Endogenous Erythropoietin Signaling Is Required for Normal Neural Progenitor Cell Proliferation*

Erythropoietin (Epo) and its receptor (EpoR), critical for erythropoiesis, are expressed in the nervous system. Prior to death in utero because of severe anemia EpoR-null mice have fewer neural progenitor cells, and differentiated neurons are markedly sensitive to hypoxia, suggesting that during development Epo stimulates neural cell proliferation and prevents neuron apoptosis by promoting oxygen delivery to brain or by direct interaction with neural cells. Here we present evidence that neural progenitor cells express EpoR at higher levels compared with mature neurons; that Epo stimulates proliferation of embryonic neural progenitor cells; and that endogenous Epo contributes to neural progenitor cell proliferation and maintenance. EpoR-null mice were rescued with selective EpoR expression driven by the endogenous EpoR promoter in hematopoietic tissue but not in brain. Although these mice exhibited normal hematopoiesis and erythrocyte production and survived to adulthood, neural cell proliferation and viability were affected. Embryonic brain exhibited increased neural cell apoptosis, and neural cell proliferation was reduced in the adult hippocampus and subventricular zone. Neural cells from these animals were more sensitive to hypoxia/glutamate neurotoxicity than normal neurons in culture and in vivo. These observations demonstrate that endogenous Epo/EpoR signaling promotes cell survival in embryonic brain and contributes to neural cell proliferation in adult brain in regions associated with neurogenesis. Therefore, Epo exerts extra-hematopoietic function and contributes directly to brain development, maintenance, and repair by promoting cell survival and proliferation independent of insult, injury, or ischemia.

Epo 2 is a hypoxia responsive cytokine required for production of erythroid cells. It triggers erythroid progenitor cell proliferation and differentiation by binding to its specific membrane receptor EpoR. EpoR-null mice and Epo-null mice die in utero because of lack of mature red blood cell production (1,2). However, functional EpoR has been identified in non-erythroid cells such as endothelial, muscle, and neural cells, and there is increasing evidence that Epo can act to stimulate cell proliferation, cell-specific function or promote cell survival in these tissues (3)(4)(5)(6)(7)(8). In culture, astrocytes and neurons up-regulate Epo and EpoR expression in response to hypoxia (8 -10). Epo activates Jak2/Stat5 and NF-B pathways to protect neurons from glutamate and hypoxic damage (8,11,12). In EpoR-null embryonic mice, increased apoptosis in brain is observed and EpoR-null embryonic cortical neurons in culture do not survive after 24-h exposure to hypoxia (8). Prior to death because of severe anemia in EpoR-null mice and Epo-null mice, the number of embryonic neural progenitor cells (NPCs) is reduced, and brain development appears underdeveloped, although no major structures are absent (8,13). In vitro, Epo can stimulate the differentiation of NPCs toward neurons. Animal models suggest that exogenous Epo may also be active in adult central nervous system. Epo has been shown to be neuroprotective for ischemia blunt force trauma and UV light damage (14 -16). Direct infusion of Epo into the lateral ventricles in gerbils provided neuroprotection to hippocampal CA1 neurons in experimental cerebral ischemia and prevented ischemia-induced learning disability, whereas infusion of soluble EpoR promoted neuronal degeneration (16). The protection to CA1 neurons and cognitive impairment is dose-dependent on the local Epo administration (17).
The lack of mature erythrocyte production in EpoR-null mice results in severe anemia and death in utero before day 13.5 (1,2). To determine whether the increased apoptosis in embryonic brain and affected brain development are due to anemia in utero and lack of oxygen delivery or due to endogenous Epo signaling in brain or neural cells, we developed mice that expressed EpoR driven by the endogenous EpoR promoter in hematopoietic tissue but not in the central nervous system. Homozygous mice can survive to adulthood, have normal hematocrit, and exhibit no gross morphologic defects. Mice with selective rescue of EpoR showed undetectable EpoR expression in brain. We present evidence for novel functions of endogenous Epo on NPCs and neural cell survival. In culture, EpoR from wild type (WT) mice is expressed at higher levels on NPCs than mature neurons in culture, and NPC proliferation is elevated in the presence of Epo. Examination of brain during development of the mice with selective EpoR expression revealed increased apoptosis during embryogenesis, reduced proliferation in the hippocampus, and increased sensitivity of neurons to hypoxia and glutamate toxicity. In addition, injec-tion of glutamate showed increased toxicity in these mice. These observations exemplify the neural protective activity of endogenous Epo and indicate that Epo directly stimulates proliferation in the hippocampus and subventricular zone, regions associated with adult neurogenesis.
For neurosphere cultures, dissociated cells were plated in uncoated T25 flask with supplemental basic fibroblast growth factor (bFGF) (PeproTech, Inc., Rocky Hill, NJ) until NPCs aggregated and proliferated to form neurospheres (2-3 days) (18). To test the mitogenic effect of Epo on NPCs, Epo was used in place of bFGF. The neurospheres were then collected for Western blotting. Bromodeoxyuridine (BrdUrd) (Sigma) at 25 M was added to the medium for proliferation assay. The cultures were maintained for 24 -48 h, fixed with 4% paraformaldehyde in phosphate-buffered saline, then immunocytochemically stained with rabbit anti-mouse EpoR (M-20) (1:200; Santa Cruz Biotechnology, Santa Cruz, CA) and mouse anti-nestin antibodies (1:100; Chemicon, Temecula, CA) using fluorescence-conjugated secondary antibodies (1:200; Chemicon). The stained neurons were examined under a Nikon Eclipse TE-2000-U fluorescent microscope, and the images were captured using PerkinElmer confocal image system. Fluorescence density was analyzed by Image-Pro Plus program (Media Cybernetics Inc., Silver Spring, MD).
Western Blotting-Tissues or cells were lysed with radioimmune precipitation assay buffer buffer (10 mM Tris-HCl, 1 mM EDTA, 0.1% SDS, 0.1% Na 3 VO 4 , 1% Triton X-100) and protease inhibitor (Roche Diagnostics, Indianapolis, IN), incubated on ice for 30 min, and centrifuged at 17,000 ϫ g for 10 min. The protein sample was run on NuPAGE 4 -12% BisTris gel (Invitrogen) for 1 h at 200 V. Protein was transferred to nitrocellulose by standard methods. The blot was blocked with 5% nonfat milk in Tween 20 Tris-buffered saline at room temperature for 1 h, probed with primary antibodies for EpoR, nestin, and ␤-actin (1:1000; Santa Cruz Biotechnology) at 4°C overnight, washed in Tris-buffered saline, probed with corresponding horseradish peroxidase-conjugated secondary antibody at room temperature for 1 h, and rinsed in Tris-buffered saline for chemiluminescent detection.
Development of Conditional Rescue Mice-The mouse EpoR gene was replaced by the human EpoR gene by substituting the genomic region between exon 1 to exon 8 of the endogenous mouse EpoR gene with exon 1 to exon 8 of the human EpoR gene. The human EpoR gene was inactivated using a neo cassette flanked by LoxP sites inserted into intron 7 (19). Mouse and human EpoR are functionally similar and in vivo the EpoRnull mouse can be rescued with a human transgene (8). Only mice heterozygous for the disrupted EpoR gene survive, as homozygous mice lack EpoR expression and die in utero. These heterozygous mice were crossed with transgenic mice expressing Cre recombinase under the direction of the endothelial cellspecific receptor tyrosine kinase Tek (Tie2) promoter/enhancer (20). Expression of Cre in cells that normally express Tie2, such as embryonic endothelium, which give rise to hematopoietic stem cells, results in recombination of the two LoxP sites and excision of the neo cassette that restores appropriately regulated expression of the EpoR gene in Cre-expressing cells and subsequent generations of cells derived from these cells. Mice from resultant litters were screened and crossed to obtain mice homozygous for the disrupted EpoR gene that carry the Tie2-Cre transgene (designated ⌬EpoR). All animal protocols were approved and carried out according to the guidelines of the NIDDK Animal Care and Use Committee.
Reverse Transcription-PCR (RT-PCR)-Adult brains were harvested and put in Qiagen RNAlater TM RNA stabilization reagent (Qiagen, Germantown, MD). Total RNA was isolated using the Qiagen RNeasy. First-strand cDNA was synthesized using murine leukemia virus reverse transcriptase and oligo(dT) 16 (PE Applied Biosystems, Foster City, CA). The EpoR transcripts in the brain were determined with RT-PCR using primers m-hEpoR-F2, 5Ј-GCAGTGAGCATGCCCAGGA-3Ј and m-hEpoR-R2, 5Ј-GCTTCACCAATCCCGTTCAAG-3Ј. The primers are specific for both mouse and human EpoR. Reaction conditions were optimized to give amplification of PCR product for even low levels of expression. PCR reaction was performed for 40 cycles of 30-s denaturation at 94°C, 1 min annealing at 58°C, and 1-min extension at 72°C.
Blood Analysis-For analysis of red cell indices, blood was drawn into a K 2 EDTA coated micro-hematocrit tube from the orbital sinus of anesthetized adult mice. The red blood cell count, hemoglobin, and hematocrit were determined for each sample.
Physical Characteristics and Gross Morphology-Adult wild type and ⌬EpoR (six-month-old) male mice (n ϭ 5 each) were weighed, deeply anesthetized, and perfused. Brains were dissected, weighed, length and width measured, and frozen sections prepared.
TUNEL Analysis of Conditional Rescue Embryonic Brains-Wild type and ⌬EpoR E16 embryonic brains were dissected and fixed with 4% paraformaldehyde in phosphate-buffered saline, prepared for frozen section, sliced, and then analyzed with TUNEL labeling reagent (Roche Diagnostics) followed by DAPI counter staining of nuclei. Serial sections were examined under a Nikon Eclipse TE-2000-U fluorescent microscope, and the images were captured using PerkinElmer confocal Image system.
BrdUrd Immunohistochemistry and Quantification-Threemonth-old mice were injected intraperitoneally with BrdUrd (100 mg/kg body weight). After 2 weeks of injection, mice were perfused, and the brains were dissected. Coronal sections (12 m in 2:20 series) were cut with a cryostat and mounted on Superfrost/Plus slides. Sections containing the hippocampal and subventricular zone (SVZ) regions were selected and treated sequentially with trypsin (0.1%) and HCl (2 M) followed by overnight incubation with sheep anti-BrdUrd antibody (1:100; BD Biosciences) at 4°C. After exposure to fluorescenceconjugated secondary antibody (1:100; BD Biosciences), the sections were mounted with vector DAPI mounting medium (Vector Laboratories Inc., Burlingame, CA). Serial sections were then examined with a fluorescent microscope, and the images were captured. All BrdUrd-positive cells in the hippocampus and SVZ were counted as described previously (21).
Glutamate Exposure-To assess glutamate toxicity in adult animals, 100 l of glutamate solution (25 M) was stereotaxically injected into the right ventricle of the mice. After 24 h the mice were euthanized, and the brains were dissected. Coronal sections (12 m in 2:20 series) were collected on Superfrost/Plus slides for apoptosis analysis by an indirect TUNEL labeling assay (Roche Diagnostics). For in vitro studies, cultured neural cells were exposed to different dose (0 -500 M) glutamate for 5 min. The glutamate was washed out completely after the exposure. The culture was examined under phase-contrast microscope after 24 h of glutamate exposure.
Statistical Analysis-Standard deviations and p values determined by the Student's t test were calculated by standard methods (Microsoft Excel; Microsoft Corporation, Redmond, WA).

Neural Progenitor Cells Express Higher Level of Epo Receptors
than Mature Neurons-We used a primary hippocampal cell culture system to evaluate whether EpoR is expressed in mature neurons or NPCs. Immunohistochemical colocalization studies showed that EpoR was colocalized with both mature neurons (MAP-2ϩ) ( Fig. 1) and NPCs (Fig. 2, A-D). Qualitatively, we found that NPCs gave more intense staining for EpoR than mature neurons (Fig. 2E). Thus, we compared the level of EpoR in two cell culture systems; one system contained predominantly neurons, and the other contained predominantly a NPC population (primary hippocampal culture and neurosphere culture, respectively). EpoR level was found higher in neurospheres than in primary cultures (Fig. 2F). These data suggest that EpoR is down-regulated as NPC terminally differentiates to mature neurons. We have reported that the expression of EpoR in the mouse brain peaks at mid gestation and then subsequently decreases to modest levels in the adult (22,23). This difference in EpoR expression level on NPC and mature neurons may account in part for the decrease of EpoR during development (22,23) because of the reduction of the NPC ratio.
Epo Stimulates Neural Progenitor Cell Proliferation-To investigate whether NPCs exhibit a direct growth response to Epo, Epo was added to NPC culture medium (Fig. 3). The number of NPCs is increased in the presence of Epo, and dose response indicated that 10 units/ml of Epo gave the optimum effect. Epo at 10 units/ml increased the number of nestin positive cells by 3-fold (Fig. 3, E and G). To test the proliferation response to Epo, the culture medium was supplemented with BrdUrd, a thymidine analogue incorporated into newly synthesized DNA of replicating cells. Double labeling of BrdUrd and nestin of 5 day cultures showed that about 50% NPC prolifer-ated in response to 10 units/ml Epo (Fig. 3, E and H). The Epo response is about half as strong as that of bFGF (Fig. 3, F, H, and I). In the presence of bFGF (20 ng/ml), the number of NPC increased by 4-fold, and almost all NPC show BrdUrd incorporation. However, the combination of Epo and bFGF treatment showed no additive response on NPC proliferation (cell number) compared with bFGF alone.
Generation of Brain EpoR-null Mice-Insertion of the neo cassette in intron 7 of the EpoR gene disrupts EpoR expression, and homozygous mice exhibit a phenotype similar to mEpoR Ϫ/Ϫ mice and die in utero because of lack of mature red blood cell production (19). Control mice were obtained by unconditional deletion of the neo cassette using a cytomegalovirus/Cre transgene (19). Conditional expression of EpoR in hematopoietic tissue but not in brain was obtained by breeding mice carrying the disrupted EpoR genetic background with the Tie2-Cre transgenic mouse (⌬EpoR) (20). Embryonic expression of Tie2-Cre in endothelium that gave rise to endothelial cells and hematopoietic stem cells rescued the EpoR-null phenotype by allowing for EpoR expression in early erythroid progenitor cells and restoring production of mature erythrocytes. ⌬EpoR mice exhibited no gross morphologic defects. Red cell indices demonstrated that the ⌬EpoR selective rescued mice had similar red blood cell content as their control peers. The red blood cell count was 8.8 Ϯ 0.9 M/l for ⌬EpoR mice and 8.8 Ϯ 1.6 M/l for control mice. The hemoglobin values were 13.2 Ϯ 1.0 g/dL for ⌬EpoR mice and 14.2 Ϯ 2.4 g/dL for control mice, and the hematocrit was 42.3 Ϯ 1.5% for ⌬EpoR mice and 44.2 Ϯ 7.1% for control mice. RT-PCR confirmed the down-regulation  of EpoR expression in brain in the ⌬EpoR mice (Fig. 4A). We used Western blotting to confirm the marked reduction of EpoR protein in brain as expected for the selective expression of EpoR only in tissue derived from Tie2-expressing cells such as hematopoietic and endothelial cells (Fig. 4A).
Epo Signaling Affects Neuron Development-Previously we demonstrated that embryonic brain of EpoR-null mice exhibited increased apoptosis, and neural cultures from these mice at E10.5 showed increased sensitivity to hypoxia suggesting an intrinsic defect during embryonic neural development as an indirect consequence of disrupted EpoR expression in hematopoietic/erythroid progenitor cells or possibly due to loss of Epo signaling in neural cells (8). The selective expression of EpoR in the ⌬EpoR mice restores normal hematopoiesis and the mice appear normal confirming that whereas EpoR expression in hematopoietic tissue is required for production of mature red blood cells and survival, EpoR expression in the central nervous system is not critical for life. We observed that during embryo-genesis, ⌬EpoR mice showed increased apoptosis in the brain compared with control mice (Fig. 4, B and C). The extent of apoptosis indicated by TUNEL positive cells was 2-fold greater in ⌬EpoR embryonic brain (Fig. 4D). These data indicate that endogenous Epo signaling is required for normal brain development. The increased apoptosis in brain observed in EpoRnull mice prior to death in utero is not only a consequence of disrupted erythroid progenitor cell differentiation and lack of oxygen delivery (8). A reduction of nestin protein in E16 ⌬EpoR mouse brain detected by Western blotting (Fig. 4A) is consistent with the increased apoptosis observed in ⌬EpoR embryonic brain.
To determine that lack of Epo signaling on neuronal cells resulted in an intrinsic defect rather than an indirect interaction with the microenvironment, neural cultures from ⌬EpoR embryos were prepared and exposed to adverse challenges. We observed poor cell survival from ⌬EpoR hippocampus at E16 compared with comparable cultures from wild type hippocampus (Fig. 5). After 24 h in cultures, the number of hippocampal cells from ⌬EpoR mice was reduced to two-thirds that from mice with targeted inactivation of EpoR in brain (⌬EpoR). Right, Western blotting for E16 brain EpoR, nestin, and ␤-actin in WT and ⌬EpoR mice. B-C, apoptosis analysis of the cortex from E16 embryo brain from WT (B) and ⌬EpoR (C) mice. Blue indicates DAPI staining to visualize nuclear DNA, and red indicates TUNEL positive cells. Scale bar, 100 m. D, the % of apoptotic cells in the region of the cortex from E16 embryos is shown for ⌬EpoR and wild type mice. wild type mice. Furthermore, under hypoxic conditions (5% O 2 ) for 24 h, less than 10% of the hippocampal cells from ⌬EpoR mice survive compared with 25% survival of wild type hippocampal cells. These results are consistent with the increased sensitivity to hypoxia of embryonic neurons from EpoR-null mice earlier in development (8). Neither ⌬EpoR nor wild type E16 cultures survived for 24 h at 2% O 2 in the absence of trophic factors, suggesting that overall, E16 cultures are more susceptible to low oxygen tension than our E10.5 cultures studied previously (8). These data demonstrate the neuroprotective effects of erythropoietin receptor expression during development and suggest that endogenous erythropoietin signaling is an important intrinsic response contributing to neuron survival in the embryo. In untreated adult animals, apoptotic cells in brain were not detected in either ⌬EpoR or wild type mice.
Brain Epo Affects Adult Neural Cell Proliferation and Glutamate Sensitivity-In adult mice, in vivo neurogenesis is restricted to the SVZ of the lateral ventricle and the hippocampus (24). To examine the Epo signaling requirement for adult NPC proliferation, BrdUrd was injected intraperitoneally into ⌬EpoR and control animals and immunoassayed. In the hippocampus proliferating cells in the region of the dentate gyrus that incorporated BrdUrd were visualized (Fig. 6, A-D). The number of BrdUrd-labeled cells in both hemispheres is decreased in the ⌬EpoR mice to one-thirds that of control mice (Fig. 6I), which reveals less NPC proliferation because of lack of EpoR expression in neural cells and the corresponding lack of endogenous Epo signaling. This suggests fewer proliferating NPC or slower cycling time for NPC from ⌬EpoR mice. BrdUrd incorporation in the SVZ was similarly visualized (Fig. 6, E-H). A reduction in the number of proliferating cells in the SVZ to about 33% of the control value, similar to the decrease found in the hippocampus, was observed (Fig. 6J). These data suggesting fewer NPCs in ⌬EpoR brain are consistent with the reduction in ⌬EpoR brain nestin at E16 detected by Western blotting (Fig.  4A). In contrast to the reduction in proliferating NPC in the ⌬EpoR mice, no difference was observed between wild type mice and control mice expressing exclusively human EpoR. Body weight, brain weight, length and width and dimensions of the olfactory bulb, cerebral cortex, and cerebellum were not different between ⌬EpoR and wild type mice. The brain/ body weight ratio was 0.0125 Ϯ 0.0006 for wild type mice and 0.0118 Ϯ 0.0005 for ⌬EpoR mice. This is a small but statistically significant reduction in the brain/body weight ratio for ⌬EpoR mice (p Ͻ 0.05).
In addition to protection against hypoxic or ischemic challenge, Epo can protect cultured neurons from glutamate-induced neuronal death (25). Glutamate is an excitatory neurotransmitter that can act as a neurotoxin in stroke, the central nervous system trauma, and epilepsy (26). The mechanisms of glutamate toxicity can signal through both ionotropic glutamate receptor such as NMDA (immediate, excitotoxicity), and metabotropic glutamate receptor (mGluR) (delayed, apoptosis) (27). We observed that loss of EpoR increased sensitivity to glutamate in neural cultures from ⌬EpoR mice (Fig. 7, A-D) and that the dose response curve for glutamate toxicity was shifted to the left for ⌬EpoR cultures compared with wild type cultures (Fig. 7E). In adult mice Epo is reported to dramatically reduce tissue damage and improve recovery of cognitive function in rat, gerbil, and mouse ischemic models (16,28). The potential for Epo to provide protection in traumatic or metabolic-toxic injury models has also been demonstrated (14 -16). We find that loss of EpoR in brain compromises tissue maintenance/survival demonstrated by increased sensitivity to glutamate toxicity in vivo. Injection of glutamate into adult brain of ⌬EpoR mice increased the number of apoptotic cells compared with control animals (Fig. 7, F-H). This is consistent with the left-shift in the glutamate toxicity curve characterizing ⌬EpoR hippocampal cell cultures (Fig. 7E).

DISCUSSION
During development, NPCs continuously generate neurons in formation of the nervous system. Although most regions of the adult central nervous system lose the neurogenesis ability in vivo, continual generation of new neurons in the central nervous system has been localized to the SVZ and the hippocampus. EpoR localization to regions of proliferating neural cells during development and in adult (7,8,29,30) combined with the neurotrophic activity of Epo suggest that Epo may contribute to neurogenesis. The proliferation and differentiation of NPC are regulated by their specialized surrounding microenvironment or "niche" (31). Many developmental signal molecules, such as BMP, Eph/ephrin, Noggin, Shh have been identified to act on the maintenance of adult NPC niches and play roles on neurogenesis and migration of newly generated neurons (32). The vasculature contributes to the neurogenic niche and regulates NPC proliferation, promoting self-renewal and production of neurons via vascular endothelial growth factor secretion (33,34). Both vascular endothelial growth factor and Epo are hypoxia inducible and neuroprotective to ischemic injury (16,35,36). Neuroprotection by these two cytokines may reflect both stimulation of NPC proliferation as well as survival or antiapoptotic effects on mature neurons. Therefore, Epo can affect neural cell proliferation, maintenance, and/or survival indirectly through regulation of the red cell mass or stimulation of endothelium or directly through EpoR expression in neural cells enabling an Epo response.
Expression of EpoR in Brain-The active erythroid form of EpoR is not restricted to hematopoietic cells, and expression in non-hematopoietic tissue includes the nervous system and endothelial cells. EpoR expressed on cultured hippocampal and cortical neurons is functional and protects neurons against glutamate and hypoxia-induced apoptosis. We now provide evidence that as in differentiating erythroid progenitor cells and myoblasts, EpoR is expressed in NPC and is down-regulated with differentiation. Unlike mature red blood cells that do not express EpoR, EpoR is retained on mature neurons. We selec- (red line) E16 embryos were counted. F-G, to determine the differential sensitivity to glutamate in adult brain in vivo, wild type (F), and ⌬EpoR (G) mice received brain injections of glutamate followed by analysis of TUNEL positive cells. Blue indicates DAPI staining to visualize nuclear DNA, and red indicates TUNEL positive cells. H, TUNEL positive cells in the brain in the region of the glutamate injection in wild type and ⌬EpoR mice were enumerated and expressed as % total cell number determined by DAPI staining. *, p Ͻ 0.001 compared with wild type control.
tively rescued the embryonic lethal EpoR-null phenotype by restoring EpoR expression driven by its own promoter in hematopoietic/endothelial cells resulting in ⌬EpoR mice that can survive through adulthood in the absence of brain EpoR. These mice exhibit increased apoptosis in the developing brain consistent with the increased apoptosis observed between E10.5 and E12.5 in the EpoR-null mice (8). These data indicate that endogenous Epo signaling plays a role in survival of neural cells during embryonic development in mice and that increased apoptosis in mice that lack EpoR expression in brain is not solely due to increased hypoxia from interruption of definitive erythropoiesis.
Epo Signaling Is Involved in Neural Progenitor Cell Proliferation and Neurogenesis-Several studies indicate that hypoxia stimulates the proliferation and differentiation of NPCs. In cultures of E14 rat primary (mesencephalic) neurons and PC12 cells, hypoxia increases tyrosine hydroxylase, the rate-limiting enzyme for catecholamine synthesis (37,38). In addition, embryonic precursors from both the peripheral nervous system and the central nervous system exhibit enhanced neuronal proliferation and differentiation in response to lowered oxygen tension (6,39). Epo signaling contributes in part to these hypoxiainduced responses. Hypoxia-induced dopamine content in differentiated dopaminergic mesencephalic neurons can be blocked by Epo antibody. Epo treatment of PC12 cells also increases intracellular monoamine levels (40). In EpoR knockout mice, the number of NPCs is decreased as well as the proportion of neurons generated, and neurons from these mice exhibit an intrinsic defect and undergo a marked increase in apoptosis when cultured under hypoxia in the absence of trophic factors and do not survive beyond 24 h (8). We find an analogous sensitivity to hypoxia in cultures of neural cells from embryonic ⌬EpoR mice. The data from these ⌬EpoR mice further support the notion that Epo is involved in neural cell development and/or maintenance of the NPC pool.
Although we find increased apoptosis during ⌬EpoR mouse brain development and a reduction in nestin in ⌬EpoR embryonic brain, adult ⌬EpoR mice show normal brain morphology, indicating that Epo is not a critical growth factor for brain remodeling or development. We observe that Epo stimulation of NPC mitosis is about half as potent as bFGF, a trophic factor useful for promoting NPC proliferation in vitro, and it is likely that bFGF or other cytokines may mask the lack of Epo signaling in brain. Similarly, the EpoR-null mice rescued by a transgene containing EpoR cDNA driven by the promoter for erythroid-specific transcription factor GATA-1 show no gross brain defects (41). However, we do find a small but statistically significant decrease in the brain/body weight ratio in ⌬EpoR mice. Although EpoR expression in neuronal cells is not an absolute requirement for life, our data suggest that Epo signaling does affect the plasticity of the central nervous system in adulthood, as indicated by our assay of neural cell proliferation and glutamate toxicity. The reduction in ⌬EpoR whole embryonic brain nestin and the reduction of BrdUrd uptake in ⌬EpoR adult mouse brain in regions associated with neurogenesis may be because of an overall decrease in the number of progenitor cells as a consequence of increased apoptosis during brain development or a decrease in proliferation rate because of loss of EpoR in neural cells. Brain endothelial progenitor cells can also proliferate. However, the ⌬EpoR mouse model was designed to retain EpoR expression in endothelial cells, and the difference in BrdUrd uptake is likely not because of variation in endothelial self-renewal or response. Therefore, the decrease in neural cell proliferation and increase in sensitivity to glutamate in the ⌬EpoR mice described here appear to be independent of Epo activity in erythroid or endothelial cells and rather relate directly to changes in EpoR expression on neural cells. Although endothelial response to Epo may contribute to Epo neuroprotection, these data demonstrate that rescue of the EpoR-null genotype by restoring EpoR expression in hematopoietic and endothelial cells is not able to completely restore the protective effects of endogenous Epo in brain. Consistent with a direct effect of Epo on neural cells is that targeted deletion of EpoR in neural cells in mice leads to impaired migration of neuroblasts post-stroke (13).
Endothelial-mediated Effects of Epo-Epo protective effects during brain ischemia have been attributed in part to endothelial response (42). Hence, the retention of EpoR expression on endothelial cells of the ⌬EpoR mice may reduce the adverse effects because of loss of EpoR in brain. Although exogenous Epo administration provides dramatic neuroprotection and reduction in infarct size (16,43,44), in the absence of exogenous Epo, no significant difference was observed in infarct size following stroke between control mice and mice with targeted loss of EpoR in neural cells only (13). EpoR expression in endothelium may contribute indirectly to the neuroprotective or sparing effect associated with Epo as suggested by Epo activation of endothelial cells during treatment of stroke in animals (42,45).
Neural Progenitor Cell Proliferation/Differentiation-Epo activity to stimulate cell proliferation and survival has proven to be pleiotropic and extends beyond response of erythroid progenitor cells to include brain, heart, vasculature, pancreas, and reproductive organs (4). We provide evidence for a direct effect of endogenous Epo on normal brain function adding support to the potential benefits of Epo intervention for protection in response to brain insult or injury. Although previous findings suggest that Epo regulates NPC differentiation (6,7,46), observations shown here suggest that Epo may be a novel trophic factor, which acts on maintenance of the NPC population or repair via neural cell proliferation. In culture, although Epo was about half as potent as bFGF in promoting NPC proliferation, the combination of Epo with bFGF did not increase the response beyond that of bFGF alone, indicating that Epo had little effect on cell growth when cultured under optimal growth conditions. Similarly, Epo did not increase the production of secondary neurospheres in adult NPC cultures supplemented with EGF (7) suggesting that Epo response may be more apparent during stress. In muscle and endothelial progenitor cell cultures, we observed maximal effects of Epo when proliferating progenitor cells were cultured under stress or suboptimal conditions of low serum or low oxygen tension (3,5), providing further evidence that Epo may be an important stress responder. In addition, anti-Epo antibodies reduced NPC neurogenesis in cultures exposed to hypoxia, but had little or no effect in control cultures (7).
The importance of endogenous Epo activity in NPC maintenance and brain repair is also suggested in animal models of ischemic damage. Infusion of soluble EpoR directly into the brain (left ventricle) of gerbils increased sensitivity to mild ischemia causing increased neuronal degeneration and impaired learning ability (16). In animal models of hypoxic preconditioning, mild hypoxia preconditioning provides tolerance against subsequent, more severe ischemic insult. Increased expression of endogenous hypoxia-inducible factor-1 and its target genes including Epo provides protection in adult mice to focal permanent cerebral ischemia and reduces infarct volume (47). Administration of soluble Epo receptor to the cerebral ventricle to block endogenous Epo activity significantly reduces the protective effect of hypoxic preconditioning (48). Similarly, hypoxia induction of hypoxia-inducible factor-1 and Epo protects against light-induced retinal degeneration in mice that is mediated via EpoR expression on photoreceptor cells (49). In brain, the high EpoR expression on NPC, the decrease of NPC proliferation in the absence of brain EpoR, and the increased neural cell sensitivity to hypoxia or glutamate toxicity in the ⌬EpoR mice provide direct evidence for a role of Epo signaling in brain for maintenance of NPC or repairing of NPC activity.