T H E JOURNAL OF BIOWCICAL CHEMISTRS 0 1994 by The American Society for Biochemistry and Molecular Biology, Inc

Vol. 269, No. 30, Issue of July 29, pp. 19488-19493, 1994 Printed in U.S.A.

A Novel Site of Erythropoietin Production OXYGEN-DEPENDENT PRODUCTION IN CULTURED RAT ASTROCYTES*

(Received for publication, March 18, 1994)

Seiji Masuda, Masaki Okano, Kenji Yamagishi, Masaya Nagao, Masatsugu UedaS, and Ryuzo SasakiO From the Department of Food Science and Technology, Faculty of Agriculture, Kyoto University, Kyoto 606, Japan and the Wesearch Institute of Life Science, Snow Brand Milk Products Company, Ltd., Tochigi 329-05, Japan

It has been shown that neurons express erythropoi- etin (Epo) receptor, but the production of Epo protein in neural tissues has not been demonstrated. Cerebral cells of rat fetuses were cultured, and Epo in the spent me- dium was measured with an enzyme-linked immunoas- say. Production of the immunoreactive Epo was depend- ent on 0, tension for cell culture; hypoxia enhanced the production. The immunoreactive Epo purified from the spent medium stimulated the growth of Epo-dependent myeloid cells and formation of fetal liver erythroid colo- nies. These biological activities were completely inhib- ited by the anti-Epo antiserum and the extracellular do- main of the Epo receptor capable of binding with Epo. When brain Epo was compared with serum Epo, brain Epo was smaller in size and more active in vitro at low ligand concentrations. These differences appear to be caused by the different extent of sialylation. Analyses with the reverse transcription-polymerase chain reac- tion method indicated that the regulation of Epo pro- duction by oxygen operates at the level of its mRNA. Immunochemical staining of the immortalized clonal cells revealed that astrocytes produced brain Epo. These results provide a novel site of Epo production and suggest that Epo acts on neurons in a paracrine fashion.

Epo’ is the glycoprotein that plays a major role in the regu- lation of erythropoiesis by stimulating the proliferation and differentiation of erythroid precursor cells (1 ,2) . The kidney is the major organ of Epo production in adult (3) whereas the liver is the primary site in fetus (4). Peritubular interstitial cells in the kidney (5, 6) and more recently peritubular interstitial fibroblasts (7) have been shown to express Epo mRNA. In the liver, hepatocytes and also non-parenchymal cells express Epo (8-10). Oxygen controls the Epo production; hypoxic conditions activate expression of the Epo gene in animals (11-15) and Epo-producing hepatic cells (10, 16-18).

Epo has been thought to exclusively act on erythroid precur- sor cells in vivo, but there is growing evidence that neurons respond to Epo. Neuronal cells express Epo-R, and binding of Epo increases intracellular concentrations of Ca” and mono- amines (19). Epo augments choline acetyltransferase in cul- tured embryonal neurons and supports the in vivo survival of

*This work was supported by grants-in-aid from the Ministry of Education, Science and Culture of Japan. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.

Fax: 81-75-753-6274. 0 To whom correspondence should be addressed. Tel.: 81-75-753-6271;

The abbreviations used are: Epo, erythropoietin; Epo-R, erythropoi- etin receptor; PBS, 10 m~ phosphate-buffered saline, pH 7.4; sEpo-R, soluble Epo-R; BSA, bovine serum albumin; mAb, monoclonal antibody; PCR, polymerase chain reaction; bp, base pair(s).

cholinergic neurons in adult rats (201, suggesting that Epo may function as a neurotrophic factor. Although the nucleotide se- quence of Epo-R cDNA from neuronal cells is identical to that from erythroid cells (191, the Epo affinity of Epo-R expressed on neurons is much lower than that on erythroid cells (19, 20). Further the blood-brain barrier formed in the late embryonic and early postnatal period (21) insulates the nerve system from the circulation, preventing the access of blood Epo to the neu- rons. Therefore, if the Epo action on neurons is physiological, a novel site for Epo production should be found in the brain. Tan et al. (22) reported the hypoxia-induced increase in Epo mRNA in brain. In the present study, we examined biosynthesis of Epo by the primary cultured cerebral cells and identified Epo-pro- ducing cells. Acornparison of in vitro activity and size was made between brain and serum Epo. The regulation of brain Epo production by 0, tension for cell culture is also presented.

EXPERIMENTAL PROCEDURES Cell Culture-Cerebral cells were cultured according to the method

reported (23). Briefly, the brain was removed from Wistar fetal rats on day 18 under sterile conditions and washed three times with PBS to remove the red blood cells. Meninges were carefully removed, and the resulting cerebral tissue was thoroughly washed with PBS. The tissue was dissociated by passing it through a 70-pm nylon mesh (cell strainer, Falcon). The cell suspension was washed twice with Dulbecco’s modified Eagle’s medium (Life Technologies, Inc.) supplemented with 10% fetal calf serum (Whitaker M. A. Bioproduct). The cells were plated with cells from one fetal braid50 cm2 of cell culture vessels. The cell cultures were carried out at 21, 5, or 1% 0,, 5% CO,, balance N,. The medium was changed every 3 days. When 0, tension was shifted in the course of cell culture, cultures were started with the fresh media equilibrated with the desired 0, tension by prior incubation for 24 h.

Epo Assay with Enzyme-linked Zmmunoassay-Epo was measured by an improved method, of a sandwich-type enzyme-linked immunoassay (24) using two mAbs, R2 and R6, that recognize different epitopes of Epo. This improved method measures Epo as low as 100 microunitdml (1 pg/ml). The amount of Epo was calculated from the standard curve drawn using recombinant human Epo (25) as the standard, which had been standardized with the Second International Standard Reference Preparation for human Epo.

Purification of Brain and Serum Epo-Cerebral cells were cultured in 175-cmZ T-flasks for 30 days in 5% 0,, and Epo in the spent medium (800 ml containing about 16 ng of Epo) was purified with the gel (600 111) to which the Epo-directed mAb R2 (24) was fixed. Bound Epo was eluted under an acidic condition, and the fractions containing Epo were neu- tralized (26). To the neutralized solution (7.5 ml), 500 pl of PBS con- taining 0.1% BSA was added. The resulting solution was ultrafiltrated (Molcut L, Millipore) to concentrate up to 500 p1 and exchange the buffer with PBS. Rat serum Epo was isolated from anemic animals (27). Epo from both sources was stored in PBS containing 0.1% BSA.

Deatment of Epo with Neuraminidase-For treatment of Epo with neuraminidase, PBS was exchanged with 0.2 M acetate buffer, pH 5.0, using Molcut L. Epo (about 16 ng) was incubated with or without 0.5

M. Okano, S. Masuda, H. Narita, S. Masushige, S. Kato, S. Ima- gawa, and R. Sasaki, submitted for publication.

19488

Brain Erythropoietin 19489

unit of neuraminidase from Arthrobacter ureafaciens (Nacalai Tesque) in a total volume of 210 pl at 37 "C. After incubation for 24 h, the reaction mixture was neutralized with NaOH, and the volume was brought to 500 pl with PBS containing 0.1% BSA.

Western Blotting of Brain and Serum Epo-For analyses of Epo with Western blotting, it was necessary to concentrate Epo and remove BSA added for the stabilization of Epo. Epo (16-25 ng in 0.51.5 ml of PBS containing 0.1% BSAand 0.05% Tween 20) was attached to the gel (5 p1) to which the Epo-directed mAb R6 (24) was fixed. The gel was washed with 1.5 ml of 0.05% Tween 20 in PBS to remove BSA and centrifuged. To the gel suspension (15 pl) was added the sample buffer (20 pl) for SDS-polyacrylamide gel electrophoresis, and the mixture was heated at 100 "C for 5 min to elute the bound Epo. The gel suspension was cen- trifuged, and the supernatant (20 pl) was heated again after the addi- tion of 1.3 p1 of P-mercaptoethanol. After centrifugation, proteins in the supernatant (20 pl) were separated by SDS-polyacrylamide gel electro- phoresis (12% acrylamide gel). Separated proteins were transferred to the nitrocellulose filter and detected using the affinity-purified rabbit anti-Epo antibody, anti-rabbit IgG biotinylated antibody (Vector Labo- ratories, Inc.), avidin-biotin-peroxidase complex (Vector Laboratories, Inc.), and the ECL Western blotting detection system. The purified rabbit anti-Epo antibody was prepared from the anti-Epo antiserum using the recombinant human Epo-fured gel (24).

Detection of Epo mRNA-Total RNA was prepared from cultured cerebral cells and the kidneys of the anemic rats by the acid guani- dinium thiocyanate-phenol-chloroform extraction method (28). Reverse transcription-PCR was performed according to the method described (29). The reverse transcription reaction was done using a random hex- amer and 4 pg of heat-denatured RNA in a volume of 20 pl. The reac- tions without reverse transcriptase or RNA were run as controls. PCR amplification of random hexamer-primed first-strand cDNA was done using 5 pl of reverse transcription mixture in a total volume of 25 p1 containing 200 nM primers of cDNA of rat kidney Epo (30) as follows: sense primer (REP62F, 5'-TCCTTGCTACTGATTCCTCTGG-3') and an- tisense primer (REP512R, 5'-AAGTATCCGCTGTGAGTGTTCG-3'). Each of 30 cycles of PCR consisted of incubation for 1 min at 95 "C for denaturation, 2 min at 65 "C for annealing, and 3 min at 72 "C for elongation. For PCR amplification of p-actin, the sense primer (5'-TC- CTATGTGGGTGACGAGGC-3') and antisense primer (5"TACATG- GCTGGGGTGTTGAAGGTCT-3') (31) were used. The amplified prod- ucts were detected by staining using ethidium bromide and by Southern blotting using the 32P-labeled HhaI-BstEII fragment, 498 bp, of rat Epo cDNA (30) as a probe. The multiprime DNA labeling system for prepa- ration of the probe and Hybond N for blotting were used according to the instructions of the manufacturer (Amersham Corp.). Temperature for hybridization was 60 "C, and the final washing of the filter was done with 0.1 x SSC at 60 "C.

Immortalization, Cloning, and Identification of Brain Cells Produc- ing Epo-Cerebral cells were cultured for 30 days in 5% 0, with changes of the medium every 3 days and then trypsinized. The detached cells were inoculated with the cell density of one-fifth of that on trypsinization and cultured for 5 days at 5% 0,. The cultured cells were cotransfected by the calcium-phosphate method with two plasmids, pM- TIOD (Japanese Cancer Research Resources Bank) containing SV40 T-antigen gene for cell immortalization and pKSVlONeo containing neo gene for resistance to G418. Transfected cells were cultured for 2 days, and then G418 (400 pg/ml) was added. The colonies resistant to G418 were picked up, and cells of the randomly selected colonies were cloned by limiting dilution. The clonal cells were cultured to examine Epo production and used for immunochemical staining with anti-glial fibril- lary acidic protein antiserum ( D M 0 Japan Co., Ltd.). This acidic pro- tein is a specific marker of astrocytes (32).

Immunochemical staining was done as follows. Cells were fixed in 4% paraformaldehyde for 30 min and in 5% acetic acid in 90% methanol for 10 min. After preincubation with 3% normal goat serum for 30 min, cells were incubated with rabbit anti-glial fibrillary acidic protein antiserum or non-immune rabbit serum for 2 h a t room temperature. Then cells were incubated with biotinylated goat anti-rabbit antiserum (1500) for 1 h. Visualization was made with a Vectastain ABC kit (Vector Labora- tories, Inc.) and diaminobenzidine.

Determination of Cellular Protein-The cells were washed thor- oughly with PBS and then denatured by adding 1 ml of 0.1 M HCl. The precipitate was dissolved in 1 ml of 1 M NaOH. Protein in the cell lysates was determined with a protein assay kit (Bio-Rad) using bovine immu- noglobulin as a standard.

A

B

0 5 10 15 20 25 30 culture day

C

30

25

20

15

10

5

n J " 18 21 24 27 18 21 24 27

culture day culture day

FIG. 1. Oxygen-dependent production of Epo by primary cul- tures of rat cerebral cells. The cells were cultured with changes of the medium every 3 days, and Epo in the spent medium was measured with an enzyme-linked immunoassay. Each point is the mean -c S.D. of sextuplicate cultures. In A, the cells were cultured in 21 (0) and 5% (0) 0,. In B, the cells were cultured in 5% 0, for 18 days, and then 0, tension was changed to 1,5, and 21% at 3-day intervals. At the 3rd day of culture, the medium was harvested to assay Epo, and the subsequent culture at the different 0, tension was started in the fresh medium. The value at day 18 indicates Epo produced for days 16-18 in 5% 0,. In C, cultures were done as in B except that 0, tension was changed to 21,5, and 1% in that order.

RESULTS

Oxygen-dependent Production of Epo by Cultured Rat Brain Cells-Cerebral cells from 18-day-old fetuses of rats were cul- tured in an 0, tension of 21 and 5%, and Epo in the spent media was assayed with an enzyme-linked immunoassay (Fig. lA). When the cells were cultured in 21% 0,, Epo production was first detected after 18 days, but its level was low, and this low production continued later. In 5% 0,, the high production was observed after 12 days of culture, and hereafter the production increased gradually. To determine whether the regulation of the Epo production by 0, was reversible, the cells cultured for 18 days in 5% 0, were placed in different 0, tensions for 3 days, and Epo in the spent medium was assayed (Fig. 1, B and C). A decrease of 0, from 5 to 1% caused a 3.9-fold increase in Epo production. Subsequently Epo production was reduced as the 0, tension was increased from 1 to 5 and 21% 0,. Likewise, when the cells cultured for 18 days in 5% 0, were exposed to 21% 0,, the Epo production was reduced, and the following decrease of 0, from 21 to 5 or 1% resulted in increases of Epo production (Fig. E ) . The Epo production in 1% 0, was approxi- mately 20 times higher than that in 21% 0,.

19490 Brain Erythropoietin

A 0.6 I T I

0 0.2 0.4 0.6 0.8 1.0 1.2 Epo ( ng/ml )

B

E e w 100

0‘ I 0 0.1 0.2 0.3 0.4

Epo ( ng/ml )

FIG. 2. Brain Epo stimulates the growth and differentiation of erythroid cells. Brain Epo was partially purified (as described under “Experimental Procedures”), and serum Epo was isolated from anemic rats (27). In A, Epo was assayed with Epo-dependent growth of EP- FDC-P2 cells measuring the increased absorbance at 600 nm due to 3-(4,5-dimethyl thiazol-2-yl)-2,5-diphenyl tetrazolium bromide reduc- tion (33), and in B , with Epo-dependent erythroid colony formation of fetal mouse liver cells (34). 0, 0, brain Epo; A, A, serum Epo; 0, A, in the presence of sEpo-R capable of binding with Epo (35); 0, A, in the absence of sEpo-R. When assayed in the presence of sEpo-R, Epo was preincubated with sEpo-R (100 pg/ml) for 1 h. Each point in A is the mean f S.D. in quintuplicate, and that in B is the mean in duplicate.

In Vitro Biological Activity of Brain Epo-We partially puri- fied the immunoreactive material (brain Epo) from the spent medium using the mAb-fixed gel. Recovery of the purification was about 90%. There was little immunoreactive Epo in the preparation that was obtained from the fresh medium by the same purification procedure. The purified brain Epo showed dose-dependent biological activities when assayed with two in vitro criteria, stimulation of Epo-dependent cell growth (Fig. 2 A ) and erythroid colony formation of fetal mouse liver cells (Fig. 2B ). The soluble Epo-R (sEpo-R), an extracellular domain of the receptor that is capable of binding with Epo (35), abro- gated the biological activities of brain and serum Epo (Fig. 2, A and B) whereas the heat-denatured sEpo-R (5 min at 100 “C) did not (data not shown). These activities were also inhibited by the rabbit anti-Epo antiserum but not by the non-immune rab- bit serum (data not shown).

Assays of Epo-dependent cell growth provide more precise data than those of colony formation. When assayed with the cell growth in low concentrations of Epo, brain Epo was more active than the Epo isolated from rat serum (Fig. 2 A ) . Although we have not done the receptor binding assay because it is difficult

to prepare brain Epo in a necessary amount, this result indi- cates that brain Epo binds to Epo-R with an affinity higher than serum Epo. Higher affinity of brain Epo may be attribut- able to the incomplete sialylation as described below.

Western Blotting of Brain and Serum Epo-Brain Epo mi- grated on the polyacrylamide gel with a molecular size of 33 kDa, which was smaller than serum Epo (35 kDa) (Fig. 3 A , lanes 1 and 2 ) . Treatment of both Epo with neuraminidase at pH 5.0 for 24 h yielded a product with a reduced size of 31 kDa (Fig. 3B, lanes 2 and 4 1, but the treatment without enzyme did not change the size, which confirmed that the size reduction was due to enzymatic removal of sialic acids (Fig. 3 A , lane 1 and Fig. 3B, lane 1; Fig. 3A, lane 2 and Fig. 3B, lane 3). These results indicate that the difference in size between brain and serum Epo is mainly due to poor sialylation of brain Epo. De- sialylation yields Epo with higher affinity to Epo-R and conse- quently an increased in vitro activity (36), which may account for the higher activity of brain Epo compared with serum Epo (see Fig. 2 A ) . The two components with 67 and 55 kDa, which are visible in all samples, are the residual BSA added to sta- bilize Epo and the heavy chain of the antibody detached from the mAb-fixed gel, respectively. These components are present in the control sample (PBS containing 0.1% BSA), which was treated with the same procedures as the Epo preparations (Fig. 3 A , lane 3).

Oxygen-dependent Expression of Epo mRNA-We next de- tected brain Epo mRNA by the reverse transcription-PCR method. Cerebral cells were cultured in 21 and 1% 0, (three dishes for each 0, tension), and RNA isolated from the indi- vidual dishes was used for reverse transcription-PCR. Staining of amplified DNAwith ethidium bromide revealed a sharp band with a size of 451 bp (Fig. 4A), the exact size expected from mRNA of rat kidney Epo (30). The band signal derived from the culture in 1% 0, was approximately 20-fold stronger than that from the culture in 21% 0,. Similar findings were obtained when the amplified DNA was detected with Southern blotting using Epo cDNA as a probe (Fig. 4B). A diffuse band with a size of 150 bp (Fig. 4A) is a PCR artifact because this band does not hybridize with the cDNA probe (Fig. 4B ). The signal of p-actin mRNA was unchanged at different 0, tensions (Fig. 4A). The restriction fragments of the product derived from cerebral RNA are consistent with those expected from the cDNA of rat kidney Epo (30) and also identical to those of the product from the kidney RNA of rats made anemic for accumulation of Epo mRNA, confirming that the 451-bp band is a specific product of Epo mRNA (Fig. 4C).

Identification of Brain Cells Producing Epo-In our cerebral cell culture (Fig. 11, the cells proliferated during a lag period with an undetectable production of Epo, and the proliferated cells appeared to produce Epo. Since neurons from 18-day-old fetuses of rats do not proliferate, it is unlikely that they produce Epo. Proliferating microglial cells appeared atop the astrocytic layer or floated in the medium, and they could be removed by shaking the culture vessels. The isolated microglial cells were recultured, but the production of Epo was undetectable.

From microscopic inspection, astrocytes appeared to be the major population (>95%) in our cultures. These cells were CO- transfected with two plasmids conferring drug resistance and cell immortalization. The colonies resistant to G418 were picked up. The cells propagated from most colonies (95/100) produced Epo. Five colonies producing Epo were randomly se- lected, and the cells were cloned by limiting dilution. All of the established five clones showed hypoxia-induced production of Epo, but the response to hypoxia as well as the production in normoxia varied from clone to clone (Table I).

The clonal cells were immunochemically stained with the

Brain Erythropoietin 19491

B A

kDa. kDa.

97 - 67- 0 C- 1) -67 46 - 0 0 - - 5 5

1 2 3

1 2 3 4

30 -

FIG. 3. Western blotting of brain and serum Epo. Epo was stored in PBS containing 0.1% BSAuntil use. To concentrate Epo and remove RSA before Western blotting, the samples (brain Epo, serum Epo, and PBS containing 0.1% BSA) were treated as described under “Experimental Procedures.”A, Epo stored at pH 7.4. Lane 1 , brain Epo (33 kDa); lane 2, serum Epo (35 kDa); lane 3, PBS containing 0.1% BSA, which was treated with the same procedures as Epo. B, Epo incubated with or without neuraminidase a t pH 5.0 for 24 h a t 37 “C. L a m I, brain Epo (33 kDa) incubated without neuraminidase; lane 2, brain Epo (31 kDa) incubated with neuraminidase; lane 3, serum Epo (35 kDa) incubated without neuraminidase; lane 4, serum Epo (31 kDa) incubated with neuraminidase. Proteins with 67 and 55 kDa are the residual BSAand heavy chain of IgG, respectively.

A B

M no 21 % 0 2 1 % 0 2 M no 2 1 % 0 2 1 % 0 2

RTase + - + + + - + + + RTase + - + + + - + + + ” ”

C

brain kidney

- S E H - S E H

~t 451 bp

p-actin

FIG. 4. Detection of Epo mRNA. The cells from rat fetal cerebrum were cultured for 18 days in 5% 0, and then exposed in 21% 0, for 2 days to repress Epo production. Then the cells were cultured for 8 h in 21 and 1% 0, (three dishes for each 0,). Epo mRNA in the total RNA from each dish (60 cm’containing about 10’ cells) was amplified with reverse transcription-PCR. In A, the amplified products of cerebral RNAwere visualized with agarose gel electrophoresis in the presence of ethidium bromide. M, marker; no, without RNA in the reverse transcriptase (RTuse) reaction;

blotting using the :’YP-labeled fragment of rat Epo cDNA (30) as a probe. In C, the amplified product of cerebral and kidney RNA was precipitated -, without reverse transcriptase in the reverse transcription reaction. In B , the amplified products of cerebral RNA were detected by Southern

by polyethylene glycol, digested with Sac1 (S) , EcoT14I ( E ) , or HincII (H), and the resulting fragments were detected with Southern blotting. From the sequence of rat Euo cDNA (30). digestion with SucI. EcoTl4I. and HincII should yield the hybridizable fragment with 377, 356, and 276 bp, respectively. Agarose concentrations were 1% in A and b and 30’in C.

TAULE I Oxygen-dependent production of Epo by cloned astrocytes

dium (IO? cells/well) using 6-well multidishes. After culture for 60 h in Cells of the astrocyte clones were inoculated in 2 ml of culture me-

5% 0,. the medium was changed with 1 m l of fresh medium, and oxygen tension for culture was shifted to 1 or 219 0,. After culture for 24 h, Epo produced in the spent medium and cellular protein was determined as described under “Experimental Procedures.” Values are the means k S.D. in triplicate.

Epo produced for 24 h Clones Medium Cell protein

21% 0, 1% Or 21% 0, 1c7r 0,

PR / 111 / PR I w z1 5.4 f 0.3 14.8-c 1.2 3.6-c 0.1 13.1 -c 1.3 22 9.520.4 31.2 k3.9 4.920.2 23.922.7 D2 7.8 2 0.4 18.4 2 1.2 6.4 2 0.3 18.6 fO.5 D3 9.1 20.5 26.5 k2.6 6.5k0.4 24.5k 1.1 D7 4.920.7 26.225.1 3.220.1 22.323.8

antiserum against glial fibrillary acidic protein, a specific marker of astrocytes (32). Fig. 5 shows some representative results. Staining with the antiserum clearly indicated expres-

sion of this marker protein in the cloned cells (Fig. 5, A, C, and D) and C-6 cells (37), a glial cell line derived from rat astrocy- toma (Fig. 5 E ) whereas staining with a non-immune serum gave few positive cells (Fig. 5, B and F ) . All other clones estab- lished here expressed this marker protein (data not shown). Epo production by C-6 cells was undetectable in either 21 or 1% 0,.

DISCUSSION

In fetal stage, Epo is produced in the liver where erythropoi- esis takes place. In adult, Epo produced by the kidney travels in the circulation to reach erythropoietic tissues, bone marrow, and spleen (in rat, mouse, etc.). So far the erythroid lineage has been thought to be a sole physiological target of Epo. The pres- ent study provides a new site for Epo production by demon- strating that astrocytes produce the bioactive Epo. Taken to- gether with the previous finding that neurons express the functional Epo-R (19, 20), our present results support the idea that Epo exerts a novel function in brain.

A much higher concentration of Epo is required to show ac- tivity in neurons than in erythroid precursor cells. In neuronal

19492 Brain Erythropoietin been found. Hepatoma cell lines producing Epo are extensively used for in vitro studies of the expression of Epo gene (16, 17). Analyses with reporter genes revealed that the 3'-flanking re- gion of the Epo gene contained an enhancer element respon- sible for hypoxia-induced transcriptional activation (42-46). The 5'-promoter region is also involved in the hypoxia induc- ibility (45, 47, 48). Proteins binding to the 3'- and 5"flanking regions have been reported (42,4943). These proteins are also hypoxia-inducible and may be trans-activators in the Epo gene transcription. Astrocyte cell lines established in the present study would be very useful for identification of the trans-acting proteins that regulate expression of the Epo gene in brain.

FIG. 5. Immunochemical detection of glial fibrillary acidic pro- tein, a specific marker of astrocytes. The clonal cells established here (21, D2, and D3) and C-6 cells, rat astrocytoma cells, were stained with the antiserum against glial fibrillary acidic protein (A, C, D, and E) or with a non-immune serum ( B and F). A and B, clone Z1; C, clone D2; D, clone D3; E and F, C-6.

cells, Epo acts at a nanomolar range (19, 20) whereas Epo at 100 PM shows maximum activity in erythroid cells (1). Indeed, the ligand affinity of Epo-R in neuronal cells (h, = 10-16 nM) is much lower than that in erythroid cells (h, = 95 PM) (19). It has been proposed that neuronal and other cells express an acces- sory protein that alters the ligand affinity through interaction with Epo-R (19, 38, 39). Glial cells including astrocytes are located very close to or interact directly with neurons. Produc- tion of Epo by astrocytes indicates that Epo acts on neurons through a paracrine mechanism that would overcome the low ligand affinity of Epo-R in neurons and also the unavailability of blood Epo in the central nervous system due to the blood- brain barrier.

Epo has been shown to elevate monoamine contents and induce rapid incorporation of Ca2+ in the rat cell line PC12, which expresses more neuronal properties with exposure to nerve growth factor (19). Furthermore Epo augments choline acetyltransferase activity in mouse embryonic primary septal neurons and in cholinergic cell line SN6 and promotes the survival of septal cholinergic neurons in adult rats that re- ceived fimbria-fornix transections (20). It appears from these results that Epo functions as a neurotrophic factor. Immuno- chemical experiments with mouse postimplantation embryo have indicated that Epo and Epo-R are expressed in the neural plate, implying that Epo is involved in neurogenesis (40). Fur- ther studies are necessary to know a true function of Epo in nervous system.

In accordance with a widely accepted notion that oxygen plays a key role in regulating Epo expression (1, 2, 41), Epo production by astrocytes was dependent on oxygen tension for culture, and the regulation operated at the level of mRNA. No renal cell lines producing Epo in an oxygen-regulated way have

REFERENCES

2. Jelkmann, W. (1992) Physiol. Rev. 72,449-489 1. Krantz, S. B. (1991) Blood 77,419-434

3. Jacobson, L. 0.. Goldwasser, E., Fried, W., and Plzak, L. (1957) Nuture 179,

4. Zanjani, E. D., Poster, J., Burlington, H., Mann, L. I., and Wasserman, L. R.

5. Koury, S. T., Bondurant. M. C., and Koury, M. J. (1988) Blood 71,524-527 6. Lacombe, C., DaSilva, J. L., Bruneval, P., Fournier, J. G., Wendling, E, Cas-

J . Clin. Invest. 81, 620-623 adevall, N., Camilleri, J. P., Bariety, J., Varet, B., and Tambourin, P. (1988)

7. Bachmann, S., Hir, M. L., and Eckardt, K. (1993)J. Histochem. Cytochem. 41, 335-341

8. Koury, S. T., Bondurant, M. C., Koury, M. J., and Semenza, G. L. (1991) Blood

9. Schuster, S. J., Koury, S. T., Bohrer, M., Salceda, S. , and Caro, J. (1992) BE J. 77,2497-2503

Hoemutol. 81, 153-159 10. Eckardt, K.-U., Pugh, C. W., Ratcliffe, P. J., and Kurta,A. (1993)PflugersArch.

Eur: J . Physiol. 423,356-3fi4 11. Bondurant, M. C., and Koury, M. J. (1986) Mol. Cell. Bid. 6,2731-2733 12. Beru, N., McDonald, J. , Lacombe, C., and Goldwasser, E. (1986)Mol. Cell. Biol.

13. Koury, S. T., Koury, M. J., Bondurant, M. C., Caro, J.. and Graber, S. E. (1989)

14. Semenza, G. L., Koury, S. T., Nejfelt, M. K., Gearhart, J. D., and Antonarakis,

15. Schuster, S. J., Badiavas, E. V., Costa-Giomi, P., Weinmann, R., Erslev, A. J.,

16. Goldberg, M. A,, Glass, G . A,, Cunningham, J. M., and Bum, H. F. (1987)Proc.

17. Goldberg, M.A., Dunning, S. P., and Bum, H. F. (1988)Science 242,1412-1415 18. Costa-Giomi, P., Caro, J., and Weinmann, R. (1990) J . Bid. Chem. 265,10185-

633-634

(1977) J. Lab. Clin. Med. 89,640-644

6, 2571-2575

Blood 74,645-651

S. E. (1991) Proc. Nutl. Acud. Sci. U. S. A. 88,8725-8729

and Caro, J. (1989) Blood 73, 13-16

Notl. Acud. Sci. U. S. A. 84, 7972-7976

10188

24.

25.

26.

27.

28. 29.

30.

31.

32.

33.

34.

35.

36.

37.

38. 39.

19. Masuda, S., Nagao, M., Takahata, K., Konishi, Y., Gallyas, F., Jr., Tabira, T.,

20. Konishi,Y., Chui, D.-H., Hirose, H., Kunishita. T., and Tabira, T. (1993)Brain

21. Mollgrad, K., and Saunders, H. R. (1986) Neuropathol. Appl. Neurobiol. 12,

22. Tan, C. C., Eckardt, K.-U., Firth, J. D., and Ratcliffe, P. J. (1992) Am. J.

23. Suzumura, A,, Bhat, S., Eccleston, P. A,, Lisak, R. P., and Silberberg, D. H.

and Sasaki. R. (1993) J. Bid . Chem. 268, 11208-11216

Res. 609,29-35

337-358

Physiol. 263, F474-F481

(1984) Brain Res. 324. 379-383 Goto, M., Murakami, A.,'Akai, K., Kawanishi, G., Ueda, M., Chiba, H., and

Sasaki, R. (1989) Blood 74,1415-1423 Goto, M.,Akai, K., Murakami, A,, Hashimoto, C., Tsuda, E., Ueda, M., Kawan-

BiolTechnology 6, 67-71 ishi, G., Takahashi, N., Ishimoto, A,, Chiba, H., and Sasaki, R. (1988)

Sasaki, R., Yanagawa, S., and Chiba, H. (1987) Methods Enzymol. 147, 328-

Okano, M., Suga, H., Masuda, S., Nagao. M., Narita, H., Ikura, K., and Sasaki, 340

Chomczynski, P., and Sacchi. N. (1987)Anul. Biochem. 162, 156-159 R. (1993) Biosci. Biotechnol. Rinchem. 57, 1882-1885

Kawasaki, E. S. (1990) in PCR Protocols (Innis, M. A,, Gelfand, D. H., and

Nagao, M., Suga, H., Okano, M., Masuda, S., Narita, H., Ikura, K., and Sasaki,

Alonso, S., Minty, A,, Bourlet, Y., and Buckingham, M. E. (1986)J. Mol. Evol.

Raff, M. C., Fields, K. L., Hakomori, S., Mirsky, R., Pruss, R. M., and Winter,

Yamaguchi, K., Akai, K., Kawanishi, G., Ueda, M., Masuda, S., and Sasaki, R.

Iscove, N. N., Sieber. F., and Winterhalter, K. H. (1974) J. Cell Physiol. 83,

Nagao, M., Masuda, S., Abe, S., Ueda, M., and Sasaki, R. (1992) Biochem.

Tsuda, E., Kawanishi, G., Ueda, M., Masuda, S., and Sasaki, R. (1990) Rur -1

Sninsky, J. J., eds) pp. 21-27, Academic Press, New York

R. (1992) Biochim. Biophys. Acta 1171,99-102

23, 11-22

J. (1979) Bruin Res. 174,283-308

(1991) J. Bid. Chem. 266,20434-20439

309-320

Hiophys. Res. Commun. 188, 888-897

Biochem. 188,405-411 _ .

Bissell, M. G., Rubinstein, L. J., Bignami, A,, and Herman, M. M. (1974) Brain .~ ~

Dong, Y. J., and Goldwasser, E. (1993) Exp. Hpmulol. 21,483-486 Nagao, M., Matsumoto, S., Masuda, S., and Sasaki, R. (1993) Blood 81,2503-

Res. 82 ,7749

2510

Brain Erythropoietin 19493 40. Yasuda, Y., Nagao, M., Okano, M., Masuda, S., Sasaki, R., Konishi, H., and

41. Porter, D. L., and Goldberg, M. A. (1993) Exp. Hematol. 21,399-404 42. Semenza, G. L., Nejfelt, M. K., Chi, S. M., andAntonarakis, S. E. (1991)Proc.

43. Beck, I., Ramirez, S., Weinmann, R., and Caro, J. (1991) J. B i d . Chem. 266,

44. Pugh, C. W., Tan, C . C., Jones, R. W., and Ratcliffe, P. J. (1991) Proc. Natl.

45. Blanchard, K. L., Acquaviva, A. M., Galson, D. L., and Bunn, H. E (1992) Mol.

46. Madan, A., and Curtin, P. T. (1993) Proc. Natl. Acad. Sci. U. S. A. 90, 3928-

Tanimura, T. (1993) Deu. Growth & Differ: 36,711-722

Natl. Acad. Sci. U. S. A. 88, 5680-5684

15563-15566

Acad. Sci. U. S. A. 88, 10553-10557

Cell. Biol. 12, 5373-5385

47. Beru, N., Smith, D., and Goldwasser, E. (1990) J. B i d . Chem. 285, 14100-

48. Imagawa, S., Goldberg, M. A,, Doweiko, J., and Bunn, H. F. (1991) Blood 77,

49. Semenza, G. L., and Wang, G. L. (1992) Mol. Cell. B i d . 12, 5447-5454 50. Beck, I., Weinmann, R., and Cam, J. (1993) Blood 82,706711 51. Wang, G. L., and Semenza, G. L. (1993) Proc. Natl. Acad. Sci. U. S. A. 90,

52. Wang, G. L., and Semenza, G. L. (1993) J. Biol. Chem. 268,21513-21518 53. Tsuchiya, T., Ueda, M., Ochiai, H., Imajoh-Ohmi, S., and Kanegasaki, S. (1993)

3932

14104

278-285

4304.4308

J. Biochem. (Zbbkyo) 113,395-400