CA2670003A1 - Genetic ablation of the prp gene cells using a targeted promoter trap strategy for production of serum-free recombinant proteins as therapeuticals - Google Patents

Abstract

The present invention provides a prion protein (PrP)-free immortalized somatic human cell line wherein both alleles of the PrP gene have been completely deleted by homologous recombination. The invention further provides a method for the production of said cell line and its use for producing human recombinant proteins which are suitable as biopharmaceuticals.

Description

Ihr Zeichen: Pat 5/ Unser Zeichen: 072819W0 CS/gn - Entwurf vom 07. November Genetic ablation of the PRP gene cells using a taraeted promoter trap strategy for production of serum-free recombinant proteins as therapeuticals The present invention provides a prion protein (PrP)-free immortalized somatic human cell line wherein both alleles of the PrP gene have been completely deleted. The invention further provides a method for the production of said cell line and its use for producing human target proteins which are suitable as biopharmaceuticals.

Introduction Prions are infectious pathogens that cause central nervous system spongiform encephalopathies in animals. In contrast to viruses and viroids, prions are fully devoid of nucleic acid and are resistant to proteases. The infectious particle has been identified as PrPs`, an isoform of PrP`, the normal cellular prion pro-tein. The prion hypothesis (Prusiner, Proc. Natl. Acad. Sci. 95, 13363-13383, (1988)) proposes that the PrPs` molecule itself converts PrP` to the abnormal conformation, either through a process of heterodimerisation or through nu-clear polymerisation.

The most common prion diseases of animals are Scrapie in sheep and goats and bovine spongiform encephalopathy (BSE) in cattle. In humans, four prion diseases have been identified: (1) Kuru, (2) Creutzfeldt-Jakob disease (CJD), (3) Gerstmann-Straussler-Scheinker disease (GSS) and (4) fatal familial in-somnia (FFI). The human prion diseases may have sporadic, genetic or infec-tious origin.

The first infectious prion disease described was Scrapie as a disease of sheep and goats over 250 years ago. Scrapie was demonstrated to be experimentally transmissible 50 years ago. There is no evidence that scrapie has ever been transmitted to man. BSE was first described in UK cattle in 1985 (Wells et al., Veterinary record 121, 419-420, (1987)) and is thought to have spread through oral consumption of ruminant-derived meat and bone meal (Wilesmith et al., Veterinary Record 123, 638-644, (1988)). The disease spread widely, peaking in 1992 with over 180,000 clinical cases in the UK, although mathe-matical estimates suggest that 1-2 million cattle could have been infected but slaughtered and entered the human food chain before they were old enough to demonstrate evidence of clinical disease (Anderson et al., Nature 382, 779-788, (1996)). BSE has crossed into up to 20 other species, including domestic and exotic cats (Wyatt et al., Veterinary Record 129:233-236, (1991); Kirk-wood and Cunningham, Veterinary Record 135, 296-303 (1994)) and exotic ungulates in British zoos. In July 1988, the spread of BSE led the UK Go-vernment to restrict the use of ruminant-derived meat and bone meals as an animal feed, and in November 1989 to specify that bovine offal were banned for human consumption.

The transmission of human prion diseases was first reported in the Fore people of Papua New Guinea in the late 1950s (Gajdusek & Zigas, New England Jour-nal of Medicine 257, 974-978 (1957)) and is thought to have been transmitted during ritualistic cannibalism and sacrificial funeral rites. Iatrogenic transmis-sion of CJD has been well documented by direct inoculation of the CNS
through contaminated medical instruments and grafts. Iatrogenic transmission has also occurred via cadaverous human pituitary growth hormone and go-nadotropins administered by intramuscular injection (Buchanan et al., British Medical Journal 302, 824-824, (1991); Brown et al., Transfusion 38:810-816, (1992)).

Variant CJD (vCJD) is a human prion disease apparently resulting from expo-sure to the bovine spongiform encephalopathy (BSE) agent. VCJD was first described 10 years ago (Will et al., Lancet 347:921-5, (1996)) as a result of systematic monitoring of the incidence and clinical phenotype of CJD in the UK.

For vCJD, in contrast to the other human prion diseases, the disease-associated form of the prion protein and infectivity is readily detectable in lymphoid tissues throughout the body (Hill, A.F. et al., Lancet 353:183-9 (1999); Head, M.W. et al., Am. J. Pathol. 164:142 (2004)), even before the onset of clinical disease (Hilton, D.A. et al., Lancet 352:703-4 (1998)). This gives rise to concerns that blood and blood products may also contain infec-tious particles, representing a possible source of iatrogenic spread of variant CJ D.

This concern has been reinforced following the experimental transmission of BSE in sheep model by transfusion of blood and buffy coat from animals in the preclinical phase of the illness (Hunter et al., J. Gen. Virol. 83: 2897-905 (2002)).

Studies in animal models suggest that most prion infectivity in blood may be cell associated, with lower levels in the plasma (Brown P et al., Transfusion 38:810-816 (1998)), and there is evidence to suggest that any infectivity present may be reduced during the process of plasma fractionation (Stenland, C.J. et al., Transfusion 42:1497-500 (2002); Gregori, L. et al., Biologicals 32:1-10 (2004)). In response to the blood transfusion transmission of vCJD a deferral of donors who themselves have been recipients of blood components since 1980 has been instituted to reduce the risk of tertiary or higher-order transmissions leading to a self-sustaining outbreak. Nevertheless, the possibil-ity that plasma or blood products may transmit the disease cannot be ex-cluded, since an appropriate blood test is not available yet (Aguzzi, A. &
Glatzel M., Nat. Clin. Pract. Neurol. 2:321-329 (2006)).

The risk of transmitting prion-related disorders through human products is a serious health concern. In case of blood products, the method of prevention is on one hand the handling of vCJD donators combined with f. e. clearing the blood by filtration.

On the other hand there is the safe production of recombinant human proteins preventing the risk of prion transmission through blood transfusion by con-taminated donors. The production of human proteins in organisms such as Escherichia coli and Saccharomyces cerevisiae allows the production of many human proteins in large scale synthesis, but factors such as plasmid stability and insolubility of the desired protein product may limit the usefulness of these systems. Most human proteins require additional post-translational modifications to perform the function of the endogenous protein and thus require their synthesis in mammalian cells or even species-specific cell lines for the plasma-like functioning of the produced protein.

For example for factor VIII, the use of non-human cell lines, such as Chinese hamster ovary cells (CHO), faces critical disadvantages in that the purified therapeutic proteins are contaminated with cellular trace components causing antigenic reactions in the patients (Refacto, Wyeth, package insert).

Moreover, proteins expressed by non-human expression systems may have non-human glycosylation patterns also giving rise to antigenic reactions in the patient. Biological stability and efficacy of clotting factors is substantially influ-enced by their 0-and N-glycosylation pattern. Especially peripheral and termi-nal monosaccharides are important, because they are detected by specific receptors from cells which are responsible for their degradation. Clotting fac-tors carry sialic acid residues as terminal monosaccharides. Modification in the composition of sialic acids in the antennae of glycoproteins can result in het-erogeneous glycosylation patterns. Thus, biological stability and efficacy are crucially affected when modification occurs. Hence, evaluation of the influence of glycosylation in non-human production cell lines is an important considera-tion in the production of recombinant clotting factors. Generally spoken, hu-man cell lines are more qualified for the production of recombinant clotting factors than non-human cell lines. The reason for this is probably that no ex-traneous oligosaccharides will be incorporated into the oligosaccharide moie-ties during synthesis of recombinant proteins.

For these reasons mammalian, especially human, systems are preferable for the production of recombinant human proteins. In particular the immortalized cell line HEK293 and its derivates, f. e. FreeStyle HEK293F, are capable of expressing recombinant human proteins (EP 05 105 965.7). Since many therapeutics are produced in mammalian systems, there is a need for ensuring the safety of these products isolated from such systems regarding complete absence of prion proteins. Taking into account that on one hand no appropri-ate test for prion infectivity is available and that on the other hand prion propagation and infectivity is dependent on the expression of the normal cellu-lar prion protein, which might convert to the infectious prion protein, the most secure and promising way to get rid of prion infectivity in such systems is to completely prevent the expression of the prion protein gene in the producing systems by complete knockout of the prion gene.

Such a knockout cell line production system as the one described in this inven-tion, completely prevents expression of the prion protein and will provide re-combinant human proteins absolutely free from infectious prion proteins and free of a risk of prion contamination for the patients, who receive recombinant pharmaceuticals produced in these prion-free cells.

In order to specifically prevent expression of a given gene in a eukaryotic cell, several approaches have been developed employing genetic engineering and molecular biology technologies. Genetic ablation (also commonly known as "genetic knockout" in the jargon used by experts in the field) implies removal of the DNA sequence encoding for a particular protein. As a consequence of this removal, the resulting "knockout cell" is completely unable to express the knocked-out gene. In contrast, "interference RNA" (RNAi) technologies do not remove the DNA sequence but instead introduce an anti-sense or complemen-tary sequence, which then prevents translation of the suppressed gene. In most cases the resulting, "knocked-down" gene is only partially suppressed, allowing some residual expression level.

Because of the wide use of the mouse as a model for human disease, and because of the early availability of murine embryonic stem (ES) cells which are easily amenable to genetic and embryonic manipulations, knockout technolo-gies have been mostly developed in mouse ES cells (Joyner A.L., Oxford Univ.
Press, UK 1993). Today, the knockout technology in the mouse has developed far enough to allow not only complete knockout ("null alleles"), but also condi-tional mutants. The targeted mutation may be temporally induced or sup-pressed, e.g. by supplementing the mouse chow with tetracycline (tet-on, tet-off systems, Gossen, M. and Bujard, H. Proc. Nati. Acad. Sci. USA 89:5547--5551. (I.992)); it may be driven by a cell- or tissue-specific promoter, or by an ubiquitous but developmentally regulated promoter which may be shut off during embryonic development but quickly activated after birth.

The generation of a knockout ES cell line always requires the same initial steps: (1) generation of a targeting construct specifically designed to knock out a particular gene; (2) introduction of this construct in the ES cells and (3) selection and screening of cells bearing a targeted deletion of the desired gene.

The targeting construct must always contain a selection cassette and a signifi-cant region of homology to the gene which will be knocked out. These two elements are absolutely necessary to achieve and detect integration only in the desired locus: the region of homology to the endogenous gene triggers homologous recombination between the two homologous DNA fragments, whereas the selection cassette allows screening of cells bearing the integrated construct. The most widely used selection cassette in ES knockout technology is the neomycin phosphotransferase cassette ("neo"), consisting of the ORF
encoding this enzyme driven by a mouse phosphoglycerate kinase (pGK) pro-moter (Soriano et al., Cell 64, 693-702 (1991)) and flanked downstream by a polyA transcription termination signal. Upon introduction of the targeting con-struct in ES cells and selection with G418, only those cells where the construct has stably integrated will be resistant to the antibiotic. Among these resistant cells there will be some where the targeting construct has replaced the en-dogenous original sequence. Targeted events in isolated clones may then be screened either by genomic PCR or by genomic Southern blot analysis. In mouse ES cells, the frequency with which the targeting construct integrates in the desired region of the genome is much lower than that of random integra-tion, which makes it usually necessary to screen several hundreds of antibi-otic-resistant clones before a correctly targeted one is identified.

One possibility to increase the targeting frequency of a given construct is to increase the length of the region of homology, since there is a linear relation-ship between both parameters (Hasty et al., Mol Cell Biol. 1991 Nov;
11(11):5586-91 (1991)). Alternatively, a number of "trap-approaches" have been developed which rely on the use of elements of the target locus for effi-cient expression of the selection cassette. For example, in the promoter trap approach, the targeting vector is designed in such a way that the transcrip-tional machinery of the endogenous target gene will drive expression of the selection cassette cloned in the targeting vector. In this case the vector con-tains homologous regions to the target gene which do not have any promoter activity, and therefore most of the clones in which the integration has occurred at random cannot survive antibiotic selection. Typically promoter trap selec-tions achieve enrichments of 100-fold for targeted clones.

Using RNAi methodologies, expression of the PrP gene has been so far only transiently and partially silenced in scrapie-infected, mouse neuroblastoma cells in culture (Daude et al., J. Cell. Sci.. 2003 Jul 1;116(Pt 13):2775-9 (2003)).

Genetic targeting of the prion gene in the mouse genome has been success-fully reported using different approaches (Table 1) in order to elucidate the functional consequence of its loss-of-function and its possible relationship to TSEs.

Table 1: PrP mouse knockout data available from the literature Name of Homologous Frequency of tar-the knock- Method Antibiotic selection arms (kb) Host cells geted recombina-out tion Zurich I Conventional G418 0.3 mg/mi n/a ES AB1 1/5000 Edinburgh Conventional G418 0.2 mg/ml Left: 1.2 E14 1/800 Ganciclovir 2 pM Right: 2.4 Nagasaki Conventional G418 0.2 mg/ml Left: 3.0 11 ES 1/321 FIAU 200 nM Right: 9.0 RcmO Conventional G418 1 mg/mi n/a Neuroblastoma n/a N2a Zurich II Conditional G418 0.4 mg/ml Left: 1.4 E14.1 82/1046(PCR), Right: 3.7 5/82 As expected from the "prion protein-only theory" (Prusiner Nobel Lecture, Dec.
8, 1997), mice homozygous for the inactivated gene (i.e., fully devoid of PrP
expression) are indeed resistant to prion infection. Most remarkably however, all of the above mentioned PrP knockout mice are viable and developmentally normal. Only a relatively mild neurological phenotype has been characterized in the Nagasaki, RcmO and Zurich II knockouts (Rossi et al., EMBO J. 20, 4, 694-702, (2001)), indicating that the function of the PrP gene is not absolutely essential. Alternatively, as it has been suggested, there may be some mecha-nism of functional gene compensation by Dpl, a gene showing considerable homology to PrP and mapping only 16 Kb downstream of it which is upregu-lated in the CNS of PrP knockout_mice (Moore et al., J. Mol. Biol. 292, 797-(1999)).

The knockout technology described above has encountered two major prob-lems when applied to somatic cells. First, knocking out a gene in these cells only provides limited information about the functional consequences of the disruption, since the phenotype of knockout cells in culture does not necessar-ily reflect the final effect in a mutant organism such as the mouse derived from knockout ES cells. The second, and more critical problem, is that the targeting frequency of homologous recombination in somatic cells is about two orders of magnitude lower than in ES cells (Hanson and Sedivy, Mol. Cell Biol.
15(1):45-51 (1995)). For efficient gene targeting in somatic cells, promoter trap approaches utilising promoterless vectors are absolutely essential (Sedivy and Dutriaux, Trends Genet. 15(3):88-90 (1999)), since they can typically achieve enrichments of 100-500 fold. Upon introduction into somatic cells and antibiotic selection, the selection cassette can only be expressed after homolo-gous recombination from the promoter of the target gene. Gene targeting can also be improved by the ClonePix technology available from Genetix, UK.

It was now found that a somatic human cell line (e.g. a cell line derived from the HEK293F cell line) could be established in which the gene coding for the prion protein sequence has been inactivated in at least one, preferably in both alleles or three alleles in triploid cells. This was achieved by homologous re-combination of a knockout vector carrying a promoterless selection marker which after targeted integration into the genome enables expression of said selection marker by the endogenous PrP promoter. Such resulting prion-ablated cells are suitable for recombinant production of human target proteins after transfection with a suitable expression vector encoding the target pro-tein. In combination with suitable protein purification and virus inactivation, this method provides an effective system to produce safe and highly active recombinant human proteins free from prion proteins for therapeutic applica-tion in humans.

Summary of the invention The present invention relates to (1) a prion protein(PrP)-free, immortalized, somatic, cell line wherein both alleles of the PrP gene were completely deleted;
(2) a method for producing the PrP free immortalized cell line as defined in (1) above, which comprises subsequent deletion of both alleles of the PrP
ORF in a respective starting cell by homologous recombination with PrP
knock-out constructs;
(3) a PrP knock-out construct as defined in (2) above;
(4) the use of the PrP-free immortalized cell line as defined in (1) above for PrP-free recombinant production of a human protein or a derivative or mutant thereof (hereinafter "target protein");
(5) a method for preparing a cell line for PrP-free recombinant production of a target protein, which comprises transfecting a PrP-free immortalized host cell line as defined in (1) above with a transfection vector compris-ing an origin of replication, and a gene encoding said human target pro-tein, whereby the gene for the human target protein is linked at its 5' end with a promoter and its 3' end with a polyA signal;

(6) a PrP-free immortalized cell line stably transfected, preferably under serum-free conditions, with the transfection vector as defined in (5) above; and (7) a method for PrP-free recombinant production of a human target protein to be used as pharmaceuticals which comprises culturing a PrP-free, im-mortalized, human cell line as defined in (6) above.

The method of embodiment (7) of the invention is particularly suitable for the production of recombinant human proteins and therapeutical antibodies, in-cluding clotting factors like factor VII/a, factor VIII, factor IX, von Willebrand factor (vWF) and Adamtsl3 and growth factors like granulocyte colony stimu-lating factor (G-CSF) or granulocyte macrophage colony stimulating factor (GM-CSF), free from prion protein contamination. For this method an immor-talized human cell line, e.g. a HEK 293 cell line, with an ablated prion protein sequence of the invention is utilized. This cell line is obtainable by transfecting an immortalized human cell line with a vector containing a selectable or selec-tion marker such as for example the neomycin phosphotransferase ORF, de-void of its own promoter and translation initiation.

In a preferred embodiment, the cell line is a human cell line, like a cell line derived from HEK 293F or Per.C6 cells (immortalized human foetal Retina cells). Other suitable cells are CHO (Chinese Hamster Ovary cells) and BHK
(Baby Hamster Kidney cells) cells.

Brief description of the Figures Fiaure 1: Promoter trap strategy to ablate the human PrP gene Fiaure 2: 1. PrP K.O. Construct pBS_Neo_P-_R+L_Arm_2B
Fiaure 3: Cloning strategy to generate construct pBS_Neo_P-_L+R_Arm_2B. Sequence of the region cloned into the pBluescript vector, i) Neomycin gene without its own promoter and ATG, ii) Left arm for homologous recombination, iii) Right arm for homologous recombination Fiaure 4: PCR-based screening strategy after G418 selection Figure 5: Genomic Southern strategy to characterize clones after stable integration of the targeting construct pBS_Neo-_P-_L+R_Arm_2B
Figure 6: Genomic PCR screening after integration of targeting construct p8S Neo_P-_R+L Arm_28. DNA marker: GeneRuler DNA Ladder Mix; Positive control: genomic DNA from a targeted PrP cell mix population. A1-8 and B1-8 are the cell clones picked by Clone-PixFL. The clones labelled with green circle were identified as PCR
positive clones due to the 2,3 kb bands.
Figure 7: Southern analysis of PrP K.O. cells after targeted integration of the first K.O. construct p8S Neo_P-_R+L Arm_28: the following is utilized: DNA marker: GeneRuler DNA Ladder Mix; WT: 293F
wild type: K.O.: a K.O. clone identified after targeted integration of the 1st K.O. construct pBS_Neo_P-_R+L_Arm_2B. Genomic DNA from K.O. clone and 293F WT cells was hybridized with the 5'-, 3'- and Neo-probes. As expected, WT 293F cells show a 10,8 kb WT band with with both 5'- and 3'-probes but no signal with the Neo-probe. For K.O. clones, a 4,2 kb band with 3'-probe and a 6,5 kb band with 3'-probe could be detected, additionally a 6,5 kb band was seen with the Neo-probe.
Figure 8: FISH analysis of PrP K.O. cells with one targeted PrP allele: pat-tern a: two signals of chr2O and Bac; Pattern b: 3 signals chr20 and Bac; pattern c: 3 signals of chr20 but only 2 signals of Bac.
Fi a u re 9: ELISA analysis of PrP K. O. cells with one targeted PrP allele and 293F cells: 2 K.O. cell lines bearing one PrP targeted allele (K.O.
1 and K.O. 6) were analyzed and compared to wild type 293F
cells.
Figure 10: 2. PrP K.O. Construct p8S Zeo_P=R+L_Arm.
Fi a u re 11: PCR screening for cell clones or mixed cell populations with two PrP K.O. integrations: In the gel the following is utilized: DNA
marker: GeneRuler DNA Ladder Mix; Positive control: genomic DNA from a targeted PrP cell mix population. T3-1 is a mixed cell population selected with zeocin. T3-2 is mixed cell population se-lected with zeocin and G418. Both mixed cell populations were identified as positive.
Fiaure 12: Genomic Southern strategy to characterize clones after stable integration of the targeting construct p8S Neo-_P-_L+R_Arm_.
Fi a. 13.: Transfection efficiency of PrP KO cells with one PrP targeted allele.
The same amount of cells were transfected with a expression plasmid and transfection efficiency was compared.

Fi g. 14: Expression of FVIII (Fig. 14a), FIX (Fig. 14b) and G-CSFb (Fig.
14c) in PrP K.O. cells with one targeted PrP allele. Expression of arbitrary units of FVIII, FIX and G-CSFb per 10E6 cells compared to 293F wild type cells.

Detailed description of the Invention The following definitions and abbreviations are utilized in the present applica-tion:

"BAC" means bacterial artificial chromosome. "bp" means base pairs. "G418"
and "zeocin" are two different selection antibiotics; stably transfected cells with constructs carrying them as selection markers become resistant ("G418R"
or "zeocinR") to these antibiotics. "Homologous recombination" refers to a mechanism whereby two DNA fragments of homologous sequence recombine with each other. "Left arm" refers to the intronic region of the PrP gene lo-cated immediately upstream of exon3. "Right arm" refers to the region imme-diately downstream of exon3 of the PrP gene. "Neo" refers to the neomycin phosphotransferase gene. "ORF" means open reading frame. "PCR" means polymerase chain reaction, "PrP" means prion gene or the prion protein.
"HEK293F" refers to a specific human embryonic kidney cell line.

Thus embodiment (1) of the invention pertains to a prion protein (PrP)-free, immortalized, somatic, human cell line wherein both alleles of the PrP gene have been completely deleted.

In accordance of the invention said cell line is capable of being transfected and being cultured under serum-free conditions.

Further it is preferred that the cell line has been rendered immortal by integra-tion of adenoviral sequences into its genome. The cell line may be derived from a starting cell selected from the group of kidney, bladder, liver, lung, cardiac muscle, smooth muscle, ovary and gastrointestinal cells. Preferred is that the starting cell is a human kidney cell line, such as a human foetal kid-ney cell. Particularly preferred is that the foetal human kidney cell is either a FreeStyle 293 (293F cells; Invitrogen R79007), a HEK 293 (293 cells; ATCC
CRL-1573; DSM ACC 305), or a 293T cell(DSM ACC 2494), preferably is a 293F cell (Invitrogen R79007).

In another preferred embodiment of the invention, the PrP ORF has been com-pletely deleted by homologous recombination with knockout traps carrying selectable or selection marker genes so that expression of the selectable or selection maker is driven by the endogenous PrP promoter.

In a particular preferred embodiment the prion-free cell line of embodiment (1) is prion-free 293F cell line pf293F, which includes all intermediate mixed populations, isolated clones necessary to isolate said final knock-out cell line and modifications derived therefrom.

In the method of embodiment (2) of the invention the knock-out constructs may be suitable to delete the entire PrP ORF in both alleles. Further, the knock-out constructs may carry the same or different promoterless selection marker genes flanked by two sequences homologous to the insertion site within the PrP gene of the starting cell. It is, however, preferred that the knockout constructs carry different selection marker genes or selectable mark-ers. The knock-out constructs may further carry one of the following functional sequences: a poly A sequence, recombinase recognition sequences, IRES and the like.

The homologous sequences of the knock-out construct may have a length of 1 to 10 kb, preferably of about 6 kb and do preferably correspond to sequences upstream and downstream of the PrP ORF of the starting cell line. Particularly preferred is that the homologous sequences are those shown in SEQ ID NOs:2 and 3. Suitable selection markers encode positive selection markers including, but not limited to, neomycin phosphotransferase, zeocin, hygromycin; and the selectable marker includes fluorescence marker such as GFP and Dsred and enzymes such as LacZ.

It is particularly preferred that the knock-out constructs have one or more of the sequences shown in SEQ ID NOs:1 and 16.

The invention is furthermore described by reference to HEK 293 or HEK 293F
cells. In such cell lines of the invention, the coding region of PrP gene is com-pletely deleted by means of a promoter trap. Three consecutive steps are necessary for deleting the coding region of the PrP gene in HEK 293F cells:

1. Targeting the coding region of PrP on one allele with a PrP knockout (here-inafter "K.O."-) construct containing a neomycin selection marker.

2. Identification and isolation of clones bearing only one targeted PrP allele and one wild-type PrP allele. This step is necessary due to the genetic hetero-geneity of the parental HEK 293F cells, which carry 3 copies of the PrP gene in 75% of the population and 2 copies in 25% of the population.

3. Targeting the coding region of PrP on the remaining allele with a second PrP
K.O. construct, this time containing a zeocin selection marker.

Following stable transfection of the first PrP K.O. construct (carrying neomycin as the selectable marker, Figure 2) into HEK293F cells, G418R clones were isolated at several different antibiotic concentrations and screened by a PCR-based strategy which identifies targeted events (Figure 4, Figure 6). Clones bearing PrP-targeted integrations were then characterized by genomic South-ern blot analysis in order to evaluate:

(a) whether the integration of the targeting construct is correct both at its 5' and and its 3' end (b) how many PrP alleles have been targeted (one, two or three), and how many are still intact (wild-type: one or two) In the second K.O. round, the second PrP K.O. construct (carrying zeocin in-stead of neomycin as the selectable marker) was then used to knock out the remaining PrP allele in those clones where one allele had been targeted with neomycin and the second PrP allele was still intact. After PCR screening and genomic Southern analysis of isolated, zeocinR clones, the complete ablation of the PrP gene was confirmed by RT-PCR analysis to demonstrate the lack of PrP
mRNA and by Western blot analysis to show complete absence of the PrP pro-tein.

The resulting, full PrP K.O. cell line described in this invention can then be used for guaranteed prion-free expression of human recombinant proteins.

During the entire process of transfection, antibiotic selection, clone isolation, screening and expansion, the cells were cultured under serum-free conditions (f. e. in FreeStyle media or Octapharma in-house media). Following establish-ment and complete genetic and phenotypic characterization of the final PrP
K.O. cell clone a Research Cell Bank was generated (see cell culture methods of 293F cells in 6.1. of Materials and Methods). This PrP K.O. cell line, hereaf-ter named "prion-free 293F cell line" can then be stably transfected with any gene of interest (f. e. Factor VIII, factor IX, G-CSF, vWF, GM-CSF, Factor VII/VIIa or antibodies) completely under serum-free conditions according to patent application (see copending EP 05 105 965, the disclosure of which is hereby incorporated in its entirety).

Cells resulting from such a stable transfection performed with prion-free 293F
cells routinely growing in serum-free medium and f. e. with a pcDNA3.1 con-struct carrying the gene of interest are seeded in semi-solid, methyl-cellulose-based medium containing an appropriate antibiotic for clone selection and a labelled antibody for detection of the highest producer clones via fluorescence.
Large numbers (for example ten-thousands) of clones are then analyzed using ClonePixFL (Genetix) with respect to cell number and secreted recombinant protein in order to subsequently pick only a few hundred best producer clones.
In contrast to other known methods, where non-producer clones and mixed clones are all randomly picked, ClonePixFL allows simultaneous identification and picking of the fastest growing clones which are also the highest producers, and which originate from single cells. The picked cells are expanded in micro-titer plates and later in spin tubes, cell culture flasks and fermenters under serum-free conditions for the complete procedure.

Here as well the whole stable transfection procedure is generated under se-rum-free conditions. Additionally, during the whole following expansion and cell culture procedure, the cells do not have any contact with serum or animal-derived proteins.

During expansion, the best clones are selected with respect to robustness, high growth rate, viability, scalability and production of f. e. active recombi-nant protein as measured by ELISA test. During this selection the number is reduced again to only a few best producing clones. Additional to the productiv-ity, correct cDNA sequence, mRNA content and cell behaviour upon fermenta-tion are the criteria to identify the best clone(s) for subcloning. For this, cells of the selected clone(s) are re-plated, analyzed and picked with ClonePixFL, and then expanded and selected as described above. Subcloning is an essen-tial step in order to select only the best producer clones eliminating possible genetic variations in the plated subpopulation of the clone. After subcloning, the selected clone(s) are banked again under serum-free conditions. The re-combinant human protein expressed by the final selected subclone(s) is char-acterized biochemically in more detail.

Furthermore, the K.O. clones can be isolated in semi-solid media using Clone-PixFL with a fluorescent labelled antibody detecting clones in which Prp gene has been knocked out completely.
Examples Materials and Methods 1. Devices used for Molecular Bioloav techniques Device Supplier Type Catalogue Comments Number Agarose gel electro- BioRad SUB-CELL GT
phoresis chamber Power supply for BioRad PAC 3000 electro horesis UV-transiluminator Vilber Lourmat Centrifuge Heraeus Biofuge fresco Max. 13,000 rpm Thermomixer Eppendorf 5436 Waterbath HAAKE Type 002-9917 37 C incubator Menmert Modell 300 Refrigerator Liebherr - +2-8 C
Freezer Liebherr - -20 C
Pure water system Millipore Milli-Transfer pipets Gilson P2, P20, P200, - -Yellow Tips Josef Peske oHG -200 pl Blue Tips Josef Peske oHG -1000 Pl Filter Tips, 5-250pl Peske (Mplti) im Rack, sterile 1491-11 Filter Tips, 100-1000 Peske (Mplti) im Rack, sterile 1420-111 S pectro photometer Beckman DU 530 Gel documentation BioSciTec. Science Gelscript ver-system Group sioni. i lpSlide Ibidi ibiTreat 81826 2. Reaaents and Kits for Molecular Bioloav techniques Reagent Supplier Order Num- Storage conditions Comments ber Pfx polymerase Invitrogen 11708-013 -20 C
NotI NEB R0189S -20 C
EcoRI NEB R0142S -20 C
KpnI NEB R0142S -20 C
NotI NEB R0146S -20 C
ConcertTM Rapid Plasmid GIBCO 11453-024 RT
Miniprep System IA uick PCR purification kit Qiagen 28104 RT
Qia Quick Gel Extraction Kit Qiagen 28704 RT
DNA Ladder Mix MBI SM0331/3 -20 C
SmartLadder Eurogentec MW-1700-07 -20 C
agarose Sigma A9539 RT

Tryptone Sigma T9410 RT
Yeast extract GIBCO 30393-029 RT
NaCI Roth 9265.1 RT
Agar Sigma A5054 RT
T41i ase NEB #M0202S -20 C
SURE 2 Supercompetent Cells Stratagene 200152 -80 C
One shot Top10F' competent Invitrogen 44-0300 -80 C
cells HiSpeed Plasmid Midi Kit Qiagen 12643 RT
EndoFree Plasmid Maxi Kit Qiagen 12362 RT
DNAeasy Blood & Tissue kit QIAGEN 69504 RT
Gentra Puregene Cell Kit QIAGEN 158767 RT
3. Bacterial arowth media LB medium: Tryptone 10 g, Yeast extract 5g, NaCI 10 g, dissolved into 1 I
H20, then autoclaved.

LB/amp agar plates: 1.5% agar into LB medium containing 100pg/ml ampicil-lin 4. Materials for 293/293F cell culture 4.1. Cell lines 293 cell line: The 293 cell line is a permanent cell line which grows adherently in the presence of serum. It was established from primary embryonic human kidney transformed with sheared human adenovirus type 5 DNA (Graham et al., 1977; Harrison et al., 1977). The E1A adenovirus gene is expressed in these cells and participates in transactivation of some viral promoters, allow-ing these cells to produce very high levels of protein.

293F cell line: The FreeStyle 293F cell line is a variant of the 293 cell line that has been adapted to suspension growth in FreeStyle 293 expression medium.
The original cell line was obtained from Robert Horlick at Pharmacopeia in 1988. The FreeStyle 293F cell line is derived from the 293 cell line and is in-tended to be used with the FreeStyle 293 expression system (Invitrogen Cata-log no. K9000-01) available from Invitrogen, or other serum-free media, such as the Octapharma in-house media. FreeStyle 293F cells are adapted to sus-pension culture in FreeStyle 293 expression medium. The FreeStyle 293F cell line exhibits the following characteristics:

- Prepared from low passage Master Cell Bank cultures derived from parental 293F cells that were re-cloned by limiting dilution. The 293 clonal derived cultures are maintained in serum-free conditions for only 30 to 35 total pas-sages.
- Adapted to high density, serum-free, suspension growth; may be main-tained in FreeStyle 293 expression medium.
- High transfection efficiency with 293fectin - Suspension cultures may be transfected in FreeStyle 293 Expression Me-dium without the need to change media.
- Permits transfection of cells at large volumes.
4.2. Devices for 293/293F cell culture Device Supplier Type Catalogue Comments Number Sterile hood Heraeus HeraSafe Incubator Kendro BDD6220 - 8% C02, 37 C
Orbital shaker GFL 3005 placed in a COZ
incubator COZ shaker incuba- Kuhner AG ISF-W-1 SM1503 tor Microscope Zeiss Axiovert 25 - -Centrifuge Kendro Me afu e 1.0 -Refrigerator Liebherr 200381 - +2-8 C
-20 C freezer Liebherr -80 C freezer Heraeus HFC 586 basic - -Water bath Memmert GmbH WB 14 - -Haemocytometer Peske Neubauer im- - 0.100 mm (Neubauer) proved depth 0.0025 mmZ
Haemocytometer Peske 03-0000 20x26x0.4 mm glass cover Manual cell counter Rexel, UK ENM - -Isopropanol bath Nalgene 1 C freezing 5100- to be filled with container 0001 250 ml isopro-panol Biostore (liquid Air liquide Arpege 110 nitro en submerse) Erlenmeyer flasks, Corning 125, 250, 500, 431143- Polycarbonate single-use 1000 ml with 431147 filter vent cap Single-use pipettes Schubert & Weiss 9.380.431 1 mL
Single-use pipettes Falcon/Costar - 356507 -2 mL
Single-use pipettes Nunc - 159625 -mL
Single-use pipettes Nunc - 159633 -mL
Single-use pipettes Nunc - 159641 -25 mL
Single-use pipettes Nunc - 159668 -50 mL
Pipetting aid Sigma-Aldrich Accujet 356555 -ml centrifugation Peske 17-1200 Sterile packed tubes 50 ml centrifugation Peske 17-1020 Sterile packed tubes Adhesive labels for CILS Thin self- LSL7-W-5-cryovials laminating TN
labels Cryovials Simport plastics T-311-2 1.8 ml (CAN) Manipulation rack TPP 100 x 200 x 25 99016 See Fig. 1 of for cryovials mm SOP
Cryoboxes Simport 136x136x50 T314-2100 mm Photometer Dynex MRX
S pectro photometer Beckman DU530 Fluorescence Mi- Olympus BX-41 croscope 4.3. Reaaents for 293/293F cell culture Reagent Supplier Order Num- Storage Comments ber conditions Serum-free culture medium Invitrogen 12338-026 +8 C in the -FreeStyle293 dark Lipofectamine2000CD Invitrogen 12566-014 4 C
OptiProSFM Invitrogen 12309-019 4 C
DMSO Sigma D-2650 room tem- do not store perature, at +8 C or dark colder as material gets crystallised 80% Ethanol + 1% MEK Zefa ADR /36 Room tem- For disinfec-erature tion Isopropanol 100% p.a. Sigma 10398 room tem- For isopropa-Aldrich perature, nol bath closed vented store for burnable solvents Opti-MEM I Invitrogen 51985-034 +8 C in the -dark Zeocin Invitrogen R250-01 At -20 C -G418 PAA P11-012 At 4 C -Lipofectamin 2000CD Invitrogen 12566-014 4 C
OptiPRO SFM Invitrogen 12309-019 4 C
Colcimid Biochrom L6221 -20 C
AG
Tween 20 Sigma P9416 RT
Albuminativ 40g/I Octapharma Batch No.:

4.4. Reaaents for FISH, Immunostaininci and ELISA
Reagent Supplier Order Number Storage Comments conditions Chromosome 20 whole Aquarius LPP20G -20 C
chromosome painting probe DAPI Antifade Aquarius DES 150L -20 C
Protein Assay (Bradford) BioRad 500-0006 4 C
Prion Protein EIA Kit SpiBio 589751 -20 C
3F4 (first Ab) Sigma 054K1525 -20 C
Goat anti mouse - FITC Novus NB720-F 4 C
(second Ab) SAF32FITC 1,3m /ml) Spibio No odernr. -20 C
SAF32 (200pg) Spibio No ordernr. -20 C
POM 17 (1mg/ml) Dr.Aguzzi Lab, -20 C
Zurich, ETH
POM 12 (1mg/ml) Dr.Aguzzi Lab, -20 C
Zurich, ETH

4.5. Initiation and standard arowth medium for 293F cells FreeStyle293 medium, no additives FreeStyle 293 expression medium allows to grow, maintain, and transfect FreeStyle 293F cells. FreeStyle 293 expression medium available from Invitro-gen is a defined, serum-free formulation specifically developed for the high density, suspension culture and transfection of 293 cells. The medium contains no human or animal origin components and is formulated with Glutamax-I to increase stability and maximize shelf life.

4.6. Freezina medium for 293F cells FreeStyle293 Medium + 10% DMSO

5. Molecular Bioloay Methods 5.1 Isolation of aenomic DNA from cell pellets with QIAGEN DNAeasy Blood &
Tissue kit (cat No. 69504) or Gentra cell kit (QIAGEN Cat No. 158767) for PCR
and Southern analysis.

5.2 Isolation of aenomic DNA from 96-well plate for PCR analysis:
= The cells were rinsed twice with PBS.
= Add 50p1 of lysis buffer (0,6m1 proteinase K should be added to 12m1 lysis buffer [10mM Tris, 10mg EDTA, 10mM Nacl, 0,5% sarcosyl] to each well according to the following table = Incubate the plates overnight at 50 C, the plates should be covered with Parafilm.
= Spin down at 2500rpm for 1min. Add 100p1 of NaCI/ETOH (150pl of 5M
NaCI to 10m1 of cold absolute ethanol) to each well and shake the plate for 30 min at room temperature. The nucleic acids precipitate as a filamentous network.
= Spin down at 2500 rpm for 1 min, invert the plate carefully to discard the solution; the nucleic acids remain attached to the plate. Blot the excess liq-uid on paper towels.
= Rinse the nucleic acid 3 times by dripping 100p1 of 80% ethanol per well using multi-channel pipette. Shake the plate for 30min during each washing step. Spin down at 2500rpm for lmin and discard the alcohol carefully by inverting of the plate each time.
= Dissolve the genomic DNA in 50 pl of TE and covered with Parafilm. The plate was put in the 37 C incubator for overnight or 1-2 hrs at 50 C for complete dissolve.
= Store the plate with genomic DNA at 4 C for use or at -20 C for storage.
6. Transfectina Cells with Lipofectamin 2000CD:

Before beginning, make sure the cells are healthy and >90% viable.

1. On the day before transfection prepare a suspension culture with a cell density of 0.8-1.1x106 cells/ml in growth media without antibiotics.
2. On the day of transfection: prepare a suspension culture with 1.1x106 viable cells/ml.
3. For each transfection prepare complexes using the following reagent amounts and volumes for every ml of cells transfected with or without Al-buminative as supplement in growth medium:
= Dilute 0.5-1.5pg DNA in 34pl of OptiMEMTM SFM or OptiPROTM SFM
= Dilute 1-10p1 of LipofectaminTM 2000 CD in 34pl of OptiMEMTM SFM or OptiProTM SFM
4. Add the complexes to the flask/plates containing cells and media.
5. Incubate the cells on plates at 37 C or suspension culture on an orbital shaker rotating at 125 rpm for 24-96 h in a C02 incubator.

6.3. Selection of the transfection, expansion of the culture, seedina cells on 96-well plate and preparina for clone pickina with ClonePixFL:

1. After transfection, anbiotics zeocin or G418 were added into transfected culture for selection.
2. After selection, the antibiotic-resistant cells were collected for isolation of genomic DNA and screened by PCR analysis.
3. The culture with positive PCR screening result was expanded to reach 5x105 cells/ml.
4. Seed the cells in a 96 well plate with 1000 cells / well with growth media.
5. Prepare replica plates for PCR and cell freezing, if cell density reaches 50%.
6. In case that 80% cell density was reached, the plate was ready for PCR
screening.
7. Pool the PCR positive cells and seed it again in 200 - 250 cells per well in a 96 well plate; or expand the PCR positive cells separately and seed 200-250 cell/well in a 96-well plate.
8. Prepare replica plates for PCR and cell freezing if cell density reaches 50%.

9. In case that 80% cell density was reached, the plate was ready for PCR
screening.
10. Expand the PCR positive cells to 5x105cells/ml for seeding test.

11. Seed the cells in different densities into Semi-Solid-Medium to find out the best cell density, so that single colonies can be easily picked by ClonePixFL
without touching other colonies.

12. Pick single cell colonies with ClonePixFL.

13. Synchronize and expand the colonies picked in 96-well plate, prepare replicas from cells for freezing and PCR analysis.

7. FISH analysis: Determination of number of chromosome 20 in 293F cells and pf293F cells:

Before beginning, make sure the cells are healthy and >85% viable.

1. Prepare a suspension culture with a cell density of 1-3x106 cells in 5-10m1 growth medium.
2. Add 0,2pg/ml Colcemid in medium and incubate on an orbital shaker rotating at 125 rpm for 30 minutes to 1 hour 3. Harvest the cells by centrifugation at 1100 rpm for 10 minutes.
4. Remove supernatant except for 500p1 and resuspend the pellet .
5. 8 - 10 minutes incubation in 5 - 10m1 75mM KCI at room temperature.
The first ml of KCI solution should be given drop wise to the cells.
6. Add 2-3m1 fixative solution and centrifuge at 1100 rpm for 10 minutes.
7. Transfer the cell pellet to a small microfuge vial and do all subsequent fixative wash in this vial. It is important that cells are completely resus-pended.
8. Cells are washed 3-6 times in fixative solution, after each wash step centrifuge the cells for 1 minute at 6000 rpm in a microfuge.
9. Slides must have room temperature, pass slide through hot steam for 2-3 seconds to moisturize the surface (water bath at 75-80 C) 10. Place 10-30pl cell suspension on the slide, don't let the liquid dry 11. After the surface becomes grainy pass the slide again for 1-2 seconds through hot steam.
12. Immediately dry on a metal plate carrying a gradient of temperature across its surface.
13. Incubate slides for 2-3 days at room temperature or incubate slides overnight at 65 C

14. Immerse slide in 2XSSC, pH 7.0 for 2 minutes 15. Dehydrate the slides in an ethanol series (70%, 85% and 100%) each for 2 minutes 16. Remove probe (Chromosome painting probe, Cytocell/Aquarius, LPP20G) from -20 C and allow warming to room temperature.

17. Ensure probe solution is uniform by repeated pipette mixing 18. Remove 5-10p1 probe and place on slide, cover with a 24x24mm glass cover slip and seal with rubber solution glue or clear nail polish 19. Denature at 75 C (+/- 1 C) for 2 minutes and hybridise at 37 C (+/-1 C) over night.

20. Remove cover slip and all traces of glue carefully 21. Wash slide in 0.4XSSC (pH 7.0) at 72 C (+/- 1 C) for 2 minutes.

22. Drain slide and wash in 2XSSC, 0.005% Tween20 (pH 7.0) at room tem-perature for 30 seconds.

23. Drain slide and apply 10p1 of DAPI antifade 24. Cover with a cover slip and allow colour to develop in the dark for 10 minutes.

25. Observation with fluorescence microscope.

8. Immunostainina: Analysis of the amount of PrP protein on the cell surface of 293F cells and pf293F cells:

Before beginning, make sure the cells are healthy and >85% viable.

1. Seed 1-3x105 cells per well (1 Slide, Ibidi) in 30-100p1 growth medium and incubate over night at 37 C.
2. If the cells growth well and become adherent, start staining 3. Carefully aspirate culture medium and rinse cells carefully with PBS, do not shake.
4. Add ice-cold fixative (50%Ethanol, 50%Methanol) for 1 minute at room temperature.
5. Immediately wash cells twice for 5 minutes with PBS
6. Cover cells with 8% BSA/PBS and incubate for 1 hour at room tempera-ture. Perform the incubation in a sealed humidity chamber to prevent air-drying of the fixed cells.
7. Wash cells twice for 5 minutes with 1xPBS
8. Gently remove excess PBS and cover cells with primary antibody (3F4, POM12, POM17, SAF32, SAF32FITC, 6H4) diluted in 1% BSA/PBS and in-cubate for 1-2 hour at room temperature. Perform the incubation in a sealed humidity chamber to prevent air-drying of the tissue sections.
9. Wash cells twice with PBS for 5 minutes 10. Gently remove excess PBS and cover cells with secondary antibody (Goat anti mouse - FITC) diluted in 1% BSA/PBS for 1-2 hours at room tem-perature. Perform the incubation in a sealed humidity chamber to prevent air-drying of the tissue sections.
11. Wash cells with PBS three times for 5 minutes in the dark.
12. Add Dapi-anti-fade to the slide, mount cover slip and examine specimen under fluorescence microscope.

9. ELISA Assay: Quantification of PrP protein levels in 293F cells and pf293F
cells containina one or two PrP alleles:

1. Collect 2x10' cells and centrifuge at 800 rpm for 5 minutes 2. Remove medium 3. Wash cell pellets with 5ml cold PBS
4. Centrifuge the cells at 800 rpm for 5 minutes.
5. Remove PBS
6. Wash cell pellet with lml cold PBS and centrifuge at 6000rpm for 5 min, remove PBS
7. Cell pellet can be lysed directly or frozen for storage at -80 C

8. 1000 pl cold lysis buffer is added to the cell pellet 9. Pass the lysates through 20, 23 and 26 gauge syringe needles 10. The protein concentration of the cell lysates was measured by Bradford Assay Example 1: Detailed description for the first knock-out round with the PrP
K.O.
construct pBS_Neo_P-_R+L_Arm_2B (SEQ ID NO:1) containing neomycin.

A. Taraetina Strateay: PrP expression will be completely obliterated in human HEK 293 or HEK 293F cells by means of a promoter trap. This strategy (out-lined in Figure 1) was designed to specifically select those cells in which the ORF of the PrP gene is replaced by the neomycin phosphotransferase ORF
(except the translation initiation codon, which belongs to the PrP gene).
Plasmid construct 2B (depicted in Figure 2) consists of 4 main components: 1.
The vector backbone (not depicted in the figure): pBluescript, 2. A "neo"(=
neomycin phosphotransferase) truncated cassette, consisting of the complete ORF for this gene except its own translation initiation codon followed by a transcription termination signal (see SEQ ID NO. 1). Critical issues: this neo truncated cassette does not carry any promoter of its own. The neo ORF was designed to be translated from the initiation ATG codon which belongs to the PrP gene. 3. A "left arm" region upstream of neo, with sequence identical to the PrP intron located between E2 and E3 (see SEQ ID NO. 2). 4. A "right arm"
region downstream of neo, with sequence identical to the PrP region down-stream of the PrP ORF (see SEQ ID NO. 3).

Upon (a) transfection of this construct into the host 293F cells, (b) subsequent integration into their genome and (c) selection with G418, the goal of this approach was to enrich for those cells in which homologous recombination (symbolized in figure 1 by two black "X") results in integration of the neo ORF
downstream of the PrP transcription and translation regulatory sequences (P, El, E2).

Only those transfected cells in which the construct has integrated downstream of a functional promoter will survive the G418 selection. Therefore all surviving cells are expected to carry integrations downstream from functional promoters (hence "promoter trap"). Because of the long homology of the construct arms to the PrP gene (in total N6 kb homologous sequence), homologous recombi-nation was expected to specifically target this integration to the PrP locus, resulting in the integration depicted in the third raw of figure 1.

B: Generation of the Taraetina PrP K.O. Construct (pBS Neo P-R+L Arm 2B): The generation of the targeting construct consists of three consecutive cloning steps (Figure 3), each of which includes: a PCR amplifica-tion step to produce an "insert", a ligation step of such an insert into a vector, and transformation of the ligation mixture into E. coli and selection of correct clones by restriction enzyme digest.

1. Generation of the pBS_Neo_P-_2B Construct: The neo ORF was amplified by PCR using as a template the pcDNA3.1+ vector (Invitrogen) using the fol-lowing synthetic oligonucleotides:

2B-Neo-F: 5'-GGCAAGAATTCGCAGAGCAGTCATTATGATTGAACAAGATGGATTGCAC- GCAG -3' (SEQ ID NO. 4) 2B-Neo-R: 5'- GGACCGCTCGAG-ATGCTTCCGGCTCGTATGTTGT-3' (SEQ ID NO.
5), which have EcoRI and XhoI restriction sites (underlined). The amplified prod-uct (1244 bp) was digested with EcoRI and XhoI and ligated into EcoRI/XhoI-digested pBluescript II KS+ (Stratagene). The ligation mixture was then trans-formed into One shot ToplOF' competent cells (Invitrogen). Screening was performed by EcoRI/XhoI restriction digest of plasmid DNA prepared from individual colonies.

2. Generation of the pBS_Neo_P-_L_Arm_2B Construct: The left arm of the targeting construct was amplified by PCR using as a template the BAC DNA
clone 186 (BACPAC Resources Center, BPCR, http://bacpac.chori.org/) and the following synthetic oligonucleotides:

2B-L-F: 5'- GGCAAGCGGCCGC-CTCTGTCTAGGAACACTGCTGTG-3' (SEQ ID NO.
6) 2B-L-R: 5'- GGCAAGAATTC-AAAATGAAGAGGAGAACGTCAGAGTC-3' (SEQ ID
NO. 7), which have NotI and EcoRI restriction sites, respectively (underlined). The amplified product (1929 bp) was digested with EcoRI and NotI and ligated into EcoRI/ Notl-digested pBS_Neo_P-_2B. The ligation mixture was then trans-formed into One shot Top10F' competent cells (Invitrogen). Screening was performed by EcoRI/NotI restriction digest of plasmid DNA prepared from individual colonies.

3. Generation of the pBS_Neo_P-_R+L_Arm_2B Construct: The right arm of the targeting construct was amplified by PCR using as a template the same BAC DNA clone 186 as in B.2 above and the following synthetic oligonucleo-tides:

2B-R-F: 5'- GGACCGCTCGAG-TGTGTACCGAGAACTGGGGTGATG-3' (SEQ ID
NO. 8) 2B-R-R: 5'- GGCGGGGTACC-GCAGAATCTCTGAGCTCACCTCAG-3' (SEQ ID NO.
9), which have XhoI and KpnI sites, respectively (underlined).

The amplified product (4.6 Kb) was digested with XhoI and KpnI and ligated into XhoI / KpnI-digested pBS_Neo_P-_L_Arm_2B. The resulting ligation mix-ture was then transformed into SURE 2 supercompetent cells (Stratagene).
Screening was performed by XhoI/KpnI restriction digest of plasmid DNA pre-pared from individual colonies. In the resulting construct, pBS_Neo_P-_R+L_Arm_2B (sequence of the cloned region presented in SEQ ID NO. 1, the neo ORF (selection cassette without any promoter sequence and without its own translation initiation codon) was flanked:

- upstream by the intronic region which in the human genome precedes exon3 of the PrP gene; and - downstream by the PrP region which in the human genome follows the PrP
ORF.

4. Linearization of the pBS_Neo_P-_R+L_Arm_2B Construct: The targeting construct was linearized by restriction digest with KpnI, purified using the QIA
quick gel extraction kit (see Molecular Biology Methods, 5.2) and then quanti-fied by running 1pl aliquot on a 0.8% agarose gel along with 5pl of the SmartLadder marker (Eurogentec).

C. Introduction of the pBS Neo P- R+L Arm 2B Construct into Human Cells:
The targeting construct pBS_Neo_P-_R+L_Arm_2B was introduced into the host cells (293F, Invitrogen) using the Lipofectamin 2000CD reagent (Invitro-gen, see transfection method, 6.2). 48 h following transfection, cells were plated onto 10 cm dishes at a density between 1.25 and 1.5 x 106 cells. Anti-biotic selection was started at the time of seeding out at concentrations rang-ing between 30 and 120 pg G418/ml. Medium exchange was performed every second/third day for 14-21 days.

D. Screenina for Taraeted Clones: When cells reached confluence under anti-biotic selection, genomic DNA was prepared, either from 1 x 106 - 1 x 10' G418R cells using the QIA DNAeasy tissue kit (see Molecular Biology Methods, 5.3), or by using a previously described method developed to prepare genomic DNA from ES cells plated on 96 well plates (Ramirez-Solis et al.,1992). From each genomic DNA preparation, a PCR mix was prepared containing 80-300 ng genomic DNA, lx PCR buffer, 200 nM each oligonucleotide primer, 5 mM
MgCl2r 200 nM dNTP and 1,33 units Expand Fidelity Polymerase (Roche). The sequence of the synthetic oligonucleotides used for screening was as follows:

K.O.-F1: 5' CGACTCAGTGTCATTCCCTGCAGTCTC 3' (SEQ ID NO. 10) K. O.- R 1: 5' CATAG CCGAATAG CCTCTCCACCCAAG 3' ( S EQ ID NO. 11) The cycle parameters were as follows: 94 C 2 min; [94 C, 15 s; 71,6 C, 30 s;
72 C 100 s] x16 cycles; [94 C, 15 s; 71,6 C, 30 s; 72 C 100 s + 3 s longer in each successive cycle]x26 cycles; 72 C 7 min. Genomic DNA samples yielding a 2.3 kb PCR product indicate the presence of one or several targeted alleles in the cell population. Targeted clones were further analyzed with Southern-blot analysis.

E. Characterization of Taraeted Clones Bearina One Taraeted PrP Allele:

1. Genomic PCR screening (Figure 4): Genomic DNA was isolated from mixed cell populations or isolated cell colonies as described in D above. All PrP
K.O.
screening PCR reactions were performed with positive and negative controls.
The template for the positive control was 200 ng genomic DNA from a cell population mix in which a targeted PrP allele had been previously detected with the PrP K.O. PCR screening method; the negative control used water instead of genomic DNA as a template. The appearance of a 2.2 kb PCR prod-uct on agarose gel after electrophoresis indicates that the corresponding cell colony bears at least one targeted PrP cell allele. With this method the clones picked by ClonePixFLwere analyzed and the results are shown in Figure 6.

2. Genomic Southern analysis (Figure 7): PrP-targeted 293F cell clones identi-fied by genomic PCR as described above were further genetically characterized by Southern blot analysis. Following electrophoretic separation of EcoRI-digested genomic DNA and capillary transfer to Hybond+ membranes, the blots were radioactively hybridized with specifically designed DNA probes (de-picted as red arrows in Figure 5) in order to verify correct and intact integra-tion of the neo ORF. 5'- and 3'- probes are homologous to the 5'- and 3'-external regions of the expected integration site. Neo-probe is used to verify whether the K.O. construct is integrated in the target locus or just randomly.

5'- and 3'- probes were generated by PCR amplification using the BAC DNA
clone 186 described in section B.2 of this Example. The Neo probe was ampli-fied from plasmid pcDNA3.1(+) which includes a neomycin cassette, summa-rized in the following table:
Probe name Length Sequence of PCR primers for DNA Expected Southern of DNA probes pattern for PrP tar-probe geted cells on one allele 5'-probe 1 251 bp 5'-K.O.-Fl (SEQ ID No. 12): 4,2 kb and 10,8 kb 5'-AGCTTTACCGTCCAGTCTTC- 3' 5'-K.O.-Rl (SEQ ID No. 13):
5'-GGTCTTGATGG CGATAACTC- 3' 5'-probe 2 252 bp 5'-K.O.-F2 (SEQ ID No. 14): 4,2 kb and 10,8 kb 5'- GAGTTATCGCCATCAAGACC-3' 5'-K.O.-R2 (SEQ ID No.15):
5'- CATGAGAACCAACGCTAGAG-3' New 3'- 288 bp New-3'-probe-forward (SEQ ID No. 6,5 kb and 10.8kb probe 28):
5'- CTAGAGGTCCAGGTCATCTTG -3' New-3'-probe-reverse (SEQ ID No.
29):
5'- TCAGGGAAATTGGGGATCCTG-3' Neo-probe 304 bp Neo-Probe-F (SEQ ID No. 21): 6,5 kb 5'-AGCGAGCACGTACTCGGATG-3' Neo-probe-R (SEQ ID No.22):
5'-AAGCACGAGGAAGCGGTCAG-3' The genomic Southern strategy with construct pBS_Neo-_P-_L+R_Arm_2B is depicted in Figure 5. After EcoRI genomic digest and radioactive hybridization, targeted PrP clones show a 4.2 kb band with the 5'-DNA probe and a 6.5 kb band with the 3'-DNA probe, while the wild-type band from the non-targeted allele , detected with either probe, is 10.8 kb. With the Neo-probe a 6,5kb DNA band should be detected for PrP targeted allele and no band should be detected for wild-type 293F cells.

Clones showing correct pattern in both PCR and Southern blot with all 3 probes were identified as targeted PrP K.O. cells on one allele. These positive clones were further analyzed with FISH, ELISA and Immunostaining.

3. FISH analysis (Figure 8):

Human PrP gene is located on chromosome 20. 293F cells were hybridized with a WCP (whole chromosome painting) probe for painting of chromosome 20. this FISH analysis shows that 25% of the 293F wild type cells have 2 cop-ies of chromosome 20, while 75% of 293F cells bear 3 copies of chromosome 20. Technically it is very difficult and almost not possible to knock out 3 PrP
genes using the current technologies. Therefore, FISH analysis with chr20 WCP painting probe helps to distinguish if the cell clones already bearing one PrP targeted allele have two or three chromosome 20. Additionally, a BACDNA
clone 186 probe, which spans over 100kb, including the PrP gene, and is a part of chromosome 20, was also used for FISH hybridization for detection of possible chromosome deletions and translocations.

Three different FISH patterns could be observed for PrP K.O. cell lines after the targeted integration of the 1st K.O. construct pBS_Neo_P-_R+L_arm_2B:

a. Two chromosomes/metaphase were painted with the chromosome 20 probe and both showed the BAC signal (Figure 8a).
b. Three chromosomes/metaphase were painted and all three showed the BAC signal (Figure 8b).
c. Three chromosomes/metaphase were painted (one chromosome only partially due to a translocation of an arm of chromosome 20 to another chromosome) and only the two complete chromosomes 20 showed the BAC signal (Figure 8c).

The statistical analysis for several PrP K.O. cell lines is summarized in the following table:

Cell line 2 x chr. 20 3 x chr20 3 x chr. 20 others Analyzed total-2xBAC 3 x BAC 2 x Bac Cell Nr.
(pattern a) (pattern b) (pattern c) K.O. 1 4(12%) 1(3%) 29 (85%) ~ 34 K.O. 2 -- 22 (69%) 10 (31 %) ~ 32 K.O. 3 2(5%) 35 (87%) -- =E(8%)i 40 -11 K.O. 4 11 (27,5%) 28 (70%) -- 1(2,5%) 40 K.O. 5 -- 4(15%) 21(78%) 2(7%) 27 K.O. 6 -- 1(4%) 23 (96%) ~ 24 K.O. cells with two BAC signals have been taken for knocking out the second PrP gene, regardless of the existence of two or three copies of chr20 per cell.
Therefore, the following two K.O. cell lines have been chosen:

K.O. 1: this cell line had one population with two copies of chromosome 20 painted and both with the FISH signal of the BAC (12% of the cells detected).
A larger population (85%) showed three copies of chromosome 20 painted.
One of these chromosomes was only painted partially and translocated to another chromosome. Only the two completely painted chromosomes 20 showed the FISH signal of the BAC.

K.O. 6: 96% of the cells had three signals of chromosome 20 and 2 signals of BAC.

K.O. 1 and K.O. 6 were further analyzed by ELISA to determine if their PrP
protein levels are about half of wild type 293F cells.

4. ELISA test (Figure 9) ELISA was performed to quantify PrP protein concentration in PrP K.O. cells compared to wild-type 293F cells. Cell lines K.O. 1 and K.O. 6 bearing one PrP
K.O. allele consistently expressed only half of PrP protein compared to WT
293F cells (Figure 9). This data additionally supports and verifies additionally the conclusion after FISH analysis.

Example 2: Detailed description for the second knockout round with a PrP K.O.
construct containing zeocin.

The targeting strategy is the same as described in Example 1, except that the antibiotic is different. Namely zeocin instead of neomycin is included in the second PrP K.O. construct and therefore selection of antibiotic-resistant clones is performed with zeocin instead of with G418.

A. Cloning of the 2nd PrP K.O. construct containing zeocin cassette: The neo-mycin ORF in first PrP K.O. construct pBS_Neo_P-_L+R_Arm_2B was replaced by the zeocin ORF without its own promoter and ATG translation initiation codon. The zeocin cassette was amplified per PCR without its own ATG from plasmid pcDNA3.1-Zeo(+) with the following PCR primers.

Zeo-F: 5'-GGCAAGAATTCGCAGAGCAGTCATTATGGCCAAGTTGACCAGTGCCGTTCC-3' (SEQ
ID No. 27) Zeo-R: 5'-GGACCGCTCGAGTCAGTCCTGCTCCTCGGCCAC-3' (SEQ ID No. 18) The resulting PCR product is 412 bp long.

The pBS_Neo_P-_R+L_Arm_2B plasmid was digested with EcoRI and XhoI and ligated with PCR-amplified Zeocin cassette, which contains on its ends restric-tion sites of EcoRI and Xhol. The plasmid generated from this ligation was called pBS_Zeo_P-_R+L_Arm (Fig. 10) (SEQ ID NO. 16) and used for target-ing the second PrP allele. This second PrP K.O. construct pBS_Zeo_P-R+L Arm was linearized with BamHI for transfection as described in B.4.

B. Introduction of the second PrP K.O. construct, pBS Zeo P R+L Arm into PrP
K.O. cells with one allele of PrP aene: The second PrP K.O. construct, pBS_Zeo_P_R+L-Arm was introduced into identified taraeted cells with one taraeted PrP allele using Lipofectamin 2000 CD reagent (see transfection method page 19). 48h after transfection, cells were plated onto 10 cm dishes or further cultured in suspension in 125 ml shaker flasks at a density between 1.25 and 1.5x106 cells/ml. For antibiotic selection, G418 and zeocin were added to culture at the following concentrations: G418 0-30pg/ml and zeocin of 0-30 pg/ml. Medium exchange was performed every second/third day for 14 - 30 days.

C. Screening for zeocin-targeted clones: When cells reached confluence under antibiotic selection, genomic DNA was prepared exactly as described for the screening of the first K.O. (page 25,D).

The PCR screening strategy was similar to that of Example 1D except that the reverse primer for PCR was designed to be homologous to the zeocin ORF.

A PCR mix was prepared containing 20-300 ng genomic DNA, lx PCR buffer, 200 nM each oligonucleotide primer, 1,25 mM MgC12, 200 nM dNTP and 1,33 units Expand Fidelity Polymerase (Roche). The sequence of the synthetic oli-gonucleotides used for screening was as follows:

Zeo-K.O.-F2: 5' CTCCTCTTCCTCCCATCTTACC 3' (SEQ ID NO. 19) Zeo-K.O.-R2: 5' CGAAGTCGTCCTCCACGAAGTC 3' (SEQ ID NO. 20) The cycle parameters were as follows: 94 C 4 min; [94 C, 15 s; 63 C, 30 s;
72 C 100 s] x16 cycles; [94 C, 15 s; 63 C, 30 s; 72 C 100 s + 3 s longer in each successive cycle] x24 cycles; 72 OC 7 min. Genomic DNA samples yield-ing a 2.3 kb PCR product indicate the presence of one or several targeted alleles in the cell population.

D. Characterization of targeted clones with two targeted PrP allele:

1.Genomic PCR screening (Figure 11): To identify if the cells or cell mix popu-lation bear two targeted PrP alleles, two independent PCR screening should be performed. PCR primer pair Zeo-K.O.-F2 and Zeo-K.O.-R2 was used to prove targeted integration of the second PrP Ko.O. construct (pBS_Zeo_P-_R+L_Arm), while primer pair K.O.-F1 and K.O.-F2 proves targeted integration of the first PrP K.O. construct (pBS_Neo_P-_R+L_Arm_2B).

Genomic DNA was isolated from mixed cell populations or isolated cell colonies as described in C above. All PrP K.O. screening PCR reactions were performed with positive and negative controls. The template for the positive control was 200 ng genomic DNA from a cell population mix in which a targeted PrP allele had been previously detected with the PrP K.O. PCR screening method at first with primer pair Zeo-K.O.-F2 and Zeo-K.O.-R2; the negative control used water instead of genomic DNA as a template. The appearance of a 2.3 kb PCR
product on agarose gel after electrophoresis indicates that the corresponding cell colony bears cells carring a zeocin cassette instead of the PrP gene as the result of the second K.O. round. The results are shown in Figure 11. Clones identified as PCR positive with primers Zeo-K.O.-F2 and Zeo-K.O.-R2 were further analyzed by PCR with K.O.-F1 and K.O.-R1 in order to determine if the target integration of the first K.O. construct is still present or whether it has been replaced by the second K.O. construct (Figure 11).

In the case that both PCR are positive, this clone will be further analyzed with Southern-blot analysis.

2. The Southern blot screening strategy (Figure 12) was similar to the one described in Example 1.

The genomic Southern strategy with construct pBS_Zeo-_P-_L+R_Arm is de-picted in Figure 11. After EcoRI genomic digest and radioactive hybridization, targeted PrP clones showed a 4,2 kb band with the 5'-DNA probe and a 5,7 kb band with the 3'-DNA probe, while the wild-type band from non-targeted al-leles, detected with either probe, is 10.8 kb. With the Zeo-probe a 5,7 kb DNA
band should be detected for PrP targeted allele and no band can be detected for wild-type 293F cells.

The difference of Southern pattern for the targeted integration of both K.O.
constructs could only be detected with the 3'-probe and Zeo- or Neo-probe.

= With the 3'-probe: With the first K.O. construct a 6,5 kb band should be detected, while with second K.O. construct a 5,7 kb band should be addi-tionally shown. In both cases the WT band is 10,8 kb.
= With the Neo-probe: A 6,5kb band should be detected upon integration of the first K.O. construct.

= With the Zeo-probe: No signal should be detected for targeted integration of the first K.O. construct, while for the second K.O. construct a 5,7kb band should light up.
= DNA probes for Southern analysis and their expected pattern are summa-rized in the following table:
Probe Length Sequence of PCR primers for DNA Expected Expected South-name of probes Southern ern pattern for DNA pattern for PrP PrP targeted probe targeted cells cells on both on one allele allele 5'-probe 251 5'-K.O.-Fl (SEQ ID No. 12): 4,2 kb and 4,2 kb and 10,8 1 bp 5'-AGCTTTACCGTCCAGTCTTC-3' 10,8 kb kb 5'-K.O.-Rl (SEQ ID No. 13):
5'-GGTCTTGATGG CGATAACTC- 3' 5'-probe 252 5'-K.O.-F2(SEQ ID No. 14): 4,2 kb and 4,2 kb and 10,8 2 bp 5'- GAGTTATCGCCATCAAGACC-3' 10,8 kb kb 5'-K.O.-R2 (SEQ ID No. 15):
5'- CATGAGAACCAACGCTAGAG-3' New 3'- 288 New-3'-probe-forward (SEQ ID No. 6,5 kb and 5,7 kb, 6,5 kb probe bp 28): 10.8kb and 10.8kb 5'- CTAGAGGTCCAGGTCATCTTG -3' New-3'-probe-reverse (SEQ ID No.
29):
5'- TCAGGGAAATTGGGGATCCTG-3' Neo- 304 Neo-Probe-F (SEQ ID No. 21): 6,5 kb 6,5 kb probe bp 5'-AGCGAGCACGTACTCGGATG-3' Neo-probe-R (SEQ ID No.22):
5'-AAGCACGAGGAAGCGGTCAG-3' Zeo- 376bp Zeo-Probe-F (SEQ ID No.23) No band 5,7 kb probe 5'-ATGGCCAAGTTGACCAGTGCCG-3' Zeo-Probe-R (SEQ ID No. 24) 5'-GTCAGTCCTGCTCCTCGGCCAC:-3' 3. FISH analysis: It was observed for in house serum-free adapted wild type 293F cells that the copy number of chromosome 20 increased with culture time. It is important to exclude that the K.O. cells bearing two PrP targeted alleles still contain a third intact chromosome 20 with a wild locus PrP.
There-fore, the identified K.O. cells bearing two targeted PrP alleles should be further hybridized with the BAC 186 probe and chr20 WCP painting probe. For a com-plete PrP K.O. cell line, no more than 2 BAC signals should be observed.

4. ELISA Assay: Performed as described in method 9. No PrP protein should be detectable in pf293F cells.

5. Immunostaining: No signal with antibody staining on the cell surface in case of pf293F cells.

Example 3: Serum-free transfection of the prion-free 293F cell line pf293F for generation of prion-free recombinant proteins.

3.1 Transfection of the pf293F cells with one PrP taraeted allele.

The PrP KO clones with one PrP targeted allele were transfected with pcDNA3.1-FVIII (SEQ ID NO. 26), pcDNA3.1-FIX (SEQ ID NO. 17) and pcDNA3.1-hyg(+)-G-CSFb (SEQ ID NO. 25) vectors to express human FVIII, FIX and G-CSFb proteins.

In Figure 13 it was shown that the transfection efficiency upon deletion of the first PrP allele is comparable to wt 293F cells (Fig.10). Expression of active units of FVIII, FIX and G-CSFb per 10E6 cells are comparable to 293F wild type cells (Fig.14). This indicates knock out of one PrP gene neither influence transfection ability of the cells nor decrease the amount of produced recombi-nant protein.

3.2 Serum-free transfection and production of human recombinant proteins in complete prion-free 293F cell lines Successful ablation of the prion ORF resulted in the generation of the novel cell line pf293F, which is capable of producing completely prion-free therapeu-tics. In order to prevent contamination of the new cell line with prions result-ing f. e. from media substances, the whole procedure (including stable trans-fection, cell culturing and fermentation) was performed under serum-free conditions (see PCT/EP2006063705 and utilizing the ClonePixFL approach.

In particular, from the vector pTG36 as disclosed in WO01/70968 and pcDNA3.1-FIX, a 1,4 kb fragment containing the open reading frame of the human clotting factor IX was cut out by double-digestion with HindIII and NotI. This fragment was ligated to the 5,6 kb fragment of the HindIII and NotI
double-digested vector pcDNA3.lHygro(+)-zz (derived from V870-20, Invitro-gen) resulting in the vector pcDNA3.1-FIX shown in SEQ ID NO. 17. 28 ml suspension culture was prepared with a cell density of 106 viable pf293F
cells.
A lipid-DNA complex was prepared by diluting 30 pg of plasmid DNA in Opti-MEM I (Invitrogen) to a total volume of 1 ml, and 40 pl of 293fectin was diluted in Opti-MEM I to a total volume of lml. After the 5 min incubation at room temperature, diluted DNA was added to 293fectin to obtain a total volume of 2 ml. The transfected samples were incubated for 20 min at room temperature in the dark. 2 ml of the transfection mix was added to the 28 ml pf293F suspension culture (final cell density is 1 x 106 cells/ml). The trans-fected pf293F cells were incubated at 37 C/humidified atmosphere of 8% CO2 in air on an orbital shaker rotating at 125 rpm for 72 h.

pf293F cells transfected as described above were seeded in semi-solid methyl-cellulose based medium containing an appropriate antibiotic, for selection of clones, and a labelled antibody for detection of the highest producer clones via fluorescence. Large numbers (thousands) of clones were analyzed using ClonePixFL (Genetix) with respect to the cell number and to secretion of the target FIX protein in order to subsequently pick only a few hundred best FIX
producer clones. In contrast to other known methods, where non-producer clones and mixed clones are randomly picked as well, the use of ClonePixFL
allows picking of fast growing clones, which are high producers only, origi-nated from single cells. The picked cells are expanded in microtiter plates and later in spin tubes, cell culture flasks and fermenters under serum-free condi-tions for the complete procedure.

FIX production clones identified by the method described above were cultured in serum-free FreeStyle 293 expression medium. The target proteins were isolated and purified according to standard procedures. Additionally, for pro-duction of safe therapeutics, the optimised purification procedure as disclosed in PCT/EP 2006/061148 including f. e. SD-treatment could be utilized.

Pf293F cells will also be used for serum-free production of recombinant human FVIII and G-CSF with expression vector pcDNA3.1-hyg(+)-G-CSFb (SEQ ID
No. 25) and pcDNA3.1-FVIII (SEQ ID No. 26).

Sequence Listing:
SEQ ID NO. 1: K.O. vector pBS_Neo_P-_R+L_Arm_2B
Molecule: 10623 bps DNA, circular Type Start End Name/Description GENE 673 2580 PrP left arm GENE 2608 3807 Neo GENE 3817 8411 PrP right arm GENE 9637 10494 AmpR

SEQ ID Nos. 2 and 3: PrP left and right arm SEQ ID Nos. 4-15: primer SEQ ID NO. 16: K.O. vector pBS_Zeo_P-_R+L_Arm Molecule: 9790 bps DNA, circular Type Start End Name/Description GENE 673 2581 PrP left arm GENE 2605 2976 Zeo GENE 2983 7583 PrP right arm GENE 8805 9662 AmpR

SEQ ID NO. 17: pcDNA3.1-FIX, Molecule: 6960 bps DNA, circular Type Start End Name/Description REGION 209 863 CMV promoter REGION 895 911 MCS"
GENE 939 2324 hFIX
GENE 2328 2339 SV40'/SV40 polya + intron REGION 2340 2370 'MCS
REGION 2381 2595 BGH pA
REGION 2658 3071 fl origin REGION 3136 3460 SV40 promoter GENE 3478 4501 HygR
REGION 4514 4886 SV40 pA
REGION 5819 5146 C PUC origin GENE 6824 5964 C AmpR(complementary strand) SEQ ID Nos 18-24: primer SEQ ID NO. 25: pcDNA3.1-hyg(+)-G-CSFb Molecule: 6237 bps DNA Circular 209 863 CMV promoter Region 970 1584 GCSFb Gene 1658 1872 BGH pA Region 1935 2348 fl origin Region 2413 2737 SV40 promoter Region 2755 3778 HygR Gene 3791 4163 SV40 pA Region 5096 4423 C PUC origin Region 6101 5241 C AmpR Gene SEQ ID NO. 26: pcDNA3.1-FVIII
Molecule: 9975 bps DNA Circular 1 655 CMV promoter Region 783 3082 hFVIII Gene human FVIII domains Al and A2 783 839 signal peptide Region signal peptide for hFVIII
840 1826 Al Region hFVIII Al domain 1977 2972 A2 Region hFVIII A2 domain 3084 3107 hinge Region 3105 5162 hFVIII Gene human factor FVIII domains A3, Cl, C2 3243 4226 A3 Region hFVIII A3 domain 4227 4670 C1 Region hFVIII C3 domain 4683 5141 C2 Region hFVIII C2 domain 5188 5402 BGH pA Region 5465 5878 fl origin Region 5943 6267 SV40 promoter Region 6285 7308 HygR Gene 7321 7693 SV40 pA Region 8626 7953 C pUC origin Region 9631 8771 C AmpR Gene SEQ ID Nos 27-29: primer

Claims (17)

1. A prion protein (PrP)-free, immortalized, somatic, human cell line wherein both alleles of the PrP gene have been completely deleted.

2. The cell line of claim 1 which (i) is capable of being transfected and being cultured under serum-free conditions; and/or (ii) has integrated adenoviral sequences into its genome; and/or (iii) is derived from a starting cell selected from the group of kidney, bladder, liver, lung, cardiac muscle, smooth muscle, ovary and gastroin-testinal cells, preferably the starting cell is a human kidney cell line;
and/or (iv) is suitable for the production of recombinant proteins.

3. The cell line of claim 2, wherein the kidney cells are human foetal kidney cells, preferably the foetal human kidney cells are selected from 293 cells (ATCC CRL-1573; DSM ACC 305), FreeStyle 293 cells (293F cells; Invi-trogen R79007) and 293T cells (DSM ACC 2494), preferably are 293F
cells (Invitrogen R79007).

4. The cell line of any one of claims 1 to 3, wherein the alleles of the PrP
gene have been completely deleted by homologous recombination with knock-out traps carrying selectable or selection marker genes so that ex-pression of the selectable or selection maker is driven by the endogenous PrP promoter.

5. The cell line of any one of claims 1 to 5, including but not limited to the final prion-free 293F cell line pf293F and all the intermediate mixed popu-lations and isolated clones necessary to isolate it.

6. A method for producing the PrP-free, immortalized, somatic human cell line of any one of claims 1 to 5 which method comprises subsequently de-leting the PrP ORF in a respective starting cell by homologous recombina-tion with several different PrP knock-out constructs or with the same con-struct, and performing antibiotic selection at progressively increasing concentrations.

7. The method of claim 6, wherein the knock-out constructs carry the same or different promoterless selection marker genes or selectable markers flanked by two sequences homologous to the insertion site within the PrP
gene of the starting cell.

8. The method of claim 7, wherein the knock-out constructs (i) carry different selection marker genes; and/or (ii) further carry one of the following functional sequences: a poly A se-quence, recombinase recognition sequences, IRES; and/or (iii) the homologous sequences have a length of 1 to 10 kb, preferably about 6 kb, and/or correspond to sequences upstream and downstream of the PrP ORF of the starting cell line, most preferably the homologous sequences are those shown in SEQ ID NOs:2 and 3; and/or (iv) the selection markers encode positive selection markers including, but not limited to, neomycin, zeocin, hygromycin; and the selectable marker include fluorescence markers such as GFP and Dsred and en-zymes such as LacZ.

9. The method of claim 8, wherein the knock-out constructs have one or more of the sequences shown in SEQ ID NOs:1 and 16.

10. A PrP knock-out construct as defined in any one of claims 6 to 9.

11. Use of the PrP-free immortalized human cell line of any one of claims 1 to for PrP-free recombinant production of a human protein, or antibody or a derivative or mutant thereof (target protein).

12. A method for preparing a human cell line for PrP-free recombinant pro-duction of a human protein, or antibody or a derivative or mutant thereof (target protein), which comprises transfecting a PrP-free immortalized human host cell line of any one of claims 1 to 5 with a transfection vector comprising an origin of replication, and a gene encoding said human tar-get protein, whereby the gene for the human target protein is linked at its 5' end with a promoter and its 3' end with a polyA signal.

13. The method of claim 12, wherein (i) the transfection is performed under serum-free conditions; and/or (ii) the transfection vector is derived from pcDNA3.1 vector from Invitro-gen; and/or (iii) the human target protein is a blood clotting factor, such as blood clotting factor VIII (including wt factor VIII or a B domain-deleted factor VIII), blood clotting factor IX, factor VII/VIIa, a human growth factor like f. e. G-CSF or GM-CSF, vWF or alpha-1-antitrypsin (A1AT) or a human antibody.

14. A PrP-free immortalized human production cell line stably transfected, preferably under serum-free conditions, with the transfection vector as defined in claim 12 or 13.

15. A method for PrP-free recombinant production of a human target protein which comprises culturing, preferably under serum-free conditions, the PrP-free immortalized human production cell line of claim 14.

16. A prion protein (PrP)-free, immortalized cell line wherein both alleles of the PrP gene have been completely deleted, selected from HEK 293F or Per.C6 cells (immortalized human foetal Retina cells, CHO (Chinese Hamster Ovary cells) and BHK (Baby Hamster Kidney cells) cells.

17. Use of the PrP-free immortalized cell line of claim 16 for PrP-free re-combinant production of a human protein, or antibody or a derivative or mutant thereof (target protein).