What Evidence Supports Use of Erythropoietin as a Novel Neurotherapeutic?

ABSTRACT: In its hormonal role, erythropoietin is produced by the kidney in response to hypoxic stress and signals the bone marrow to increase the number of circulating erythrocytes. It has become clear in recent years, however, that erythropoietin and its receptor are members of a cytokine superfamily that mediates diverse functions in nonhematopoietic tissues. Nonhormonal erythropoietin actions include a critical role in the development, maintenance, protection, and repair of the central nervous system (CNS). Our group has found serendipitously that recombinant human erythropoietin administered into the systemic circulation is not strictly excluded from the brain. Human recombinant erythropoietin appears within the cerebrospinal fluid in neuroprotective concentrations, probably by translocation initiated by binding to the erythropoietin receptor on the luminal surface of the endothelium. This observation suggested that recombinant human erythropoietin could be therapeutic for CNS diseases, a possibility further supported by positive findings in a model of ischemic stroke. Recombinant human erythropoietin administered systemically either in advance of, or up to 3 hours after, a cerebral arterial occlusion in rats prevents apoptosis of neurons within the ischemic penumbra and reduces infarction volume by 75%. Erythropoietin also dramatically reduces postinfarct inflammation in this model. Other brain and spinal cord injuries such as mechanical trauma, experimental autoimmune encephalitis or subarachnoid hemorrhage also respond favorably to erythropoietin administered within a similar window of time. In addition to ameliorating neuronal injury, erythropoietic therapy also directly modulates neuronal excitability and acts as a trophic factor for neurons in vivo and in vitro. Erythropoietin may therefore provide benefit in epileptic or degenerative neurologic diseases. Given the outstanding safety record for recombinant human erythropoietin after more than a decade in widespread clinical use, the results of multiple preclinical investigations suggest that this cytokine or its derivatives may be useful for treatment of a variety of nervous system diseases. [ONCOLOGY 16(Suppl 10):79-89, 2002]

For nearly a century, itwasknown that a substance withinthe systemic circulation regulates the serum hemoglobin concentration ofmammals. Following many years of futile attempts to isolate this factor, whichis present in only trace amounts in the plasma, the glycoprotein erythropoietinwas finally isolated from the urine of patients with aplastic anemia.[1]Erythropoietin DNA probes were constructed from tryptic fragments of the isolateand the gene cloned in 1985.[2,3] In the normal adult, the kidney produceserythropoietin for control of the circulating erythrocyte mass in a classichormonal feedback loop responding to low tissue oxygenation.[4,5] The targets ofthis renal erythropoietin are the colony-forming units-erythroid (CFUe) thatreside within the bone marrow for which erythropoietin acts as a survival anddifferentiation factor.

Determination of the structure of erythropoietin and itsreceptor identified these molecules as members of a large family of relatedmacromolecules with pleiotropic activities involving cell survival, growth,differentiation, and inflammation.[5] Structurally, erythropoietin is mostclosely related to growth hormone and prolactin and more distantly to severalinterleukins— important molecules in the modulation of inflammation.[6]Following identification of the erythropoietin and the erythropoietin-receptorgenes, tissue surveys showed surprisingly that these proteins were alsoexpressed in other tissues,[7] including the brain.[8] These findings clearlysuggested that the target(s) of erythropoietin extend beyond the bone marrowwhere its action is likely not hormonal in nature.

Erythropoietin Action Within the Nervous System

Effects of Direct Administration Into the Brain

The first studies evaluating the function of brain-derivederythropoietin suggested that it possessed trophic activities. Specifically,recombinant human erythropoietin supported the growth and differentiation ofseptal cholinergic forebrain neurons in vitro, reminiscent of its activity inthe bone marrow. However, this erythropoietic therapy also enhanced the survivalof injured neurons in vivo after intrathecal injection.[9] Initial study of thecell-type specific expression of these proteins concluded that astrocytes werethe source of "brain" erythropoietin and the erythropoietin-receptor-expressingneurons were the presumed target.[10,11]

Primary rodent astrocytes maintained in vitro produceerythropoietin as a response to hypoxic stress.[12] Notably, when theseastrocytes are cocultured with neurons, the erythropoietin induced by hypoxicstress protects adjacent neurons from injury.[10] More extensive study of thecell-type-specific expression patterns of the erythropoietin/erythropoietin-receptorgenes in the nervous system has clarified that a much more complicatedrelationship exists as both neurons and astrocytes can express either protein(or both), with a pronounced variation depending upon the anatomic region andthe developmental stage that is examined.[13]

The relevance of these findings for the adult mammalian brainwas confirmed by the results of in vivo experiments performed using rodentmodels. These findings clearly demonstrated that ischemic neuronal injury couldbe impressively reduced by the direct administration of erythropoietin into thebrain or the lateral ventricles.[14-16] Immunocytochemical studies additionallyshowed that the central nervous system (CNS) responds to hypoxic insults byup-regulating the expression of erythropoietin and erythropoietin receptor by cells within the border zone (penumbra)surrounding the necrotic core over a period of 12 to 24 hours.[15-17] Thesechanges are adaptive, as neutralizing the endogenous erythropoietin produced bythis up-regulation markedly amplifies injury.[14]

The mechanism(s) of action of recombinant erythropoietin intissue protection could involve either the protection of neurons from death, thereduction of inflammatory responses, or both. These possibilities have not beenadequately investigated to date. Erythropoietin has also been shown to affectstem cell recruitment and differentiation,[18] which could provide importantrestorative benefits. In the case of ischemic brain injury, acuteneuroprotection appears similar to the role of erythropoietin in erythropoiesis.Specifically, in the control of red cell production, erythropoietin does notfunction strictly as a growth factor, but rather, prevents death of erythrocyteprecursors by inhibiting programmed cell death or apoptosis.[4] To accomplishthis, erythropoietin induces a gene expression program that includes activationof members of the bcl-2 family of antiapoptotic proteins, especially the longform of BCLx (Bcl-xL).

Regardless of type, brain injury commonly initiatesapoptosis. Using an in vitro model, prior study has demonstrated that theneuroprotective effects of recombinant erythropoietin following hypoxic orneurotoxin stress requires protein synthesis,[10] which is required formodulation of programmed cell death. Erythropoietin protects neurons fromchemical hypoxia and cerebral ischemic injury by up-regulating Bcl-xLexpression.[19] However, in addition to its acute effects on neuronal programmedcell death, erythropoietin also provides beneficial effects through modulationof the inflammation reaction to brain injury. This action appears to account fora major component of its neuroprotective activity (see below).

One way erythropoietin could affect inflammation is bymodulation of members of the NFkB family that are principal regulators ofinflammatory genes.[20,21] For example, previous work has shown that NFkBitself is strongly up-regulated following traumatic spinal cord injury, aproduct of macrophages/microglia, endothelial cells, and neurons.[22] One NFkB-dependentgene product is inducible nitric oxide synthase, which has been shown toincrease the first day after injury, and peaks by day 7.[23] Erythropoietinis known to modulate nitric oxide production in the brain,[24] as well in othertissues.[25] Recently, erythropoietin has been shown to signal through the NFkBpathway,[26] as well as by the Jak2/STAT5 system.[27-29] Further study will benecessary to determine whether the prominent effects of recombinanterythropoietin on both primary and secondary injury depend in part uponregulation of NFkB or other pathways.

In addition to neurons and astrocytes, capillary endothelialcells forming the blood-brain barrier (BBB) also express erythropoietin receptorat high levels.[17,30] Similar to the response of brain cells to hypoxia, thecapillary endothelium responds to hypoxia by up-regulating erythropoietinreceptor. In experimental models, administration of erythropoietin not only canprevent endothelial cell death after hypoxic injury, but also is effective inpreventing apoptosis of the endothelium following nonischemic injuries.[31] Theprotective effects of erythropoietin on the vasculature may not be limited tocapillaries, however, as erythropoietin also potently protects vascular smoothmuscle cells from injury-induced apoptosis.[32]

In sum, the results of multiple studies of animal models have definitively established that an endogenous erythropoietin/erythropoietin-receptor system functions within the CNS outside of hormonal (feedback) control. This local system is activated by hypoxic stressors, is protective of neurons and endothelial cells, and can be beneficially activated by the direct intrathecal administration of erythropoietin (Figure 1).

Central Neurologic Effects of Peripheral Administration

In the course of studying the activities of proinflammatorycytokines on cognitive behavior, results in our laboratory suggested thatsystemically administered erythropoietin also has central neurologic effects.Using rodent models of conditioned taste aversion and spatial orientation, weobserved potent behavioral effects of systemically administered recombinanthuman erythropoietin. (unpublished experiments). These findings were unexpectedin light of the prevailing concept that the large size of the erythropoietinmolecule (> 30,000 kD) and its extensive glycosylation would exclude itstransfer from the systemic circulation across the BBB.

Of note, several human studies had previously concluded thatsystemically administered erythropoietin was in fact excluded from the brain inthe presence of an intact BBB.[12,23] This was clearly not true in ourexperiments using normal animals. In retrospect, these diametrically opposedconclusions can be reconciled by the knowledge that significant BBB transferrequires doses of recombinant erythropoietin significantly higher than thoseused for conventional treatment of anemia in humans.

Employing immunohistochemical techniques, we further discovered that certain capillaries, especially within white matter, prominently express erythropoietin receptor (Figure 2). After hypoxic injury, erythropoietin receptor is greatly up-regulated within the microvessels surrounding the region of injury (the penumbra)[16,17] to a level similar to that of white matter capillaries. Other stressors, such as epileptic seizures, also dramatically stimulate up-regulation of erythropoietin receptor within the microvasculature (Figure 3). In follow-up studies, we examined expression of erythropoietin receptor within the microvasculature of the BBB. The results show that both the endothelial cell and the specialized astrocyte that completely invests the brain side of the capillary with endfeet possess abundant cytoplasmic and surface membrane erythropoietin-receptor protein. Evaluation of anti-erythropoietin-receptor labeling by transmission electron microscopy also showed that the protein is present on the surface of the endothelial cell, within vesicles in the cytoplasm of the endothelial cell, and within the astrocytic endfeet (Figure 4).

The capillaries of the BBB are located at a strategicinterface between the systemic circulation and the relatively privilegedenvironment of the brain. The presence of erythropoietin receptor upon theendothelial cell provides a potential and specific access route forerythropoietin into the brain. This mechanism has been previously documented fora number of large proteins, including other cytokines,[34] insulin,[35]transferrin,[36] and leptin.[37] Specifically, upon binding to its cognatereceptor upon the luminal surface of the endothelial cell, the macromolecule isinternalized and carried across to the brain side for release, in a processtermed transcytosis.[38] Obviously, breakdown of the integrity of the BBB, asfrequently but not universally occurs after injury (although often with a delaybeyond the useful therapeutic time window), is also a route for otherwiseimpermeant macromolecules within the circulation to gain access into the brain.

As a direct test of the hypothesis that recombinant human erythopoietin can cross the intact BBB, large doses were administered systemically (5,000U/kg body weight, intraperitoneally [IP]) to normal rats. Erythropoietin, in fact, did appear within the CSF but with a delay of approximately 1 hour, peaking by 3 to 4 hours (Figure 5), but at concentrations within the effective therapeutic range for ischemic stroke. The observed transfer of 0.8% to 1% of administered dose compares well with what has been observed for other macromolecules (eg, insulin).[35] In other animal models, including nonhuman primates, similar results for systemically administered erythropoietic therapy have also been reported.[39]

The erythropoietin receptor of the endothelial cell possessesa lower affinity for erythropoietin than the receptor expressed by erythrocyteprecursors. However, both receptor subtypes are products of the same gene. Thereis a much higher density of receptors present on the surface of the endothelialcell, as expected for mediation of transcytosis. Given that in its hormonal roleerythropoietin normally circulates in the low picomolar range, endogenouserythropoietin can interact appreciably with the capillary receptor only undersevere and chronic hypoxic stress that stimulates increases of the serumerythropoietin concentration to more than 100-fold above normal. Theseconsiderations may explain why high doses of recombinant human erythropoietinare required for significant BBB transport to occur.

The existence of a BBB/erythropoietin-receptor relationship,in conjunction with the demonstration that erythropoietin administered into theperipheral circulation results in levels of erythropoietin in the CSF atconcentrations well within those observed to provide neuroprotection in in vitromodels, predicted that recombinant human erythropoietin administeredperipherally could be useful in the treatment of neurologic injury. A number ofanimal models have been used to evaluate this possibility, including ischemicstroke, mechanical brain trauma, neurotoxic injury, experimental autoimmuneencephalitis,[40] subarachnoid hemorrhage,[41-43] spinal cord ischemia,[44]spinal cord trauma,[45] and acute retinal ischemia.[46]

Ischemic Stroke—Observations derived from a three-vessel transient occlusion model of focal ischemia (stroke) within the middle cerebral artery of rats is illustrative of the markedly neuroprotective action of recombinant erythropoietin.[40] In these experiments, a very large penumbra (area at risk) is produced by permanently ligating the common carotid and distal middle cerebral artery on the same side, followed by a 1-hour reversible occlusion of the contralateral common carotid. This procedure produces a small necrotic core and a large surrounding volume of potentially viable neurons within the frontal cortex. Ischemia, however, triggers apoptosis within the susceptible neuronal population residing in the penumbra. Untreated, this region is nonviable when evaluated after 24 hours (Figure 6).

If erythropoietin is administered before, or up to 3 hoursafter, the induction of ischemia, however, the entire penumbra is rescued. Ifadministration is delayed for 6 hours, only partial penumbral salvage isobserved. By 9 hours, erythropoietin is without apparent benefit when assessedat 24 hours. This temporal relationship for efficacy is fully consistent with amodulation of apoptosis, which by 9 hours is irreversibly activated. Theseeffects of recombinant human erythropoietin can be directly visualized by usinga staining method that identifies the DNA degradation products specificallyproduced during the course of programmed cell death; however,virtually none areseen within the penumbral region of animals having received recombinanterythropoietin.[47]

Spinal Cord Ischemia—Recombinanthuman erythropoietin also prevents neuronal apoptosis in a rabbit model oftransient spinal cord ischemia[44] that reproduces injury associated with aorticsurgical procedures such as aneurysm repair.[48] In this study, a single dose oferythropoietin given IP at 350 to 1,000 U/kg body weight immediately upon therestoration of flow following a 20-minute occlusion of the abdominal aorta isassociated with a complete recovery of motor function within 24 hours followinginjury.[44] In contrast, animals that received saline exhibited only minimalrecovery of motor function.

Mechanical Trauma orAutoimmune Activation— Histologicanalyses of the brain after focal ischemia identified another major benefit ofrecombinant human erythropoietin administration: a much reduced infiltration ofthe injured region by inflammatory cells. This observation suggested that oneimportant action of erythropoietic therapy might be to temper the intensereactive inflammation following any injury that tends to expand the volume ofaffected tissue. We therefore used three animal models to determine whetherrecombinant erythropoietin was beneficial in the setting of more purelyinflammatory injuries caused by mechanical trauma or autoimmune activation.

First, the effect of systemic recombinant erythropoietin on the extent of brain injury arising from blunt trauma was studied in the mouse.[40] For this experiment, anesthetized mice were subjected to a precisely calibrated blow to the intact calvarium and subsequently received either saline or recombinant human erythropoietin (5,000 U/kg body weight, IP) before, immediately following, or up to 9 hours after injury. These animals were allowed to recover for 10 days, at which time their brains were removed and examined histologically. Animals that had received only saline after injury exhibited extensive cavitary lesions (Figure 7) with abundant inflammatory cell infiltrates. In contrast, those that received erythropoietin within 3 hours following injury exhibited virtually no cavitation or active inflammation.

Spinal cord injury also is characterized by intense inflammation, frequently with a devastating clinical outcome. In these cases, much of the neurologic dysfunction actually arises from subacute disruption of white matter (ie, the ascending and descending long tracts), particularly due to the apoptosis of oligodendrocytes.[49] This process is driven to a large extent by reactive inflammation.[22] A rat model study of spinal cord injury produced by direct contusion to the dura overlying the cord showed large cavitary lesions when evaluated 7 days later (Figure 8), plus profound motor dysfunction (Figure 9). Administration of a single dose of erythropoietin (5,000 U/kg body weight, IP) at the time of injury almost completely prevented cavitation and provided for nearly normal motor function by several weeks after injury. Quantification of oligodendrocyte apoptosis in erythropoietin-treated animals using TUNEL labeling within the white matter fasciculus cuneatus demonstrated virtually no apoptosis of the oligodendrocytes that produce the myelin sheaths, in contrast to an extensive presence in saline-treated animals.

Effects of erythropoietic therapy were also assessed in the primary inflammatory injury associated with experimental autoimmune encephalitis.[40] In this model, guinea pig myelin basic protein is injected into a footpad of female Fisher rats along with complete Freund’s adjuvant. A stereotyped inflammatory reaction occurs within 10 days, producing a severe neurologic disability, including paraplegia and death. This intense inflammatory reaction and its neurologic sequelae can be completely prevented by administration of the anti-inflammatory agent dexamethasone. However, after terminating glucocorticoid-induced immunosuppression, the symptoms subsequently flare. Thus, steroids only delay the development of clinical experimental autoimmune encephalitis. In contradistinction, recombinant erythropoietin administered beginning at day 3 after immunization not only significantly blunts the neurologic symptoms but also does not exhibit a rebound effect upon discontinuation (Figure 10).[49a]

In addition to its effects on apoptosis and inflammation,erythropoietin also affects the excitability of neurons both acutely andchronically.[9,50,51] One common form of potentially lethal neuronal stress isthought to occur during excessive excitatory activity. For example, theglutamate analog kainate produces seizures and subsequent neuronal injury andcell death. These seizures are limbic in nature, involving the hippocampus.Kainate has therefore been used as one model of temporal lobe epilepsy. Whenadministered intraperitoneally in this model, erythropoietin (5,000 U/kg bodyweight) delays both the onset of clinical seizures and reduces the severity ofthe seizures. However, unlike its action in ischemia and inflammatory injurymodels, recombinant human erythropoietin must be administered in advance ofkainate to be effective. If the erythropoietin dose is delayed until the onsetof seizures, it is completely without effect.[40]

In summary, recombinant erythropoietin does not behave in this model as a conventional antiepileptic (ie, via an immediate effect mediated at the cell membrane), but rather, it initiates a specific and protective gene expression program. Further evidence for gene activation produced by erythropoietin is the observation that the effects of a single dose at seizure threshold persist for up to 1 week (Figure 11).

Beneficial Effects inAcute Retinal Ischemia

Although erythropoietin is expressed in the human fetalretina,[13] it was previously unknown whether its expression persists into theadult retina, and whether it has a physiologic role in the retina. Usingimmunohistochemistry and Western blot studies, we observed that botherythropoietin and erythropoietin receptor were present in the adult, andfurther, that expression of the erythropoietin receptor could be significantlyincreased in the retina by ischemia,[46] similar to the brain. This findingsuggested that the ischemic retina can up-regulate recovery pathways that mightbe enhanced by the exogenous application of erythropoietin.

Studies evaluating the potential utility of erythropoietin were carried out using a model of transient global retinal ischemia produced by raising intraocular pressure to above systolic blood pressure reversibly.[46] This is a relevant model for retinal diseases such as acute glaucoma. Systemic administration of recombinant erythropoietin before or immediately after retinal ischemia not only greatly reduced histopathologic damage (Figure 11), but also promoted functional recovery as assessed by electroretinography. Neutralization of endogenous erythropoietin with soluble erythropoietin receptor exacerbated ischemic injury, which supports a crucial role for an endogenous erythropoietin/erythropoietin-receptor system in the survival and recovery of retinal neurons after an ischemic insult. Erythropoietin therapy also significantly diminished TUNEL labeling of neurons in the ischemic retina, implying an antiapoptotic mechanism of action.

Erythropoietin as Adjuvant Therapy of Cancer?

In addition to primary treatment of neurologic disease,recombinant human erythropoietin could also prove useful as adjuvant therapy toreduce injury to normal tissue following cancer treatment using surgery,radiation, or chemotherapy. A rationale for evaluating such use arises from theconsideration that primary treatments are associated with localized ischemia andinflammation. Critical exploratory work needs to be performed to assess therelative contributions of erythropoietin, not only to reducing the severity ofthe adverse effects of primary therapy, but also to reducing treatment efficacyby stimulating resistance in tumor cells.

Further Insights Into Mechanisms of Action

Some Hurdles to Be Cleared

Several theoretical issues concerning recombinant humanerythropoietin therapy may temper enthusiasm for its use in treating some typesof neurologic disease.

• First,any condition that requires multiple dosing, eg, Parkinson’s disease, willinevitably be associated with an increased serum hemoglobin concentration. Therheological consequences may be problematic. For example, recent study oftransgenic animal models with high levels of endogenous expression oferythropoietin shows unequivocally that the hematocrit markedly worsens clinicaloutcome after ischemia.[52]

• Second,even in the absence of significant effects on the hematocrit, however,erythropoietin has been shown to possess procoagulant properties. For example,thrombocytes respond to recombinant erythropoietin administration, becoming morereactive, providing for adverse effects in canine models[53] as well as inhemodialysis patients.[54] The effects on platelets may depend on the dose oferythropoietin as well.[55]

• Third,erythropoietin has been reported to drive angiogenesis and growth ofexperimental tumor models.[55a] Whether these observations also have relevancefor administration of high-dose recombinant human erythropoietin in patientswith neurologic disease needs to be determined.

• Fourth,recombinant erythropoietin possesses well-known vasoconstrictive properties thatlead to hypertension in some patients. This could limit its clinical use incertain populations.

These considerations suggest that recombinant humanerythropoietin may not be the ideal therapy for chronic clinical situations.Work is currently under way to develop analogs of erythropoietin lackingerythropoietic stimulation, while retaining full neurologic andanti-inflammatory actions.

Phase II Clinical Trials

Until recently, most clinical trials attempting to extend thepositive findings obtained in animal models of neurologic injury to humandisease have failed to show a benefit in humans. In the case of acute stroke,for example, more than 28 major clinical trials have evaluated a wide variety oftherapeutic categories, but all have failed except for tissue plasminogenactivator (tPA).[56] The clinical utility of tPA, however, has beenunfortunately limited due to its contraindicated use in a sizable proportion ofpatients. A more generally applicable therapy for acute ischemic and hemorrhagicstroke is still needed.

On the positive side, successful results of the first phaseII trial of the use of recombinant human erythropoietin in the treatment ofnonhemorrhagic, acute ischemic human stroke have been reported.[57] Inclusioncriteria included age less than 80 years, ischemic stroke in the middle cerebralartery, onset of symptoms less than 8 hours before the first dose of recombinanterythropoietin, clear deficit on stroke scales, and stroke confirmed bydiffusion-weighted magnetic resonance imaging (MRI).

End points were neurologic scoring (National Institutes ofHealth and Scandinavian stroke scales), clinical outcome rating on day 30 (Barthelindex and modified Rankin scale), and MRI to estimate evolution of lesion size.Recombinant erythropoietin (33,333 IU) or saline were administered intravenouslyafter inclusion and repeated 24 hours and 48 hours later. In 33 patients givenhigh-dose erythropoietin, no adverse effects were identified. Neurologicfollow-up scoring and outcome scales yielded significantly better results forpatients treated with erythropoietin compared with patients who received placebo(< .05). Thus, recombinant human erythropoietin, a well-tolerated and safecompound, appears to be the first neuroprotective agent with significantbeneficial effects in human stroke.

In conclusion, these studies suggest potential and majorroles for recombinant erythropoietin as a novel neurologic therapeutic agent.New analogs lacking erythropoietic but retaining neurologic activity aredesirable. Further studies are currently under way in advance of anticipatedclinical trials.

References:

1. Miyake T, Kung CK and Goldwasser E: Purification of human erythropoietin. J Biol Chem 252:5558-5564, 1977.

2. Jacobs K, Shoemaker C, Rudersdorf R, et al: Isolation andcharacterization of genomic and cDNA clones of human erythropoietin. Nature313:806-810, 1985.

3. Lin FK, Suggs S, Lin CH, et al: Cloning and expression ofthe human erythropoietin gene. Proc Natl Acad Sci U S A 82:7580-7584,1985.

4. Fisher JW: Erythropoietin: Physiologic and pharmacologicaspects. Proc Soc Exp Biol Med 216:358-369, 1997.

5. Lacombe C, Mayeux P: Biology of erythropoietin. Haematologica 83:724-732, 1998.

6. Frank SJ: Receptor dimerization in GH and erythropoietinaction—It takes two to tango, but how? Endocrinology 143:2-10, 2002.

7. Chikuma M, Masuda S, Kobayashi T, et al: Tissue-specificregulation of erythropoietin production in the murine kidney, brain, and uterus.Am J Physiol Endocrinol Metab 279:E1242-1248, 2000.

8. Tan CC, Eckardt KU, Firth JD, et al: Feedback modulationof renal and hepatic erythropoietin mRNA in response to graded anemia andhypoxia. Am J Physiol 263:F474-481, 1992.

9. Konishi Y, Chui DH, Hirose H, et al: Trophic effect oferythropoietin and other hematopoietic factors on central cholinergic neurons invitro and in vivo. Brain Res 609:29-35, 1993.

10. Morishita E, Masuda S, Nagao M, et al: Erythropoietinreceptor is expressed in rat hippocampal and cerebral cortical neurons, anderythropoietin prevents in vitro glutamate-induced neuronal death. Neuroscience76:105-116, 1997.

11. Masuda S, Okano M, Yamagishi K, et al: A novel site oferythropoietin production. Oxygen-dependent production in cultured ratastrocytes. J Biol Chem 269:19488-19493, 1994.

12. Marti HH, Gassmann M, Wenger RH, et al: Detection oferythropoietin in human liquor: Intrinsic erythropoietin production in thebrain. Kidney Int 51:416-418, 1997.

13. Juul SE, Anderson DK, Li Y, et al: Erythropoietin anderythropoietin receptor in the developing human central nervous system. PediatrRes 43:40-49, 1998.

14. Sakanaka M, Wen TC, Matsuda S, et al: In vivo evidencethat erythropoietin protects neurons from ischemic damage. Proc Natl Acad SciU S A 95:4635-4640, 1998.

15. Sadamoto Y, Igase K, Sakanaka M, et al: Erythropoietinprevents place navigation disability and cortical infarction in rats withpermanent occlusion of the middle cerebral artery. Biochem Biophys Res Commun253:26-32, 1998.

16. Bernaudin M, Marti HH, Roussel S, et al: A potential rolefor erythropoietin in focal permanent cerebral ischemia in mice. J CerebBlood Flow Metab 19:643-651, 1999.

17. Siren AL, Knerlich F, Poser W, et al: Erythropoietin anderythropoietin receptor in human ischemic/hypoxic brain. Acta Neuropathol (Berl)101:271-276, 2001.

18. Shingo T, Sorokan ST, Shimazaki T, et al: Erythropoietinregulates the in vitro and in vivo production of neuronal progenitors bymammalian forebrain neural stem cells. J Neurosci 21:9733-9743, 2001.

19. Wen TC, Sadamoto Y, Tanaka J, et al: Erythropoietinprotects neurons against chemical hypoxia and cerebral ischemic injury byup-regulating Bcl-xL expression. J Neurosci Res 67:795-803, 2002.

20. Baeuerle PA, Henkel T: Function and activation ofNF-kappa B in the immune system. Annu Rev Immunol 12:141-179, 1994.

21. Baeuerle PA, Baltimore D: NF-kappa B: Ten years after. Cell87:13-20, 1996.

22. Bethea JR, Castro M, Keane RW, et al: Traumatic spinalcord injury induces nuclear factor-kappaB activation. J Neurosci18:3251-3260, 1998.

23. Xu J, Kim GM, Chen S, et al: iNOS and nitrotyrosineexpression after spinal cord injury. J Neurotrauma 18:523-532, 2001.

24. Calapai G, Marciano MC, Corica F, et al: Erythropoietinprotects against brain ischemic injury by inhibition of nitric oxide formation. EurJ Pharmacol 401:349-356, 2000.

25. Squadrito F, Altavilla D, Squadrito G, et al: Recombinanthuman erythropoietin inhibits iNOS activity and reverts vascular dysfunction insplanchnic artery occlusion shock. Br J Pharmacol 127:482-488, 1999.

26. Digicaylioglu M, Lipton SA: Erythropoietin-mediatedneuroprotection involves cross-talk between Jak2 and NF-kappaB signalingcascades. Nature 412:641-647, 2001.

27. Lacombe C, Mayeux P: The molecular biology oferythropoietin. Nephrol Dial Transplant 14:22-28, 1999.

28. Yoshimura A, Misawa H: Physiology and function of theerythropoietin receptor. Curr Opin Hematol 5:171-176, 1998.

29. Bittorf T, Seiler J, Ludtke B, et al: Activation of STAT5during EPO-directed suppression of apoptosis. Cell Signal 12:23-30, 2000.

30. Yamaji R, Okada T, Moriya M, et al: Brain capillaryendothelial cells express two forms of erythropoietin receptor mRNA. Eur JBiochem 239:494-500, 1996.

31. Carlini RG, Alonzo EJ, Dominguez J, et al: Effect ofrecombinant human erythropoietin on endothelial cell apoptosis. Kidney Int55:546-553, 1999.

32. Akimoto T, Kusano E, Inaba T, et al: Erythropoietinregulates vascular smooth muscle cell apoptosis by a phosphatidylinositol 3kinase-dependent pathway. Kidney Int 58:269-282, 2000.

33. Buemi M, Allegra A, Corica F, et al: Intravenousrecombinant erythropoietin does not lead to an increase in cerebrospinal fluiderythropoietin concentration. Nephrol Dial Transplant 15:422-423, 2000.

34. Banks WA, Kastin AJ, Broadwell RD: Passage of cytokinesacross the blood-brain barrier. Neuroimmunomodulation 2:241-248, 1995.

35. Baura GD, Foster DM, Porte D, Jr, et al: Saturabletransport of insulin from plasma into the central nervous system of dogs invivo. A mechanism for regulated insulin delivery to the brain. J Clin Invest92:1824-1830, 1993.

36. Descamps L, Dehouck MP, Torpier G, et al:Receptor-mediated transcytosis of transferrin through blood-brain barrierendothelial cells. Am J Physiol 270:H1149-1158, 1996.

37. Golden PL, Maccagnan TJ, Pardridge WM: Human blood-brainbarrier leptin receptor. Binding and endocytosis in isolated human brainmicrovessels. J Clin Invest 99:14-18, 1997.

38. Broadwell RD: Endothelial cell biology and the enigma oftranscytosis through the blood-brain barrier. Adv Exp Med Biol331:137-141, 1993.

39. Farrell F, Juul S, Elliott C, et al: Erythropoietincrosses the blood brain barrier: An analysis in a nonhuman primate model(abstract 4265). Blood 98(suppl 1):148b, 2001.

40. Brines ML, Ghezzi P, Keenan S, et al: Erythropoietincrosses the blood-brain barrier to protect against experimental brain injury. ProcNatl Acad Sci U S A 97:10526-10531, 2000.

41. Springborg JB, Ma X, Rochat P, et al: A singlesubcutaneous bolus of erythropoietin normalizes cerebral blood flowautoregulation after subarachnoid haemorrhage in rats. Br J Pharmacol135:823-829, 2002.

42. Grasso G: Neuroprotective effect of recombinant humanerythropoietin in experimental subarachnoid hemorrhage. J Neurosurg Sci45:7-14, 2001.

43. Grasso G, Buemi M, Alafaci C, et al: Beneficial effectsof systemic administration of recombinant human erythropoietin in rabbitssubjected to subarachnoid hemorrhage. Proc Natl Acad Sci U S A99:5627-5631, 2002.

44. Celik M, Gokmen N, Erbayraktar S, et al: Erythropoietinprevents motor neuron apoptosis and neurologic disability in experimental spinalcord ischemic injury. Proc Natl Acad Sci U S A 99:2258-2263, 2002.

45. Gorio A, Gökmen N, Erbayraktar S, et al: Recombinanthuman erythropoietin counteracts secondary injury and markedly enhancesneurological recovery from experimental spinal cord trauma. Proc Natl AcadSci USA 99:9450-9455, 2002.

46. Junk A, Mammis A, Savitz SI, et al: Erythropoietinadministration protects retinal neurons from acute ischemia-reperfusion injury. ProcNatl Acad Sci U S A. In press.

47. Siren AL, Fratelli M, Brines M, et al: Erythropoietinprevents neuronal apoptosis after cerebral ischemia and metabolic stress. ProcNatl Acad Sci U S A 98:4044-4049, 2001.

48. Gharagozloo F, Larson J, Dausmann MJ, et al: Spinal cordprotection during surgical procedures on the descending thoracic andthoracoabdominal aorta: Review of current techniques. Chest 109:799-809,1996.

49. Liu XZ, Xu XM, Hu R, et al: Neuronal and glial apoptosisafter traumatic spinal cord injury. J Neurosci 17:5395-5406, 1997.

49a. Agnello D, Bigini P, Villa P, et al: Erythropoietinexerts an anti-inflammatory effect on the CNS in a model of experimentalautoimmune encephalomyelitis. Brain Res. In press.

50. Kawakami M, Sekiguchi M, Sato K, et al: Erythropoietinreceptor-mediated inhibition of exocytotic glutamate release confersneuroprotection during chemical ischemia. J Biol Chem 276:39469-39475,2001.

51. Kawakami M, Iwasaki S, Sato K, et al: Erythropoietininhibits calcium-induced neurotransmitter release from clonal neuronal cells. BiochemBiophys Res Commun 279:293-297, 2000.

52. Wiessner C, Allegrini PR, Ekatodramis D, et al: Increasedcerebral infarct volumes in polyglobulic mice overexpressing erythropoietin. JCereb Blood Flow Metab 21:857-864, 2001.

53. Wolf RF, Peng J, Friese P, et al: Erythropoietinadministration increases production and reactivity of platelets in dogs. ThrombHaemost 78:1505-1509, 1997.

54. Sayinalp N, Erdem Y, Ilhami Ozcebe O, et al: Recombinanthuman erythropoietin increases platelet aggregation in chronic hemodialysispatients. Thromb Res 90:195-198, 1998.

55. Beguin Y: Erythropoietin and platelet production. Haematologica84:541-547, 1999.

55a. Yasuda YT, Musha T, Tanaka H, et al: Inhibition oferythropoietin signalling destroys xenografts of ovarian and uterine cancers innude mice. Br J Cancer 84:836-843, 2001.

56. Tissue plasminogen activator for acute ischemic stroke.The National Institute of Neurological Disorders and Stroke rt-PA Stroke StudyGroup. N Engl J Med 333:1581-1587, 1995.

57. Ehrenreich H, Hasselblatt M, Lewczuk P, et al: oietin treatmentfor acute stroke: A randomized double-blind proof-of-concept trial in man. MolecularMedicine. In press.