35, 936±943. 119, 523±527.
Fairbairn, L. J., Cowling, G. J., Reipert, B. M., and Dexter, T. M. Hellstrom-Lindberg, E., Ahlgren, T., Beguin, Y., Carlsson, M.,
(1993). Suppression of apoptosis allows differentiation and Carneskog, J., Dahl, I. M., Dybedal, I., Grimfors, G., Kanter-
development of a multipotent hemopoietic cell line in the Lewensohn, L., Linder, O., Luthman, M., Lofvenberg, E.,
absence of added growth factors. Cell 74, 823±832. Nilsson-Ehle, H., Samuelsson, J., Tangen, J. M., Winqvist, I.,
Faquin, W. C., Schneider, T. J., and Goldberg, M. A. (1992). Oberg, G., Osterborg, A., and Ost, A. (1998). Treatment of
Effect of inflammatory cytokines on hypoxia-induced erythro- anemia in myelodysplastic syndromes with granulocyte colony-
poietin production. Blood 79, 1987±1994. stimulating factor plus erythropoietin: results from a random-
Fukuda, M. N., Sasaki, H., Lopez, L., and Fukuda, M. (1989). ized phase II study and long-term follow-up of 71 patients.
Survival of recombinant erythropoietin in the circulation: the Blood 92, 68±75.
role of carbohydrates. Blood 73, 84±89. Henry, D. H., Beall, G. N., Benson, C. A., Carey, J., Cone, L. A.,
Gallagher, N. J., McCarthy, J. M., and Lange, R. D. (1960). Eron, L. J., Fiala, M., Fischl, M. A., Gabin, S. F., Gottlieb, M. S.,
Observations on erythropoietic stimulating factor (ESF) in the Galpin, J. E., Groopman, J. E., Hooton, T. H., Memsek, J. G.,
plasma of uremic and non-uremic anemia patients. Ann. Intern. Levin, R. L., Miles, S. A., Riehart, J. J., Rios, A., Robbins, W. J.,
Med. 52, 1201±1212. Ruckdeschel, J. C., Smith, J. A., Spruance, S. L., Starrett, B.,
 Erythropoietin 959

Toney, J., Zalusky, R., Abels, R. I., Bryand, E. C., Larholt, K. M., Keller, G., Kennedy, M., Papayannopoulou, T., and Wiles, M. V.
Sampson, A. R., and Rudnick, S. A. (1992). Recombinant (1993). Hematopoietic commitment during embryonic stem cell
human erythropoietin in the treatment of anemia associated differentiation in culture. Mol. Cell. Biol. 13, 473±486.
with human immunodeficiency virus (HIV) infection and zido- Keyes, W. G., Donohue, P. K., Spivak, J. L., Jones, M. D., and
vudine therapy. Ann. Intern. Med. 117, 739±747. Oski, F. A. (1989). Assessing the need for transfusion of pre-
Henry, D., Abels, R., and Larholt, K. (1995). Prediction of mature infants and role of hematocrit clinical signs and erythro-
response to recombinant human erythropoietin (r-HuEPO/ poietin level. Pediatrics 84, 412±417.
epoetin) therapy in cancer patients. Blood 85, 1676±1678. Kickler, T., and Spivak, J. L. (1988). Effect of repeated phlebot-
Higuchi, M., Oh-eda, M., Kuboniwa, H., Tomonoh, K., omy on serum immunoreactive erythropoietin levels in autolo-
Shimonaka, Y., and Ochi, N. (1992). Role of sugar chains in gous blood donors. J. Am. Med. Assoc. 260, 65±67.
the expression of the biological activity of human erythropoie- Kieran, M. W., Perkins, A. C., Orkin, S. H., and Zon, L. I. (1996).
tin. J. Biol. Chem. 267, 7703±7709. Thrombopoietin rescues in vitro erythroid colony formation
Hochberg, M. C., Arnold, C. M., Hogans, B. B., and Spivak, J. L. from mouse embryos lacking the erythropoietin receptor.
(1988). Serum immunoreactive erythropoietin in rheumatoid Proc. Natl Acad. Sci. USA 93, 9126±9131.
arthritis: impaired response to anemia. Arthritis Rheum. 31, Klassen, D. K., and Spivak, J. L. (1990). Hepatitis-related hepatic
1318±1321. erythropoietin production. Am. J. Med. 89, 684±686.
Hoelktke, R. D., and Streeten, D. H. P. (1993). Treatment of Kling, P. J., Dragsten, P. R., Roberts, R. A., Dos Santos, B.,
orthostatic hypotension with erythropoietin. N. Engl. J. Med. Brooks, D. J., Hedlung, B. E., and Taetle, R. (1996a). Iron
329, 611±615. deprivation increases erythropoietin production in vitro, in nor-
Holland, K., and Spivak, J. L. (1990). Haematological manifesta- mal subjects and patients with malignancy. Br. J. Haematol. 95,
tions of the acquired immunodeficiency syndrome. Clin. 241±248.
Haematol. 3, 103±109. Kling, P. J., Schmidt, R. L., Roberts, R. A., and Widness, J. A.
Huang, L. E., Ho, V., Arany, Z., Krainc, D., Galson, D., (1996b). Serum erythropoietin levels during infancy; associa-
Tendler, D., Livingston, D. M., and Bunn, H. F. (1997). tions with erythropoiesis. J. Pediatr. 128, 791±796.
Erythropoietin gene regulation depends on heme-dependent Kobayashi, M., Laver, J. H., Kato, T., Miyazaki, H., and
oxygen sensing and assembly of interacting transcription fac- Ogawa, M. (1995). Recombinant human thrombopoietin (Mpl
tors. Kidney Int. 51, 548±552. ligand) enhances proliferation of erythroid progenitors. Blood
Huang, L. E., Gu, J., Schau, M., and Bunn, H. F. (1998). 86, 2494±2499.
Regulation of hypoxia-inducible factor 1 is mediated by an Koury, M. J., and Bondurant, M. C. (1990). A survival model of
O2-dependent degradation domain via the ubiquitin-proteasome erythropoietin action. Science 248, 378±381.
pathway. Proc. Natl Acad. Sci. USA 95, 7987±7992. Koury, S. T., Bondurant, M. C., and Koury, M. J. (1988).
Ifudu, O., Matthew, J. J., Mayers, J. D., Macey, L. J., Brezsnyak, W., Localization of erythropoietin synthesizing cells in murine kid-
Reydel, C., McLendon, E., Surgrue, T., Rao, S., and neys by in situ hybridization. Blood 71, 524±527.
Friedman, E. A. (1997). Severity of AIDS and the response to Koury, S. T., Koury, M. J., Bondurant, M. C., Caro, J., and
EPO in uremia. Am. J. Kidney Dis. 30, 28±35. Graber, S. E. (1989). Quantitation of erythropoietin-producing
Imagawa, S., Yamaoto, M., and Miura, Y. (1997). Negative reg- cells in kidneys of mice by in situ hybridization: correlation with
ulation of the erythropoietin gene expression by the GATA hematocrit, renal erythropoietin mRNA and serum erythro-
transcription factors. Blood 89, 1430±1439. poietin concentration. Blood 74, 645±651.
Jacobs, K., Shoemaker, C., Rudersdorf, R., Neill, S. D., Koury, S. T., Bondurant, M. C., Koury, M. J., and Semenza, G. L.
Kaufman, R. J., Mufson, A., Seebra, J., Jones, S. S., (1991). Localization of cells producing erythropoietin in murine
Hewick. R., Fritsch, E. F., Kawakita, M., Shimizu, T., and liver by in situ hybridization. Blood 77, 2469±2471.
Miyake, T. (1985). Isolation and characterization of genomic Krieg, M., Marti, H. H., and Plate, K. H. (1998). Coexpression of
and cDNA clones of human erythropoietin. Nature 313, 806±810. erythropoietin and vascular endothelial growth factor in ner-
Jacobson, L. E., Goldwasser, E., Fried, W., and Plzak, L. (1957). vous system tumors associated with von Hippel-Lindau tumor
Role of the kidney in erythropoiesis. Nature 179, 633±634. suppressor gene loss of function. Blood 92, 3322±3389.
Jelkmann, W., Pagel, H., Wolff, M., and Fandrey, J. (1992). Krystal, G. (1983). A simple microassay for erythropoietin based
Monokines inhibiting erythropoietin production in human hepa- on 3H-thymidine incorporation into spleen cells for phynylhy-
toma cultures and in isolated perfused rat kidneys. Life Sci. 30, drazine treated mice. Exp. Hematol. 11, 649±660.
301±308. Kumar, P., Shankaran, S., and Krishnan, R. G. (1998).
Jelkmann, W., Pagel, H., Hellwig, T., and Fandrey, J. (1997). Recombinant human erythropoietin therapy for treatment of
Effects of antioxidant vitamins on renal and hepatic erythro- anemia of prematurity in very low birth weight infants; a ran-
poietin production. Kidney Int. 51, 497±501. domized, double-blind, placebo-controlled trial. J. Perinatol.
Jensen, J. D., Jensen, L. W., Madsen, J. K., and Poulsen, L. 18, 173±177.
(1995). The metabolism of erythropoietin in liver cirrhosis Lacombe, C., Da Silva, J. L., Bruneval, P., Fournier, J. G.,
patients compared with healthy volunteers. Eur. J. Haematol. Wending, F., Casadevall, N., Camilleri, J. P., Bariety, J.,
54, 111±116. Varet, B., and Tambourin, P. (1988). Peritubular cells are the
Kallio, P. J., Pongratz, I., Gradin, K., McGuire, J., and site of erythropoietin synthesis in the murine hypoxic kidney.
Poellinger, L. (1997). Activation of hypoxia-inducible factor 1 J. Clin. Invest. 81, 620±623.
: posttranscriptional regulation and conformational change by Law, M. L., Cai, G.-Y., Lin, K., Wei, Q., Huang, S. Z.,
recruitment of the Arnt transcription factor. Proc. Natl Acad. Hartz, J. H., Morse, H., Lin, C. H., Jones, C., and Kao, F. T.
Sci. USA 94, 5667±5672. (1986). Chromosomal assignment of the human erythropoietin
Kaufman, J. S., Reda, D. J., Fye, C. L., Goldfarb, D. S., gene and its DNA polymorphism. Proc. Natl Acad. Sci. USA
Henderson, W. G., Kleinman, J. S., and Vaamonde, C. A. 83, 6920±6924.
(1998). Subcutaneous compared with intravenous epoetin in Lim, V. F., Degrown, R. L., Zavala, D., Kirchner, P. T., Abels, R.,
patients receiving hemodialysis. N. Engl. J. Med. 399, 578±583. Perry, P., and Fangman, J. (1989). Recombinant human
960 Jerry L. Spivak

erythropoietin treatment in pre-dialysis patients. Ann. Intern. patients with the anemia of cancer. N. Engl. J. Med. 322,
Med. 110, 104±108. 1689±1692.
Lin, F.-K., Suggs, S., Lin, C.-H., Browne, J. K., Smalling, R., Miller, C. B., Jones, R. J., and Zahurak, M. L. (1992). Impaired
Egrie, J. C., Chen. K. K., Fox, G. M., Martin, F., Stabinsky, Z., erythropoietin response to anemia after bone marrow trans-
Badrawi, S. M., Lai, P.-H., and Goldwasser, E. (1985). Cloning plantation. Blood 80, 2667±2682.
and expression of the human erythropoietin gene. Proc. Natl Misaizu, T., Matsuki, S., Strickland, T. W., Takeuchi, M.,
Acad. Sci. USA 82, 7580±7584. Kobata, A., and Takasakai, S. (1995). Role of antennary struc-
Ludwig, H., Fritz, E., Leitgeb, C., Perchestorfer, M., ture of N-linked sugar chains in renal handling of recombinant
Samonigg, H., and Schuster, J. (1994). Prediction of response human erythropoietin. Blood 86, 4097±4104.
to erythropoietin treatment in chronic anemia of cancer. Blood Mittleman, M., and Lessin, L. S. (1994). Clinical application of
84, 1056±1063. recombinant erythropoietin in myelodysplasia. Hematol./Oncol.
McDonald, J. D., Lin, F.-K., and Goldwasser, E. (1986). Cloning, Clin. N. Am. 8, 993±1009.
sequencing and evolutionary analysis of the mouse erythropoie- Miyake, T., Kung, C. K.-H., and Goldwasser, E. (1977).
tin gene. Mol. Cell. Biol. 6, 842±848. Purification of human erythropoietin. J. Biol. Chem. 252,
MacDougall, I. C., Neubert, P., Coles, G. A., Roberts, D. E., 5558±5564.
Dharmasena, A. D., and Williams, J. D. (1989). Pharmaco- Mladenovic, J., Eschback, J. W., Koup, J. R., Garcia, J. F., and
kinetics of recombinant human erythropoietin in patients on Adamson, J. W. (1985). Erythropoietin kinetics in normal and
continuous ambulatory peritoneal dialysis. Lancet 1, 425±427. uremic sheep. J. Lab. Clin. Med. 105, 659±663.
McGary, E. C., Rondon, I. J., and Beckman, B. S. (1997). Post- Moia, M., Mannucci, P. M., Vizzotto, L., Masati, S.,
transcriptional regulation of erythropoietin mRNA stability by Cattaneo, M., and Ponticelli, C. (1987). Improvement in the
erythropoietin mRNA-binding protein. J. Biol. Chem. 272, haemostatic defect of uraemia after treatment with recombinant
8628±8634. human erythropoietin. Lancet 2, 1227±1229.
McMahon, F. G., Vargas, R., Ryan, M., Jain, A. K., Abels, R. I., Moore, R. D., Keruly, J. C., and Chaisson, R. E. (1998). Anemia
Perry, B., and Smith, I. L. (1990). Pharmacokinetics and effects and survival in HIV infection. J. AIDS Hum. Retrov. 19, 29±33.
of recombinant human erythropoietin after intravenous and sub- Morishita, E., Masuda, S., Nagao, M., Yasuda, Y., and Sasaki, R.
cutaneous injections in healthy volunteers. Blood 76, 1718±1722. (1997). Erythropoietin receptor is expressed in rat hippocampal
Maier, R. F., Obladen, M., Kattner, E., Natzschka, J., Messer, J., and cerebral cortical neurons, and erythropoietin prevents in vitro
Regazzoni, B. M., Speer, C. P., Fellman, V., Grauel, E. L., glutamate-induced neuronal death. Neuroscience 76, 105±116.
Groneck, P., Wagner, M., Moriette, G., Salle, B. L., Mudido, P. M., Georges, D., Dorazio, D., Yen-Lieberman, B.,
Verellen, G., and Scigalla, P. (1998). High-versus low-dose Base, S., O'Brien, W. A., Spritzler, J., and Lederman, M. M.
erythropoietin in extremely low birth weight infants. J. (1996). Human immunodeficiency virus type 1 activation after
Perinatol. 132, 866±870. blood transfusion. Transfusion 36, 860±865.
Maier, R. F., Obladen, M., Scigalla, P., Linderkamp, O., Duc, G., Nakahata, T., Gross, A. J., and Ogawa. M. (1982). A stochastic
Hieronimi, G., Halliday, H. L., Versmold, H. T., Moriette, G., model of self-renewal and commitment to differentiation of the
and Jorch, G. (1994). The effect of epoietin  (recombinant primitive hemopoietic stem cells in culture. J. Cell. Physiol. 113,
human erythropoietin) on the need for transfusion in very 455±458.
low-birth-weight infants. N. Engl. J. Med. 330, 1173±1178. Narhi, L. O., Arakawa, T., Aoki, K. H., Elmore, R., Rohde, M. F.,
Marti, H. H., Gassmann, M., Wenger, R. H., Kvietikova, I., Boone, T., and Strickland, T. W. (1991). The effect of carbohy-
Morganti-Kossmann, M. C., and Kossmann, T. (1997). drate on the structure and stability of erythropoietin. J. Biol.
Detection of erythropoietin in human liquor: intrinsic erythro- Chem. 266, 23022±23026.
poietin production in the brain. Kidney Int. 51, 416±418. Negrin, R. S., Stein, R., Vardiman, J., Doherty, K., Cornwell, J.,
Masuda, S., Okano, M., Yamgishi, K., Nago, M., Ueda, M., and Krantz, S., and Greenberg, P. L. (1993). Treatment of the ane-
Saski, R. (1994). A novel site of erythropoietin production. mia of myelodysplastic syndromes using recombinant human
Oxygen-dependent production in cultured rat astrocytes. granulocyte colony-stimulating factor in combination with
J. Biol. Chem. 269, 19488±19493. erythropoietin. Blood 82, 737±743.
Matthews, D. J., Topping, R. S., Cass, R. T., and Giebel, L. B. Negrin, R. S., Stein, R., Doherty, K., Cornwell, J., Vardimann, J.,
(1996). A sequential dimerization mechanism of erythro- Krantz, S., and Greenberg, P. L. (1996). Maintenance treatment
poietin receptor activation. Proc. Natl Acad. Sci. USA 93, of the anemia of myelodysplastic syndromes with recombinant
9471±9476. human granulocyte colony-stimulating factor and erythropoie-
Means Jr, R. T., and Krantz, S. B. (1992). Progress in understand- tin: evidence for in vivo synergy. Blood 87, 4076±4081.
ing the pathogenesis of the anemia of chronic disease. Blood Ness, P. M., Bourke, D. L., and Walsh, P. C. (1992). A random-
80, 1639±1647. ized trial of perioperative hemodilution versus transfusion of
Middleton, S. A., Johnson, D. L., Jin, R., McMahon, F. J., preoperatively deposited autologous blood in elective surgery.
Collins, A., Tullai, J., Gruninger, R. H., Jolliffe, L. K., and Transfusion 32, 226±230.
Mulcahy, L. S. (1996). Identification of a critical ligand binding Ohls, R. K., Harcum, J., Schibler, K. R., and Christensen, R. D.
determinant of the human erythropoietin receptor. Evidence for (1997). The effects of erythropoietin on the transfusion require-
common ligand binding motifs in the cytokine receptor family. ments of preterm infants weighing 750 grams or less; a random-
J. Biol. Chem. 271, 14045±14054. ized, double-blind, placebo-controlled study. J. Pediatr. 131,
Miles, S. A., Mitsuyasu, R. T., Moreno, J., Baldwin, G., 653±655.
Alton, N. K., Souza, L., and Glaspy, J. A. (1991). Combined Okano, M., Masuda, S., Narita, H., Masushige, S., Kato, S.,
therapy with recombinant granulocyte colony-stimulating fac- Imagawa, S., and Saski, R. (1994). Retinoic acid up-regulates
tory and erythropoietin decreases hematologic toxicity from erythropoietin production in hepatoma cells and in vitamin A-
zidovudine. Blood 77, 2109±2117. depleted rats. FEBS Lett. 349, 229±233.
Miller, C. B., Jones, R. J., Piantadosi, S., Abeloff, M. D., and Osterborg, A., Boogaerts, M. A., Cimino, R., Essers, U.,
Spivak, J. L. (1990). Decreased erythropoietin response in Holowiecki, J., Juliusson, G., Jager, G., Najman, A., and
 Erythropoietin 961

Peest, D. (1996). Recombinant human erythropoietin in trans- Sawada, K., Krantz, S. B., Dessypris, E. N., Koury, S. T., and
fusion-dependent anemic patients with multiple myeloma and Sawyer, S. T. (1989). Human colony-forming units-erythroid do
non-Hodgkin's lymphoma ± a randomized multicenter study. not require accessory cells, but do require direct interaction with
The European study group of erythropoietin (epoetin ) treat- insulin-like growth factor I and/or insulin for erythroid devel-
ment in multiple myeloma and non-Hodgkin's lymphoma. opment. J. Clin. Invest. 83, 1701±1709.
Blood 87, 2675±2682. Sawada, K., Krantz, S. B., Dai, C.-H., Koury, S. T., Horn, S. T.,
Phair, J. P., Abels, R. I., McNeill, M. V., Pharm, D., and Glick, A. D., and Civin, C. I. (1990). Purification of human blood
Sullivan, D. J. (1993). Recombinant human erythropoietin burst-forming units-erythroid and demonstration of the evolu-
treatment; investigational new drug protocol for the anemia tion of erythropoietin receptors. J. Cell. Physiol. 142, 219±230.
of the acquired immunodeficiency syndrome. Arch. Intern. Sawyer, S. T., and Hankins, W. D. (1993). The functional form of
Med. 153, 2669±2675. the erythropoietin receptor is a 78-kDa protein: correlation with
Philo, J. S., Aoki, K. H., Arakawa, T., Narhi, L. O., and Wen, J. cell surface expression, endocytosis, and phosphorylation. Proc.
(1996). Dimerization of the extracellular domain of the erythro- Natl Acad. Sci. USA 90, 6849±6853.
poietin (EPO) receptor by EPO; one high-affinity and one low- Scandurro, A. B., Rondon, I. J., Wilson, R. B., Tenenbaum, S. A.,
affinity interaction. Biochemistry 35, 1681±1691. Garry, R. F., and Beckman, B. S. (1997). Interaction of eryth-
Potts, B., Hogan, M. D., Lamperth, L., and Spivak, J. L. (1992). ropoietin RNA requires an association with heat shock protein
Replication of HIV-1 and HIV-2 in human bone marrow cul- 70. Kidney Int. 2, 579±584.
tures. Virology 188, 840±849. Schade, S. G., and Fried, W. (1976). Suppressive effect of endo-
Popovski, M. A. (1998). Infection and America's blood supply: a toxin on erythropoietin-responsive cells in mice. Am. J. Physiol.
1998 status report. Am. J. Clin. Pathol. 109, 659±661. 231, 73±76.
Powell, J. S., Berkner, K. L., Lebo, R. V., and Adamson, J. W. Schmitt, R. M., Bruyns, E., and Snodgrass, H. R. (1991).
(1986). Human erythropoietin gene: high level expression in Hematopoietic development of embryonic stem cells in vitro:
stably transfected mammalian cells and chromosome localiza- cytokine and receptor gene expression. Genes Dev. 5, 728±740.
tion. Proc. Natl Acad. Sci. USA 83, 6465±6469. Schneider, H., Chaovapong, W., Matthews, D. J., Karkaria, C.,
Rachmilewitz, E. A., Aker, M., Perry, D., and Dover, G. (1995). Cass, R. T., Zhan, H., Boyle, M., Lorenzini, T., Elliott, S. G.,
Sustained increase in haemoglobin and RBC following long- and Giebel, L. B. (1997). Homodimerization of erythropoietin
term administration of recombinant human erythropoietin to receptor by a bivalent monoclonal antibody triggers cell prolif-
patients with homozygous -thalassaemia. Br. J. Haematol. 90, eration and differentiation of erythroid precursors. Blood 89,
341±345. 473±482.
Ratcliffe, P. J., Ebert, B. L., Ferguson, D. J., Firth, J. D., Schreiber, S., Howaldt, S., Schnoor, M., Nikolaus, S., Bauditz, J.,
Gleadle, J. M., Maxwell, P. H., and Pugh, C. W. (1995). Gasche, C., Lochs, H., and Raedler, A. (1996). Recombinant
Regulation of the erythropoietin gene. Nephrol. Dial. erythropoietin for the treatment of anemia in inflammatory
Transplant. 10, 18±27. bowel disease. N. Engl. J. Med. 334, 619±623.
Recny, M. A., Scoble, H. A., and Kim, Y. (1987). Structural Schuster, S. J., Wilson, J. H., Erslev, A. J., and Caro, J. (1987).
characterization of natural human urinary and recombinant Physiologic regulation and tissue localization of renal erythro-
DNA-derived erythropoietin, identification of des-arginine 166 poietin messenger RNA. Blood 70, 316±318.
erythropoietin. J. Biol. Chem. 262, 17156±17163. Semenza, G. L., and Wang, G. L. (1992). A nuclear factor induced
Reissmann, K. R. (1950). Studies on the mechanism of erythro- by hypoxia via de novo protein synthesis binds to the human
poietic stimulation in parabiotic rats during hypoxia. Blood 5, erythropoietin gene enhancer at a site required for transcrip-
372±380. tional activation. Mol. Cell. Biol. 12, 5447±5454.
Reissmann, K. R., and Udupa, K. B. (1972). Effect of erythro- Semenza, G. L., Traystman, M. D., Gearheart, J. D., and
poietin on proliferation of erythropoietin-responsive cells. Cell Antonarakis. S. E. (1989). Polycythemia in transgenic mice
Tissue Kinet. 5, 481±489. expressing the human erythropoietin gene. Proc. Natl Acad.
Revicki, D. A., Brown, R. E., Henry, D. H., McNeill, M. V., Sci. USA 86, 2301±2305.
Rios, A., and Watson, T. (1994). Recombinant human erythro- Semenza, G. L., Koury, S. T., Nejfelt, M. K., Gearhart, J. D., and
poietin and health-related quality of life of AIDS patients with Antonarakis, S. E. (1991). Cell-type-specific and hypoxia-indu-
anemia. J. AIDS 7, 474±484. cible expression of the human erythropoietin gene in transgenic
Rondon, I. J., MacMillan, L. A., Beckman, B. S., Goldberg, M. A., mice. Proc. Natl Acad. Sci. USA 88, 8725±8729.
Schneider, T., Bunn, H. F., and Malter, J. S. (1991). Hypoxia Semenza, G. L., Agani, F., Booth, G., Forsythe, J., Iyer, N.,
up-regulates the activity of a novel erythropoietin mRNA bind- Jiang, B. H., Leung, S., Roe, R., Wiener, C., and Yu, A.
ing protein. J. Biol. Chem. 25, 16594±16598. (1997). Structural and functional analysis of hypoxia-inducible
Rutherford, C. J., Schneider, T. J., Dempsy, H., Kirn, D. H., factor 1. Kidney Int. 51, 553±555.
Brugnara, C., and Goldberg, M. A. (1994). Efficacy of dif- Shannon, K. M., Naylor, G. S., Torkildson, J. C., Clemons, G. K.,
ferent dosing regimens for recombinant human erythro- Schaffner, V., Goldman, S. L., Lewis, K., Bryant, P., and
poietin in a simulated perisurgical setting: the importance Phibbs, R. (1987). Circulating erythroid progenitors in the ane-
of iron availability in optimizing response. Am. J. Med. 96, mia of prematurity. N. Engl. J. Med. 317, 728±733.
139±145. Shannon, K. M., Keith, J. F., 3rd Mentzer, W. C.,
Sakanaka, M., Wen, T. C., Matsuda, S., Mauda, S., Morishita, E., Ehrenkranz, R. A., Brown, M. S., Widness, J. A.,
Nagao, M., and Sasaki, R. (1998). In vivo evidence that erythro- Gleason, C. A., Bifano, E. M., Millard, D. D., and Davis, C. B.
poietin protects neurons from ischemic damage. Proc. Natl (1998). Recombinant human erythropoietin stimulated erythro-
Acad. Sci. USA 95, 4635±4640. poiesis and reduces erythrocyte transfusions in very low birth
Sasaki, H., Bothner, B., Dell, A., and Fukuda, M. (1987). weight preterms infants. Pediatrics 95, 1±8.
Carbohydrate structure of erythropoietin expressed in Chinese Shaw, G., and Kamen, R. (1986). A conserved AU sequence from
hamster ovary cells by a human erythropoietin cDNA. J. Biol. the 30 untranslated region of GM-CSF mRNA mediates selec-
Chem. 262, 12059±12076. tive mRNA degradation. Cell 46, 659±667.
962 Jerry L. Spivak

Shoemaker, C. B., and Mitsock, L. D. (1986). Murine erythropoie- Sullivan, P. S., Hanson, D. L., Chu, S. Y., Jones, J. L., and
tin gene: cloning, expression and human gene homology. Mol. Ward, J. W. (1998). Epidemiology of anemia in human immu-
Cell. Biol. 6, 849±858. nodeficiency virus (HIV)-infected persons: results from the
Siciliano, M., Tomasello, D., Milani, A., Ricerca, B. M., Storti, S., multistate adult and adolescent specturm of HIV disease sur-
and Rossi, L. (1995). Reduced serum levels of immunoreactive veillance project. Blood 91, 301±308.
erythropoietin in patients with liver cirrhosis and chronic ane- Syed, R. S., Reid, S. W., Li, C., Cheetham, J. C., Aoki, K. H., Liu,
mia. Hepatology 22, 1132±1137. B., Zhan, H., Osslund, T. D., Chirino, A. J., Zhang, J., Finer-
Silva, M., Grillot, D., Benito, A., Richard, C., Nunez, G., and Moore, J., Elliott, S., Sitney, K., Katz, B. A., Matthews, D. J.,
Fernandez-Luna, J. L. (1996). Erythropoietin can promote Wendoloski, J. J., Egrie, J., and Stroud, R. M. (1998).
erythroid progenitor survival by repressing apoptosis through Efficiency of signalling through cytokine receptors depends cri-
Bcl-XL and Bcl-2. Blood 88, 1576±1582. tically on receptor orientation. Nature 395, 511±516.
Singh, A., Eckardt, K. U., Zimmerman, A., Gotz, K. H., Takeuchi, M., Takasaki, S., Shimada, M., and Kobata, A. (1990).
Hamann, M., Ratcliffe, P. J., Kurtz, A., and Reinhart, W. H. Role of sugar chains in the in vitro biological activity of human
(1993). Increased plasma viscosity as a reason for inappropriate erythropoietin produced in recombinant Chinese hamster ovary
erythropoietin formation. J. Clin. Invest. 91, 251±256. cells. J. Biol. Chem. 265, 12127±12130.
Sowade, O., Warnke, H., Scigall, P., Sowade, B., Franke, W., Tan, C. C., Eckardt, K.-U., and Ratcliffe, P. J. (1991). Organ
Messinger, D., and Gross, J. (1997). Avoidance of allogeneic distribution of erythropoietin messenger RNA in normal and
blood transfusions by treatment with epoetin  (recombinant uremic rats. Kidney Int. 40, 69±76.
human erythropoietin) in patients undergoing open-heart sur- Tan, C. C., Eckardt, K.-U., and Firth, J. D. (1992). Feedback
gery. Blood 89, 411±418. modulation of renal and hepatic erythropoietin mRNA in
Spivak, J. L. (1993). The clinical physiology of erythropoietin. response to graded anemia and hypoxia. Am. J. Physiol. 263,
Semin. Hematol. 30, 2±11. F474±F481.
Spivak, J. L., and Hogans, B. B. (1987). Clinical evaluation of a Walker, F., Nicola, N. A., Metcalf, D., and Burgess, A. W. (1985).
radioimmunoassay for serum erythropoietin using reagents Hierarchical down-modulation of hemopoietic growth factor
derived from recombinant erythropoietin. Blood 70, 143a. receptors. Cell 43, 269±276.
Spivak, J. L., and Hogans, B. B. (1989). The in vivo metabolism of Walle, A. J., Wong, G. Y., Clemons, G. K., Garcia, J. F.,
recombinant human erythropoietin in the rat. Blood 73, 90±99. and Niedermayer, W. (1987). Erythropoietin-feedback circuit
Spivak, J. L., and Seiber, F. (1983). Erythropoietin. Horm. Norm. in the anemia of end-stage renal disease. Kidney Int. 31, 1205±
Abnorm. Hum. Tissues 3, 63±96. 1209.
Spivak, J. L., Barnes, D. C., Fuchs, E., and Quinn, T. C. (1989). Wang, G. L., and Semenza, G. L. (1993). Desferrioxamine induces
Serum immunoreactive erythropoietin in HIV-infected patients. erythropoietin gene expression and hypoxia-inducible factor 1
J. Am. Med. Assoc. 261, 3104±3107. DNA-binding activity: implications for models of hypoxia sig-
Spivak, J. L., Pham, T., Isaacs, M., and Hankins, W. D. (1991). nal transduction. Blood 82, 3610±3615.
Erythropoietin is both a mitogen and a survival factor. Blood Wang, Z., Goldberg, M. A., and Scadden, D. T. (1993). HIV-1
77, 1228±1233. suppresses erythropoietin production in vitro. Exp. Hematol. 21,
Spivak, J. L., Fisher, J., Isaacs, M. A., and Hankins, D. (1992). 683±688.
Protein kinases and phosphatases are involved in erythro- Wang, G. L., Jiang, B.-H., Rue, E. A., and Semenza, G. L.
poietin mediated signal transduction. Exp. Hematol. 20, 500± (1995a). Hypoxia-inducible factor 1 is a basic-helix-loop-helix-
504. PAS heterodimer regulated by cellular O2 tension. Proc. Natl
Spivak, J. L., Ferris, D. K., Fisher, J., Noga, S. J., Isaacs, M., Acad. Sci. USA 92, 5510±5514.
Connor, E., and Hankins, W. D. (1996). Cell cycle-specific Wang, G. L., Jiang, B. H., and Semenza, G. L. (1995b). Effect of
behavior of erythropoietin. Exp. Hematol. 24, 141±150. protein kinase and phosphatase inhibitors on expression of
Srinivas, V., Zhu, X., Salceda, S., Nakamura, R., and Caro, J. hypoxia-inducible factor 1. Biochem. Biophys. Res. Commun.
(1998). Hypoxia-inducible factor 1 (HIF-1) is a non-heme iron 216, 669±675.
protein. J. Biol. Chem. 273, 18019±18022. Wasley, L. C., Timony, G., Murtha, P., Stoudemire, J., Dorner,
Steinberg, S. E., Garcia, J. F., Matzke, G. R., and Mladenovic, J. A. J., Caro, J., Krieger, M., and Kaufman, R. J. (1991). The
(1986). Erythropoietin kinetics in rats: generation and clear- importance of N- and O-linked oligosaccharides for the bio-
ance. Blood 67, 646±649. synthesis and in vitro and in vivo biologic activities of erythro-
Stopka, T., Zivny, J. H., Goldwasser, E., Prchal, J. F., Nescas, E., poietin. Blood 77, 2624±2632.
and Prchal, J. T. (1998a). Guinea pig serum erythropoietin Watson, A. J., Gimenez, L. F., Cotton, S., Walser, M., and
(EPO) selectively stimulates guinea pig erythroid progenitors: Spivak, J. L. (1990). Treatment of the anemia of chronic renal
human or mouse erythroid progenitors do not form erythroid failure with subcutaneous recombinant human erythropoietin.
burst-forming unit colonies in response to guinea pig serum Am. J. Med. 89, 432±435.
EPO. Exp. Hematol. 26, 910±914. Wedzicha, J. A., Cotes, P. M., Empey, D. W., Newland, A. C.,
Stopka, T., Zivny, J. H., Goldwasser, E., Prchal, J. F., Nescas, E., Royston, J. P., and Tam, R. C. (1985). Serum immunoreactive
and Prchal, J. T. (1998b). Human hematopoietic progenitors erythropoietin in hypoxic lung disease with and without poly-
express erythropoietin. Blood 91, 3766±3772. cythemia. Clin. Sci. 69, 413±422.
Strauss, R. (1994). Erythropoietin and neonatal anemia. N. Engl. Wen, D., Boissel, J.-P. R., Tracy, T. E., Gruninger, R. H.,
J. Med. 330, 1227±1228. Mulcahy, L. S., Czelusniak, J., Goodman, M., and Bunn, H. F.
Sui, X., Tsuji, K., Tajima, S., Tanaka, R., Muraoka, K., (1993). Erythropoietin structure±function relationships: high
Ebihara, Y., Ikebuchi, K., Yasukawa, K., Taga, T., degree of sequence homology among mammals. Blood 82,
Kishimoto, T., and Nakahata, T. (1996). Erythropoietin- 1507±1516.
independent erythrocyte production; signals through gp130 Wenger, R. H., Kvietikova, I., Rolfs, A., Gassmann, M., and
and c-kit dramatically promote erythropoiesis from human Marti, H. H. (1997). Hypoxia-inducible factor-1  is regulated
CD34‡ cells. J. Exp. Med. 183, 837±845. at the post-mRNA level. Kidney Int. 51, 560±563.
 Erythropoietin 963

Wide, L., Bengtsson, C., and Birgegard, G. (1989). Circadian structures, with implications for receptor binding. Protein Sci.
rhythm of erythropoietin in human serum. Br. J. Haematol. 2, 1373±1382.
72, 85±90. Wojchowski, D. M., Orkin, S. H., and Sytkowski, A. J. (1987).
Widness, J. A., Sawyer, S. T., Schmnidt, R. L., and Active human erythropoietin expressed in insect cells using a
Chestnut, D. H. (1991). Lack of maternal to fetal transfer of baculovirus vector: a role for N-linked oligosaccharide.
125
I-labelled erythropoietin in sheep. J. Dev. Physiol. 15, Biochim. Biophys. Acta 910, 224±232.
139±143. Wu, J., Liu, X., and Jaenisch, R. (1995). Generation of committed
Widness, J. A., Veng-Pedersen, P., Peters, C., Pereira, L. M., erythroid BFU-E and CFU-E progenitors does not require
Schmidt, R. L., and Lowe, L. S. (1996). Erythropoietin phar- erythropoietin or the erythropoietin receptor. Cell 83, 59±67.
macokinetics in premature infants:developmental, nonlinearity, Wurnig, C., Windhager, R., Schwameis. E., Kotz, R., Zoubek, A.,
and treatment effects. J. Appl. Physiol. 80, 140±148. Stockenhuber, F., and Kurz, R. W. (1996). Prevention of che-
Winearls, C. G., Pippard, M. J., Dowinging, M. R., Oliver, D. O., motherapy-induced anemia by the use of erythropoietin in
Reid, C., and Cotes, P. M. (1986). Effect of human erythropoie- patients with primary malignant bone tumors (a double-blind,
tin derived from recombinant DNA on the anaemia of patients randomized, phase III study). Transfusion 36, 155±159.
maintained by chronic haemodialysis. Lancet 2, 1175±1178. Yamaguchi, K., Akai, K., Kawanishi, G., Ueda, M., Masuda, S.,
Winkelmann, J. C., Penn, L. A., Deaven, L. L., Forget, B. G., and and Sasaki, R. (1991). Effects of site-directed removal of
Jenkins, R. B. (1990). The gene for the human erythropoietin N-glycosylation sites in human erythropoietin on its produc-
receptor: analysis of the coding sequence and assignment of tion and biological properties. J. Biol. Chem. 266, 20434±
chromosome 19p. Blood 76, 24±30. 20439.
Witthuhn, B. A., Quelle, F. W., Silvennoinen, O., Yi, T., Tang, B., Youssoufian, H., Longmore, G., Neumann, D., Yoshimura, A.,
Miura, O., and Ihle, J. N. (1993). JAK2 associates with the and Lodish, H. F. (1993). Structure, function and activation of
erythropoietin receptor and is tyrosine phosphorylated and the erythropoietin receptor. Blood 81, 2223±2236.
activated following stimulation with erythropoietin. Cell 74, Zanjani, E. D., Ascensao, J. L., McGlave, P. B., Banisadre, M.,
227±236. and Ash, R. C. (1981). Studies on the liver to kidney switch of
Wlodawer, A., Pavlovsky, A., and Gustchina, A. (1993). erythropoietin production. J. Clin. Invest. 67, 1183±1188.
Hematopoietic cytokines: similarities and differences in the