PROCESS FOR PRODUCING AND PURIFYING FACTOR VIII

AND ITS DERIVATIVES

Background Art

Factor VIII is a plasma glycoprotein involved in blood coagulation.

Deficiency or abnormality in its function results in severe hereditary disease called

hemophilia A (Eaton, D. et al., 1986, Biochemistry 25: 505-512; Toole, J. J. et al, 1984,

Nature 312: 342-347; Vehar, G. A. et al., 1984, Nature 312: 337-342). Up to now, the

only treatment for hemophilia A has been intravenous administration of factor VIII

prepared from human blood or a recombinant source. Due to the safety reason,

recombinant factor VIII has been preferred to plasma derived factor VIII. However,

since expression level of factor VIII is 2-3 order magnitudes lower than other

molecules in the same expression system (Lynch C. M., 1993, Human Gene Therapy

4: 259-272), recombinant factor VIII production has not met its demand.

Several attempts have achieved an improved expression of factor VIII by

removing B-domain which has been known not to have any function in the cofactor

activity of factor VIII (Eaton et al., 1986, Biochemistry 25:8343-8347; Burke, R. L. et al.,

1986, J. Biol. Chem., 261: 12574-12578; Toole, J. J. et al., 1986, Proc. Natl. Acad. Sci.

USA, 83: 5939-5942; Fay et al., 1986, Biochem. Biophys. Acta, 871:268-278). U. S. Pat.

No. 5,661,008 and WO-A-91/ 09122 described B-domain deleted versions of factor

VIII, which is similar to the shortest form of plasma factor VIII. U. S. Pat. No.

5,112,950 and U. S. Pat. No. 7,041,635 disclosed the single chain forms of B-domain

deleted factor VIII molecules.

As a choice for the mammalian cell expression, Chinese Hamster Ovary

(CHO) cell expression system has been used in producing many therapeutic proteins

including factor VIII (Chu, L et al., 2001, Curr. Opin. Biotehnol., 12: 180-187). The

characteristics of CHO cell line are elucidated. It can grow either in anchorage

dependent manner or in suspension manner, adapt to either serum-containing

medium or serum-free medium, and especially support post-translational

modifications of proteins nearly identical to the human glycosylation patterns

(Brooks S.A., 2004, MoI. Biotechnol., 28: 241-255; Jenkins, N., et al., 1996, Nat.

Biotechnol., 14: 975-981; Chen, Z., et al., 2000, Biotechnol. Lett., 22: 837-941; MoIs, J.,

et al., 2005, 41: 83-91). CHO cell lines producing therapeutic proteins have been

usually cultured in the animal-derived protein free medium for the purpose of

addressing safety concerns about transmission of animal derived virus or prion and

for the purpose of easier purification. (Chu, L., et al., 2001, Curr. Opin. Biotehnol., 12:

180-187). However, removal of serum from the cultivating media also deprives the

naturally contained protease inhibitors in a serum supplement and makes it difficult

to maintain the viability of the cells during the production processes (MoIs, ]., et al.,

2005, 41: 83-91; Sandberg, H., et al., 2006, Biotechnol Bioeng., 95: 961-971).

Reduced viability and stressful conditions seem to increase the production of

secreted or released proteases from dead cells which can attack the therapeutic

proteins and cause heterogeneity. Heterogeneity caused by internal cleavages of

therapeutic protein might be the major problem because cleaved proteins can be

inactive and make it difficult to maintain "lot to lot" consistency during the

production and purification processes. Therefore, it is important to maintain a

relatively low level of protease or to prevent protease activity during production.

A few successful efforts to prevent this proteolysis caused by released

proteases from CHO cell line during culture have been reported, even though

universal inhibitor(s) which could apply to all therapeutic proteins produced in

CHO cell line has not yet been found. Satoh M et al. reported the presence of cystein

and serine proteases released from CHO cell. Chotteau et al. (Chotteau, V., et al.,

2001, in Animal cell technology: from target to market, Kluwer Academic publishers, pp.

287-292) found that an unidentified, extracellular metal-dependent protease from

CHO cell culture medium was responsible for the proteolysis of truncated factor VIIL

In WO- A- 90/ 02175, it is disclosed that some serine or cysteine proteases from CHO

cell culture can be blocked by the inhibitor peptides, which increase factor VIII

productivity. EP A 0 306 968 discloses addition of aprotinin to culture medium

increased expression of factor VIII in CHO cell medium by three times.

In U. S. Pat. No. 5,851,800, inventors claimed the inhibitors of

metalloproteases and chymotrypsins could reduce detrimental effect on factor VIII

production in cell culture. Sandberg H. et al. characterized two types of proteolytic

activities released by CHO cells in a cell culture. One was originated from

metalloproteinases, and the other from serine protease. Only metalloproteinases was

found to have a strongly negative effect on the factor VIII activity. However, even

though inhibitor of metalloproteases such as EDTA and 1,10 o-phenantroline could

block the factor VIII cleavage as described by Sandberg H. et al., these inhibitors

cannot be directly added into the CHO cell culture medium due to its toxic effect on

cells, judged from our experiments.

AU the above-mentioned protease inhibitors and commercially available

protease cocktail which contain inhibitors against serine, cystein, aspartic and

aminopeptidases such as aprotinin, bestatin, leupeptin, E-64 and pepstatin A have

been applied to our single chain factor VIII derivative (described in U. S. Pat. No.

7,041,635) culture, but we found that none of them were effective in protecting our

factor VIII derivative from cleavage by released protease (s) from CHO cell culture

during the culture.

U. S. Pat. No. 6,300,100 discloses sulfated polysaccharide such as heparin

protected an intact Tissue Factor Pathway Inhibitor (TFPI) from cleavage by

proteases present in the culture medium. In addition, U. S. Pat. No. 5,112,950

discloses sulfated dextran to substitute the stabilizing effect of Von Willebrand factor

on factor VIII in serum free media. However, to our knowledge, there has been no

report on the inhibitory effect of dextran sulfate against proteases in connection with

factor VIII molecules.

The present invention aims to demonstrate the protective effect of dextran

sulfate on the cleavage of Factor VIII or its derivatives from proteases produced

during CHO cell culture.

Advantageous Effects

In one aspect of this invention, there is provided a process for the production

of Factor VIII or its derivatives in a mammalian host cell line adapted to serum-free

media which is supplemented with dextran sulfate. Addition of dextran sulfate in

culture media effectively reduced or blocked factor Vlll-cleaving activities of (a)

certain protease(s) originated from CHO cell culture media and concurrently

increased homogeneity of the produced factor VIII molecules. In another aspect of

this invention, there is provided an efficient method for purifying factor VIII

molecules from dextran sulfate-containing media using monoclonal antibody-based

purification steps.

This invention relates to an effective inhibitor which can protect our single

chain factor VIII derivatives described in U. S. Pat. NO. 7,041,635 from cleavage by

protease(s) released during a mammalian host cell cultivation and to increase the

homogeneity of the produced factor VIII derivative. Also this invention relates to a

method of purifying the factor VIII without being affected by addition of the

protease inhibitor.

The mammalian host cell may be any animal cell which can express

recombinant factor VIII ₇ and is preferably an animal cell where a desired

transformed cell can be easily separated, for example, a Chinese hamster ovary

(CHO) cell, BHK cell, or COS cell, and more preferably a CHO cell.

In the previous patent U. S. Pat. No. 6,300,100, there was a description about

the protective effects of sulfated polysaccharides on a target protein against certain

proteases, in which a target protein was Tissue Factor Pathway Inhibitor (TFPI).

Therefore, we tested whether those sulfated polysaccharides could protect our target

molecule-factor VIII. This invention showed that only dextran sulfate possesses a

very strong protective effect on factor VIII cleavage when added to culture medium

during cultivation process.

Dextran sulfate can be obtained from bacterial fermentation or chemical

synthesis. The molecular weights of dextran sulfate can vary from 20 to 5,000 kDa in

molecular weight, and is preferably 50 to 2,000 kDa.

Sulfur content of dextran sulfate can also vary depending on its source

material. Regardless of the sulfur content of dextran sulfate, it can be employed to

this invention only if the dextran sulfate can protect factor VIII from cleavage by

ρrotease(s) released from a cell cultivation process. The sulfur content of dextran

sulfate is preferably in the range of 5 to 20 wt% of sulfated saccharide, more

preferably more than 17 wt%,

Depending on the expression level of factor VIII and its host cell line, the

amount of dextran sulfate added to a growing media can be adjusted and not limited

to those showed in preferred embodiments of this invention.

In one preferred embodiment of this invention, factor VIII molecule is one

of the factor VIII derivatives, named dBN(64-53)(hereafter called 12GdBN) ₇ which is

disclosed in U. S. PAT. NO. 7,041,635. This factor VIII derivative has internal

deletion in part of B-domain and N-terminal part of A3 and was designed to have a

new N-glycosylation recognition sequence in its fusion site. As the method described

in example 6 in U. S. PAT. NO. 7,041,635, a CHO cell line stably expressing the

12GdBN was prepared and adapted to commercially available serum-free media.

Hereinafter this clone is designated as "#39 clone" and all cells mentioned in

examples are referred to this CHO cell line (#39).

This invention also relates to a process for purifying factor VIII or derivatives

expressed in mammalian host cell line from culture media supplemented with

dextran sulfate using an affinity chromatography. The affinity chromatography

includes affinity column which contains affinity molecules coupled to solid support

such as agarose or sepharose. The affinity molecules can be anti-factor VIII

antibodies which can be monoclonal or polyclonal and can be peptides with high

affinity to factor VIII.

Description of Drawings

Figure 1 shows the comparative effects of different sulfated polysaccharides

on protecting the cleavage of intact factor VIII.

Figure 2 shows the effects of the molecular weight and concentration of

dextran sulfate on the fragmentation of a B-domain deleted factor VIII, 12GdBN.

Figure 3 shows the effects of sulfate, dextran, and dextran sulfate on the

fragmentation of 12GdBN.

Figure 4 shows the effect of dextran sulfate on the fragmentation of 12GdBN

in perfusion culture in accordance with an embodiment of the present invention.

Figure 5 shows a Coomassie Brilliant blue R250-statined SDS-PAGE gel of

the elution fractions from an immunoaffinity chromatography in accordance with an

embodiment of the present invention.

Best Mode

This invention is further illustrated with reference to the following Examples,

but can be applied to other factor VIII molecules and other cell lines as it will be

understood by the skilled person in the art. Therefore, the following examples

should not be construed as limiting the scope of this invention.

Heparin, Low molecular weight of heparins (~3 kDa and 4~6 kDa),

Dermatan sulfate, dextran (500 kDa), sodium sulfate, dextran sulfate (500 kDa, 10

kDa, 8 kDa) were purchased from Sigma. Dextrans were derived from Leuconostoc

mesenteroides, strain B 512. Different molecular weights of dextran sulfate were

produced by limited hydrolysis and fractionation. Sulfate groups were added by

esterification with sulfuric acid under mild conditions. This dextran sulfate

contained approximately 17 % of sulfur, (http://www.sigmaaldrich.com/sigma-

aldrich/ product_inf ormation_sheet/ d6001pis.pdf)

Plating #39 CHO cell line

The above-described #39 clone, which is harboring DNA fragment encoding

12GdBN _/ was cultured in serum free media (ProCHOδ media purchased from

Cambrex). At two passages of subculture after thawing, 4 x 10 ⁵ cells were seeded in

each well of a 6-well plate.

Western blot assay

The culture medium containing expressed factor VIII was subjected to 7.5%

SDS-PAGE gel and blotted to PVDF membrane. Blotted membrane was probed with

an A2 domain-specific antibody called #26-1 which was generated by the inventors

of this invention. Secondary mouse IgG coupled with horse-radish peroxidase was

used to visualize the factor Vlll-anti factor VIII antibody complex on the blot.

Example 1

Comparison of protective effect of various sulfated polysaccharides on

fragmentation of 12GdBN

High molecular weight of dextran sulfate (about 500 kDa), heparin, two

kinds of low molecular weight heparin (~3 kDa and 4~6 kDa), and dermartan sulfate

were purchased from Sigma Co., Ltd. and resuspended in water and filter sterilized.

Cells were plated as mentioned above. Twenty-four hours after seeding, the medium

was replaced with a fresh one and five kinds of sulfated polysaccharides were added

in each well at a final concentration of 25 mg/L, 50mg/L, 100mg/L, or 200mg/L,

respectively. After 48 hours incubation, culture supernatants were collected and

analyzed through Western blot assay. As shown in figure 1, there was little

protection effect of three kinds of heparin which were effectively protecting TFPI

described in other patents. However, dextran sulfate can provide efficient protection

of cleavage and in a concentration-dependent manner. More than 92% of Factor VIII

in the culture supernatant (lane 2 in figure 1-(D)) remained intact compared to the

factor VIIIs in a culture medium with no additives (41%; lanel in figure 1-(D)) and

the factor VIIIs in a culture medium with heparins or dermatan sulfate (52%~67%;

lane 3~6 in figure 1-(D)). This shows that not all sulfated polysaccharides can protect

all the target proteins and the protective effect of a certain sulfated polysaccharide is

very specific to a target protein.

Example 2

Effect of molecular weight of Dextran sulfate on cleavage of 12GdBN

If dextran sulfate with a lower molecular weight can be applied to protect the

cleavage, a lower molecular weight of dextran sulfate may be easily separated from

factor VIII more based on its difference in size. So, to see if lower molecular weight

of dextran sulfate can protect the cleavage of expressed 12GdBN during cell

cultivation, 8 kDa, 10 kDa and 500 kDa dextran sulfate, which have the same content

of sulfur and originated from the same source, were added to the medium at varying

concentrations of 100 mg/L, 200 mg/L, 400 mg/L and 1000 mg/L. At 72 hours after

addition of dextran sulfate, culture medium was harvested and analyzed by Western

blotting assay. As shown in figure 2, although an increasing amount of dextran

sulfate with low molecular weight (lane 1 to lane 8 in figure 2) was added into the

cell culture medium, there was not observed any efficient protective effect on

12GdBN cleavage. Only 500 kDa dextran sulfate (lane 9 to lane 12) was shown to

protect the fragmentation of single chain 12GdBN.

Example 3

Only sulfated dextran can protect the cleavage.

To- see if a separate functional group of dextran sulfate has the inhibitory

effect of cleavage, equimolar amounts of dextran (500 kDa), sodium sulfate, and

dextran sulfate (500 kDa) were added in the culture medium. Cells were seeded as

described in experiments. At 24 hours after seeding, several concentrations ranging

from 250 mg/L to 1000 mg/L of dextran sulfate (500 kDa) and dextran (500 kDa) or

several concentrations ranging from 71 g/L to 384 g/L of sodium sulfate were

added to the medium. At 48 hours after addition, medium was collected from each

Il

well and analyzed by Western blot assay. As shown in figure 3, only 500 kDa

dextran sulfate (lane 5 to lane 7) showed protective effect on cleavage of 12GdBN as

depicted in figure 3. Neither dextran only (lane 2 to lane 4) nor sodium sulfate (lane

8 to lane 10) was shown to inhibit the protease activities of released protease(s)

during CHO cultivation. Fragmentation pattern of culture medium with either

dextran or sodium sulfate only was similar to that of the culture medium with no

additives (control, lane 1).

Example 4

Application of dextran sulfate to suspension culture

Dextran sulfate (500 kDa) was applied to a perfusion culture system. One

vial of cell was thawed and expanded in T75 flask and further expanded in T125

flask. Cells in T125 flask were transferred into 250 ml, 1 L and 3 L spinner flasks

serially and maintained as suspension culture on a magnetic stirrer plate at 37°C in

5% of CO ² / air mixture with a rotation speed of 100 rpm. Exponentially growing

cells in 3 L spinner flask were collected and inoculated into 7.5 L bioreactor with a

working volume of 5 L. Dextran sulfate (500 kDa) was added at the concentration of

200 mg/L in the serum free medium in the bioreactor. From the fourth day after

inoculation, the culture medium was collected every second day for 20 days. Cell

viability was maintained above 92.7% during culture perfusion, and factor VIII

fragment was detected less than 5% judged by densitometric analysis of each band

in Western blot during fermentation process. Exemplary Western blots of culture

media collected on day 6, day 12 and day 18 were included in figure 4.

Example 5

Purification of 12GdBN from culture media by immuno affinity chromatography

Due to the highly negative characteristic of dextran sulfate, ionic exchange

chromatography cannot be applied to purify secreted factor VIII in culture media

supplemented with dextran sulfate. Therefore, culture media produced in example 4

were concentrated by the tangential flow ultrafiltration and subjected to an

immunoaffinity column pre-equilibrated with equilibration buffer (20 mM of Tris-

HCl [pH 7.0], 400 mM NaCl, 5 mM CaCl ₂ , 3 mM EDTA). Immunoaffinity column

was prepared by coupling monoclonal anti-factor VIII that recognizes A2 region of

factor VIII heavy chain, to CNBr-activated sepharose resins. Factor Vlll-bound

immunoaffinity column was washed with 2.5 bed volume of equilibration buffer and

2.5 bed volume of washing buffer(20 mM of Tris-HCl [pH 7.0], 400 mM NaCl, 5 mM

CaCl ₂ , 3 mM EDTA, 10% ethylene glycol). Elution was performed in a stepwise

gradient elution with elution buffers containing ethylene glycol at the concentration

ranging from 40-60%. Nine of 12 elution fractions (elution fraction number 1 to 9

corresponded to El to E9 in figure 5) were sampled and subjected to 7.5% of SDS-

PAGE and stained with Coomassie brilliant blue R 250 dye. Only one-step

purification of the culture medium gave more than 95% of highly pure single-chain

factor VIII as illustrated in figure 5.