IE65986B1 - Method for preparing proteins - Google Patents

Abstract

A method for increasing the production of protein, especially t-PA, in cultures of cells producing this protein, where a substance which induces the production of the protein is added to the cell culture, is characterised in that thioglycolic acid, thiodiglycolic acid, L-cysteine, glutathione, butyrylcholine bromide, butyrylcholine chloride, nonactic acid, furan fatty acid, ascorbic acid, aphidicolin, 6-hydroxy-4,6-dimethyl-3-hepten-2-one, fusaric acid, mevalonic acid, trans-anhydromevalonic acid, anhydromevalonolactone, cis-anhydromevalonolactone, D- alpha -hydroxy-substituted (C3 or C4) aliphatic mono- or dicarboxylic acid or the salts thereof is used as protein-inducing substance.

Description

Method for preparing proteins This invention relates to a method tor preparing proteins, particularly plasminogen activators such as t-PA (tissue plasminogen activator) and the mutants thereof, in general, the invention provides a method of increasing the productivity of cell cultures (e.g. CHO cells) for the synthesis of proteins.

Plasminogen activators are a class of serine proteases 10 which activate the proenzyme plasminogen by cleaving the peptide bond between Arg-560 and Val-561 to produce the active enzyme plasmin. Plasmin, in turn, is the last stage of the fibrinolysis system of the blood stream which splits the fibrin structure of a clot into soluble peptides.

In pathogenetic disorders such as coronary disease, blood clots are not broken up fafft enough to ensure that the tissue receives an adequate supply of oxygen. The consequence of occlusion of the coronary arteries by a blood clot is an infarction, whilst the deficiency in the supply of oxygen to the heart muscle results in necrosis of the affected tissue.

Rapid reopening of the blood vessels can ensure that there is a blood supply to the heart muscle in a very short time and thus prevent whole sections of the heart muscle from dying off.

Up till now, streptokinase and urokinase have been used in therapy. However, the properties of t-PA I, 59 8 6 - 2 have some advantages over these known plasminogen activators: the fibrin-specific local lysis, the absence of systemic lytic effects caused by fibrinogen breakdown, high rates of reperfusion and the absence of antibody formation. t-PA is a glycoprotein with a molecular weight of about 65,000 Daltons which occurs as a singlechained and two-chained enzyme and is made up of 527 amino acids.

The affinity of t-PA for plasminogen is about 100 times greater in the presence of fibrin than in the absence of fibrin. This high affinity of t-PA for plasminogen in the presence of fibrin ensures effective activation without activating any free plasminogen in the periphery.

Human t-PA was first obtained in pure form from the uterus. t-PA has also been detected in other tissues and cells, e.g. endothelial cells, including arterial and venous endothelial cells. However, the amounts of t-PA formed are so small that it is impossible to obtain them on a commercial scale.

Another natural source of t-PA is cell culture.

A number of cell lines have been investigated for possible t-PA production (Gronow M. et al., Trends in Biotechnology 1: 26-29, 1983). Fairly large quantities of t-PA were obtained for limited clinical studies and for investigation of their structure from a human cell line, namely Bowes melanoma cells (Wallen et al., European Journal of Biochemistry 132: 681-686, 1983).

However, these yields are too small for large scale production of t-PA for wide ranging clinical use.

Only by genetic engineering has it been possible to produce larger quantities of recombinant t-PA for therapeutic use.

In 1983, t-PA was cloned and the amino acid sequence was determined (Pennica, Nature 301: 214-221, 1983).

Attempts to obtain t-PA from E. coli in which the genetic information for t-PA from melanoma cells had been integrated were unsuccessful since the resulting protein is not glycosylated and consequently does not correspond to the native t-PA.

For this reason, various research groups looking into the production of human t-PA have sought out permanent mammalian cell lines from various sources.

Some of these cell lines are human cell lines with an apparently improved yield of t-PA. Another cell line is the normal Chinese hamster ovary cell.

EP-A-93,619 describes the preparation of t-PA from transformed Chinese hamster ovary cells (CHO cells).

The resulting t-PA (r-tPA) does not differ from naturally occurring t-PA in any way.

Several publications describe mutants of t-PA and the preparation thereof, e.g. EP-A 241,208, 241,209, 241,210, 240,334, 234,051, 233,013, 231,624, 225,286, 213,794, 201,153, 199,574, 196,920 and DE-A 3,708,681.

The present application refers to all t-PA derivatives as mutants, regardless of whether they contain one or more amino acids which differ from the amino acid in the same position in the naturally occurring t-PA or wherein . one or more amino acids which occur in the natural t-PA do not appear in these mutants. A t-PA in which a number of amino acid groups of the naturally occurring t-PA are absent, as in EP-A-196,920, for example, is sometimes referred to as a degraded species.

EP-A 199,574, for example, describes a t-PA mutant which has, in positions 270 to 279, certain amino acids which differ from the amino acids in the corresponding positions of the natural t-PA.

In DE-A 3,08,681, we prepared some t-PAs which have an amino acid in position 117 which differs from the amino acid in the corresponding position of the natural t-PA, but unlike the natural t-PA the mutant t-PA is not glycosylated at this position.

A number of publications have described the influence of various substances on the productivity of cell cultures for the synthesis of proteins.

In various cell systems, it is reported that there is a connection between intracellular c-AMP levels and t-PA synthesis (see for example Kooistra et al. in Thrombosis and Haemostasis, 54(1): 192, Abstract P 1133) and it is reported that dibutyryl c-AMP increases the production of t-PA in human endothelial cells by a factor of "2 to 3, but has no affect on the synthesis of t-PA in Bowes melanoma cells. In kidney cells the synthesis of t-PA was increased by means of phosphodiesterase inhibitors (Journal of Cell Biology 91: 195-200, 1981). On the other hand, c-AMP inhibits t-PA synthesis in macrophages (Vassalli et al., Cells, 8: 271-281, 1976) and is inert in teratocarcinoma cells (Nishimune et al., Experimental Cell Research 146: 439-444, 1983).

. Butyric acid induces the synthesis of t-PA in teratocarcinoma cells (Nishimune - loc.cit.) but has an inhibitory effect in renal tumour cells (Nelson . et al., Proc. Am. Ass. for Cancer Research, 26:35, 1985) and Brouty-Boye et al., report in Biotechnology 12: 10, 1984 that it has a neutral effect in embryonic lung cells.

In human endothelial cells, t-PA synthesis can be stimulated by thrombin (Levin et al., Thombosis and Haemostasis, 56(2): 115-119, 1986), Alcohol (Laug, Journal of the American Medical Association, 250(6): 772-776, 1983 B) , Vitamin C and Vitamin A (Inada et al., Biochemical and Biophysical Communications 130(1): 182-187, 1985). However, t-PA secretion in bovine endothelial cells is inhibited by thrombin (Loskutoff, Journal of Clinical Investigations, 64: 329-332, 1979) and Endotoxin (Crutchley et al., Journal of Biological Chemistry, 261(1): 154-159, 1986).

The secretion of t-PA from the pig endothelium is affected by means of heparin (Marchwardt et al., Thrombosis Research, 8: 217-223, 1976), whereas in human embryonic lung cells heparin has a minimal effect.

The synthesis of t-PA is also positively affected by concanavalin A in embryonic lung cells (BroutyBoye - loc.cit.) and epithelial cells (Electricwala et al., Thrombosis and Haemostasis, 53: 200-203, 1985), 5-azacytidine in epithelial cells (Electricwala - loc.cit.) and melanoma cells (Silagi et al., Biological Medicine, 57: 418, 1984), Tizabrine in melanoma cells (Roba et al., Thrombosis and Haemostasis 50(1): 83, 1983) and aphidicolin in hepatoma cells (Orfanoudakis et al., Biological Chemistry, Hoppe-Seyler, 336(9): 832, 1985).

It is stated in EP-A-219,270 that a combination of heparin and endothelial cell growth factor (ECGF) increases the production of t-PA and single-chained urokinase plasminogen activator from normal human diploid lung fibroblast cells in serum-free medium.

The influence of butyric acid on Chinese hamster ovary cells was reported by Storrie gt al., Journal of Cell Physiology, 94: 69-76, 1978, and Wright, Experimental Cell Research 78: 456-460, 1978. However, these investigations were carried out with untransformed cell lines and the results were by and large concerned with the changes in the morphology and growth rate of the cells.

Moreover, EP-A-0,239,292 describes the use of sodium butyrate for increasing protein production in the cultivation of cells modified by genetic engineering or of hybridomas, although only the use of sodium butyrate in antibody production is described in so many words.

However, the successful use of the aliphatic monocarboxylic acid and the salts thereof in conjunction with CHO cells could not have been predicted on the basis of the existing statements, particularly as Gorman et al. (Nucleic Acids Research, 11: 7631-7648, 1983) describe the use of sodium butyrate to increase the production of CAT only in serum-containing medium.

It has now been found that the addition of a aliphatic carboxylic acid or the salts thereof to a culture of transformed CHO cells in serum-free medium increases the protein yield.

The present invention thus relates to a method of increasing the production of a protein from cultures of transformed CHO cells, which is characterised in that a C^^-aliphatic monocarboxylic acid or a salt thereof is added to the culture and cultivation is carried out in a serum-free medium.

The present invention further relates to a method of increasing the production of proteins in cultures of transformed CHO cells, which is characterised in that the cells are cultivated in the presence of thioglycolic acid, thiodiglycolic acid, L-cysteine, glutathione, butyrylcholine bromide, butyrylcholine chloride, nonactic acid, furan fatty acid, ascorbic acid, aphidicolone, 6-hydroxy-4, 6-dimethyl-3-hepten-2-one, D-«C-hydroxy substituted (C^ or C^) aliphatic mono- or dicarboxylic acid or the salts thereof, the cells are removed from the culture and placed in fresh, serum-free culture medium and cultivated in the presence of a C^^-aliphatic monocarboxylic acid or salts thereof.

The proteins in question are preferably the plasminogen activators, particularly rt-PA and the mutants thereof. However, the effect of the method is independent of the proteins and the DNA coding for these proteins and can therefore be applied to CHO cells which are used to express any foreign protein.

Mutant rt-PAs are derivatives of t-PA which have one or more amino acid groups which differ from the amino acid groups in the corresponding position of natural t-PA, or in which one or more amino acid groups of natural t-PA do not occur. Some t-PA mutants of this kind are described in the literature, especially in the patent applications mentioned above, more particularly in EP-A 199,574 and DE-A 3,708,681.

Preferred protein production-increasing substances are thioglycolic acid, nonactic acid, butyric acid and propionic acid.

Nonactic acid and processes for the preparation thereof are described for example in Helv. Chim. Acta, Vol. XLV, (1962), No. 15-16, P. 129-138; Tetrahedron, Vol. 36, No, 1, p. 46-49 and J. Org. Chem., 1980, 45, p. 4259-4260.

The protein production -increasing substances are commercially available products or are known from the literature or may be prepared by known methods.

Butyryl choline bromide and chloride, are commercially available products and may be obtained, for example, from Sigma Chemical Co., St.

Louis, Missouri, USA.

Various publications describe furan fatty acids, such as Lipids, 1977, 12(10), 828-36, J. Chem.

Soc. Chem. Commun., 1976, (16), p. 630-1, Lipids, 1975, 10(11), p. 695-702, and Fette, Seifen, Anstrich mittel, No. 8, 1986, p. 290-292.

Useful acids of this type may also be prepared by carrying out a reaction according to the following scheme: a) TOA wherein m = 2 to 5, preferably 4, n = 7 to 12, preferably 8 or 10, The preparation of Compounds and 6 is described in Voss et al., Helv. Chim. Acta, 1983, 66, 2294.

Preferably, m is 4 and n is 8 or 10.

The preparation of trans-anhydromavalonic acid is described by H. Dieckmann in Die Isolierung und Darstellung von trans-5-hydroxy-3-methyl-pentan2-sflure, Archiv fflr Mickrobiologie, 62, 322-327 (1986). The compounds anhydromavalonic acid lactone and cis-anhydromavalonic acid lactone and the preparation thereof were described by Keder et al., in Die Bildung von /^-Anhydronmavalonsaurelacten durch verschiedene Pilze, Biochem Zeitschrift, 341, 378-386 (1965).

The compound 6-hydroxy-4,6-dimethyl-3-hepten-2one may be prepared by methods known from the art.

It is also a natural substance and can be isolated from Capiscum annuns var. angulosum, as described in C.A. 97: 159552f and Eiyo to Shokuryo 1982, (2) 95-101. This compound is also described in dissertation Neue Sekund&rstoffe aus Streptomycesten: Isolierung und Strukturaufklflrung der Colabomycine, Pyrrolame und des Pyridazomycins by R. Grote, University of GOttingen, 1987, as well as its preparation by the cultivation of Streptomyces olivaceus (strain Tfl 3082, deposited at Deutsche Sammlung f«r Rikroorganism^on 23rd November 1987 in accordance with the Treaty of Budapest under No. 4309) and its isolation. It has the following NMR spectrum: C9H16°2 <156·23) EI-MS: m/e = 138 (M+-H2O, high resolution, ΟθΗ^Ο, 6%); 123 (M+-CH5O, 8%) ; 98 (high resolution.

C6H10°' 48%)' 83 (1θ0%) 1H-NMR (200 MHz, CDClg): 6= 1.27 (s, 7-Hg u. 8-H3); 1.43 (s, broad, HO, exchanges with MeOD)j 2.21 (s, 1-H3); 2.35 (d. J = 1.3 Hz, 9-H3); 2.32 (s, broad, 5-H2); 6.13 (s, broad, 3-H) ppm 13C-NMR (50.3 MHz, CDC13): 6= 22.0 (o, C-9); 30.0 (O, 2C, C-7 u. C-8); 32.0 (o, C-l); 54.4 (u, C—5); 71.1 (u, C—6), 127.2 (o, C-3); 155.1 (u, C-4); 198.6 (u, C-2) ppm Examples of the salts of the acids specified include, in particular, salts of the alkali metals, preferably the sodium and potassium salts.

The method according to the invention may be carried out in a serum-containing medium or preferably in a serum-free medium.

If the method is carried out in a serum-containing medium, it is preferable to use, for example, thioglycolic acid, nonactic acid or furan fatty acid or the salts thereof.

If the method is carried out in a serum-free medium, it is preferable to use a C33 aliphatic monocarboxylic acid, more particularly butyric or propionic acid, or the salts thereof.

It has been found that the compound which increases protein production exhibits effectiveness at concentrations of from 0.005 mcM to 500 mM, preferably from 10 racm to 500 mM, e.g. from 1 mM to 10 mM.

(C3_5) aliphatic monocarboxylic acid, thioglycolic acid, thiodiglycolic acid, L-cysteine and glutathione or the salts thereof are particularly effective at 1 mM to 10 mM.

The substances which increase protein production may be added on the corresponding day of cultivation, e.g. on day 0, i.e. at the start of cultivation, up to the end of the third day of cultivation.

The inducer substances which increase' protein production may also be used in several stages.

Thus, the cells may be cultivated in the presence of a first one or more inducers, the cells are then removed from the culture, placed in fresh cultivation medium and cultivated in the presence of a second one or more inducers which increases protein production. In this multi-stage process, Cg_5 aliphatic monocarboxylic acids or the salts thereof, preferably butyric acid and thioglycolic acid or their salts, in particular, show an enhancing effect.

With multi-stage use, it is not necessary to use the same inducer for all stages; rather, one of the inducers may be used in one stage and another inducer in another stage. This multi-stage process also shows an enhancing effect, such as, in particular, the use of aphidicolin in one stage and one of the other inducers e.g. butyric acid, propionic acid, butyryl choline chloride or butyryl choline bromide or the salts of any of these, in another stage.

Combinations of two or more substances may be used. Preferably, not more than one substance is used which inhibits cell growth. For example, a aliphatic monocarboxylic acid, preferably butyric acid and thioglycolic acid or the salts thereof may be used, the roonocarboxylic acid or the salt thereof preferably being added to the cell culture as the second substance.

Transformed Chinese hamster ovary cells are CHO cells which are transformed with a vector coding for a desired protein, these cells being capable of synthesising and expressing the protein under cultivation.

The transformed Chinese hamster ovary cells used may be, for example, the cells described in EPA 93,619 and EP-A-199574 and in DE-A-3,708,681.

By cultivating these transformed CHO cells it is possible to prepare t-PA (EP-A 93,619) and certain t-ΡΆ mutants (EP-A 199,574 and DE-A 3,708,681).

The transformation of CHO cells with vectors whilst the cells express other proteins is effected by methods known from the literature, e.g. in the European and German patent applications referred hereinbefore.

The host cell used may be the cell known by the name CHO-Kl which was isolated in 1957 from biopsy material by Puck and was included in 1970 in the American Type Culture Collection (ATCC) under the name CCL-61. This cell line was deposited again by the ATCC on 23rd December 1987 in accordance with the Treaty of Budapest under No. CRL 9618.

The DNA sequence coding for the desired protein may be prepared by the method known from the literature, e.g. in some of the above-mentioned patent applications. Vectors which contain this sequence and controlling sequences as well may be constructed by methods known per se, e.g. using E. coli K 12294 (ATCC 31446) and E. coli X 1776 (ATCC 31537), which were re-deposited by the ATCC on 23rd December . 1987 in accordance with the Treaty of Budapest under Nos. 53704 and 53705; examples of such vectors include the plasmids pETPER and pETPFR, which are described in EP-A 93,619. The CHO cells may be transformed by the methods known from EP-A 93,619, DE-A 3,708,681, EP-A 117,059 and EP-A 199,574 or by other methods known from the literature.

The vector for the t-PA production was integrated in a subclone, CHO-Kl-DUX Bll (Urlaub et al., Proceedings of the National Academy of Science, USA, 77: 4216-4220, 1980). This cell line, a dihydrofolate reductase negative mutant (DHFR-), is suitable for a DHFR-dependent selection of the transfected cell clones. The vector for human c-DNA may be a known plasmid vector pBR-322-or the Escherichia coli derivative thereof. The promoter region for the gene in question and for the DHFR gene comes from SV 40 and the termination region for both genes comes from a hepatitis B surface antigen.

In the Examples, CHO-Kl-DUX Bll cells are used which are transformed by the plasmids pETPER or pEPEFR (see EP-A 117,059).

Various known products are available for use as the serum-free medium (see for example BMFT—Status Seminar 12, held on 13th November 1985, Federal Ministry of Research and Technology, 5300 Bonn 2, page 111-120). In the following Example, a DMEM/Ham’s F 12 Medium (1:1) (Gibco) is used as the serum-free medium. As the additive, foetal calves' serum (FCS) is used in a concentration of 7.5% or 1-2%.

The cell cultures are incubated at 37°C in a Hereaus incubator and supplied with a mixture of carbon dioxide and air containing 7.5% CO2 with a relative humidity of 100%. Depending on the batch, different culture vessels are used: spinner vessels for 100-1000 ml of medium, Roux dishes (Nunc) 175 cra^ (70-100 ml 2 of medium), 75 cm (30-50 ml of medium), 25 cm (7-10 ml of medium) and multiple culture dishes made by Costar with 2 cm for 1 ml cultures. For maintaining the strain, the cells are used in a density of 0.2-0.3 x 10® cells/ml and subcloned every 2-3 days.

The substances which stimulate protein production may be added in a 1:100 dilution at the various stages, i.e. -they are first dissolved in medium or, if they do not dissolve readily, they are first dissolved in ethyl alcohol and then added to the medium and then the medium containing this substance is added to the cell-containing medium in the ratio 1:100 v/v.

In the Examples which follow, cells were taken from an exponentially growing culture in order to stimulate the synthesis of t-ΡΆ and centrifuged in plastic test tubes (Falcon) for 10 minutes at 1000 x g in a Beckmann T 3-6 centrifuge. After the supernatant had been decanted the cell pellet was resuspended with fresh medium.

For the production of t-PA in serum-containing medium, F12/DMEM + 7.5% FCS + gentamycin was used as the fresh medium. The resulting suspension 2 may be prepared in 2 cm 24-well dishes at a cell density of 0.2 to 0.3 x 10® cells/ml.

For the production of t-PA in serum-free medium, the cells were first adapted for 3 to 4 days in a medium of 1 to 2% foetal calves* serum, centrifuged again and the cell pellet was resuspended in serum2 free medium and then transferred into 2 cm 24well dishes in a cell density of 0.4-0.5 x 10® cells/ml.

Quantitative t-PA determinations ml Aliquots of cell supernatants and cell extracts as appropriate were taken or prepared at various times. % increases given below and in the figures are compared with the control (0%). The samples were measured either immediately or after storage at -20°C using the following methods: METHOD OF DETERMINATION - Direct chromogenic assay (DCA) p-Nitroaniline is formed as the reaction product from the synthetic peptide substrate S-2288 (D-H-IlePro-Arg-p-nitroaniline) of Messrs. Kabi Diagnostica using t-PA. The activity of t-PA is proportional to the formation of p-nitroaniline, measured at 405 mn, 37°C.

Spectrophotometric determination of the enzyme kinetics was carried out automatically with an ACP-5010 analyser made by Messrs. Eppendorf.

The activity of the enzyme was calculated using this given standard. A t-PA standard series of 1-15 mcg/ml was prepared using a laboratory standard consisting of 72% of 1-chained material.

The substrate solution (100 mM tris/106 mM NaCl) was used in a concentration of 50 mM.

- ELISA Using an ELISA (Enzyme linked immuno-sorbent assay) the t-PA samples of the supernatant and the cell extract were quantitatively determined. The standard used was a laboratory standard in a concentration series of 0.038-5 mcg/ml. The samples were diluted beforehand to dilutions of 1:10 to 1:1000.

Characterisation of the transformed CHO cells Cell growth The CHO cells are inoculated with 0.2-0.3 x 10® celIs/ml in culture dishes (in medium containing 7.5% FCS) and initially replicate very slowly.

Only after a dormant phase lasting 24 hours does logarithmic cell growth begin. With a cell density of 1.3 + 0.04 x 10® cells/ml, the first non-vital cells appear in the culture on the fourth day; their proportion in terms of the total number of cells is 13.2 + 1.3%. The maximum cell density is reached on the 5th day at 1.8 + 0.05 x 10® cells/ml (Fig. 1).

The number of living cells is still 72 + 3% of the total cell mass after 7 days.

The stationary phase which follows the exponential cell growth is initiated by various factors. As a result of a high population density and the using up of essential nutrients, the cells no longer divide. The duration of the stationary phase varies considerably and depends on components of the medium, such as serum. An accumulation of acids and toxic metabolic products and changes in factors of the medium (acid pH, low partial pressure, inadequate ventilation) also cause the cells to die off. t-PA production in CHO cells There is a linear relationship between cell growth and t-PA synthesis for CHO cells cultivated in medium containing 7.5% FCS.

The pattern of t-PA production in time shows two phases. In the first phase, the t-PA concentration in the supernatant increases constantly. In the second phase, enzyme synthesis comes to a virtual standstill and the t-PA concentration in the medium remains substantially constant (Fig. 2A. Values given relate to 1 ml of cell supernatant). The maximum t-PA concentration on the 4th day is 11.1 + 0.7 mcg/ml in the Elisa. The t-PA concentrations measured are somewhat higher in the chromogenic assay, presumably caused by 2-chained t-PA, but with the Elisa method the curves run parallel in both methods of detection up to the 4th day.

The intracellular t-PA concentration also runs parallel to the cell growth and is relatively low compared with the total quantity of t-PA synthesised.

On the 4th day, the proportion thereof is 7% of the total activity (Fig. 2B. Values given relate to 1 ml of cell suspension).

As can be seen from Fig. 3Ά, the t-PA synthesis increases slightly up to the 4th day, based on the number of cells; with the Elisa method, a maximum of 8.1 + 1.2 meg t-PA/1 x 10® cells is determined. Cellularly bound t-PA was determined only by the Elisa method since these low t-PA concentrations (less than 1 meg) cannot be determined by DCA.

The intracellular t-PA content, based on the cell number, is somewhat higher during the logarithmic phase of cell growth and is between 500 and 900 ng of t-PA/1 x 10® cells, and from the 4th day the t-PA value is about 450 ng/1 x 10® cells (Fig. 3B> .

Influence of serum and serum factors in the medium on cell growth and the t-PA production in CHO cells The cells were prepared in quantities of 0.25 x 10® cells/ml and at different serum concentrations in the medium.

A serum concentration of below 5% in the medium also has a negative effect on cell division, but the productivity if affected only slightly.

The influence of the serum concentration in the medium on cell growth and the t-PA production after 48 hours' incubation of the cells is shown in Table 1. The values are given in percent of the maximum values (7.5% FCS in the medium). The maximum value for the cell number is 0.56 + 0.03 x 10® cells and, for the t-PA concentration, 5.7 + 0.5 racg/ral.

The t-PA determinations were carried out by means of DCA (average + SE, n = 4) .

TABLE 1 % FCS Number of cells t-PA 40 61.5 + 2.8 59.2 + 3.9 20 79.4 + 3.8 81.2 + 2.4 10 88.6 + 5.3 90.3 + 6.7 7.5 100.0 + 6.4 100.0 + 7.9 5 72.4 + 4.9 87.0 + 11.3 1 43.1 + 1.5 79.2 + 6.7 0 37.5 + 1.8 75.4 + 4.6 J· Cell culture in a serum-free medium The cells were prepared in serum free medium in a density of 0.4 to 0.5 x 10® cells/ml. The cells divide until the 4th day and the maximum number of living cells is also achieved on the 4th day at 1.0 + 0.2 x 10® cells/ml (Fig. 4). t-PA production in serum-free medium The t-PA concentration in the supernatant increases constantly until the 7th day and, by the Elisa method, reaches a maximum of 13.6 + 0.54 mcg/ml.

The t-PA values determined by DCA are below the t-PA values determined by Elisa on the 1st to 3rd days and from the 4th day onwards they are above these t-PA values (Fig. 5).

The t-PA synthesis per cell in serum free medium is approximately the same as in medium containing serum and is between 7.2 + 3.6 and 10.3 + 2.8 mcg/1 x 10® cells (Elisa) (Fig. 6).

The percentage of t-PA in serum free medium based on the total quantity of protein in the supernatant is between 10.9 + 1.0 and 12.5 + 1.4%. The results are shown in Table 2. The t-PA values were determined by Elisa (average + SE, n = 4).

TABLE 2 Day Protein mcg/lxlO® cells t-PA mcg/lxlO® cells 1 71.0 + 8.0 7.2 + 3.6 3 81.8 + 9.4 9.0 + 2.3 5 82.0 + 12.2 9.3 + 0.8 Effect of aliphatic monocarboxylic acids The increase in t-PA production in CHO cells in medium containing 7.5% FCS as a result of different aliphatic monocarboxylic acids is indicated in Table 3. All the acids were used in a concentration range of from 10 mM to 500 mM. The optimum dosage range for all monocarboxylic acids is between 1 and 10 mM. The effectiveness of the acids depends on the chain length.

All effective monocarboxylic acids inhibit cell growth. Propionic, butyric, isobutyric and isovaleric acids are effective in a wider concentration range since the concentration-dependent inhibition of the primary metabolism which results in inhibition of cell growth coincides with the increase in t-PA synthesis.

The values given in Table 3 relate to 1 ml of cell supernatant and are given in % compared with the control (0%). The t-PA was determined by DCA (average + SE, n - 3-24). The maximum value relates to an individual sample. Data for the individual samples are shown in brackets.

TABLE 3 Monocarboxylic acid Optimum < range iosage % increase in t-PA synthesis Maximum Average Formic acid 100 mcM 7.4 6.5 + 0.9 Propionic acid 5-10 mM (10 mM) 45.5 28.0 + 2.5 Butyric acid 1-5 mM ( 1 mM) 68.3 41.9 + 3.8 Isobutyric acid 0.6-10 mM (10 mM) 32.9 16.6 + 2.8 Isovaleric acid 1-20 mM ( 5 mM) 39.3 18.7 + 2.4 The average values relate to several production runs lasting 3 to 4 days, the monocarboxylic acids having been added on day 0 to day 3 of cultivation, as shown in Table 3A below. Data for the individual / 5 _ samples are shown in brackets.

TABLE 3A Monocarboxylic acid Days of addition Sample days n Formic acid 10 (0) 2-3 (2) 4 Propionic acid 0-3 (2, 1-4 (3) 24 Butyric acid 1 (3) 2-4 (3) 18 Isobutyric acid 0-2 (3) 2-4 (3) 11 Isovaleric acid 0-2 (4) 1-4 (4) 19 n = number of production runs Fig. 7 shows the time curve for the t-PA production (determined by DCA) increase in medium containing 7.5% FCS brought about by means of C3-C5 monocarboxylic acids. The concentration of butyric acid and isobutyric acid was ImM and the concentration of proprionic acid and isovaleric acid was 5mM. After 24 hours it is already possible to detect a stimulation of t-PA production from 18.1 + 5r7% (C4) to 24.8 + 0.5% (C5). In the case of isobutyric acid, there was no increase in production until 48 hours had elapsed. A maximum increase in t-PA production was achieved for the C3 and C4 carboxylic acid on the 3rd day, with 37.2 + 5.1% for Na-butyrate.

The increase in production is only short-lived for butyric and isobutyric acid and on the 4th day the effect can no longer be observed. The increasing effect of propionic acid and isovaleric acid, on the other hand, is still present on the 4th day and manifests itself in an increase in t-PA production of 31.8% for C3 and 8.5% for C5.

The effects are also dependent on the physiological state of the cells. The cells are most sensitive to stimulation during exponential growth.

Table 4 shows the increase in t-PA production in CHO cells in medium containing 7.5% FCS by means of C3-C5 monocarboxylic acids depending on the day of application of the carboxylic acids are 48 hours' growth of the cells in medium containing 7.5% FCS. The concentration was 1 mM for butyric and isovaleric acid and 5 mM for propionic and isovaleric acid. The values refer to 1 ml of cell supernatant and are given in % compared with the control (0%) (average + SE, 3 = 3-5). The t-PA was determined by means of DCA.

TABLE 4 Day of application C3 C4 C4-I C5 0 28.7 + 9.3 33.0 + 5.4 20.1 + 10.7 17.2 + 1.0 1 24.4 + 4.1 40.2 + 12.2 25.3 + 5.1 19.4 + 1.5 Effects of butyric acid on t-PA production in CHO cells The stimulating effect can be monitored by an increase both in the intracellular t-PA concentration and also in the extracellular t-PA concentration.

As is clear from Fig. 8, the increase in intracellular t-PA content in Na-butyrate-treated cultures is greatest after 24 hours incubation compared with the control (43.8 + 9.8%). In the supernatant, the percentage increase is somewhat less after 24 hours, on average 19.2% (DCA and Elisa) compared with the control values. This effect lasts for 72 hours in the supernatant. The Na-butyrate was added to a medium containing 7.5 % FCS at a concentration of ImM.

Fig. 9A-C shows the synthesis of t-PA based on the cell number and shows that the stimulation of t-PA synthesis by Na-butyrate lasts for more than 7 days in spite of the inhibition of cell growth. As above, the medium contained 7.5% FCS and ImM Na-butyrate. During the first 4 days, enzyme synthesis per cell increases constantly.

The t-PA concentration increases from 7.5 + 1.1 meg to 28.9 + 5.3 meg t-PA/1 x 10θ cells (Elisa) in the supernatant (Fig. 8A). Using the DCA method, a maximum t-PA content of 47.4 + 7.4 meg t-PA/1 x 10θ cells is determined (Fig. 9B). At the same time the intracellular t-PA content increases and falls £ from 1.5 + 0.2 meg of t-PA/1 x 10 cells on the 5th day (Fig. 9C).

Table 5 shows the stimulation of t-PA synthesis by Na-butyrate in medium containing 7.5% FCS based on the cell number. Na-butyrate was added on day 0 in a concentration of 1 mM. The values on days 1-4 are given in % compared with-the control (0%) and relate to 1 x 10^ cells. The t-PA content in the supernatant and in the cell extract was determined by Elisa (average + SE, n = 5-10).

TABLE 5 Day Intracellular Extracellular 1 90.9 + 9.5 68.5 + 12.4 2 99.7 + 26.3 118.8 + 16.1 3 93.9 + 31.4 223.4 + 77.4 4 209.8 + 41.4 280.7 + 85.9 An optimum increase in production was achieved when Na-butyrate was added to the cells during exponential growth. The maximum effect of Na-butyrate always occurs on the 3rd day when there is an application interval from day 0 to 2 (Fig. 10).

By applying Na-butyrate every 2-3 days after changing the medium in the same cell population, it was possible to maintain the t-PA production at a higher level for a fairly long time.

Table 6 shows the increase in t-PA production in CHO cells which were cultivated in medium with 7.5% FCS. 1 mM of Na-butyrate was added each time the medium was changed or only on day 0. The values of 2, 4, 7 or 9 days, respectively, after the medium had been changed are given in % compared with the control (0%) and relate to 1 ml of cell supernatant. The t-PA determination was effected by means of DCA (average + SE, n = 3).

TABLE 6 Change of medium Application Day Day 0, 2, 4, 7 2 33.0 + 5.4 4 107.4 + 7.6 7 96.6 + 5.8 9 62.0 + 4.8 Effects of butyric acid on the t-PA production in CHO cells in a serum free medium ' > By contrast with the short-term effect of butyrate25 stimulated t-PA synthesis in medium containing serum, the effect in serum-free medium lasts longer and, 1 day after the application of Na-butyrate (at a concentration of ImM) , it is 44.1 + 7.1% compared with the control. The maximum increase was achieved 3 days after the application of Nabutyrate, at 143.3 + 12.8%. The increasing effect can still be detected 6 days after the addition of Na-butyrate, at 49.9 + 7.0% (Figs. 11 and 12).

The t-PA synthesis, based on 1 x 10® cells, is also higher in serum-free medium than in serumcontaining medium and rises to above 50 mcg/1 x 10® cells (Fig. 13). Na-butyrate was added to give a concentration of ImM after 24 hours cultivation in serum-free medium.

Effects of propionic acid C3-carboxylic acid exhibits very similar effects on t-PA synthesis to those produced by butyric acid. However, the percentage increase is about 14% less and higher concentrations (5-10 mM) are needed in order to stimulate. t-PA. Furthermore, propionic acid inhibits cell growth to a lesser extent and therefore leads to a longer-lasting increase in the t-PA synthesis.

Stimulation of t-PA synthesis by propionic acid appears to be very specific, in contrast to butyric acid, since in serum-free medium the extracellular protein content was increased only slightly with the control and the incorporation of [ H]-leucine was increased only slightly.

Effects of dicarboxylic acids, hydroxycarboxylic acids and ketocarboxylic acids on t-PA production in CHO cells Hydroxycarboxylic acids such as lactic, glyceric, malic and tartaric acid show only a slight increase in t-PA synthesis of between 6 and 10%, whilst tartaric acid in a concentration of 10 mM stimulated the t-PA synthesis best (10.3 + 2.9% compared with the control,. The cell division is not affected by hydroxycarboxylic acids and the effect on t-PA synthesis can still be detected after 96 hours.

Effects of ascorbic acid The effect of ascorbic acids in a concentration of 10 mM on t-PA production lasts for 72 hours and when applied on day 0 it results in a 15.2 + 3.1% increase in t-PA synthesis after 24 hours (Fig. 14). The medium contained 7.5% FCS.

Effects of long-chained fatty acids on t-PA production in CHO cells It was found that long-chained fatty acids such as nonactic acid with 10 carbon atoms and furan fatty acids were able to cause an increase (Table 7). In this example, a furan fatty acid according to formula 4 was used wherein m - 4 and n - 8.

However, in serum-free medium there was no evidence of an increase in t-PA synthesis or protein synthesis.

The increase in t-PA production in CHO cells caused by long-chained fatty acids is shown in Table 8.

The cells were cultivated in medium containing 7.5% FCS. The acids were tested in a concentration range from 10 mcM to 10 mM, a concentration of mM being optimum. The values relate to 1 ml of cell supernatant and are given as a percentage compred with the control (0%) (average + SE, n = 4-9). The maximum value refers to an individual sample.

The t-PA was determined by DCA.

TABLE 7 Fatty acid % increase Maximum in t-PA synthesis Average Nonactic acid Furan fatty acid 58.3 32.1 30.3 + 6.5 19.0 + 3.2 The average values relate to several production runs of 4 days, the acids being added on day 0, as shown in Table 8A. Data for the individual samples are shown in brackets. TABLE 7A Fatty acids Day of addition Day of sample n Nonactic acid Furan fatty acid 0 (0) 0 (0) 1-4 (2) 1-4 (2) 4 9 Effects of thiols and sulphides on t-PA production in CHO cells For sulphur-containing compounds which are derivatives or substitution products of carboxylic acids showed an increase in t-PA production of between 16% and 31.2%, without having a negative effect on cell growth (Table 8). The substituted carboxylic acids show their optimum activity in the same concentration range (1-10 mM) as the other active carboxylic acids. Only glutathione was effective in a lower concentration range of 1-10 mcM. Thioglycolic acid proved to be the preferred stimulator of t-PA synthesis in CHO cells.

In this Example, the cells were cultivated in medium 10 containing 7.5% PCS. All the substances were tested in a concentration range of from 1 mcM to 10 mM.

The values given in Table 9 relate to 1 ml of cell supernatant and are given in % compared with the control (0%) (average value + SE, n = 6-19).

The maximum value relates to an individual sample. The t-PA was determined by DCA. The concentration of the substance is shown in brackets.

TABLE 8 Thiols and sulphides Optimum dosage range % increase -synthesis Maximum in t-PA Average Thioglycolic acid 1 mM 66.4 31.2 + 3.2 Thiodiglycolic acid 1 mM 31.0 18.0 + 3.8 L-cysteine 1-10 mM ( lmM) 41.4 20.9 + 3.4 Glutathione 1-10 uM (lOuM) 30.6 16.0 + 2,0 The average values refer to several production runs of 3 days, the substances having been added from day 0 to day 2 as shown in Table 9A. Data for the individual samples are shown in brackets.

TABLE 8A Thiols and sulphides Day of addition Day of sample n Thioglycolic acid 0-2 (1) 1-3 (2) 19 Th iod iglycolic acid 1 (1) 2-3 (3, 9 L-cysteine 0-1 (0) 1-3 (2) 11 Glutathione 0 (0) 2-3 (2, 6 One advantage of thiocarboxylic acids over raonocarboxylic acids is the absence of any inhibitory effect on cell proliferation, but the production of t-PA could not be increased further with these substances than with butyric acid. The duration of the increasing effect is also short and attains its maximum value after 24 hours for thioglycolic acid and after 48 hours for cysteine. After 4 days incubation of the cells the effect on t-PA production is still small (Fig. 15).

However, in serum-free medium, the stimulating effect on t-PA synthesis is less than in medium containing serum since all the substances inhibit cell division between 10% and 20%.

Effects of thioglycolic acid t-PA production is increased for 72 hours after treatment with thioglycolic acid and is independent of the day of application. The percentage increase in the enzyme synthesis is at its maximum after 24 hours and then falls continuously.

Table 9 shows the increase in t-PA production in CHO cells cultivated for 24 to 96 hours in medium containing 7.5% FCS and 1 mM thioglycolic acid, depending on the day of application. The values relate to 1 ml of cell supernatant and are given in % compared with the control (0%) (average + SE, n = 4-6). The t-PA was determined by DCA.

TABLE 9 Time (h) Day of application 0 1 2 24 27.6 + 5.6 34.7 + 10.9 34.5 + 7.5 48 16.9 + 2.3 17.1 + 3.6 16.3 + 8.6 72 14.7 + 1.0 10.2 + 2.7 9.4 + 1.7 96 3.4 + 1.0 2.7 + 0.8 0.5 + 1.0 t-PA production may be increased by the application of thioglycolic acid (ImM) twice rather than once (Fig. 16) .

In Fig. 17Ά (DCA) and Fig. 17B (Elisa) the curve of t-PA synthesis is shown based on 1 x 10® cells.

The t-PA synthesis is increased compared with the control by means of 1 mM thioglycolic acid for the first 3 days (DCA). By contrast, the t-PA values determined by Elisa are increased for the entire period. The rate of increase is still about 35% on the 4th day. The intracellular t-PA content per cell is also increased by means of thioglycolic acid. The maximum intracellular t-PA concentration is 1.1 + 0.1 mcg/1 x 10® cells after 24 hours and falls within 96 hours to approximately half the original concentration (Fig. 17C).

The extracellular t-PA concentration is increased by 10.7% compared with the control on the 5th day.

The percentage increase in t-PA synthesis by means of thioglycolic acid is somewhat lower in serumfree medium since cells division is slightly inhibited (10-20%).

Effects of derivatives of monocarboxylic acids on t-PA production in CHO cells Within the group of C4 carboxylic acid derivatives, other substances were found which have an effect on t-PA synthesis in CHO cells. These substances include, primarily, butyryl choline bromide (BCB) and chloride (BCC) which increase t-PA production by between 30 and 40%. These C4 derivatives show the same effects as Na-butyrate in every respect and are effective in the same concentrations.

Table 10 shows the increase in t-PA production in CHO cells achieved by means of the two abovementioned derivatives of monocarboxylic acids.

The cells were cultivated in a medium containing 7.5% PCS. All the substances were tested at concentrations ranging from 10 mcM to 10 mM. The values are given in % compared with the control (0%) and relate to 1 ml of cell supernatant (average + SE, n = 4-15). The values which relate to the individual samples are given in brackets. The maximum value refers to an individual sample. The t-PA was determined by DCA.

TABLE 10 Derivatives of monocarboxylic acid Optimum dosage range % increase in t-PA synthesis Maximum Average BCC 5-10 mM ( 10 mM) 63.9 38.8+4.4 15 BCB 5-10 mM (2.5 mM) 49.6 29.1 + 4.4 9 The average values relate to several production runs of 4 days as shown in Table 10Ά. Values for the individual samples are shown in brackets.

TABLE 10A Substance Day of addition Sample days n BCC 1 (1) 2-4 (4) 15 BCB 1 (1) 2-4 (4) 9 Combined effects of different t-PA synthesis stimulators When combinations of substances are used at least one substance should have the property of not inhibiting cell growth.

An increased yield was achieved, for example, with a combination of thioglycolic acid and butyric acid. The application interval of the individual substances may play an important role, with butyric acid, for example, always being added as the second substance in order to reduce the inhibitory effect on cell growth.

When the substances are administered simultaneously on day 0 the effect of the butyric acid is not increased.

The effect is cumulative for a combination of butyric and thioglycolic acid. A maximum increase of 70% may be achieved.

In serum-free medium the effects of butyric acid could not be increased any further by the addition of another substance.

The increase in t-PA production in CHO cells which are cultivated for 3 days in medium containing 7.5% FCS is shown in Table 11. 1 mM of thioglycolic acid (TG) and 1 mM of butyric acid were added to the cells in combination at various times. The values relate to 1 ml of cell supernatant and are given as a percentage compared with the control (0%, (average + SE, n = 4-8). The t-PA was determined by DCA.

TABLE 11 Day of Application TG Butyric acid 0 1 2 31.3 + 1.7 67.9 + 5.2 68.8 + 4.4 .3 + 3.2 31.4 + 5.5 41.4 + 5.5 Effects of aphidicolin Aphidicolin, a diterpene obtained from Cephalosporium aphidicola, specifically inhibits DNA alpha-polymerase and stimulates t-PA synthesis in CHO cells in a concentration of from 1 to 10 mcM. After 24 hours the t-PA yield has been increased by 44.9 + 13.9%.

Aphidicolin was found to stimulate t-PA synthesis in CHO cells. It may inhibit DNA synthesis in reversible manner but the RNA and protein synthesis is unaffected. Aphidicolin is thus suitable on the one hand for stimulating t-PA synthesis and on the other hand for synchronising the cells.

The effect of aphidicolin on t-PA synthesis when the cells are incubated for a fairly long time in medium containing aphidicolin is described hereinbefore. This passage relates to studies with CHO cells which have been exposed to aphidicolin for 24 hours and subsequently cultivated further in the medium without aphidicolin. The phases 3 of the cell cycle and the incorporation of [ H]thymidine during 24 hour labelling and thereafter were analysed in order to monitor the inhibition of DNA synthesis and resumption of synthesis after removal of the aphidicolin. Furthermore, the RNA, protein and t-PA synthesis were determined.

Cell growth The cell growth was inhibited slightly (5-10%) by aphidicolin (1 mcM) after 24 hours' treatment. However, when the medium was changed, the aphidicolintreated cultures were returned to the same cell number as the control cultures. t-PA synthesis The t-PA synthesis was stimulated in 1 mcM aphidicolinsynchronised cells. An effect could still be seen 96 hours after the changing of the medium.

The increase in t-PA production in 1 mcM aphidicolintreated CHO cells cultivated in medium containing 7.5% FCS is shown in Table 12. The values relate to 1 ml of cell supernatant and are given as a percentage compared with the control (0%) (average + SE, n = 4). The t-PA was determined using DCA.

TABLE 12 Time (h after changing % increase id t-PA synthesis of medium) 24 40.2 + 6.7 48 59.9 + 11.1 72 32.8 + 6.4 96 14.7 + 1.3 Fig. 18 shows the time curve for the increase in t-PA synthesis in lmcM aphidicholin-synchronised CHO cells cultivated in serum-free medium. The effect reaches its maximum 48 hours after the changing of the medium.

The t-PA concentration in serum-free medium was determined after 96, 120 and 144 h using Elisa, whilst the values obtained compared with the control values were substantially higher than the values determined by DCA. The results are shown in Table 13. The test was carried out with 1 mcM aphidicolin-treated CHO cells which were cultivated in serum-free medium.

The values relate to 1 ml of cell supernatant and are given as a percentage compared with the control (0%) (average + SE, n = 4). The t-PA was determined by means of Elisa.

TABLE 13 Time (h after changing of medium) % increase in t-PA synthesis 96 36.7 + 6.2 120 42.0 + 8.2 144 51.3 + 10.2 An increased intracellular t-PA concentration could only be observed during 24 hour incubation with aphidicolin and was increased by about 30%.

Butyric acid, propionic acid, butyryl choline chloride and bromide were able to increase t-PA production still further in this system. Stimulation was successful both in serum-free and serum-rich medium, whilst an additional increasing effect in serumfree medium was only achieved when the substances were applied after 24 hours adaptation of the cells to this condition of the medium.

Table 14 shows the percentage increase in t-PA production in aphidicolin-treated cultures after the application of Na-butyrate. Na-butyrate increases t-PA synthesis by a further 14% after 96 hours.

The test was carried out with 1 mcM aphidicolintreated CHO cells, which were cultivated with and without Na-butyrate, in serum-free culture. The values relate to 1 ml of cell supernatant and are given as a % compared with the control (0%) (average + SE, n = 4) . The t-PA was determined by means of Elisa.

TABLE 14 Time (h after Aphidicolin + Na-butyrate changing of Time of application (h) medium) 0 24 96 36.7 + 6.2 120 42.0 + 8.8 34.1 + 9.7 31.8 + 10.2 50.6 49.7 + 11.0 + 4.3 Effects of 6-hydroxy-4,6-dimethyl-3-hepten- 2-one (DHO) The t-PA production is increased after treatment with above compound after 144 hours. For stimulating t-PA synthesis, concentrations in micromolar range were effective, e.g. 0.00.5 to 109 racM.

Fig. 19 shows the increase in t-PA production in CHO cells which were cultivated up to 168 hours in serum-free medium and 7 mM to 7 mcM DHO. The values were determined by DCA. Fig. 20 shows the corresponding values obtained by Elisa.

Cell cultures prepared by the method described above may be worked over in known manner to isolate the t-PA, for example after the production phase the culture medium is separated from the cell mass and the cell supernatant is purified by ultrafiltration and chromatographic procedures. <· ·· 4 - 38 The isolated t-PA can then be formulated into pharmaceutical preparations, e.g. in dissolved or lyophilised form.

The above Examples may be repeated, replacing the t-PA producing CHO cells by transformed CHO cells which produce other proteins e.g. t-PA mutants or other pharmacologically effective proteins, as in the above-mentioned European and German patent applications, particularly the CHO cells in EP-A 199,574 and 196,920 and in DE-A 3,708,681.

It should be noted that, although the invention has been illustrated above with reference to a variety of inducers, the compounds fusaric acid, mevalonic acid, trans-anhydromevalonic acid, anhydromevalonic acid lactone and cis-anhydro-mevalonic acid lactone can all be used in like manner as inducers. - 39 CLAIMS

Claims (8)

1. Method of increasing the production of a protein from cultures of transformed CHO cells, characterised in that a C 3 _ 4 aliphatic monocarboxylic acid or a salt thereof is added to the culture, and cultivation is carried out in serum-free medium.

2. Method according to claim 1, characterised in that the protein is rt-PA or a mutant thereof.

3. Method according to claim 1 or 2, characterised in that butyric acid or a salt thereof is used as the protein-inducing substance.

4. Method according to claims 1 to 3, characterised in that the C3-4 aliphatic monocarboxylic acid or a salt thereof is added in a concentration of from 1 mM to 10 mM.

5. Method according to one of claims 1 to 4, characterised in that the cells are cultivated in the presence of a substance selected from thioglycolic acid, thiodiglycolic acid, L-cysteine, glutathione, butyryl choline bromide, butyryl choline chloride, nonactic acid, furan fatty acid, ascorbic acid, aphidicolin, 6. -hydroxy-4,6-dimethyl-3-hepten-2-one, D-<<- hydroxysubstituted (C^ or C^) aliphatic mono- or dicarboxylic acid or the salts thereof, the cells are removed from the culture and then placed in new serum-free culture medium and cultivated in the presence of a C 3 _ 4 aliphatic monocarboxylic acid or a salt thereof.

6. Method according to claim 5, characterised in that aphidicolin is the first-mentioned substance and the aliphatic monocarboxylic acid is butyric acid, propionic acid or the salts thereof.

7. Method according to claim 5, characterised in that butyric acid and thioglycolic acid or the salts thereof are used.

8. A method according to any one of claims 1 to 7 substantially as described herein.