AU776280B2 - Production of erythropoietin by endogenous gene activation - Google Patents

Description

t

AUSTRALIA

PATENTS ACT 1990 DIVISIONAL APPLICATION NAME OF APPLICANT: Roche Diagnostics GmbH ADDRESS FOR SERVICE: DAVIES COLLISON CAVE Patent Attorneys 1 Little Collins Street Melbourne, 3000.

INVENTION TITLE: "Production of erythropoietin by endogenous gene activation" The following statement is a full description of this invention, including the best method of performing it known to us: I I

T

Preparation of Erythropoietin by Endogenic Gene Activation The invention relates to human cells which are capable, on the basis of an activation of the endogenic human EPO gene, of producing EPO in sufficient amount and purity to permit economical preparation of human EPO as a pharmaceutical preparation. The invention furthermore relates to a method of preparing such human EPO-producing cells, DNA constructs for activating the endogenic EPO gene in human cells, and methods for the largescale production of EPO in human cells.

Erythropoietin (EPO) is a human glycoprotein which stimulates the production of red blood cells. EPO occurs in the blood plasma of healthy persons only in very low concentrations, so that preparation in large amounts is not possible in this manner. EP-B-0148 605 and EP- B-0205 564 describe the preparation of recombinant human EPO in CHO cells. The EPO described in EP-B-0148 605 has a higher molecular weight than urinary EPO and no O-glycosylation. Meantime, the EPO from CHO cells, described in EP-B- 0 205 564 is available in large amounts and in pure form, but it originates from nonhuman cells.

Moreover, the ability of CHO cells to produce is often relatively limited.

Furthermore, the harvesting of human EPO from the urine of patients with aplastic anemia is known (Miyake et al., J. Biol. Chem. 252 (1977), 5558 5564). Therein a seven-step process is disclosed which includes ion exchanger chromatography, ethanol precipitation, gel filtration and adsorption chromatography. An EPO preparation w':h a specific activity of about 70,000 U/mg of protein is obtained in a 21% yield. Disadvantages of this process and other methods of obtaining urinary EPO consist in the procurement of starting material in sufficient amounts and in repeatable quality. Furthermore, the purification from urine is difficult and even a purified product is not free of urinary contaminants.

GB-A-2085 887 describes a method for the preparation of human lymphoblastoid cells which are capable of producing EPO in small amounts. The economical production of a pharmaceutical with these cells is not possible.

WO 91/06667 describes a method for the recombinant preparation of EPO. In a first process step in primary human embryo kidney cells the endogenic EPO gene is brought by homologous recombination into operative linkage with a viral promoter and the DNA is isolated from these cells. In a second step the DNA thus isolated is transformed to nonhuman CHO cells and the expression of EPO in these cells is analyzed. No mention is found that production of EPO in human cells is possible.

WO 93/09222 describes the production of EPO in himian cells, wherein a relatively high EPO production of up to 960,620 mU/106 cells/24 h is found in human fibroblasts which had been transfected with a vector containing the complete EPO gene. These cells contain an exogenic EPO gene which is not at the correct EPO gene locus, so that problems are to be expected with regard to the stability of the cell line. No information on a constitutive EPO production is found in WO 93/092222. Furthermore, neither is'there any information as to whether the EPO produced can be obtained in a quality sufficient for pharmaceutical purposes.

Furthermore, in WO 93/09222 an activation of the endogenic EPO gene in human HT1080 cells is described. In it an EPO production is found of only 2,500 mU/10 6 cells in 24 h (corresponding approximately to 16 ng/10 6 cells/24 Such low production is entirely unsuited for the economical production of a pharmaceutical preparation.

W094/12650 and WO 95/31560 describe how a human cell with an endogenic EPO gene activated by a viral promoter is able, after amplification of the endogenic EPO gene, to produce up to approximately 100,000 mU/106 cells/24 h (corresponding to about 0.6 gg/10 6 cells/24 This amount too is still not sufficient for the economical production ofa pharmaceutical.

The problem on which the present invention is based thus consisted in eliminating at least partially the above-described disadvantage of the state of the art and to offer a S: technologically better method for the preparation of EPO in human cells.

In a preferred embodiment, this problem is addressed by activation of the endogenic EPO gene in human cells and optionally by subsequent amplification of the activated human EPO gene. In this manner it has surprisingly been possible, by the selection of suitable starting cells, DNA constructs and selection strategies, to provi( human cells which are capable of producing EPO in sufficient quantity, quality and purity to permit the economical production of pharmaceutical preparations. Especially after amplification of the activated endogenic EPO gene, cells may be obtained which have a definitely higher production output than the CHO production cells used previously for the preparation of recombinant EPO.

One embodiment of the invention is a human cell which contains a copy of an endogenic EPO gene in operative linkage with a heterologous expression control sequence active in the human cell and is capable without prior gene amplification of producing at least 200 ng of 6 cells per 24 hours. Preferably the human cell according to the invention is capable of the production of 200 to 3000 ng EPO/106 cells/24h, and most preferably for the production of 1000 to 3000 ng EPO/106 cells/24h.

Another embodiment of the present invention is a human cell which is obtained by gene amplification of the genetic material from the cell previously stated and contains several copies of an endogenic EPO gene, each in operative linkage vith a heterologous expression control sequence active in the human cell, and is capable of the production of at least 1,000 ng EPO/106 cells/24h.

With special preference the human cell line obtainable by amplification of genetic material is capable of producing 1,000 to 25,000 ng EPO/ 10 cells/24h, and most preferably for the production of 5,000 to 25,000 ng EPO/10 6 cells/24h.

The human cell is any cell, provided it can be cultured in vitro. Especially preferred are human cells which can be cultured in a serum-free medium, and especially in suspension. In this manner the production of EPO can be performed in a large fermenter with a culture capacity of, for example, 1,000 liters.

Especially preferred is a human cell which is an immortalized cell, a HT1080 cell (Rasheed et al., Cancer 33 (1974), 1027-1033), a HeLaS3 cell (Puck et al., J.Exp. Med.

103(1956), 273-286), a Namalwa cell (Nadkari et al., Cancer 23 (1969), 64-79) or a cell derived therefrom.

An example of a cell according to the invention is the clone "Aladin" which was deposited on July 1997 according to the prescriptions of the Budapest Agreement at the Deutsche Sammlung von Mikroorganismen und Zellkulturen (DSMZ), Mascheroder Weg Ib, 38124 Braunschweig, under S number DSM ACC 2320.

In the cell according to one aspect of the invention, the endogenic EPO gene is linked with a heterologous expression control sequence which is active in the human cell. The expression control sequence comprises a promoter and preferably additional expression-improving sequences, an enhancer. The promoter can be a regulatable or a constitutive promoter. Preferably the promoter is a strong viral promoter, an SV40 or a CMV promoter. The CMV promoter/enhancer is especially preferred.

Furthermore, to optimize the EPO expression it may be preferred for the endogenic EPO gene in the human cell, which is in operative association with the heterologous promoter, to have a signal peptide-coding sequence which is different from the natural signal peptidecoding sequence and codes preferably for a signal peptide with a modified amino acid sequence. Especially preferred is a signal peptide-coding sequence which codes for a signal peptide sequence modified in the area of the first four amino acids which is selected from Met-X,-X 2

-X

3 wherein X, is Gly or Ser, X 2 is Ala, Val, Leu, Ile, Ser or Pro, and X 3 is Pro, Arg, Cys or His, provided that Xi-XZ-X 3 is not the sequence Gly-Val-His, and especially from Met-Gly-Ala-His, Met-Ser-Ala-His Met-Gly-Val-Pro or Met-Ser-Val-His.

With special preference the sequence of the first four amino acids of the signal peptide is Met-Ser-Ala-His.

In an additional aspect, the present invention relates to a method for preparing a human EPO-producing cell, as previously stated, comprising the steps: Providing human starting cells which contain at least a copy of an endogenic EPO gene, Transfecting the cells with a DNA construct comprising: two flanking DNA sequences which are homologous with regions of the human EPO gene locus, to permit a homologous recombination, (ii) a positive selection marker gene, and (iii) a heterologous expression control sequence which is active in the human cell, Culturing the transfected cells under conditions in which a selection takes place for the presence of the positive selection marker gene, Analyzing the cells selectable according to step and *.o Identifying the EPO-producing cells.

The DNA construct used in preparing the human EPO-producing cell contains two flanking DNA sequences which are homologous with areas of the human EPO gene locus in order to permit a homologous recombination. The selection of suitable flanking sequences is performed, for example, by the methods described in WO 90/11 354 and WO 91/09 955.

Preferably the flanking sequences have each a length of at least 150 bp. With special preference the homologous DNA sequences are selected from the area of the sequences, exon 1 and intron I of the EPO gene. It is especially preferred to use a DNA sequence modified in the area of exon 1, which codes for a signal peptide modified in the first amino acids. The modifications in the exon 1 area are preferably as stated above.

The selection marker gene can be any selection marker gene suitable for eucaryotic cells, which upon expression leads to a selectable phenotype, antibiotic resistance, *auxotrophy, etc. An especially preferred positive selection marker gene is the neomycin phosphotransferase gene.

Optionally, a negative selection marker gene such as, say, the HSV-thymidine kinase gene, can also be present, by whose expression cells are destroyed in the presence of a selection agent.

If an amplification of the EPO gene endogenously activated in the human cell is desired, the DNA construct contains an amplification gene. Examples of suitable amplification genes are dihydrofolate reductase, adenosine deaminase, ornithine decarboxylase, etc. An especially preferred amplification gene is the dihydrofolate reductase gene, especially a dihydrofolate reductase arginine mutant which has a lower sensitivity to the selective agent (methotrexate) than the wild type gene (Simonsen et al., Proc. Natl. Acad. Sci. USA (1983); 2495).

If an amplification gene is present in the DNA construct used for activating the EPO gene, the method of the invention can also comprise the following steps: Amplification of the DNA sequence coding for EPO, and Obtaining EPO-producing cells which contain a number of copies, greater in comparison with the starting cell, of an endogenic DNA sequence coding for mature EPO in operative linkage with a heterologous expression control sequence.

Appropriate DNA constructs for the activation of the endogenic EPO gene present in the human starting cell as the plasmids listed in the examples: pl87, pl89, p190 and p192 or a plasmid derived therefrom. Especially preferred is the plasmid p189 which was deposited at the DSMZ in accord with the provisions of the Budapest Agreement on July 16 th 1997 under the accession number DSM 1 1661 or a plasmid derived therefrom. Preferably, the DNA constructs are circular plasmid molecules which are used for the transfection of the human cell in a linearized form.

An additional subject matter of the present invention is a DNA construct for the activation of an endogenic EPO gene in a human cell, comprising: two flanking DNA sequences which are chosen homologous to regions of the human EPO gene locus selected from 5'-untranslated sequences, exon I and intron 1, in order to permit a homologous recombination, a modified sequence being present in the exon 1 region coding for the amino acids: Met-X,-X2-X3 wherein X, is Gly or Ser, X 2 is Ala, Val, Leu, Ile, Ser or Pro, and X 3 is Pro, Arg, Cys or His, provided that X,-X,-X 3 is not the sequence Gly-Val-His, and especially for the amino acids: Met-Gly-Ala-His, Met-Ser-Ala-His, Met-Gly-Val-Pro or Met-Ser-Val-His, (ii) a positive selection marker gene, (iii) a heterologous expression control sequence which is active in a human cell, and (iv) optionally an amplification gene.

Still another subject of the present invention is a DNA construct for the activation of an endogenous EPO gene in a human cell, comprising; two flanking DNA sequences which are homologous to regions of the human EPO gene locus, selected from 5'-untranslated sequences, exon I and intron 1, to permit a homologous recombination, (ii) a positive selection marker gene, (iii) a heterologous expression control sequence which is active in a human cell, the distance between the heterologous expression control sequence and the translation start of the EPO gene being no greater than 1100 bp, and (iv) optionally, an amplification gene.

Surprisingly it was found that, in the case of a modification of the EPO signal sequence and/or a shortening of the distance between the heterologous expression control sequence and the start of the EPO gene, an optimized expression is obtained. Preferably the distance between the promoter of the heterologous expression control sequence and the translation start of the EPO gene is not greater than 1100 bp, with special preference not more than 150 bp, and most preferentially not more than 100 bp. An especially preferred example of a S DNA construct to be used according to the invention is the plasmid p189 DSM 11661 or a plasmid derived therefrom.

Still another aspect of the present invention is a method for preparing human EPO, wherein a human cell according to the invention is cultured in a suitable medium under conditions in which a production of EPO takes place and the EPO is harvested from the culture medium.

A serum-free medium is used preferentially. The cells are preferably cultured in suspension.

The culturing is performed especially preferentially in a fermenter, especially in a large S fermenter with a capacity of, for example, 10 1 50,000 1.

The harvesting of human EPO from the culture medium of human cell lines comprises preferably the following steps: Passing the cell supemate over an affinity chromatography medium and harvesting the fractions containing EPO, optionally passing the EPO-containing fractions over a hydrophobic interaction chromatography medium and harvesting the EPO-containing fractions, passing the EPO-containig fractions over hydroxy apatite and harvesting the EPOcontaining fractions, and concentrating and/or passing over a reverse-phase (RP) HPLC medium.

Step of the purification process includes passing the cell supemate, which in some cases can be pre-treated, over an affinity chromatography medium. Preferred affinity chromatography media are those to which a blue dye is coupled. An especially preferred example is blue sepharose.

After elution from the affinity chromatography medium the EPO-containing eluate is optionally passed over a hydrophobic interaction chromatography medium. This step is expedient if a culture medium with a serum content 2% is used. If a culture medium with a low serum content is used, 1% or a serum-free medium is used, this step can be omitted. A preferred hydrophobic interaction chromatography medium is butylsepharose.

The eluate from step or, if used, step is passed in step of the method of the invention over hydroxyapatite and the EPO-containing eluate is subjected to a concentrating step and/or a reverse-phase HPLC purification step. The concentration is performed preferably by exclusion chromatography, membrane filtration, the use of a medium, such as a membrane with an exclusion size of 10 kD having proven desirable.

By the method according to the invention an isolated human EPO with a specific activity of S: at least 100,000 U/mg protein in vivo (normocythemic mouse) is obtainable, which is free from urinary contaminants and can differ in its glycosilation from recombinant EPO from CHO cells. Preferably, the EPO of the invention has a specific activity of at least 175,000, and with special preference at least 200,000 to 400,000 or 450,000 IU/mg protein. The human EPO obtainable by the method of te invention can contain a-2, 3- and/or a-2, 6linked sialic acid residues. In studies of EPO from cells which contain an endogenically S* activated EPO gene, on the basis of the present preliminary results, the presence of a-2, 3and a-2, 6 linked sialic acid residues were found. Furthermore, it was found on the basis of the present preliminary results that the human EPO of the invention has a content of less than 0.2% N-glycol neuraminic acid based on the content ofN-acetyl neuraminic acid The purity of the human EPO of the invention amounts preferably to at least 90%, with special preference at least 95%, and most preferably at least 98% of the total protein content The determination of the total protein content can be performed by reverse phase HPLC, with a Poros R/2H column.

Furthermore, by the method of the invention, human EPO species are obtainable which differ in their amino acid sequence. Thus it was found by mass spectrometric analysis (MALDI-MS) that a human EPO can be isolated from HeLa S3 cells, which is mainly a polypeptide with a length of 165 amino acids, which is formed by C-terminal processing of an arginine residue, and in some cases includes up to 15% of an EPO with 166 amino acids.

Also, a human EPO is obtainable which includes a polypeptide with a length of 166 amino acids, a non-processed EPO. From Namalwa cells, for example, a human EPO has been isolated which comprises a mixture of polypeptides with a length of 165 and 166 amino acids.

This human EPO can be used as an active substance for a pharmaceutic preparation which can contain additional active substances, as well as pharmaceutically common adjuvants, vehicles and additives.

In still another aspect the present invention relates to an isolated DNA which codes for a human EPO with a sequence modified in the region of the first four amino acids of the signal peptide, which is selected from: Met-X,-X 2

-X

3 wherein X, is Gly or Ser, X 2 is Ala, Val, Leu, Ile, Ser or Pro, and X 3 is Pro, Arg, Cys or His, provided that XI-X,-X 3 is not the sequence Gly-Val-His, and especially for the amino acids: Met-Gly-Ala-His, Met-Ser-Ala-His Met-Gly-Val-Pro or Met-Ser-Val-His.

The DNA can be for example a genomic DNA or a cDNA.

The invention will continue to be explained by the following examples and figures and sequence listings.

Figure 1 is a schematic representation of the amplification of homology regions of the EPO gene from the area of the 5'-untranslated sequences, exon 1 and intron 1.

Figure 2 is a schematic representation of a plasmid which contains EPO homology regions from the area of the 5'-untranslated sequences, exon I and intron 1.

Figure 3 is a schematic representation of a gene-activation sequence which contains the Rous sarcoma virus promoter (RSV), the neomycin phosphotransferase gene (NEO), the early polyadenylation region of SV40 (SVI pA), the early promoter (SVI), the dihydrofolate reductase gene (DHFR), an additional early SV40 polyadenylation region, and the cytomegalovirus immediate early promoter and enhancer (MCMV).

Figure 4a The preparation of the EPO gene targeting vector p 176.

Figure 4b The preparation of the EPO gene targeting vectors p179 and p187.

Figure 4c The preparation of the EPO gene targeting vector p 18 9 Figure 4d The preparation of the EPO gene targeting vector p 190.

Figure 4e The preparation of the EPO gene targeting vector p1 92 Figure 5 A schematic representation of the preparation of EPO cDNA with signal sequence mutations.

Figure 6a The hybridization of cellular DNA with a probe from the CMV area of the gene cassette represented in Fig. 3; the lanes 1 to 4 are each of DNA from human cells cleaved with the restriction enzymes Agel and AscI; lane 1: EPO-producing HeLa S3 cell amplified with 1,000 nM MTX; lane 2: EPOproducing HeLa S3 cell amplified with 500 nM MTX; lane 3: EPO- Sproducing HeLa S3 Cell without amplification; lane 4: HeLa S3 cell without activated EPO gene; lane 5: digoxigenin-marked length marker; the size of the hybridizing fragment in lanes 1 to 3 is approximately 5,200 bp, and Figure6b The hybridization of a probe from the coding area of EPO with DNA from human cells; lane 1: digoxigenin-marked length marker, lanes 2 to 4: DNA from human cells cleaved with the restriction enzymes BamHI, HindIu and SalI; lane 2: EPO producing HeLa S3 cell amplified with 500 nM MTX (length of the band produced by the non-activated endogenic gene: 3,200 bp; length of the copy of the EPO gene activated by gene targeting: 2,600 bp); lane 3: DNA from an EPO producing HeLa S3 cell not amplified; lane 4: DNA from an HeLa S3 control cell.

SEQ ID No.

and No. 2 Nucleotide sequences of the primers used for preparing PCR Product 1 (Fig.

1).

SEQ ID No. 3 and No. 4 Sequences of the primers used for preparing PCR Product 2 (Fig. 1).

SEQ ID No. 5 Sequence of the primer EPO EX1.

SEQ ID No. 6 Sequence of the primer EX2.

SEQ ID No. 7 Sequence of the primer EX3 (Met-Gly-Ala-His).

SEQ ID No. 8 Sequence of the modified signal peptide start encoded by primer EX3.

SEQ ID No. 9 Sequence of primer EX4 (Met-Ser-Ala-His).

SEQ ID No. 10 Sequence of the modified signal peptide start encoded by primer EX4.

SEQ ID No. 11 Sequence of primer EX5 (Met-Gly-Val-Pro).

SEQ ID No. 12 Sequence of the modified signal peptide start encoded by primer SEQ ID No. 13 Sequence of primer EX6 (Met-Ser-Val-His).

SEQ ID No. 14 Sequence of the modified signal peptide start encoded by primer EX6.

SEQ ID No. 15 Sequence of primer EX genome 1.

SEQ ID No. 16 Sequence of primer EX13.

SEQ ID No. 17 Sequence of primer EPO EX 17.

Examples The activation of the EPO gene locus for protein production in an industrial scale was achieved by homologous integration of a gene activation sequence which contains the neomycin phosphotransferase (NEO) gene for the selection (G-418 resistance), the murine Sdihydrofolate reductase (DHFR) gene (for gene amplification by MTX) and the cytomegalovirus (CMV) immediate early promoter and enhancer for gene activation.

Example 1 Cloning of EPO Homology Regions Homology regions of the EPO gene were amplified by using a genomic placenta DNA (Boehringer Mannheim). Two PCR products were prepared from a homology region 6.3 kB long from the area of the 5'-untranslated sequences of the EPO gene, exon 1 and intron 1 (cf. Figure The primers used for the preparation of PCR Product l.had the following sequences: 5'-CGC GGC GGA TCC CAG GGA GCT GGG TTG ACC GG-3' (SEQ ID No.

1) and 5'-GGC CGC GAA TTC TCC GCG CCT GGC CGG GGT CCC TCA GC-3' (SEQ ID No. The primers used for the preparation of PCR Product 2 had the following sequences: 5'-CGC GGC GGA TCC TCT CCT CCC TCC CAA GCT GCA ATC-3' (SEQ ID No. 3) and 5.GGC CGC GAA TTC TAG AAC AGA TAG CCA GGC TGA GAG-3' (SEQ ID No. 4).

The desired segments were cut out of PCR Products 1 and 2 by restriction cleavage (PCR Product 1: HindIII, PCR Product 2: HindIII and Eco RV) and cloned into the vector pCRII (Invitrogen) which had been cleaved with Hind III and Eco RV. The recombinant vector obtained in this manner was named Sepopcrl000 (cf. Figure 2).

Example 2 Construction of EPO Gene Targeting Vectors 2.1 A gene activation sequence which contains the NEO gene, the DHFR gene and a CMV promoter/enhancer (cf. Fig. 3) was inserted into the Agel site of the plasmid 5epocr1000 containing the EPO homology region, and the plasmid p176 was obtained (cf. Fig. 4a). To bring the CMV promoter as close as possible to the translation start locus of the EPO gene, a segment 963 bp long was deleted between the restriction sites AscI and Agel (partial cleavage), whereupon the plasmid p179 was obtained (Fig. 4b).

2.2 To achieve an optimization of the expression, nucleotides in exon 1, which code for the beginning of the EPO leader sequence Met-Gly-Val-His, were replaced by the synthetic sequence Met-Ser-Ala-His (cf. also Example This sequence was obtained by amplification of a genomic EPO-DNA sequence, of the plasmid pEPO148, o which contains a 3,5 kB BstEII/EcoRI fragment (including the exons 1-5) of the human EPO gene sequence under the control of the SV40 promoter (Jacobs et al., Nature 313 (1985), 806 and Lee-Huang et al., Gene 128 (1993), 227) as template with the primers Ex4 (SEQ ID No. 9) and Exl7 (SEQ ID No. 17) (Table The plasmid p187 was thus obtained (Fig. 4b).

2.3 The plasmid p189 was prepared from the plasmid p187 by insertion of the herpes simplex virus thymidinekinase gene (HSV-TK) which originated from Psvtk-1 (PvuII/NarI fragment) (Fig. 4c). The HSV-TK gene is under control of the promoter at the 3' end of intron 1 (Eco RV/ClaI) in an opposite orientation relative to :the CMV promoter and should serve for negative selection for a homologous recombination.

2.4 For the construction of plasmid p190, an SfiI/BglI fragment ofpHEAVY, a plasmid which contains the cDNA of an arginine mutant of DHFR described in Simonsen et al.

(Proc. Natl. Acad. Sci, USA 80 (1983), 2495) was subcloned into the plasmid pGenak-1 cut with SfiI and BgUlI, which contains the NEO gene under control of the RSV promoter and the late SV40 polyadenylation site as terminator, the murine DHFR gene under control of the early SV40 promoter and of the early SV40 polyadenylation site as terminator (Kaufiann et al., Mol. Cell. Biol. 2 (1982), 1304; Okayama et al., Mol. Cell. Biol. 3 (1983), 280, and Schimke, J. Biol. Chem. 263 (1988), 5989) and the -11- CMV promoter (Boshart et al., Cell 41 (1995) 521). Then an Hpal fragment which contained the cDNA coding for the DHFR arginine mutant was ligated into the plasmid p189 cut with Hpal, whereupon the plasmid p190 was obtained (Fig. 4d).

To obtain a transfection vector without the HSV-TK gene, an AscI/NheI fragment of the plasmid p 190, which contained the gene activation sequence, was ligated into an AscI/NheI fragment, containing the exon 1, of the plasmid p187. The resulting plasmid was named p192 (Fig. 4e).

Example 3 Transfection of Cells Various cell lines were selected for the production of EPO and transfected with targeting vectors.

3.1 Namalwa Cells *o The cells were cultured in T150 tissue culture bottles and transfected by electroporation (1 x 107 cells/800 gl of electroporation buffer 20 mM Hepes, 138 mM NaCI, 5 mM KC1, 0.7 mM Na 2

HPO

4 6 mM D-glucose monohydrate pH 7.0, 10 pg linearized DNA, 960 gF, 260 V BioRad Gene Pulser). After the electroporation the cells were cultured in RPMI 1640, of fetal calf serum (FCS), 2 mM of L-glutamine, 1 mM of sodium pyruvate in forty-six 96-well plates. After two days the cells were cultured for 10 to 20 days in 1 mg/mi of medium containing G-418. The supemate was tested in a solid-phase ELISA for the production of EPO (see Example The EPO producing clones were expanded in 24-well plates and T-25 tissue culture bottles. Aliquots were frozen and the cells were subcloned by FACS (Ventage, Becton Dickinson). The subclones were repeatedly tested for EPO production.

3.2 HT 1080 Cells The conditions were as described for Namalwa cells, except that the HT1080 cells were cultivated in DMEM, 10% FCS, 2 mM of L-glutamine, 1 mM of sodium pyruvate. For transfection by electroporation the cells were released from the walls of the culture vessels by trypsinization. After electroporation I x 10' cells were cultured in DMEM, 10% (v/v) FCS, 2 mM L-glutamine, 1 mM sodium pyruvate in five 96-well plates.

3.3 HeLa S3 Cells The conditions were as described for Namalwa cells, except that the HeLa cells were cultured in RPM 1640, 10% FCS, 2 mM L-glutamine, 1% MEM nonessential amino acids (Sigma), and 1 mM sodium pyruvate. For transfection by electroporation the -12cells were released from the walls of the culture vessels by trypsinization. The conditions for the electroporation were 960 jiF/250 V. After the electroporation the cells were cultured in RPMI 1640, 10% FCS, 2 mM L-glutamine, 1% MEM, 1 mM sodium pyruvate in T75 tissue culture bottles. 24 hours after electroporation the cells were trypsinized and cultured for 10 to 15 days in a medium containing 600 pg/ml of G-418 in ten 96-well plates.

Example 4 Selection for EPO-Producing Clones The culture supemate of transfected cells was tested in an EPO ELISA. All steps were performed at room temperature. 96-well plates previously coated with streptavidin were coated with biotinylated anti-EPO antibodies (Boehringer Mannheim). For coating, the plates were first washed with 50 mM of sodium phosphate pH 7.2, and 0.05% of Tween 20. Then 0.01 ml of coating buffer (4 gg/ml of biotinylated antibody, 10 mM sodium phosphate pH 7.2, 3 g/1 bovine serum albumin, 20 g/1 saccharose, and 9 g/1 NaCI were added per well, and incubated at room temperature for 3 h. Then the plates were washed with 50 mM of sodium phosphate pH 7.2, dried and sealed.

Before the test, after washing the plates three times with0.3 ml of phosphate-buffered salt solution (PBS) and 0.05% Tween20 (Sigma); the plates were incubated overnight with 0.2 ml PBS and 1% Crotein (Boehringer Mannheim) per well in order to block unspecific bonds.

After removal of the blocking solution, 0.1 ml of culture supernate was added and the plates were incubated overnight. The individual wells were washed three times with 0.3 ml of PBS and 0.05% Tween 20 each time. Then 100 pl of peroxidase (POD) conjugated monoclonal antibody (Boehringer Mannheim, 150 mU/ml) was added for two hours. The wells were then again washed three times with 0.3 ml of PBS and 0.05% Tween 20 each time. Then the peroxidase reaction was performed using ABTS' as substrate in a Perkin Elmer Photometer at 405 nm. A standard calibration curve using recombinant EPO from CHO cells (Boehringer Mannheim, 100-1000 pg/well) was used to calculate the EPO concentrations.

Example 5 EPO Gene Amplification To increase the EPO expression the EPO-producing clones were cultured in the presence of increasing concentrations (100 pM 1000 nM) of methotrexate (MTX). The clones were tested at each MTX concentration by an ELISA (see Example 4) for the production of EPO.

Strong producers were subcloned by limited dilution.

Example 6 Signal Sequence Mutations -13- To optimize the leader sequence of the EPO molecule, the first amino acids coded by exon 1 were replaced. Primers with different sequences (SEQ ID No. 4-17; the 3' primer contained a Celll site for the selection of modified sequences) were used in order to obtain as template an Ascl/Xbal fragment by PCR using plasmid pEP0227 which contains a 4 kb Hindl/EcoRI fragment (including exons 1-5) of the human EPO gene sequence under control of the SV40 promoter (Jacobs et al., Nature 313 (1985), 806; Lee-Huang et al., Gene 128 (1993), 227). The resulting fragments were then cloned into the plasmid pEP0148 (Example and the plasmids pEPO 182, 183, 184 and 185 were obtained (Figure The EPO gene expression was driven by a SV40 promoter. COS-7 cells were transfected transiently with the constructs (DEAE dextran method) and the cells were tested 48 h after transfection for EPO production.

The mutated leader sequence Met-Ser-Ala-His obtained in this manner with the best EPO expression was used for constructing the gene targeting vectors (cf. Example 2.2).

Example 7 Characterization of EPO-Producing Cell Lines Three different cell lines (Namalwa, HeLa S3 and HT 1080) were selected for the EPO gene activation. EPO-producing clones were obtained by transfection with the plasmids p1 7 9, S p187, p189, p-190 or p192 (cf. Examples 2 and 3).

Approximately 160,000 NEO-resistant clones were tested for EPO production, of which 12 to 15 secreted EPO reproducibly in significant yield into the cell supemate.

Of these a total of 7 EPO clones were surprisingly identified which produced without gene amplification, by MTX, EPO in sufficient amounts for a large industrial production. The EPO production of these clones ranged from 200 ng/ml to more than 1000 cells/24h. An example of one such cell is the clone "Aladin" deposited at the DSMZ (DSM ACC 2320), which was obtained from a Namalwa cell.

After gene amplification with 500 nM of MTX the EPO production of the identified EPO clones was increased to more than 3000 ng/ml/10 6 cells/24h. An additional increase of the MTX concentration to 1000 nM led to a production of up to more than 7000 ng/ml/10 6 cells/24h.

The clones obtained also showed EPO production under serum-free culture conditions.

Example 8 Characterization of the Genome of EPO-Producing Clones 8.1 Methodology Human genomic DNA was isolated from about 108 cells and quantified (Sambrook et al., 14- 1989). After cleaving the genomic DNA with restriction enzymes, Agel and AscI, or BamHI, Hind III and Sall, respectively, the DNA fragments were separated according to their size by agarose gel electrophoresis and finally transferred to a nylon membrane and immobilized.

The immobilized DNA was hybridized with digoxigenin-marked EPO probes or gene activation sequence-specific DNA probes (DIG DNA Labeling Kit, Boehringer Mannheim) and washed under stringent conditions. The specific hybridization signals were detected by means of a chemiluminescence process using radiation-sensitive films.

8.2 Results The treatment of cells with 500 nM of MTX led to an increase of the hybridization signal at the EPO locus by a factor of 5 to 10. Upon an additional increase to 1000 nM MTX an increase by a factor >10 was obtained (Figure 6a).

In the case of hybridization with the EPO specific probe, the copies of chromosome 7 that were unaffected by homologous recombination were also detected. As it can be seen in Figure 6b, these likewise hybridizing DNA fragments have a different, clearly distinguishable size and are not changed in their signal strength by the use of MTX.

Example 9 Purification of EPO from Culture Supernates of Human Cell Lines (HeLa S3; Namalwa and HT1080) For the purification of EPO from cell culture supernates of human cell lines, basically two methods were used, which differ in number and principle of the chromatography steps and were used depending on the composition of the medium and the EPO concentration: Method 1: 1st step: blue sepharose column 2nd step: butyl-sepharose column 3rd step: hydroxyapatite column 4th step: reconcentration Method 2: 1st step: blue sepharose column 2nd step: hydroxyapatite column 3rd step: reconcentration (alternative 3rd step: RP-HPLC) Example of purification of an HeLaS3 cell culture supernate with 2% fetal calf serum (FCS) by method 1: 1_ Blue Sepharose Column: A 5 ml Hi-Trap-Blue column (Pharmacia's blue sepharose ready-to-use column) was equilibrated with at least 5 column-volumes (SV's) of buffer A (20 mM Tris-HC1, pH 7.0; mM CaCl 2 100 mM NaCI). Then 70 ml of HeLa cell supernate (containing approx. 245 Ag EPO and 70-100 mg total protein) was drawn up overnight at a flow of 0.5 ml/min by the circulatory method.

The column was washed with at least 5 SV's of buffer B (20 mM ofTris-HCl, pH 7.0; mM CaClI, 250 mM NaCI) and at least 5 SV's of buffer C (20 mM Tris-HC1, pH 7.0; 0.2 mM CaCl 2 250 mM NaC1) at 0.5 ml/min. The success of the washing was followed by measuring the protein content at OD280.

The elution of EPO was performed with buffer D (100 mM Tris-HC1, pH 7.0; 0.2 mM CaCIl; 2 M NaC1) at a flow of 0.5 ml/min. The elution solution was collected in 1 2 ml fractions.

The EPO content of the fractions, wash solutions and the flow through was determined by reverse phase (RP)-HPLC by applying an aliquot to a POROS R2/H column (Boehringer Mannheim). Alternatively, an immunological dot blot was performed for the qualitative identification of fractions containing EPO.

Fractions containing EPO (8-12 ml) were pooled and applied to a butyl-sepharose column.

The yield after the blue sepharose column was about 175 Itg EPO (corresponds to about In general the yield after blue sepharose was between 50 and 2. Butvl Sepharose Column (Hydrophobic Interaction Chromatographv) A self-made 2-3 ml butyl sepharose column (material: Toyopearl Butyl S650) was equilibrated with at least 5 SV's of buffer D (100 mM Tris-HC1, pH 7.0; 0.2 mM CaCl 2 and 2 M NaCI) and then the blue sepharose pool containing EPO from 1. (approx. 150 Ag EPO) was drawn up at a flow of 0.5 ml/min.

The column was washed at 0.5 ml/min with at least 5 SV's of buffer E (20 mM Tris-HC1, pH 7.0; 2 M NaCl and 10% isopropanol). The washing success was followed by measuring the protein content at OD280.

The elution of EPO was performed with buffer F (20 mMTris-HC1, pH 7.0; 2 M NaCI and isopropanol) at a flow of 0.5 ml/min. The elution solution was cole.cted in 1 2 ml fractions.

-16- The EPO content of the fractions, the wash solutions and the flow-through was determined by RP-HPLC by applying an aliquot to a POROS R2/H column. Alternatively, an immunological dot blot was performed for the qualitative identification of EPO-containing fractions.

Fractions containing EPO (10 15 ml) were pooled and applied to a hydroxyapatite column.

The yield of the butyl sepharose column was about 130 pg EPO (corresponds to about In general, the yield of the butyl sepharose was between 60 and 85% of the amount applied from the blue sepharose pool.

3. Hvdroxvapatite column A 5 ml hydroxyapatite column (Econo-Pac CHT II ready-to-use column from BioRAD)was .equilibrated with at least 5 SV's of buffer F (20 mM Tris-HC1, pH 7.0; 2 M NaCI; isopropanol) and then the butyl sepharose pool containing EPO from 2. (approx. 125 gg EPO) was drawn up at a flow of 0.5 ml/min.

The column was washed with at least 5 SV's of buffer G (20 mM Tris-HC1, pH 7.0; 2 M NaC1) at 0.5 ml/min. The washing success was followed by measuring the protein content at OD280.

*i The elution of EPO was performed with buffer H (10 mM sodium phosphate, pH 7.0; mM NaCI) at a flow of 0.5 ml/min. The elution solution was collected in 1-2 ml fractions.

The EPO content of the fractions, wash solutions and flow-through was determined through RP-HPLC by applying an aliquot to a POROS R2/H column.

Fractions containing EPO (3-6 ml) were pooled. The yield of the hydroxyapatite column was about 80 gg of EPO (corresponds to about In general, the yield of the hydroxyapatite column amounted to between 50 and 65% of the butyl sepharose pool applied.

4_ Reconcentration The pooled EPO fractions from the hydroxyapatite step were raised to a concentration of 0.1 0.5 mg/ml in centrifuging units with a exclusion size of 10 kD Microsep by Filtron), 0.01% of Tween 20 was added and stocked in aliquots at Table of Yields EPO (gg) Yield Start 245 100 Blue sepharose 175 Butyl sepharose column 130 53 Hydroxyapatite column 80 33 Reconcentration 60 The purity of the isolated EPO was approximately and generally even To increase the EPO yield, Method 2 was also used, in which the butyl sepharose step was missing. This method is applicable especially in the case of cell culture supernates without or with 1% FCS addition, and yields isolated EPO of approximately equal purity The presence of 5 mM CaCI 2 in the equilibration buffer (buffer F) for the hydroxyapatite column led in this method to improved binding and thus also to a reproducible elution behaviour of EPO in the hydroxyapatite step. Therefore Method 2 was practiced with basically the same procedure as in Method 1, with the following buffers: 1_ Blue Sepharose Column: r r Equilibration buffer (buffer A) Washing buffer 1 (buffer B) Washing buffer 2 (buffer C) Elution buffer (buffer D) 2_ Hydroxyapatite Column: Equilibration buffer (buffer F): Washing buffer (buffer G): Elution buffer (buffer H): 20 mM Tris-HC1, pH mM CaCI,; 100 mM NaCI 20 mM Tris-HCI, pH mM CaCI,; 250 mM NaCI 20 mM Tris-HC1, pH mM CaC12, 250 mM NaCI 100 mM Tris-HC1, pH mM CaClz; 2 M NaCl 50 mM Tris-HC1, pH mM CaCI 2 1 M NaCI 10 mM Tris-HC1, pH mM CaCIl; 80 mM NaCI 10 mM sodium phosphate, pH mM CaC1 2 80 mM NaCI Table of Yields: EPO (Ig) Yield Start 600 100 Blue Sepharose 450 Hydroxyapatite column 335 Reconcentration 310 52 The addition of 5 mM CaC1 2 to the buffers B to G in Method 1 also resulted in better binding and more defined elution of the hydroxyapatite column.

Example 10 Determination of the Specific Activity in vivo of EPO from Human Cell Lines (Bioassay on the Normocythemic Mouse) The dose-related activity of EPO on the proliferation and differentiation of erythrocyte precursor cells was determined in vivo in mice on the basis of the increase of reticulocytes in the blood after EPO administration.

For this purpose eight mice were treated parenterally with different doses of the EPO sample and to be analyzed and of an EPO standard (balanced against the WHO's standard EPO).

The mice were then kept under constant, defined conditions. 4 days after the EPO treatment blood was taken from the mice and the reticulocytes stained with acridine orange.

Determination of the number of reticulocytes per 30,000 erythrocytes was performed by microfluorimetry in the flow cytometer by analyzing the red fluorescence histogram.

The computation of the biological activity was made from the values of the reticulocyte counts of the specimen and those of the standard at the various doses by the method described by Linder of content determination in pairs with parallel straight lines Linder, Planen und Auswerten von Versuchen, 3rd ed., 1969, Birkenhiuser Verlag Basel).

Result: EPO from cell line Specific Activity U/mg HeLa S3 (Sample 1) 100,000 HeLa S3 (Sample 2) 110,000 P.'OPERFjs'iX97.9-pc doc2A'4l' Il The reference to any prior art in this specification is not, and should not be taken as, an acknowledgment or any form of suggestion that that prior art forms part of the common general knowledge in Australia.

Throughout this specification and the claims which follow, unless the context requires otherwise, the word "comprise", and variations such as "comprises" and "comprising", will be understood to imply the inclusion of a stated integer or step or group of integers or steps but not the exclusion of any other integer or step or group of integers or steps.

S* *o -19A- SEQUENCE

LISTING

GENERAL INFORMATION:

APPLICANT:

NAME: Boehringer Mannheim GmbH ADDRESS: Sandhofer Str. 112-132 CITY: Mannheim COUNTRY: Germany POSTAL ZONE NUMBER: 68305 (ii) TITLE OF THE INVENTION: Preparation of erythropoietin by endogenous gene activation (iii) NUMBER OF SEQUENCES: 17 (iv) COMPUTER-READABLE VERSION: MEDIUM: Floppy Disk COMPUTER: IBM PC compatible OPERATING SYSTEM: PC-DOS/MS-DOS SOFTWARE: Patent In Release Version #1.30 (EPA) INFORMATION ON SEQ ID NO.: 1 SEQUENCE CHARACTERISTICS: LENGTH: 32 base pairs KIND: Nucleotide STRAND FORM: Single strand TOPOLOGY: linear (xi) DESCRIPTION OF SEQUENCE: SEQ ID No. 1: CGCGGCGGAT CCCAGGGAGC TGGGTTGACC GG 32 INFORMATION ON SEQ ID NO.: 2 SEQUENCE CHARACTERISTICS: LENGTH: 38 base pairs KIND: Nucleotide STRAND FORM: Single strand TOPOLOGY: linear (xi) DESCRIPTION OF SEQUENCE: SEQ ID NO. 2: GGCCGCGAAT TCTCCGCGCC TGGCCGGGGT CCCTCAGC 38 INFORMATION ON SEQ ID NO. 3 SEQUENCE

CHARACTERISTICS:

LENGTH: 36 base pairs KIND: nucleotide STRAND FORM: single strand TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO.: 3 CGCGGCGGAT CCTCTCCTCC CTCCCAAGCT GCAATC 36 INFORMATION ON SEQ ID NO. 4 SEQUENCE CHARACTERISTICS: LENGTH: 36 base pairs KIND: Nucleotide STRAND FORM: single strand TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO.: 4 GGCCGCG AAT TCTAGAACAG ATAGCCAGGC TGAGAG 36 INFORMATION ON SEQ ID NO. SEQUENCE

CHARACTERISTICS:

LENGTH: 24 base pairs KIND: nucleotide STRAND FORM: single strand TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO.: TCACCCGGCG CGCCCCAGGT CGCT 24 INFORMATION ON SEQ ID NO. 6 SEQUENCE

CHARACTERISTICS:

LENGTH: 61 base pairs KIND: Nucleotide STRAND FORM: Single strand TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID, NO.: 6 ATGCTCGAGC GGCCGCCAGT GTGATGGATA TCTGCAGAGC TCAGCTTGCC CGCGAATTCT A 61 INFORMATION ON SEQ ID NO. 7 SEQUENCE

CHARACTERISTICS:

LENGTH: 78 base pairs KIND: Nucleotide STR.AND FORM: Single strand TOPOLOGY: Linear

FEATURE:

NAME/KEY: CDS POSITION: 49-.60 (xi) SEQUENCE DESCRIPTION: SEQ ID NO.: 7 TCACCCGGCG CGCCCCAGGT CGCTGAGGGA CCCCGGCCAG GCGCGGAG ATG Goo GCC 57 Me Oly Ala CAC GGTGAGTACT CGCQvoCCT 78 His -22- INFORMATION ON SEQ ID NO. 8 SEQUENCE CHARACTERISTICS: LENGTH: 4 amino acids KIND: amino acid TOPOLOGY: linear 0ii) KIND OF MOLECULE: Protein (Xi) SEQUENCE DESCRIPTION: SEQ ID 8 Met Gly Ala His 1 INFORMATION ON SEQ ID NO. 9: SEQUENCE CHARACTERISTICS: LENGTH: 78 base pairs KIND: Nucleotide STRAND FORM: Single strand TOPOLOGY: linear owe*S :6 (ix) FEATURE: NAME/KEY: CDS POSITION: 49-.60 .00 0 (Xi) SEQUENCE DESCRIPTION: SEQ ID NO.: 9 TCACCCGGCG CGCCCCAGGT CGCTGAGGGA CCCCGGCCAG GCOCGGAG ATG AGC GCC 57 *Met Ser Ala es CAC GGTG;AGTACT CGCGGGCT 78 is 23- IINFORMATION ON SEQ D NO. SEQUENCE CHARACTERISTICS: LENGTH: 4 amino acids KIND: amino acid TOPOLOGY: linear (ii) KIND OF MOLECULE: Protein (xi) SEQUENCE DESCRIPTION: SEQ ID NO.: Met Ser Ala His 1 INFORMATION ON SEQ ID NO. 11 SEQUENCE CHARACTERISTICS: LENGTH: 78 base pairs KIND: nucleotide STRAND FORM: Single strand TOPOLOGY: linear

FEATURE:

NAMIE/KEY: CDS POSITON: 49.. (xi) SEQUEN4CE DESCR[PTION: SEQ ID NO.: 11 TCACCCGGCG CGCCCCAGGT CGCTGAGGGA CCCCGGCCAG GCGCGGAG ATG GGG GTG 57 Met Gly Val *CCC GGTGAGTACT CGCGCGC-r 78 24 IINFORIMATION ON SEQ ID NO. 12 SEQUENCE CHARACTERISTICS: LENGTH: 4 amino acids KIND: amino acid TOPOLOGY: linear (ii) KIND OF MOLECULE: Protein (xi) SEQUENCE DESCRIPTION: SEQ ID No. 12 Met Gly Val Pro

I

INFORMATION ON SEQ ID NO. 13 SEQUENCE CHARACTERISTICS: LENGTH: 78 base pairs KID,,: Nucleotide STRAND FORM: Single strand (ix) TOPOLOGY: Linear (ix) FEATURE: NAME/IKEY: CDS POSITION: 49. (Xi) SEQUENCE DESCRIPTION: SEQ ID NO.: 13 TCACCCGGCG CGCCCCAGGT CO;CTGAGGGA CCCCGGCCAG GCGCGGAG ATO AGC GTO 57 Met Scr Val CAC GGTGAGTACT CGCGGGCT 78 His INFORMATION ON SEQ ID NO. 14 SEQUENCE CHARACTERISTICS: LENGTH: 4 amino acids KIND: amino acid TOPOLOGY: linear (ii) KIND OF MOLECULE: Protein (xi) SEQUENCE DESCRIPTION: SEQ ID NO.: 14 Met Ser Val His

I

INFORMATION ON SEQ ID NO. SEQUENCE CHARACTERISTICS: LENGTH: 59 base pairs KIND: Nucleotide STRAND FORM: single strand TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO. S. GGACATTCTA GAACAGATAT CCAGGCTGAG CGTCAGGCGG GGAGGGAGAA GGGTGGCTG 59 INFORMATION ON SEQ ID NO. 16 SEQUENCE CHARACTERISTICS: LENGTH: 48 base pairs S. KIND: nucleotide STRAND FORM: single strand TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO.: 16 GTGATGGATA TCTCTAGAAC AGATAGCCAG GCTGAGAGTC AGGCGGGG 48 -26- INFORMATION ON SEQ ID NO. 17 SEQUENCE CHARACTERISTICS: LENGTH: 39 Base pairs KIND: nucleotide STRAND FORM: single strand TOPOLOGY: linear (xi) SEQUENCE DESCRIPTION: SEQ ID NO.: 17: ATGGATATCA-TCGATTCTAG AACAGATAGC CAGGCTGAG 39

Claims (50)

1. An isolated human cell, characterized in that it contains a copy of an endogenic EPO gene in operative linkage with a heterologous promoter active in the human cell and is capable of the production of at least 200 ng 6 cells/24h. obtained by a method according to any one of claims 13 to 22.

2. An isolated human cell, according to claim 1, characterized in that it is capable of the production of 200 3000 ng EPO/10 6 cells/24h. characterized in that 15 it contains several copies of an endogenic EPO gene, each in operative linkage with a heterologous promoter active in the human cell and is capable of producing at least .:1000 ng EPO/10 6 cells/24h.

4. An isolated human cell according to claim 3, characterized in that 20 it is capable of the production of 1000 25,000 ng EPO/10 6 cells/24h.

5. An isolated human cell according to any one of claims 1 to 4, characterized in that it is an immortalized cell.

6. An isolated human cell according to any one of claims I to characterized in that that it can be cultured in serum-free medium.

7. An isolated human cell according to any one of claim 1 to 6, characterized in that it is selected from an HT 1080 cell, an HeLa S3 cell, a Namalwa cell or a cell derived therefrom.

8. An isolated human cell according to any one of claims I to 7, characterized in that the activated endogenic EPO gene is under control of a viral promoter, especially a CMV promoter. P \OPER\hxS517210-49ctmImnnd doc-16/1203 29

9. An isolated human cell according to any one of claims 1 to 8, characterized in that the EPO gene has a signal peptide-coding sequence which is different from the natural signal peptide-coding sequence. An isolated human cell according to claim 9, characterized in that the EPO gene, which is in operative linkage with the heterologous expression control sequence, has a signal peptide-coding sequence which codes for a signal peptide sequence modified in the region of the 4 first amino acids, which is selected from: Met-X,-X 2 -X3 wherein X, is Gly or Ser, X, is Ala, Val, Leu, Ile, Ser or Pro, and X 3 is Pro, Arg, Cys or His, provided that X,-X 2 -X 3 is not the sequence Gly-Val-His.

11. An isolated human cell according to claim characterized in that the first 4 amino acids are selected from: Met-Gly-Ala-His, Met-Ser-Ala-His, 15 Met-Gly-Val-Pro or Met-Ser-Val-His. 00

12. An isolated human cell according to claim 11, characterized in that the sequence of the first 4 amino acids of the signal peptide is Met-Ser-Ala-His. 20 13. Method for the preparation of a human EPO-producing cell characterised in that it I" contains a copy of an endogenic EPO gene in operative linkage with a heterologous promoter active in the human cell and is capable of the production of at least 200ng EPO/106 cells/24h comprising the steps: Provision of human starting cells which contain at least one copy of an endogenic EPO gene, Transfection of the cells with a DNA construct comprising: two flanking DNA sequences which are homologous to regions of the human EPO gene locus, to permit a homologous recombination, (ii) a positive selection marker gene and (iii) a heterologous expression control sequence which is active in the human ceU, Culturing the transfected cells under conditions in which a selection takes place for the presence of the positive selection marker gene, Analyzing the cells selectable according to step and Identification of the EPO-producing cells.

14. Method according to claim 13, characterized in that the homologous DNA sequences are selected from the regions of the sequences, exon I and intron I of the EPO gene. Method according to claim 14, characterized in that a sequence modified in the region of exon 1 is used.

16. Method according to claims 14 or characterized in that the neomycin phosphotransferase gene is used as selection marker gene.

17. Method according to any one of claims 14 to 16, characterized in that the DNA construct additionally comprises an amplification gene.

18. Method according to claim 17, characterized in that a dihydrofolate reductase gene is used as amplification gene.

19. Method according to claim 18, characterized in that the gene of a dihydrofolate reductase arginine-mutant is used as amplification gene. Method according to any one of claims 17 to 19, furthermore comprising the steps: amplification of the EPO gene and obtaining of EPO-producing cells which contain several copies of an endogenic EPO gene each in operative linkage with a heterologous expression control sequence.

21. Method according to any one of claims 13 to characterized in that a CMV promoter/enhancer is used as expression control sequence.

22. Method according to any one of claims 13 to 21, characterized in that the plasmid p189 (DSM 11661) or a plasmid derived therefrom is used in linearized form as DNA construct.

23. DNA construct for activating an endogenic EPO gene in a human cell, comprising: two flanking DNA sequences which are homologous to regions of the human EPO gene locus selected from 5'-untranslated sequences, exon 1 and intron 1, in order to permit a homologous recombination, a modified sequence being present in the region of exon 1, coding for the amino acids: Met-X,-X 2 -X 3 wherein X, is Gly or Ser, X 2 is Ala, Val, Leu, Ile, Ser or Pro, and X 3 is Pro, Arg, Cys or His, provided that X,-Xz-X 3 is not the sequence Gly-Val-His, (ii) a positive selection marker gene, (iii) a heterologous expression control sequence that is active in a human cell, and (iv) if desired, an amplification gene.

24. DNA construct according to claim 23, characterized in that that the amino acids are selected from Met-Gly-Ala-His, Met-Ser-Ala-His, Met-Gly-Val-Pro or Met-Ser-Val-His.

25. DNA construct for activating an endogenic EPO gene in a human cell, comprising: two flanking DNA sequences which are homologous with regions of the human EPO gene locus selected from 5'-untranslated sequences, exon I and intron 1, in order to permit a homologous recombination, S(ii) a positive selection marker gene, (iii) a heterologous expression control sequence which is active in a human cell, the distance between the heterologous expression control sequence and the translation start of the EPO gene being not greater than 1100 bp, and (iv) an amplification gene if desired.

26. Plasmid p189 or a plasmid derived therefrom.

27. Method for preparing human EPO, characterized in that a human cell according to any one of claims 1 to 12 is cultured in a suitable medium under conditions in which a production of EPO occurs and the EPO is harvested from the culture medium.

28. Method according to claim 27, characterized in that a serum-free medium is used. -31--

29. Method according to either one of claims 27 or 28, characterized in that the cells are cultured in suspension. Method according to any one of claims 27 to 29, characterized in that the culturing is performed in a fermenter.

31. Method according to claim characterized in that the volume of the fermenter is 10 1 50,000 1.

32. Method according to any one of claims 27 to 31, characterized in that the harvesting of the EPO from the culture medium comprises the steps: Passing the cell supemate over an affinity chromatography medium and harvesting the fractions containing EPO, if desired, passing the fractions containing the EPO over a hydrophobic interaction chromatography medium and harvesting the fractions containing EPO, passing the fractions containing EPO over hydroxyapatite and harvesting the fractions containing EPO, and concentrating or/and passing over a reverse-phase HPLC medium. S 33. Method according to claim 32, characterized in that a blue sepharose medium is used in step

34. Method according to either one of claims 32 or 33, characterized in that a butyl sepharose medium is used in step -Method according to any one of claims 32 to 34, characterized in that the concentration is performed by exclusion chromatography.

36. Method according to claim characterized in that a medium with an exclusion size of 10 kD is used.

37. Method according to any one of claims 27 to 36, 32 characterized in that a human EPO is obtained with a purity of at least

38. Method according to any one of claims 27 to 37, characterized in that a human EPO is obtained with a specific activity in vivo (normocythemic mouse) of at least 100,000 IU/mg.

39. Method according to claim 38, characterized in that a human EPO is obtained with a specific activity in vivo (normocythemic mouse) of at least 175,000 IU/mg to 450,000 IU/mg. Method according to any one of claims 27 to 39, characterized in that a human EPO is obtained with a content of less than 0.2% N-glycolneuraminic acid with respect to the content ofN-acetyl neuraminic acid.

41. Method according to any one of claims 27 to characterized in that a human EPO is obtained with a-2,3-linked sialic acid residues.

42. Method according to any one of claims 27 to 42, characterized in that a human EPO is obtained with a-2,3- and a-2,6-linked sialic, acid residues.

43. Method according to any one of claims 27 to 42, characterized in that a human EPO is obtained which comprises a polypeptide with a length of 165 amino acids.

44. Method according to any one of claims 27 to 42, characterized in that a human EPO is obtained which comprises a polypeptide with a length of 166 amino acids. Method according to any one of claims 27 to 42, characterized in that a human EPO is obtained which comprises a mixture of polypeptides with a length of 165 and 166 amino acids. -33-

46. Isolated human EPO with a specific activity in vivo of at least 100,000 U/mg protein, obtainable by the method according to any one of claims 27 to 45, which is free of urinary impurities.

47. Isolated human EPO according to claim 46, characterized in that it has a purity of at least

48. Isolated human EPO according to claim 46 or 47, characterized in that it contains less than 0.2% of N-glycol rieuraminic acid with respect to the content of N-acetyl neuraminic acid.

49. Isolated human EPO according to any one of claims 46 to 48, characterized in that it bears a-2,3-linked sialic acid residues.

50. Isolated human EPO according to any one of claims 46 to 49, characterized in that p it bears a-2,3- and a-2,6-linked sialic acid residues. S51, Isolated human EPO according to any one of claims 46 to characterized in that it comprises a polypeptide with a length of 165 amino acid residues.

52. Isolated human EPO according to any one of claims 46 to characterized in that it comprises a polypeptide with a length of 166 amino acid residues.

53. Isolated human EPO according to any one of claims 46 to characterized in that it comprises a mixture ofpolypeptides with a length of 165 and 166 amino acids.

54. Pharmaceutical preparation, characterized in that it contains a human EPO according to any one of claims 46 to 53 as active substance, together if desired with other active substances and pharmaceutically common vehicles, adjuvants or additive substances. Isolated DNA which codes for a human EPO with a sequence modified in the area of the first 4 amino acids of the signal peptide, which is selected from: .34 P 'OPER\Fs 'J)7 JX-s-p doc-)2Miutl Met-XI-X2-X3 wherein Xi is Gly or Ser, X 2 is Ala, Val, Leu, lie, Ser or Pro, and X 3 is Pro, Arg, Cys or His, on the condition that XI-X 2 -X 3 is not the sequence Gly-Val-His.

56. Isolated DNA according to claim characterized in that the amino acid sequence is chosen from: Met-Gly-Ala-His, Met-Ser-Ala-His Met-Gly-Val-Pro or Met-Ser-Val-His.

57. DNA according to claim 55 or 56, characterized in that it is a genomic DNA.

58. DNA according to claim 55 or 56, characterized in that it is a cDNA.

59. An isolated human cell according to any one of claims I to 12 substantially as hereinbefore described with reference to the figures and/or examples. A method according to any one of claims 13 to 22 and 27 to 45 substantially as hereinbefore described with reference to the figures and/or examples.

61. A DNA construct according to any one of claims 23 to 25 substantially as hereinbefore described with reference to the figures and/or examples. P \OPER\FasU\2532 11 v.doc-.27 3M12

62. Isolated human EPO according to any one of claims 46 to 53 substantially as hereinbefore described with reference to the figures and/or examples.

63. An isolated DNA according to any one of claims 55 to 58 substantially as hereinbefore described with reference to the figures and/or examples. DATED this 2 8 th day of March 2002 ROCHE DIAGNOSTICS GMBH by DAVIES COLLISON CAVE Patent Attorneys for the Applicants ee -36-