Journal of Biotechnology, 17 (1991) 155-167 © 1991 Elsevier Science Publishers B.V. (Biomedical Division) 0168-1656/91/$03.50 ADONIS 016816569100056B
155
BIOTEC 00564
Effect of free fatty acids and phospholipids on growth of and product formation by recombinant baby hamster kidney (rBHK) and Chinese hamster ovary (rCHO) cells in culture G. Schmid
1, H.
Zilg
2, U.
Eberhard 1 and R. Johannsen 1
Cell Culture Development Group and 2 Research Laboratories of Behringwerke A G, Marburg, F.R. G. (Received 16 March 1990; revision accepted 24 June 1990)
Summary Recombinant BHK and CHO cells producing human antithrombin III (rh ATIII) were used to investigate the utilization of phospholipids and free fatty acids from low-serum (0.1% FBS) culture medium. Both cell lines show distinctly different patterns of fatty acid utilization. For rBHK ATIII cells it is shown that under low serum conditions several different combinations of free fatty acids (bound to bovine albumin) elicit an identical growth stimulatory effect although individual consumption and production rates of fatty acids are different. Increased fatty acid concentrations lead to increased uptake rates without any further effect on growth rate being observed. Recombinant antithrombin II| formation is found to be a function of combinations and concentrations of fatty acids present in the culture medium. Mammalian cell culture; Free fatty acid utilization; Phospholipid utilization; Antithrombin III production; Chinese hamster ovary (CHO) cells; Baby hamster kidney (BHK) cells
Correspondence to." G. Schmid, Central Research Laboratories, ZFE/MB, F. Hoffmann-La Roche AG, CH-4002 Basle, Switzerland.
156
Introduction
Several extensive review articles (Howard and Howard, 1974; Spector, 1972; Rosenthal, 1987; Spector et al., 1981; Cornwell and Morisaki, 1984) provide an excellent overview of the knowledge of fatty acid, glyceride, and phospholipid metabolism of cultured mammalian cells in the biochemical literature. Essentially all cells examined have been found to readily take up exogenous lipids contained in the culture medium. Lipids may be present in the medium as components of added serum or they can be supplemented in the form of isolated plasma lipoprotein fractions, free fatty acids complexed to serum albumin or fatty acid/ phospholipid microemulsions. The incorporated lipids are utilized within the cell, either as a source of energy or as building blocks for cell growth. When adequate supplies of lipids are available in the medium, de novo synthesis is inhibited. In spite of this, cells accumulate excessive triglycerides and cholesterol esters in the form of cytoplasmic inclusions when exposed to an overabundance of fipids. The fatty acid composition of cellular storage lipids as well as of membrane phospholipids resembles that of lipids present in the medium. Modification of membrane phospholipid fatty acyl groups has been shown to lead to differences in cell function (membrane enzyme activity and transport properties). The specific growth rate of cultured cells can be likewise affected. Research activities towards the development of serum-free a n d / o r protein-free cell culture media have long recognized the potential growth promoting effects of exogenous fatty acids, phospholipids or lipoprotein fractions. Thus, lipids are included in most serum-free medium formulations (Barnes and Sato, 1980; Bettger and Ham, 1982; Higuchi, 1973). To our knowledge, however, there are only limited or no quantitative data available on the utilization of (or possible limitations in) e.g. free fatty acids and on the influence of different combinations and concentrations of free fatty acids on cell growth rates, product formation rates, and fatty acid metabolic quotients. This communication provides such data for two genetically engineered mammalian cell lines producing recombinant human antithrombin III.
Materials and Methods
Cell lines
Construction of recombinant Chinese hamster ovary (rCHO) and baby hamster kidney (rBHK) cells expressing human antithrombin III (rh ATIII) has been described before (Zettlmeissl et al., 1987, 1988; Wirth et al., 1988). Medium composition and culture conditions
Low-serum growth medium was developed as will be published elsewhere (Schmid and Johannsen, 1990). In this study Iscove's modified Dulbecco's medium (IMDM) was used as basal medium supplemented with 0.1% fetal bovine serum (FBS), 3 or
157 30 mg 1-1 human transferrin (partially iron-saturated), 5 mg 1-1 insulin (Hoechst, Frankfurt, F.R.G.), 0.02 mM monoethanolamine (Sigma Chemical Co., St. Louis, MO) and free fatty acids (Sigma) in various combinations and concentrations (see below) bound to fatty acid free bovine serum albumin (BSA). A concentrated stock, 50 x , was prepared by dissolving fatty acids in absolute ethanol (10 mg m1-1) and adding aliquots, typically 0.5% (v/v) per FFA, to a solution of albumin in PBS (125 mg ml-1). Additional medium components included: human fibronectin (5 mg 1-1), trypsin inhibitor (50 IU ml-1), methotrexate (10 /~mol 1-1) (Sigma), and G418 sulfate (400 mg 1-1) (Gibco, Grand Island, NY), used with rBHK cells only. Further, a commercially available growth supplement SF-X (Costar Europe Ltd., Badhoevedorp, The Netherlands) providing transferrin, insulin, monoethanolamine, inorganic trace elements plus a mixture of free fatty acids bound to albumin was tested (composition as given by Yssel et al., 1984). In some experiments, a phospholipid/cholesterol, Sigma/BSA microemulsion, was used as lipid source. Its preparation was carried out as described in the literature (Iscove and Melchers, 1978). The crude soybean lipids (Epikuron 100 P, Lukas Meyer AG, Hamburg, F.R.G.) contained 20-23% phosphatidyl choline, 21-24% phosphatidyl ethanolamine, and 18-22% phosphatidyl inositol. Human transferrin, BSA, human fibronectin, and trypsin inhibitor all were from Behringwerke AG, Marburg, F.R.G. Adherent rBHK ATIII cells were cultivated in T25 and T75 flasks in a 37°C incubator with 5% CO 2 in air. Inoculum cells were usually grown in medium containing 0.5% FBS and seeded at 2.5 x 10E4 viable cells per ml (approx. 1 x 10E4 cells per cm2). Prior to the experiments, flasks were treated with 2 ml human fibronectin (50 mg 1-1) for 1 h at room temperature. Excess fibronectin was then removed by aspiration before the addition of test medium. Multiple flasks were prepared. One flask per day was used to estimate cell numbers and to take medium samples. The trypsinizing solution contained trypsin at 0.25% (w/v) and EDTA at 0.03% (w/v) in phosphate-buffered saline (PBS). rCHO ATIII cells were grown in suspension culture using 100 ml spinner vessels (Bellco Glass, Inc., Vineland, NJ) operated at 37°C and 120 rpm. Spinner vessels were coated with Sigmacote (Sigma) prior to usage so as to prevent excessive cell adherence. Samples for analyses were taken daily. Cells were seeded at 1 x 10E5 viable cells per ml in IMDM supplemented with 0.1% FBS and 2% (v/v) SF-X (see above).
Metabolite and product analyses For determination of the viable count and percentage viability the cell sample was diluted 1:1 with 0.05% (w/v) trypan blue in saline and counted on a hemacytometer; nonviable cells stained blue. The remainder of each sample was centrifuged to remove the cells and frozen for later analysis. Total fatty acid concentrations (free and bound) in culture medium were determined by gas liquid chromatography (g.l.c.). Alkaline hydrolysis, methylation with methanol/hydrochloric acid, and extraction yielded a methyl ester fraction. This was separated and quantitated by g.l.c, on an open tubular capillary column
158 with chemically bonded phase type FS-FFAP (Machery-Nagel, Dtiren, F.R.G.) and flame ionization detection. Sorbent extraction clean-up followed by thin layer chromatography on silica gel was used to estimate the individual phospholipid concentrations in culture medium. Visualization by manganese chloride treatment and heating led to fluorescent spots quantitated by densitometry. For analysis of the product concentration culture medium was submitted to a specific ELISA assay using 98% pure antithrombin III isolated from human plasma as standard (Zettlmeissl et al., 1987).
Results and Discussion
In the course of our work on the development of low-serum growth media for recombinant adherent cell lines (Schmid and Johannsen, 1990) we observed the well-documented growth promoting effects (i.e., increased growth rates/~) of exogenous free fatty acids (FFAs) and phospholipids if included in the culture medium (Barnes and Sato, 1980; Bettger and Ham, 1982; Spector et al., 1981; Rosenthal, 1987). It seemed appropriate to investigate growth characteristics and product formation as well as utilization of lipidic supplements for model recombinant cell lines as a function of different combinations and concentrations of free fatty acids and phospholipids. Fig. la shows the growth curves for adherent rBHK cells producing human antithrombin III. Cells were grown in T25 flasks in low-serum medium as given under Materials and Methods supplemented with different combinations of free fatty acids bound to bovine albumin. In the whole series of experiments the recombinant BHK cells exhibited identical growth characteristics. After a lag period of about 24 h the cells started to grow exponentially with a maximum specific growth rate of 0.81 _+0.04 d-1 (20 + 1 h doubling time) until they reached 1 × 10E6 cells per ml (approx. 4 × 10E5 cells per cm2). Similarly, viability was found not to be a function of FFA combinations tested (data not shown). The reference experiment shown was performed in basal IMDM with 0.1% FBS only. Contrary to these results, growth stimulation was found to be a function of combinations and concentrations of free fatty acids (as assessed by monitoring tritiated thymidine incorporation into DNA) for lymphocytes (Spieker-Polet and Polet, 1981). Rintoul et al. (1978) reported that except for the saturated fatty acids stearate and palmitate, all fatty acids tested (linoleic, oleic and elaidic acids) supported the growth of CHO K1 cells equally well. Hatzfeld et al. (1982) observed a strong additive effect of saturated and unsaturated FFAs bound to bovine albumin with various cells of the human hematopoietic and immune systems (U937, K562, HL-60, Jurkat, Raji). Similar synergistic effects have been communicated by other authors (Yamane, 1978; Doi et al., 1978; Rockwell et al., 1980). Unfortunately, in these cases data on specific growth rates or concentrations of fatty acids in the culture media are not provided.
159
(a) no s u p p l e m e n t s
T ,000,000
o
..t1:
500,000 - Z O (D LLI
oleic acid
eS:
.~
~p~SJ
200,000
linoleic acid
ole[c and iinoleic acids -_~-_.
100,000
Ld
50,000
O3
155
BIOTEC 00564
Effect of free fatty acids and phospholipids on growth of and product formation by recombinant baby hamster kidney (rBHK) and Chinese hamster ovary (rCHO) cells in culture G. Schmid
1, H.
Zilg
2, U.
Eberhard 1 and R. Johannsen 1
Cell Culture Development Group and 2 Research Laboratories of Behringwerke A G, Marburg, F.R. G. (Received 16 March 1990; revision accepted 24 June 1990)
Summary Recombinant BHK and CHO cells producing human antithrombin III (rh ATIII) were used to investigate the utilization of phospholipids and free fatty acids from low-serum (0.1% FBS) culture medium. Both cell lines show distinctly different patterns of fatty acid utilization. For rBHK ATIII cells it is shown that under low serum conditions several different combinations of free fatty acids (bound to bovine albumin) elicit an identical growth stimulatory effect although individual consumption and production rates of fatty acids are different. Increased fatty acid concentrations lead to increased uptake rates without any further effect on growth rate being observed. Recombinant antithrombin II| formation is found to be a function of combinations and concentrations of fatty acids present in the culture medium. Mammalian cell culture; Free fatty acid utilization; Phospholipid utilization; Antithrombin III production; Chinese hamster ovary (CHO) cells; Baby hamster kidney (BHK) cells
Correspondence to." G. Schmid, Central Research Laboratories, ZFE/MB, F. Hoffmann-La Roche AG, CH-4002 Basle, Switzerland.
156
Introduction
Several extensive review articles (Howard and Howard, 1974; Spector, 1972; Rosenthal, 1987; Spector et al., 1981; Cornwell and Morisaki, 1984) provide an excellent overview of the knowledge of fatty acid, glyceride, and phospholipid metabolism of cultured mammalian cells in the biochemical literature. Essentially all cells examined have been found to readily take up exogenous lipids contained in the culture medium. Lipids may be present in the medium as components of added serum or they can be supplemented in the form of isolated plasma lipoprotein fractions, free fatty acids complexed to serum albumin or fatty acid/ phospholipid microemulsions. The incorporated lipids are utilized within the cell, either as a source of energy or as building blocks for cell growth. When adequate supplies of lipids are available in the medium, de novo synthesis is inhibited. In spite of this, cells accumulate excessive triglycerides and cholesterol esters in the form of cytoplasmic inclusions when exposed to an overabundance of fipids. The fatty acid composition of cellular storage lipids as well as of membrane phospholipids resembles that of lipids present in the medium. Modification of membrane phospholipid fatty acyl groups has been shown to lead to differences in cell function (membrane enzyme activity and transport properties). The specific growth rate of cultured cells can be likewise affected. Research activities towards the development of serum-free a n d / o r protein-free cell culture media have long recognized the potential growth promoting effects of exogenous fatty acids, phospholipids or lipoprotein fractions. Thus, lipids are included in most serum-free medium formulations (Barnes and Sato, 1980; Bettger and Ham, 1982; Higuchi, 1973). To our knowledge, however, there are only limited or no quantitative data available on the utilization of (or possible limitations in) e.g. free fatty acids and on the influence of different combinations and concentrations of free fatty acids on cell growth rates, product formation rates, and fatty acid metabolic quotients. This communication provides such data for two genetically engineered mammalian cell lines producing recombinant human antithrombin III.
Materials and Methods
Cell lines
Construction of recombinant Chinese hamster ovary (rCHO) and baby hamster kidney (rBHK) cells expressing human antithrombin III (rh ATIII) has been described before (Zettlmeissl et al., 1987, 1988; Wirth et al., 1988). Medium composition and culture conditions
Low-serum growth medium was developed as will be published elsewhere (Schmid and Johannsen, 1990). In this study Iscove's modified Dulbecco's medium (IMDM) was used as basal medium supplemented with 0.1% fetal bovine serum (FBS), 3 or
157 30 mg 1-1 human transferrin (partially iron-saturated), 5 mg 1-1 insulin (Hoechst, Frankfurt, F.R.G.), 0.02 mM monoethanolamine (Sigma Chemical Co., St. Louis, MO) and free fatty acids (Sigma) in various combinations and concentrations (see below) bound to fatty acid free bovine serum albumin (BSA). A concentrated stock, 50 x , was prepared by dissolving fatty acids in absolute ethanol (10 mg m1-1) and adding aliquots, typically 0.5% (v/v) per FFA, to a solution of albumin in PBS (125 mg ml-1). Additional medium components included: human fibronectin (5 mg 1-1), trypsin inhibitor (50 IU ml-1), methotrexate (10 /~mol 1-1) (Sigma), and G418 sulfate (400 mg 1-1) (Gibco, Grand Island, NY), used with rBHK cells only. Further, a commercially available growth supplement SF-X (Costar Europe Ltd., Badhoevedorp, The Netherlands) providing transferrin, insulin, monoethanolamine, inorganic trace elements plus a mixture of free fatty acids bound to albumin was tested (composition as given by Yssel et al., 1984). In some experiments, a phospholipid/cholesterol, Sigma/BSA microemulsion, was used as lipid source. Its preparation was carried out as described in the literature (Iscove and Melchers, 1978). The crude soybean lipids (Epikuron 100 P, Lukas Meyer AG, Hamburg, F.R.G.) contained 20-23% phosphatidyl choline, 21-24% phosphatidyl ethanolamine, and 18-22% phosphatidyl inositol. Human transferrin, BSA, human fibronectin, and trypsin inhibitor all were from Behringwerke AG, Marburg, F.R.G. Adherent rBHK ATIII cells were cultivated in T25 and T75 flasks in a 37°C incubator with 5% CO 2 in air. Inoculum cells were usually grown in medium containing 0.5% FBS and seeded at 2.5 x 10E4 viable cells per ml (approx. 1 x 10E4 cells per cm2). Prior to the experiments, flasks were treated with 2 ml human fibronectin (50 mg 1-1) for 1 h at room temperature. Excess fibronectin was then removed by aspiration before the addition of test medium. Multiple flasks were prepared. One flask per day was used to estimate cell numbers and to take medium samples. The trypsinizing solution contained trypsin at 0.25% (w/v) and EDTA at 0.03% (w/v) in phosphate-buffered saline (PBS). rCHO ATIII cells were grown in suspension culture using 100 ml spinner vessels (Bellco Glass, Inc., Vineland, NJ) operated at 37°C and 120 rpm. Spinner vessels were coated with Sigmacote (Sigma) prior to usage so as to prevent excessive cell adherence. Samples for analyses were taken daily. Cells were seeded at 1 x 10E5 viable cells per ml in IMDM supplemented with 0.1% FBS and 2% (v/v) SF-X (see above).
Metabolite and product analyses For determination of the viable count and percentage viability the cell sample was diluted 1:1 with 0.05% (w/v) trypan blue in saline and counted on a hemacytometer; nonviable cells stained blue. The remainder of each sample was centrifuged to remove the cells and frozen for later analysis. Total fatty acid concentrations (free and bound) in culture medium were determined by gas liquid chromatography (g.l.c.). Alkaline hydrolysis, methylation with methanol/hydrochloric acid, and extraction yielded a methyl ester fraction. This was separated and quantitated by g.l.c, on an open tubular capillary column
158 with chemically bonded phase type FS-FFAP (Machery-Nagel, Dtiren, F.R.G.) and flame ionization detection. Sorbent extraction clean-up followed by thin layer chromatography on silica gel was used to estimate the individual phospholipid concentrations in culture medium. Visualization by manganese chloride treatment and heating led to fluorescent spots quantitated by densitometry. For analysis of the product concentration culture medium was submitted to a specific ELISA assay using 98% pure antithrombin III isolated from human plasma as standard (Zettlmeissl et al., 1987).
Results and Discussion
In the course of our work on the development of low-serum growth media for recombinant adherent cell lines (Schmid and Johannsen, 1990) we observed the well-documented growth promoting effects (i.e., increased growth rates/~) of exogenous free fatty acids (FFAs) and phospholipids if included in the culture medium (Barnes and Sato, 1980; Bettger and Ham, 1982; Spector et al., 1981; Rosenthal, 1987). It seemed appropriate to investigate growth characteristics and product formation as well as utilization of lipidic supplements for model recombinant cell lines as a function of different combinations and concentrations of free fatty acids and phospholipids. Fig. la shows the growth curves for adherent rBHK cells producing human antithrombin III. Cells were grown in T25 flasks in low-serum medium as given under Materials and Methods supplemented with different combinations of free fatty acids bound to bovine albumin. In the whole series of experiments the recombinant BHK cells exhibited identical growth characteristics. After a lag period of about 24 h the cells started to grow exponentially with a maximum specific growth rate of 0.81 _+0.04 d-1 (20 + 1 h doubling time) until they reached 1 × 10E6 cells per ml (approx. 4 × 10E5 cells per cm2). Similarly, viability was found not to be a function of FFA combinations tested (data not shown). The reference experiment shown was performed in basal IMDM with 0.1% FBS only. Contrary to these results, growth stimulation was found to be a function of combinations and concentrations of free fatty acids (as assessed by monitoring tritiated thymidine incorporation into DNA) for lymphocytes (Spieker-Polet and Polet, 1981). Rintoul et al. (1978) reported that except for the saturated fatty acids stearate and palmitate, all fatty acids tested (linoleic, oleic and elaidic acids) supported the growth of CHO K1 cells equally well. Hatzfeld et al. (1982) observed a strong additive effect of saturated and unsaturated FFAs bound to bovine albumin with various cells of the human hematopoietic and immune systems (U937, K562, HL-60, Jurkat, Raji). Similar synergistic effects have been communicated by other authors (Yamane, 1978; Doi et al., 1978; Rockwell et al., 1980). Unfortunately, in these cases data on specific growth rates or concentrations of fatty acids in the culture media are not provided.
159
(a) no s u p p l e m e n t s
T ,000,000
o
..t1:
500,000 - Z O (D LLI
oleic acid
eS:
.~
~p~SJ
200,000
linoleic acid
ole[c and iinoleic acids -_~-_.
100,000
Ld
50,000
O3
