OUP user menu

Brain Erythropoietin Receptor Expression in Alzheimer Disease and Mild Cognitive Impairment

Michael I. Assaraf MSc, Zuanel Diaz MSc, Adrienne Liberman BSc, Wilson H. Miller Jr. MD, PhD, Zoe Arvanitakis MD, Yan Li PhD, David A. Bennett MD, Hyman M. Schipper MD, PhD, FRCPC
DOI: http://dx.doi.org/10.1097/nen.0b013e3180517b28 389-398 First published online: 1 May 2007


Cellular mechanisms conferring neuroprotection in the brains of patients with Alzheimer disease (AD) remain incompletely understood. Erythropoietin (Epo) and the erythropoietin receptor (EpoR) are expressed in neural tissues and protect against oxidative and other stressors in various models of brain injury and disease. Our objective in this study was to determine whether EpoR is upregulated in the brains of persons with sporadic AD and mild cognitive impairment (MCI). Postmortem hippocampus and temporal cortex from subjects with AD, MCI, and no cognitive impairment (NCI) were procured from the Religious Orders Study. Total immunoreactive EpoR protein was determined by Western blotting. Astrocytes expressing immunoreactive EpoR were quantified in 4 temporal and 6 hippocampal regions, and correlated with clinical, neuropsychologic, and neuropathologic indices. Total immunoreactive EpoR protein was markedly increased in AD and MCI temporal cortex versus NCI tissues. Composite measures of glial EpoR expression in temporal cortex layers I to IV were significantly greater in the MCI group compared with the NCI and AD groups. Hippocampal EpoR scores were increased in persons with MCI and AD relative to those with NCI. There was substantial subregional heterogeneity in disease-related EpoR expression patterns in AD and MCI temporal cortex and hippocampus. There was no association of EpoR-positive astrocytes with summary measures of global cognition or AD pathology. We conclude that upregulation of EpoR in temporal cortical and hippocampal astrocytes is an early, potentially neuroprotective, event in the pathogenesis of sporadic AD.

Key Words
  • Alzheimer disease
  • Astrocyte
  • Erythropoietin
  • Erythropoietin receptor
  • Hippocampus
  • Mild cognitive impairment
  • Temporal cortex


Alzheimer disease (AD) is a dementing illness characterized by progressive neuronal degeneration, gliosis, and the accumulation of intracellular inclusions (neurofibrillary tangles) and extracellular deposits of amyloid (senile plaques) in discrete regions of the basal forebrain, hippocampus, and association cortices (1). Mild cognitive impairment (MCI) refers to elderly individuals who experience gradual cognitive decline (usually memory) of at least 6 months' duration that fails to meet the clinical criteria for AD or other dementia. In many cases, amnestic MCI is the preclinical harbinger of sporadic AD and approximately 40% to 50% of subjects with MCI will progress to probable AD over a 3- to 5-year follow-up period (2).

The advent of effective neuroprotective therapy in AD presupposes an understanding of the mechanisms responsible for neuronal attrition in this disease and the adaptive responses of the CNS to the neurodegenerative process. Astrocytes, a major class of neuroglia, subserve a wide range of neuroprotective functions in the mammalian CNS. These cells play important roles in maintenance of the blood-brain barrier, ion and redox homeostasis, sequestration, and metabolism of potentially excitotoxic amino acids, and elaboration of immunomodulatory cytokines and neuropeptides. Astrocyte hypertrophy, accumulation of glial fibrillary acidic protein (GFAP)-positive intermediate filaments, and upregulation of heat shock proteins and other cytoprotective factors are characteristic histopathologic features of the major aging-related neurodegenerative disorders including AD, Parkinson disease, Huntington disease, and amyotrophic lateral sclerosis (3).

Erythropoietin (Epo) is a 30-kDa glycoprotein hormone that stimulates proliferation and differentiation of erythroid cells in response to decreased tissue oxygen delivery. Epo has been shown to exert important cytoprotective effects within the central and peripheral nervous systems. Studies indicated that peripherally administered recombinant Epo crosses the intact blood-brain barrier (4). Epo and its receptor are expressed in rodent and human CNS tissues (5, 6), as well as in cultured neurons and astrocytes (7, 8). Epo gene expression in brain can be regulated by hypoxia-inducible factor-1 (HIF-1) that, in turn, is activated by a variety of stressors including hypoxia (9). Preconditioning with Epo protects neurons from injury caused by neurotoxins, reactive oxygen species, and nitric oxide (10, 11). Administration of Epo to the CNS of mice, rats, and gerbils by intracerebroventricular infusion significantly reduces neuronal death and prevents learning disability resulting from cerebral ischemia (12). Epo has also been shown to limit the extent of concussive brain injury, immune damage in experimental autoimmune encephalomyelitis (13), and excitotoxicity induced by kainate (14). In addition to conferring neuronal protection, Epo is synthesized and secreted by mammalian astrocytes and acts in an autocrine fashion to defend the astroglial compartment from low O2 tension and other stressors (15-21).

The erythropoietin receptor (EpoR) is a 66-kDa member of the cytokine receptor superfamily that contains an extracellular ligand-binding portion, a single hydrophobic transmembrane domain, and an intracellular cytoplasmic sequence lacking direct kinase activity (22-24). The EpoR is expressed as a dimer on the surface of erythroid progenitor cells, neurons, and astroglia. Upon binding to Epo, the EpoR undergoes a conformational change that triggers a signal transduction (phosphorylation) cascade, resulting in proliferation, differentiation, or protection of the target cell population (25). In light of our observations that Epo/EpoR interactions mediate glioprotection in primary culture (20), we set out to determine whether cortical and hippocampal astroglia upregulate cell surface EpoR as a potential neuroprotective mechanism in patients with sporadic AD and MCI relative to its expression in normal elderly control tissues. Clinical and pathologic data from 28 deceased participants from the Religious Orders Study were used to examine the relation of EpoR to clinical diagnosis, cognition proximate to death, and AD pathology.

Materials and Methods


Subjects were older Catholic nuns, priests, and brothers enrolled in the Religious Orders Study, a longitudinal clinical-pathologic study of aging and AD funded by the National Institute on Aging. Study participants come from about 40 different groups across the United States. Each signed informed consent and anatomical gift act forms donating his or her brain to Rush investigators at the time of death. The study was approved by the Rush University Medical Center Institutional Review Board. Since January 1994, more than 1,000 persons have enrolled in the study and completed the baseline evaluation. The follow-up rate is >95% of survivors, and the autopsy rate is >90%. Details of the study have been reported previously (26-29). Data from 28 subjects (9 with no cognitive impairment [NCI], 10 with MCI, and 9 with AD) were used for the present study.

Clinical Evaluations, Neuropsychologic Testing, and Clinical Diagnoses

Each subject underwent annual structured clinical evaluations, which included a medical history, neurologic examination, and detailed neuropsychologic testing, as reported previously (26). Neuropsychologic testing was administered annually (with up to 12 years of annual testing performed to date) and included the Mini-Mental Status Examination (used for descriptive purposes only) and 19 other neuropsychologic tests that were used to create composite scores of cognition, including a global score and 5 separate cognitive domain scores (episodic memory, semantic memory, working memory, perceptual speed, and visuospatial ability), as reported previously (29). All neuropsychologic test results were reviewed by a board-certified neuropsychologist. For purposes of this study, neuropsychologic test scores proximate to death (mean time interval from final testing to death was about 8 months) were used in all analyses. Each subject was examined by a clinician with expertise in aging, who, after review of all data, classified subjects with respect to dementia status and other relevant neurologic conditions. The diagnosis of AD was made according to the criteria established by the National Institute of Neurologic and Communicative Disorders and Stroke and the Alzheimer Disease and Related Disorders Association (30). MCI was diagnosed when the neuropsychologist determined the presence of cognitive impairment on testing, but the clinician did not identify dementia (26). Annual follow-up clinical evaluations were identical in all essential components. At time of death, a board-certified neurologist with expertise in dementia, but blinded to pathologic data, reviewed all relevant clinical information to render a classification of dementia status. This diagnosis was used for the purpose of the study. The mean time interval from the last clinical evaluation to autopsy was about 8 months.

Neuropathologic Evaluation

Brains were removed in a standard fashion, as reported previously (31). The mean postmortem interval was approximately 8 hours. Uniform examination for cerebral infarction was conducted on 1-cm coronally cut hemispheres. For this study, we dichotomized chronic macroscopic cerebral infarctions as present or absent, as previously described (26). Blocks from one hemisphere were fixed in 4% paraformaldehyde for 3 to 14 days. For this study, tissue from the superior temporal gyrus and the hippocampus, 2 regions affected early in the course of AD, were used. Tissue was paraffin-embedded and cut in 6-μm sections and processed for either modified Bielschowsky silver staining to identify and quantify AD pathology (neuritic plaques, diffuse plaques, and neurofibrillary tangles) or immunolabeling as described below. A measure of overall AD pathology, based on average counts of neuritic plaques, diffuse plaques and neurofibrillary tangles, was used in this study as described previously (32). Sections of cortex and substantia nigra were also evaluated for immunoreactive α-synuclein to delineate the presence of Lewy body pathology.

Immunofluorescence Confocal Microscopy

Qualitative analysis of EpoR expression patterns in the AD, MCI, and NCI specimens was performed using immunofluorescence confocal microscopy. Paraffin sections were deparaffinized, rehydrated, and heated in EDTA solution to enhance antigen accessibility. After application of primary polyclonal and monoclonal antibodies, fluorescein isothiocyanate-conjugated secondary goat anti-mouse antibody was used for detection of GFAP-positive astrocytes and rhodamine-conjugated secondary goat anti-rabbit was used to visualize EpoR. Quenching of autofluorescence was achieved by incubating sections in 0.3% Sudan black in 70% ethanol. The fluorescent preparations were examined using a Bio-Rad MRC-600 laser scanning confocal imaging system (Bio-Rad, Mississauga, Ontario, Canada), as described previously (33). Images were scanned on 2 channels (red and green) and merged to produce a single profile. Regions exhibiting colocalization of red and green fluorophores emit yellow fluorescence in the merged images.


Tissue sections (6-μm) were immunostained with rabbit-derived polyclonal antisera directed against KLH-conjugated human EpoR peptide (AVARP-9008, 1:50; Aviva Antibody, San Diego, CA), a mouse monoclonal antibody recognizing the astrocyte marker GFAP (NCL-GFAP-GA5, 1:50, Novo Castra Laboratories, UK) and the Vectastain Elite ABC Kit (Vector Laboratories, Inc., Burlingame CA). The slides were incubated for 16 hours with anti-EpoR antibody and the reaction product was visualized as a black precipitate with Vector SG. The preparations were then incubated with the anti-GFAP monoclonal antibody for 16 hours with revelation of the reaction product using Vector Nova RED substrate. In some sections, anti-EpoR staining was performed in the presence of purified EpoR protein as an immunoabsorption control.

Quantification of Erythropoietin Receptor-Positive Astrocytes

Percentages of GFAP-positive astrocytes coexpressing immunoreactive EpoR were computed in 6 layers of the hippocampus (stratum oriens, pyramidal layer, stratum radiatum, molecular layer, granular layer, and hilus of the dentate gyrus) and in temporal cortex layers I to VI and the subcortical white matter. For each area surveyed, numbers of GFAP-positive and EpoR-positive glial cells were counted in 400x fields with the aid of an ocular grid by a single investigator unaware of the tissue source. The average number of cells counted per individual layer was 216.2 ± 40.0 for the hippocampus and 217.0 ± 14.8 for the temporal cortex.

Western Blotting

Samples of frozen temporal cortex from patients with NCI, MCI, and AD (n = 3 per group) were disrupted by 2- to 4-second bursts of a Polytron homogenizer using 10 volumes of homogenization buffer at 4°C. Cell debris was removed by centrifugation at 2,000 rpm for 5 minutes. Extracts were centrifuged at 14,000 rpm and 4°C, and supernatants were transferred to fresh tubes. Protein concentrations were determined with the Bio-Rad protein assay (Bio-Rad). Fifty micrograms of protein was added to an equal volume of 2x sample buffer and run on a 10% sodium dodecyl sulfate-polyacrylamide gel. Proteins were transferred to nitrocellulose membranes (Bio-Rad), stained with Ponceau S in 5% acetic acid to ensure equal protein loading, and blocked with 5% milk in PBS containing 0.5% Triton X-100 for 1 hour at room temperature. The membrane was incubated overnight at 4°C with rabbit-derived polyclonal antisera (2 μg/mL) directed against human EpoR (AVARP-9008, Aviva Antibody, San Diego, CA). After 3 washes with PBS and 0.5% Triton X-100, blots were incubated with a goat anti-rabbit secondary antibody (1:5000; PharMingen) for 1 hour at room temperature. Bands were visualized by enhanced chemiluminescence (Amersham Pharmacia Biotech, Baie d'Urfe, Quebec, Canada). Immunostaining for β-actin served as a protein loading control.

Statistical Analysis

Demographic, clinical, and pathologic characteristics among the 3 diagnostic groups (NCI, MCI, and AD) were compared using one-way analysis of variance for continuous variables and chi-square tests for categorical variables. We used one-way blocked analysis of variance to compare EpoR-positive astrocyte measures across different cortical regions for individuals in each of the NCI, MCI, and AD clinical diagnostic groups, where person was the blocking factor. Tukey's multiple comparison adjustment was used to control type I error rate. We conducted Pearson correlations to examine the intercorrelations between each of the cortical measures and factor analyses with varimax rotation to examine how the measures load together. Cronbach α coefficients were used to assess for internal consistency. Based on these analyses, we created composite measures of EpoR-positive astrocytes that allow for reduced random error and minimal floor and ceiling effects. To examine the relation of composite measures of EpoR-positive astrocytes to clinical diagnosis, we conducted one-way analysis of variance tests. Tukey's multiple comparison adjustment was used to control type I error rate. We repeated the analysis using one-way analysis of covariance to adjust for age, sex, and education. Next, multiple linear regression analyses were conducted to examine the relation of composite measures of EpoR-positive astrocytes to composite measures of cognitive function proximate to death in models adjusted for age, sex, and education. Similarly, we used multiple linear regression analysis to examine the relation of composite measures of EpoR-positive astrocytes to AD pathology, again adjusting for age, sex, and education. All regression models were performed in SAS, and models were validated graphically and analytically.


Demographic, Clinical, and Pathologic Characteristics

Data from 28 persons (mean age at death 84.0 years, mean education 18.5 years, and mean Mini-Mental Status Examination score 23.1) were used in this study, 9 with no NCI, 10 with MCI, and 9 with AD (Table). Mean ages were similar in all 3 groups. There were more men in the NCI group, and persons with AD had fewer years of formal education (p = 0.05).

View this table:

Mean Mini-Mental Status Examination scores of the NCI and MCI groups were similar and significantly greater than that of the AD group (p < 0.01). The MCI group exhibited lower scores on global cognition (p = 0.005), episodic memory (p = 0.004), perceptual speed (p = 0.03), and visuospatial ability (p = 0.03) compared with scores for those with NCI. Subjects with AD had lower scores than the other 2 groups on global cognition and all 5 cognitive abilities (all p values < 0.01). Braak stage V or VI was present in 44.5% of subjects with AD, 10% of those with MCI, and 0% of those with NCI. Persons with AD had the highest level of AD pathology (p = 0.004), with the MCI group exhibiting intermediate values (Table). Cerebral infarctions were present in 56% of subjects with AD, 20% of those with MCI, and 11% of those with NCI. Lewy body pathology was observed in 11% of subjects with AD, 22% of those with NCI, and 0% of those with MCI.

Western Blots

The polyclonal anti-EpoR antisera revealed 3 isoforms of EpoR in human brain (bands 1, 2, and 3 in Fig. 1). Previous studies have demonstrated that band 1 (66 kDa) corresponds to the mature, cell surface form of the EpoR, band 2 (˜64 kDa) represents the maturing, endoplasmic reticulum form of the EpoR, and band 3 (˜62 kDa) probably corresponds to the nonglycosylated form of the protein (34). As depicted in Figure 1, EpoR bands in the NCI group were faint, whereas moderate and intense EpoR bands were detected in the MCI and AD homogenates, respectively.


Anti-EpoR immunoblots of temporal cortex derived from subjects with no cognitive impairment (NCI), mild cognitive impairment (MCI), and Alzheimer disease (AD). Band 1 (66 kDa) denotes the mature, cell surface form of the erythropoietin receptor (EpoR), band 2 (˜64 kDa) represents the maturing, endoplasmic reticulum form of the EpoR, and band 3 (˜62 kDa) probably corresponds to the nonglycosylated form of the receptor.

Immunofluorescence Confocal Microscopy

In the temporal cortex and hippocampus of the NCI, MCI, and AD specimens, astrocytes were readily identified by their GFAP-positive (fluorescent green) perikarya and radiating processes (Fig. 2). GFAP staining generally appeared more intense in the AD and MCI specimens than in the NCI preparations, indicative of astrocyte hypertrophy in the former groups. The glial hypertrophy was not restricted to particular cortical or hippocampal layers and was also observed in subcortical white matter. EpoR staining (red fluorescence) was evident along the surface and peripheral cytoplasm of GFAP-positive (astrocytes) and GFAP-negative cells. Some of the GFAP-negative cells expressing EpoR were identified as neurons on the basis of size, morphology, and palisading organization within discrete tissue layers (data not shown). In both the temporal cortex and hippocampus, EpoR staining appeared more robust in the MCI and AD samples than in the NCI preparations. In the NCI samples, EpoR immunofluorescence was observed within or in close proximity to GFAP-positive astrocytes, but there was little evidence of colocalization of the 2 proteins within these cells. In contrast, EpoR expression often exhibited partial colocalization with GFAP staining (yellow fluorescence) in the MCI and AD brain sections (Fig. 2). No immunostaining was observed in immunoabsorption control preparations and in the absence of the primary antibodies.


Erythropoietin receptor (EpoR)/glial fibrillary acetic protein (GFAP) colocalization in human hippocampus. (A) Control for anti-EpoR (red) antibody. The astrocytes fluoresce green after staining with anti-GFAP antibody and fluorescein isothiocyanate. (B) Subject with no cognitive impairment. GFAP-positive astrocytes are depicted (green). EpoR staining (red) does not colocalize with the astroglia in this field (no yellow fluorescence). (C) Patient with mild cognitive impairment. Yellow fluorescence indicates extensive colocalization of EpoR (red) to astrocytes (green). (D) Patient with Alzheimer disease. There is colocalization of EpoR (red) to astroglia (green) yielding yellow fluorescence. Scale bars = 25 μm.

Quantification of Erythropoietin Receptor-Positive Astrocytes and Relation to Clinical Diagnosis

In the immunohistochemical preparations, GFAP-positive astrocytes and EpoR appeared reddish-brown and black, respectively. EpoR immunoreactivity was abrogated by EpoR protein immunoabsorption or by elimination of the primary antibodies (Fig. 3). Patterns of GFAP and EpoR staining in dual-labeled sections were similar to those noted by immunofluorescence confocal microscopy (above). As shown in Figure 4, there was marked heterogeneity in EpoR expression patterns within the temporal cortex and hippocampus of the subjects with NCI, MCI, and AD. In temporal cortex, percentages of GFAP-positive astrocytes coexpressing EpoR were 1) significantly increased in AD and MCI versus NCI in layers II to IV (p < 0.05 for each comparison), 2) increased in MCI (p < 0.05) but not AD (p > 0.05) versus NCI in layers V to VI, and 3) unrelated to diagnosis in layer I and the subcortical white matter (p > 0.05 for each comparison). In the hippocampus, percentages of GFAP-positive astrocytes coexpressing EpoR were 1) significantly increased in AD and MCI versus NCI in the stratum oriens and pyramidal layer (p < 0.05 for each comparison), 2) increased in MCI (p < 0.05) but not in AD (p > 0.05) versus NCI in the granular layer and hilus of the dentate gyrus, and 3) unrelated to diagnosis in the stratum radiatum and molecular layer (p > 0.05 for each comparison).


Dual-label erythropoietin receptor (EpoR)/glial fibrillary acetic protein (GFAP) immunohistochemical localization in human temporal cortex (A) and hippocampus (B). Clinical states are designated at the top of each column (NCI; no cognitive impairment; MCI; mild cognitive impairment; AD; Alzheimer disease). Brain subregions are indicated to the left of each row. Arrows denote EpoR-positive/GFAP-positive (black-red) astrocytes, whereas arrowheads indicate EpoR-negative/GFAP-positive (red) astrocytes. Insets depict immunoabsorption control for anti-EpoR staining (left), EpoR-negative astrocyte (middle), and EpoR-positive astrocyte (right). Scale bar = 25 μm in full panels and 10 μm in insets.


Erythropoietin receptor (EpoR) expression in human cortical (A) and hippocampal astrocytes (B). Differential expression of EpoR within the layers of the human temporal cortex (A) and hippocampus (B) is seen in subjects with no cognitive impairment (NCI; open bars), mild cognitive impairment (MCI; hatched bars), and Alzheimer disease (AD; filled bars). *, p < 0.05; **, p < 0.01; ***, p < 0.001, relative to the NCI cases. Numbers in parentheses indicate number of cases per group.

Intercorrelations of Erythropoietin Receptor-Positive Astrocytes and Relation to Clinical Diagnosis

EpoR-positive astrocytes in the different subregions were highly inter-related. Therefore, we created composite scores of temporal cortex and hippocampus measures. For the temporal cortex, Pearson correlations and factor analysis suggested that layers I and II to IV load together, and layers V to VI and the subcortical white matter load together. For the hippocampus, Pearson correlations were all positive and ranged from 0.22 to 0.73, and high internal consistency was demonstrated by a Cronbach α coefficient of 0.86. Subsequent analyses of the relation of Epo-R-positive astrocytes to clinical diagnoses, cognition, and AD pathology were conducted using these 3 groups (one based on temporal cortex layers I and II to IV, one based on temporal layers V to VI and subcortical white matter, and one based on all 6 hippocampal measures).

Figure 5 shows the relation of the temporal and hippocampal composite scores of Epo-R-positive astrocytes to clinical diagnoses. The score for Epo-R-positive astrocytes in temporal layers I to IV was related to clinical diagnosis (F2,12 = 25.18, p < 0.0001). Persons with MCI had a significantly higher score than those with NCI (p < 0.0001) or AD (p = 0.0312). Persons with AD had a significantly higher score than those with NCI (p = 0.0037). We did not find significant differences in Epo-R-positive astrocytes for the temporal cortex layers V to VI and white matter score (F2,14 = 0.50, p = 0.6167). The results were unchanged after adjustments for age, sex, and education. Similarly, we found that the composite score for Epo-R-positive astrocytes in the hippocampus was related to clinical diagnosis (F2,22 = 14.13, p = 0.0001). Persons with MCI (p < 0.0001) and AD (p = 0.0105) had significantly higher scores than those with NCI. We repeated the analysis with adjustments for age, sex, and education, and the results were unchanged.


Erythropoietin receptor (EpoR)-positive astrocyte scores (composite scores in temporal regions in A and in hippocampal regions in B in subjects with no cognitive impairment (NCI), mild cognitive impairment (MCI), and Alzheimer disease (AD). White bars denote median values, boxes delimit the 25th and 75th percentiles, horizontal black bars denote minimum and maximum values in each diagnostic group, and circles indicate outlier values (see text for significance levels). (A) Subjects with MCI had significantly higher temporal I to IV scores relative to the NCI and AD groups, and subjects with AD had significantly higher scores than persons with NCI; temporal V to VI and subcortical white matter scores were not different among the 3 diagnostic groups. (B) Subjects with MCI and AD had significantly higher hippocampal scores than the NCI group (see text for significance levels).

Relation of Erythropoietin Receptor-Positive Astrocytes to Cognitive Function Proximate to Death

Using multiple linear regression analyses adjusting for age, sex, and education, we examined the relation of the 3 summary scores of EpoR-positive astrocytes to cognition proximate to death. None of the 3 scores was related to global cognition. Overall, we did not find associations of scores with any of the 5 cognitive domains (all p values > 0.1), except for the temporal layers V and VI and subcortical white matter score being associated with visuospatial ability (parameter estimate = 0.09, SE = 0.04, p = 0.03). The inability to find associations with the linear models is consistent with the apparent upregulation of EpoR in MCI.

Relation Between Erythropoietin Receptor-Positive Astrocytes and Alzheimer Disease Pathology

Given the relation of measures to clinical diagnosis, we next examined the relation of summary scores of EpoR-positive astrocytes to a measure of overall AD pathology, using linear regression analyses adjusting for age, sex, and education. None of the 3 summary scores was related to AD pathology (all 3 p values > 0.36).


In this study, immunolabeling techniques were used to delineate and quantify glial EpoR expression in the temporal cortex and hippocampus of elderly subjects with MCI, sporadic AD, and NCI. The brain tissue specimens were procured from individuals enrolled in the Religious Orders Study, permitting meticulous correlation of neuropathology/immunohistochemistry with antemortem clinical diagnosis, cognitive function proximate to death, and AD pathology. We ascertained that EpoR is expressed in a proportion of GFAP-positive astrocytes residing in adult human temporal cortex and hippocampus, corroborating earlier reports in rodent tissues (35). In layers I to IV of the temporal cortex, the composite score of EpoR-positive astrocytes was increased in persons with AD relative to NCI, but the AD values remained lower than those recorded in individuals with MCI. In hippocampus, scores of EpoR-positive astrocytes were similar between the MCI and AD groups, with both being significantly elevated relative to NCI values. There was no apparent association of glial EpoR expression with global cognition, the 5 individual cognitive domain scores, or the extent of AD pathology. The marked regional heterogeneity in astrocytic EpoR expression among the 3diagnostic groups differs from the topographically more uniform glial hypertrophy observed in the AD and MCI neural tissues. This discordance in the patterns of GFAP and EpoR expression suggests that the latter may be regulated independently of the generalized astroglial response (reactive gliosis) to injury and disease.

A rapidly expanding body of literature attests to the fact that, distinct from their role in erythropoiesis, Epo-EpoR interactions confer significant cytoprotection to a host of neuronal and non-neuronal cell types within the mammalian CNS (11-14). Antioxidant, antiapoptotic, anti-inflammatory, neurotrophic, angiogenic, and synaptogenic activities have been implicated as potentially important mechanisms mediating Epo-related neuroprotection (36). Particularly germane to the present study are recent reports that Epo potently protects rodent hippocampus from β-amyloid toxicity, ischemia, and other noxious stimuli (37-39).

We and others have shown that Epo is synthesized and secreted by astrocytes and acts in an autocrine fashion to defend the astroglial compartment from a broad range of noxious stimuli, including low O2 tension, reactive oxygen/nitrogen species, and apoptogenic agents (15-20). By facilitating the action of Epo (derived locally or from the systemic circulation), upregulation of the astrocytic EpoR in AD-affected neural tissues may protect the host cells from the potentially injurious effects of excess β-amyloid, proinflammatory cytokines, and redox-active metals implicated in the pathophysiology of this degenerative condition (40, 41). In AD and other neurodegenerative disorders, protection of the astroglial compartment may, in turn, impact the survival of indigent neuronal populations by curtailing extracellular glutamate concentrations (excitotoxicity), replenishment of neuronal glutathione and ascorbate contents (antioxidant defenses), and provision of essential neurotrophic factors (20, 41, 42). A major finding of the present study was that enhanced glial EpoR expression in certain temporal and hippocampal regions was already apparent at the MCI stage, a frequent harbinger of incipient AD (2). This observation indicates that induction of the glial epoR gene is a relatively early event in the pathogenesis of this dementing disorder. We recently documented that expression of glial heme oxygenase-1 (HO-1), a sensitive marker of tissue oxidative stress, is significantly augmented in the MCI hippocampus and temporal cortex relative to individuals without cognitive impairment and within the range of expression observed in established AD (43). Along similar lines, Nunomura et al (44) showed that nitrotyrosine and 8-hydroxydeoxyguanosine immunoreactivities, markers of oxidative protein and nucleic acid damage, respectively, are increased in the brains of patients with very early AD and that these redox indices may actually diminish in magnitude with advancing disease. Taken together, these observations raise the possibility that cortical and hippocampal astroglia upregulate EpoR as an adaptive response to ambient oxidative stress that reaches peak intensity in preclinical (MCI) and the very earliest stages of sporadic AD. This formulation is consistent with other data garnered from the Religious Orders Study indicating that AD pathology in the temporal lobe is well advanced by the time persons meet the clinical criteria for MCI (31). A nonlinear relationship between glial EpoR expression and advancing AD may explain the lack of association of EpoR scores with global cognition and AD pathology in the current study.

Since the advent of Epo therapy in 1986, systemic administration of the hormone for the management of anemia in adult and pediatric patients with chronic kidney disease has proven to be safe, effective, and well-tolerated (45, 46). In early-phase clinical trials currently underway in Germany, recombinant human Epo administered intravenously has shown promise in limiting cerebral infarct volume in patients with acute stroke. Of note, treated patients achieved 60- to 100-fold increases in cerebrospinal fluid Epo concentrations relative to untreated controls and no significant safety issues have thus far been identified (47). Additional studies have shown that asialoEpo, a deglycosylated congener of Epo that does not significantly affect the hematocrit, also crosses the blood-brain barrier and provides potent neuroprotection (48-50). In light of the accruing clinical experience, the results of numerous whole animal and in vitro investigations, and the data reported herein, the role of Epo (or asialoEpo) as neuroprotective therapy in persons with MCI and early AD warrants further investigation.


We are indebted to the hundreds of nuns, priests, and brothers from the following groups participating in the Religious Orders Study: Archdiocesan priests of Chicago, IL, Dubuque, IA, and Milwaukee, WI; Benedictine Monks, Lisle, IL, Collegeville, MN, and St. Meinrad, IN; Benedictine Sisters of Erie, Erie, PA; Benedictine Sisters of the Sacred Heart, Lisle, IL; Capuchins, Appleton, WI; Christian Brothers, Chicago, IL, and Memphis, TN; Diocesan priests of Gary, Gary, IN; Dominicans, River Forest, IL; Felician Sisters, Chicago, IL; Franciscan Handmaids of Mary, New York, NY; Franciscans, Chicago, IL; Holy Spirit Missionary Sisters, Techny, IL; Maryknolls, Los Altos, CA, and Ossining, NY; Norbertines, DePere, WI; Oblate Sisters of Providence, Baltimore, MD; Passionists, Chicago, IL; Presentation Sisters, B.V.M., Dubuque, IA; Servites, Chicago, IL; Sinsinawa Dominican Sisters, Chicago, IL, and Sinsinawa, WI; Sisters of Charity, B.V.M., Chicago, IL, and Dubuque, IA; Sisters of the Holy Family, New Orleans, LA; Sisters of the Holy Family of Nazareth, Des Plaines, IL; Sisters of Mercy of the Americas, Chicago, IL, Aurora, IL, and Erie, PA; Sisters of St. Benedict, St. Cloud and St. Joseph, MN; Sisters of St. Casimir, Chicago, IL; Sisters of St. Francis of Mary Immaculate, Joliet, IL; Sisters of St. Joseph of LaGrange, LaGrange Park, IL; Society of Divine Word, Techny, IL; Trappists, Gethsemani, KY, and Peosta, IA; and Wheaton Franciscan Sisters, Wheaton, IL.

We thank Traci Colvin, George Hoganson, and Julie Bach for Religious Orders Study Coordination; Todd Beck for analytic programming; George Dombrowski and Greg Klein for data management; and the staff of the Rush Alzheimer's Disease Center and Rush Institute for Healthy Aging.


  • This study was supported by a grant from Ortho Biotech (HMS and WHM) and the National Institute on Aging grants R01 AG15819, P30 AG10161 (DAB), and K23 AG23675 (ZA).


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
  37. 37.
  38. 38.
  39. 39.
  40. 40.
  41. 41.
  42. 42.
  43. 43.
  44. 44.
  45. 45.
  46. 46.
  47. 47.
  48. 48.
  49. 49.
  50. 50.
View Abstract