- Email: [email protected]
Development of a serum-free medium for the production of erythropoietin by suspension culture of recombinant Chinese hamster ovary cells using a statistical design
Journal of Biotechnology 69 (1999) 85 – 93
Development of a serum-free medium for the production of erythropoietin by suspension culture of recombinant Chinese hamster ovary cells using a statistical design Gyun Min Lee a,*, Eun Jung Kim a, No Soo Kim a, Sung Kwan Yoon b, Yong Ho Ahn b, Ji Yong Song b a
Department of Biological Sciences, Korea Ad6anced Institute of Science and Technology 373 -1, Kusong-Dong, Yusong-Gu, Taejon 305 -701, South Korea b Biotech Research Institute, LG Chem. Research Park, P.O. Box 61, Yusong-Gu, Taejon 305 -380, South Korea Received 10 June 1998; received in revised form 14 December 1998; accepted 12 January 1999
Abstract In order to develop a serum-free (SF) medium for the production of erythropoietin (EPO) by suspension culture of recombinant Chinese hamster ovary (rCHO) cells, a statistical optimization approach based on a Plackett – Burman design was adopted. A basal medium was prepared by supplementing Iscove’s modified Dulbecco’s medium (IMDM) with Fe(NO3)39H2O, CuCl2 and ZnSO47H2O which are generally contained in SF medium formulations. Insulin, transferrin and ethanolamine were also supplemented to the basal medium to determine their optimal concentrations. From this statistical analysis, glutamate, serine, methionine, phosphatidylcholine, hydrocortisone and pluronic F68 were identified as positive determinants for cell growth. The SF medium was formulated by supplementing the basal medium with components showing positive effects on cell growth in suspension culture. An EPO titer in this optimized SF medium was 79% of that in IMDM supplemented with 5% dialyzed fetal bovine serum (dFBS). Furthermore, the in vitro and in vivo biological activities of EPO produced in the SF medium were comparable to those produced in the serum-supplemented medium. Taken together, the results obtained here show that a Plackett–Burman design facilitates the development of SF media for the production of EPO by suspension culture of rCHO cells. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Plackett–Burman design; Serum-free medium; rCHO; Erythropoietin
* Corresponding author. Tel.: +82-42-869-2618; fax: +82-42-869-2610. E-mail address: [email protected] (G.M. Lee) 0168-1656/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 1 6 5 6 ( 9 9 ) 0 0 0 0 4 - 8
86
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
1. Introduction Erythropoietin (EPO), a glycoprotein primarily produced by the kidney, is the principal regulator of red blood cell production (Fisher, 1983; Sytkowski, 1984). For the large-scale production of EPO, a gene coding for human EPO was cloned and the corresponding protein was expressed in Chinese hamster ovary (CHO) cells (Lin et al., 1985; Davis et al., 1987). For commercial production of EPO, recombinant CHO (rCHO) cells are frequently cultivated in roller bottles since they preferentially adhere to a surface. However, the use of roller bottles for large-scale cultivation requires intensive labor and often makes it difficult to control product quality. Alternatively, rCHO cells can be cultivated in suspension. In fact, the suspension culture-based manufacturing process has become the method of large-scale, commercial production of therapeutically important proteins from rCHO cells (Sinacore et al., 1996). Furthermore, it is desirable to use serum-free (SF) media in suspension culture because serum can cause problems in the subsequent processes (Glassy et al., 1988; Keen and Rapson, 1995). However, since there is no universal SF medium applicable to all cell lines, specific media suitable for each cell line need to be developed (Hayter et al., 1991; Keen and Rapson, 1995; Zang et al., 1995). The nutritional requirements of mammalian cells are so complex that extensive efforts have been made to identify suitable serum-substituting supplements (Castro et al., 1992). The classical approach of changing one medium component at a time is still being used but becomes impractical because it is time-consuming and has the risk of neglecting interactions among supplements (Castro et al., 1992; Freshney, 1994). The need for efficient methods for developing SF medium has led to adoption of a statistical experimental design (Ganne and Mignot, 1991; Castro et al., 1992). In this study, a Plackett – Burman statistical design, which provides an efficient way of screening a large number of medium supplements and identifying important ones, was used to develop an optimized SF medium for the production of EPO by suspension culture of rCHO cells. Furthermore,
the biological activity of EPO produced in this SF medium was examined.
2. Materials and methods
2.1. Cells, medium preparation and culture conditions The rCHO cells (LGE10-9-27) expressing high levels of human EPO were used in this study. These rCHO cells were developed by transfection of a vector containing the dihydrofolate reductase (dhfr) and human EPO genes into dhfr-deficient CHO cells (ATCC CRL 9096). Their EPO gene was coamplified with dhfr by stepwise increments in methotrexate (MTX, Sigma, St. Louis, MO) level up to 5 mM (Yoon et al., 1998). The cells were maintained using Iscove’s modified Dulbecco’s medium (IMDM, Gibco, Grand Island, NY) supplemented with 5% dialyzed fetal bovine serum (dFBS, Gibco) and 5 mM MTX. For the SF medium formulation, medium supplements, except for insulin, were dissolved in water before addition to the culture media. Insulin was dissolved in 1 M acetic acid. According to a Plackett–Burman matrix, 24 types of SF media were prepared for each set of experiments and filtered through a 0.2 mm membrane before use. MTX was not included in SF media. Exponentially growing cells in T-flasks containing IMDM with 5% dFBS and 5 mM MTX were used as an inoculum. After trypsinization, the cells were diluted in fresh media and centrifuged at 1200× g for 5 min at 4°C. After removing the supernatant, the cells were suspended in phosphate buffered saline (PBS). The cells were pelleted again by centrifugation and were resuspended in the medium under test. Cells were seeded at a density of 1.5× 105 cells ml − 1 to the spinner flasks (Bellco Glass, Vineland, NJ) containing 50 ml of prepared SF media. The experiments were carried out in triplicate. The spinners were stirred at 100 rpm and were incubated in a 5% CO2/air mixture, humidified at 37°C. After 48, 72 and 96 h cultivation, cell concentration and viability were determined by trypan blue dye exclusion. The spent media were aliquoted and kept frozen at −80°C.
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
87
Table 1 Plackett–Burman matrix of the experimental design Runs
Variable
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
A + + + + + − + − + + − − + + − − + − + − − − − −
B + + + + − + − + + − − + + − − + − + − − − − + −
C + + + − + − + + − − + + − − + − + − − − − + + −
D + + − + − + + − − + + − − + − + − − − − + + + −
E + − + − + + − − + + − − + − + − − − − + + + + −
F − + − + + − − + + − − + − + − − − − + + + + + −
G + − + + − − + + − − + − + − − − − + + + + + − −
H − + + − − + + − − + − + − − − − + + + + + − + −
I + + − − + + − − + − + − − − − + + + + + − + − −
J + − − + + − − + − + − − − − + + + + + − + − + −
K − − + + − − + − + − − − − + + + + + − + − + + −
2.2. EPO assay The EPO concentration secreted into the medium was determined by a sandwich enzyme linked immunosorbent assay (ELISA) as described previously (Yoon et al., 1998).
2.3. Determination of biological acti6ities of EPO For purification of secreted EPO into the media, the culture supernatants were loaded on a reversed phase high performance liquid chromatography (HPLC) column (Hi-Pore RP-304 column, BIO-RAD, Hercules, CA) and were eluted with a linear gradient of ethanol. The activity of purified EPO was determined by in vitro and in vivo bioassay methods. An in vitro EPO assay that used the stimulatory effect of EPO on the incorporation of [3H]thymidine into DNA in cultured mouse spleen cells was done as described by Krystal (1983). An in vivo EPO
L − + + − − + − + − − − − + + + + + − + − + + − −
M + + − − + − + − − − − + + + + + − + − + + − − −
N + − − + − + − − − − + + + + + − + − + + − − + −
O − − + − + − − − − + + + + + − + − + + − − + + −
P − + − + − − − − + + + + + − + − + + − − + + − −
Q + − + − − − − + + + + + − + − + + − − + + − − −
R − + − − − − + + + + + − + − + + − − + + − − + −
S + − − − − + + + + + − + − + + − − + + − − + − −
T − − − − + + + + + − + − + + − − + + − − + − + −
W − − − + + + + + − + − + + − − + + − − + − + − −
X − − + + + + + − + − + + − − + + − − + − + − − −
Y − + + + + + − + − + + − − + + − − + − + − − − −
activity was evaluated by reticulocyte counts in peripheral blood of normal mice as described by Barbone et al. (1994).
2.4. Statistical methodology The Plackett–Burman design, which allows the investigation of up to N− 1 variables with N experiments, was applied to develop SF media for the suspension culture of rCHO cells. The basal medium used was IMDM with Fe(NO3)39H2O (2 mg l − 1), CuCl2 (0.0025 mg l − 1) and ZnSO47H2O (1 mg l − 1). The experiments were carried out according to the design matrix shown in Table 1. Each row represents one trial (culture medium) and each column represents a single variable (medium component). Two variables (I and P in Table 1) are dummy variables which are imaginary nutrients. Since no change is made to them, they are used to determine the standard error (S.E.). The (+ ) and (−) elements represent the
88
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
Table 2 Nutritional variables and their corresponding concentrations
A B C D E F G H I J K L M N O P Q R S T W X Y
Nutritional variables
Low level (mg 1−1)
High level (mg l−1)
Alanine Arginine Asparagine/aspartate Cystine Glutamine Glutamate Glycine Serine Dummy Methionine Phenylalanine/tyrosine Threonine/valine/isoleucine Leucine/tryptophan/lysine Proline/histidine Na2SeO3 Dummy Insulin Transferrin Hydrocortisone Ethanolamine Phosphatidylcholine Glutathione Pluronic F68
25 84 25/30 91 584 75 30 42
37.5 168 50/60 182 759 97.5 60 84
30 66/104 95/94/105 105/16/146 40/42 0.02
60 132/208 190/188/210 210/32/292 80/84 0.04
2.5 10 0 3 0 0 0
high and low levels of each variable present within each trial, respectively. Independent variables selected for analysis and their corresponding concentrations are summarized in Table 2. Medium components known to be essential for cell growth were selected as independent variables. They were insulin, transferrin, hydrocortisone, ethanolamine, phosphatidylcholine, glutathione, pluronic F68, sodium selenite and amino acids (all from Sigma). According to the previous work by Castro et al. (1992), amino acids involved in the major metabolic pathways (alanine, glutamine, glutamate, glycine and serine) were tested independently because the metabolism of one amino acid may be influenced by the levels of others. Arginine, cystine and methionine were also considered independently. Due to the limited number of variables, the remaining amino acids were tested in five subgroups considering their similarities in metabolic pathways. They were: (a) asparagine and aspartate; (b) phenylalanine and tyrosine; (c)
5 20 10 6 5 1 1
threonine, valine and isoleucine; (d) leucine, tryptophan and lysine; and (e) proline and histidine. A low concentration ( − ) of each nutrient was taken as that contained in the IMDM whereas a high concentration (+ ) was established at 1.3–2fold. Statistical analyses were made to identify medium components that had a positive effect on cell growth and/or EPO production. First, the effect of each variable upon measured response (EA) was determined by subtracting the average response of the low level (− ) from that of the high level (+ ) for each parameter (i.e. cell growth and EPO production). Next, the S.E. of EA was calculated as: (S.E.)2 = (sum of squares of the dummy effect)/ (number of dummy variables). Finally, the significance level of each variable effect was determined using t distribution associated with Student’s t- test: t-value= (EA)/(S.E.)
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
Confidence levels were accepted only to 80% level. Accordingly, it was possible to rank the medium components that significantly affect cell growth and/or EPO production, either positively or negatively.
3. Results In order to develop a SF medium for the production of EPO by suspension culture of rCHO, medium supplements important for cell growth and/or EPO production need to be identified. For an efficient test of a large number of medium supplements, a statistical approach based on a Plackett–Burman design was applied. Experiments were carried out according to the Plackett–Burman matrix shown in Table 1. A last row of (−) elements is a basic assembly referring to the basal medium composition for each experiment set. The basal medium was prepared by supplementing IMDM with Fe(NO3)39H2O, CuCl2 and ZnSO47H2O, salts that are generally included in SF medium formulations (Castro et al., 1992; Keen and Rapson, 1995). This medium does not contain any nucleoside precursors, and therefore it does not bypass dhfr selection. Insulin, transferrin, ethanolamine, sodium selenite, and other medium supplements were also added to the basal medium to determine their optimal concentrations. Thus, a total 28 different supplements were selected as variables, based on their growth promoting abilities reported in previous studies (Barnes and Sato, 1980; Glassy et al., 1988; Butler and Jenkins, 1989; Castro et al., 1992; Keen and Rapson, 1995). Since some amino acids were tested in sub-groups, 21 independent variables with two levels were tested in the 24 experiments. Their randomized order and the corresponding concentrations are given in Table 2. The parameters evaluated in statistical analyses were maximum viable cell concentration (cells ml − 1) and EPO concentration (mg ml − 1) in each culture. The rCHO cells were cultivated in 24 types of SF media. Their growth and EPO production were determined. After all sets of experiments were repeated three times, statistical analyses were
89
made. Table 3 presents the results of data analyses for the effect of each variable on each parameter, and its associated t-value and level of significance. Confidence levels were accepted only to 80%. The effects of dummy variables (I and P) were never as high as to allow of significance greater than 80%. These were all included in the calculation of the variance of the effects, as a measure of the uncertainty of the results. Both positive and negative variables were found in this analysis. Only positive variables which were referred to having stimulatory effects on cell growth and/or EPO production were considered for medium optimization. Negative variables were thought to have inhibitory effects on cell growth and/or antibody production at the high level. The components included in the positive group were glutamate, serine, methionine, phosphatidylcholine, hydrocortisone and pluronic F68. In order to formulate the SF medium, these components were added to IMDM at the corresponding high level shown in Table 2. For insulin and transferrin, the increase in their concentration did not show any significant effects on cell growth and antibody production. Thus, these and other components which were not included in the positive group were supplemented to the basal medium at the corresponding low level shown in Table 2. Finally, the SF medium for suspension culture of rCHO cells was developed (Table 4). In order to make a comparison between the optimized SF medium and IMDM supplemented with 5% dFBS in regard to cell growth and EPO production, rCHO cells were cultivated in both media (Fig. 1). The maximum viable cell concentration obtained in the SF medium was 7.1 × 105 cells ml − 1, which is only 54% of that in serumsupplemented medium. However, EPO titer in the SF medium was 17.9 mg ml − 1, which is approximately 79% of that in serum-supplemented medium. This is because EPO productivity per cell in the SF medium was higher. The EPO productivity per cell in the SF and serum-supplemented media was 9.62 and 6.88 mg 106 cells − 1 day − 1, respectively. Since the type of media may influence the biological activity of EPO, the in vitro and in vivo biological activities of EPOs produced in the SF
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
90
Table 3 Nutritional variables and their effects on viable cell concentration and EPO titera Variables
Alanine Arginine Asn/Asp Cystine Glutamine Glutamate Glycine Serine Dummy Methionine Phe/Tyr Thr/Val/Iso Leu/Try/Lys Pro/His Na2SeO3 Dummy Insulin Transferrin Hydrocortisone Ethanolamine Phosphatidylcholine Glutathione Pluronic F68 a
Viable cell concentration
EPO titer
Effect (cells ml−1)
t-Value
Significance (%)
Effect (mg ml−1)
t-Value
Significance (%)
2.08×103 3.54×104 4.83×104 −1.94×104 4.56×104 7.54×104 5.11×104 9.21×104 −2.63×104 7.99×104 −4.14×104 −6.92×103 −9.89×104 −6.56×104 −1.66×104 6.58×103 −8.33×101 2.09×104 5.91×104 3.19×104 5.94×104 −8.92×103 1.25×105
0.11 1.85 2.53 −1.01 2.38 3.94 2.67 4.81 −1.37 4.18 −2.16 −0.36 −5.17 −3.43 −0.87 0.34 −0.00 1.09 3.09 1.67 3.11 −0.47 6.55
7 68 76 50 74 84 77 87 60 85 72 22 88 82 45 21 0 53 80 66 80 28 90
1.85×10−1 2.53×10−2 −3.34×10−1 1.75×10−1 4.00×10−1 −2.06×10−1 −1.56×10−1 3.13×10−1 −1.53×10−1 2.11×10−1 −1.35×10−1 −3.04×10−1 −4.69×10−1 −2.50×10−1 9.46×10−3 1.64×10−1 2.32×10−1 2.66×10−1 4.21×10−1 −8.78×10−2 2.38×10−1 −1.13×10−1 3.26×10−1
1.17 0.16 −2.10 1.10 2.52 −1.30 −0.98 1.97 −0.97 1.33 −0.85 −1.92 −2.95 −1.58 0.06 1.03 1.46 1.68 2.65 −0.55 1.50 −0.71 2.06
55 0 72 53 76 58 49 70 49 59 45 69 79 64 4 51 62 66 77 32 63 39 71
Positive and negative variables are shown in bold and italic, respectively.
and serum-supplemented media were determined. As summarized in Table 5, the biological activitity of EPO produced in the SF medium was comparable to that in serum-supplemented medium.
4. Discussion The Plackett–Burman statistical design technique used in this study was found to be very efficient in the development of a SF medium for the production of EPO by suspension culture of rCHO cells. Medium supplements which had positive effects on cell growth and/or EPO production could be identified in only 24 tests. In determining the level of components necessary for SF medium formulations, this method allows evaluation of the probability of the observed effect purely by chance. Thus, it is possible to rank the variables that affect the parameters, either
positively or negatively. The level of confidence accepted for this analysis, which is a trade-off between the probability of meaningful variables and the effect of experimental errors, was 80%. The effects of dummy variables were never as high as to allow a level of significance greater than 80% (Table 3), showing validity of this analysis. Medium supplements which had positive effects on cell growth were glutamate, methionine, serine, hydrocortisone, phosphatidylcholine, and pluronic F68. Glutamate also arises as a result of the metabolism of glutamine. Thus, a low level of glutamate was expected to satisfy glutamate requirement of cells. However, it was not enough to do so. It might be because a significant portion of glutamate nitrogen was incorporated into aspartate and alanine via transaminase reactions (Ardawi and Newsholme, 1982; Hayter et al., 1991).
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
91
Table 4 Composition of an optimized SF medium IMDM supplemented with: Concentration (mg l−1)
Components Fe(NO3)39H2O CuCl2 ZnSO 7H O 4
2 0.0025 1 0.02 22.5 42 30 2.5 10 10 3 5 1
2
Na2SeO3 Glutamate Serine Methionine Insulin Transferrin Hydrocortisone Ethanolamine Phosphatidylcholine Pluronic F68
Methionine provides sulfur and methyl groups in metabolic reactions. Moreover, it is the first amino acid to be inserted in all eukaryotic polypeptide chains. Therefore, it is likely that requirement for methionine is significant in this rCHO cell line expressing EPO. Methionine depletion was also found in batch culture of some CHO cell lines (Kurano et al. 1990). In contrast to these results, an increase from 15 to 30 mg l − 1 in the amount of methionine was found to inhibit the growth of other CHO cell lines (Castro et al., 1992). This suggests that requirement for methionine significantly varies among CHO cell lines. Serine is known to be utilized extensively by CHO cells during batch culture (Hayter et al., 1991). Thus, serine may be a growth-limiting nutrient. In addition, glycine as a precursor of puriTable 5 In vitro and in vivo activities of erythropoietins (EPOs) produced in the serum-free (SF) and serum supplemented media Medium
In vitro activitya (×105 IU mg−1)
In vivo activityb (×104 IU mg−1)
SF Serum-supplemented
1.50 9 0.18 1.58 9 0.19
7.7590.53 8.06 90.67
a b
Average 9S.D., n =5. n= 10.
Fig. 1. Cell growth and erythropoietin (EPO) production during a batch culture of recombinant Chinese hamster ovary (rCHO) cells in SF medium () and serum-supplemented medium (). (A) Viable cell concentration; (B) EPO titer.
nes in nucleic acid synthesis is formed from serine and tetrahydrofolate. Since the cells were grown in the absence of MTX, their synthesis of tetrahydrofolate may be significant. The presence of sufficient levels of tetrahydrofolate can contribute to a high rate of serine metabolism in proliferating cells and an additional requirement for serine (Snell et al., 1987). Hydrocortisone is one of the growth hormones present in serum which have mitogenic effects. Its positive effects may be explained by the previous observation that it could promote cell proliferation (Freshney, 1994). Phosphatidylcholine is one of phospholipids which are major components of the cell membranes and are critical in maintaining the integrity and regulation of various membrane functions (Castro et al., 1996). Suspension culture may require a high level of phosphatidylcholine in order to induce changes in the fluidity of the membrane
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
92
lipid bilayer as a result of changes in phospholipid fatty acyl composition. The changes in membrane fluidity may influence the shear sensitivity of cells in culture (Schmid et al., 1991). Pluronic F68, a non-ionic surfactant, also showed a positive effect on cell growth. The protective effect of pluronic F68 in adverse hydrodynamic environments has been reported (Papoutsakis, 1991). Suspension culture may require pluronic F68 for the protection of freely suspended cells against agitation damage. Based on the variables identified by the Plackett – Burman design, a SF medium for EPO production by suspension culture of rCHO cells (LGE10-9-27) has been formulated. Although the SF medium developed in this study includes expensive insulin and transferrin, their cost is less than 3% of the price of EPO on the market. In addition, by decreasing their concentration, the cost of SF medium can be reduced further. For the commercialization of a cell culture product, rapid development of production medium is important. This statistical design technique enabled the development of the SF medium rapidly in a relatively small number of experiments. While the EPO titer obtained in this SF medium was only about 79% of that in IMDM supplemented with 5% dFBS, the intrinsic benefits of using a SF medium, such as a facilitated down-stream purification and a reduced requirement for quality assurance tests, clearly outweigh the small decrease in EPO titer. Furthermore, the biological activity of EPO produced in this SF medium, one of the most important factors considered in development of SF media, was comparable to that in the serum-supplemented medium. In conclusion, a Plackett – Burman design technique enabled the efficient development of a SF medium for the production of EPO by suspension culture of rCHO cells (LGE10-9-27). This SF medium formulation will be a reference for developing a SF medium for suspension culture of other rCHO cell lines.
Acknowledgements The authors thank Ms H.S. Cheong for help
with determination of biological activities of EPO.
References Ardawi, M.S., Newsholme, E.A., 1982. Maximum activities of some enzymes of glycolysis, the tricarboxylic acid cycle and ketone-body and glutamine utilization pathways in lymphocytes of the rat. Biochem. J. 208, 743 – 748. Barbone, A.G., Aparicio, B., Anderson, D.W., Natarajan, J., Ritchie, D.M., 1994. Reticulocyte measurements as a bioassay for erythropoietin. J. Pharm. Biomed. Anal. 12, 515 – 522. Barnes, D., Sato, G.H., 1980. Methods for growth of cultured cells in serum-free medium. Anal. Biochem. 102, 255 – 270. Butler, M., Jenkins, H., 1989. Nutritional aspects of the growth of animal cells in culture. J. Biotechnol. 12, 97 – 110. Castro, P.M.L., Hayter, P.M., Ison, A.P., Bull, A.T., 1992. Application of a statistical design to the optimization of culture medium for recombinant interferon-gamma production by Chinese hamster ovary cells. Appl. Microbiol. Biotechnol. 38, 84 – 90. Castro, P.M.L., Ison, A.P., Hayter, P.M., Bull, A.T., 1996. CHO cell growth and recombinant interferon-g production: effects of BSA, pluronic and lipids. Cytotechnology 19, 27 – 36. Davis, J.M., Arakawa, T., Strickland, T.W., Yphantis, D.A., 1987. Characterization of recombinant human erythropoietin produced in Chinese hamster ovary cells. Biochemistry 26, 2633 – 2638. Fisher, J.W., 1983. Control of erythropoietin production. Proc. Soc. Exp. Biol. Med. 173, 289 – 305. Freshney, R.I., 1994. Culture of Animal Cells: A Manual of Basic Technique, 3rd edition. Wiley J. Liss, New York, pp. 85 – 98. Ganne, V., Mignot, G., 1991. Application of statistical design of experiments to the optimization of factor VIII expression by CHO cells. Cytotechnology 6, 233 – 240. Glassy, M.C., Tharakan, J.P., Chau, P.C., 1988. Serum-free media in hybridoma culture and monoclonal antibody production. Biotechnol. Bioeng. 32, 1015 – 1028. Hayter, P.M., Curling, E.M.A., Baines, A.J., Jenkins, N., Salmon, I., Strange, P.G., Bull, A.T., 1991. Chinese hamster ovary cell growth and interferon production kinetics in stirred batch culture. Appl. Microbiol. Biotechnol. 34, 559 – 564. Keen, M.J., Rapson, N.T., 1995. Development of a serum-free culture medium for the large scale production of recombinant protein from a Chinese hamster ovary cell line. Cytotechnology 17, 153 – 163. Krystal, G., 1983. A simple microassay for erythropoietin based on 3H-thymidine incorporation into spleen cells from phenylhydrazine treated mice. Exp. Hematol. 11, 649 – 660.
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93 Kurano, N., Leist, C., Messi, F., Kurano, S., Fiechter, A., 1990. Growth behavior of Chinese hamster ovary cells in a compact loop bioreactor. 2. Effects of medium components and waste products. J. Biotechnol. 15, 113–128. Lin, F.K., Suggs, S., Lin, C.H., Browne, J.K., Smalling, R., Egrie, J.C., Chen, K.K., Fox, G.M., Martin, F., Stabinsky, Z., Badrawi, S.M., Lai, P.H, Goldwasser, E., 1985. Cloning and expression of the human erythropoietin gene. Proc. Natl. Acad. Sci. USA 82, 7580–7584. Papoutsakis, E.T., 1991. Media additives for protecting freely suspended animal cells against agitation and aeration damage. Trends Biotechnol. 9, 316–324. Sinacore, M.S., Charlebois, T.S., Harrison, S., Brennan, S., Richards, T., Hamilton, M., Scott, S., Brodeur, S., Oakes, P., Leonard, M., Switzer, M., Anagnostopoulos, A., Foster, B., Harris, A., Jankowski, M., Bond, M., Martin, S., Adamson, S.R., 1996. CHO DUKX cell lineages preadapted to growth in serum-free suspension culture enable rapid development of cell culture processes for the manufacture of recombinant proteins. Biotechnol. Bioeng. 52, 518 – 528.
.
93
Schmid, G., Zilg, H., Eberhard, U., Johannsen, R., 1991. Effect of free fatty acids and phospholipids on growth of and production formation by recombinant baby hamster kidney (rBHK) and Chinese hamster ovary (rCHO) cells in culture. J. Biotechnol. 17, 155 – 167. Snell, K., Natsumeda, Y., Weber, G., 1987. The modulation of serine metabolism in hepatoma 3924A during different phases of cellular proliferation in culture. Biochem. J. 245, 609 – 612. Sytkowski, A.J., 1984. Erythropoietin: a prime regulator of red cell differentiation. Biomed. Pharm. 38, 369 – 371. Yoon, S.K., Ahn, Y.H., Kwon, I.C., Han, K., Song, J.Y., 1998. Influence of reducing agents on the secretion rate of recombinant erythropoietin from CHO cells. Biotechnol. Lett. 20, 101 – 104. Zang, M., Trautmann, H., Gandor, C., Messi, F., Asselbergs, F., Leist, C., Fiechter, A., Reiser, J., 1995. Production of recombinant proteins in Chinese hamster ovary cells using a protein-free cell culture medium. Bio/Technology 13, 389 – 392.
Development of a serum-free medium for the production of erythropoietin by suspension culture of recombinant Chinese hamster ovary cells using a statistical design Gyun Min Lee a,*, Eun Jung Kim a, No Soo Kim a, Sung Kwan Yoon b, Yong Ho Ahn b, Ji Yong Song b a
Department of Biological Sciences, Korea Ad6anced Institute of Science and Technology 373 -1, Kusong-Dong, Yusong-Gu, Taejon 305 -701, South Korea b Biotech Research Institute, LG Chem. Research Park, P.O. Box 61, Yusong-Gu, Taejon 305 -380, South Korea Received 10 June 1998; received in revised form 14 December 1998; accepted 12 January 1999
Abstract In order to develop a serum-free (SF) medium for the production of erythropoietin (EPO) by suspension culture of recombinant Chinese hamster ovary (rCHO) cells, a statistical optimization approach based on a Plackett – Burman design was adopted. A basal medium was prepared by supplementing Iscove’s modified Dulbecco’s medium (IMDM) with Fe(NO3)39H2O, CuCl2 and ZnSO47H2O which are generally contained in SF medium formulations. Insulin, transferrin and ethanolamine were also supplemented to the basal medium to determine their optimal concentrations. From this statistical analysis, glutamate, serine, methionine, phosphatidylcholine, hydrocortisone and pluronic F68 were identified as positive determinants for cell growth. The SF medium was formulated by supplementing the basal medium with components showing positive effects on cell growth in suspension culture. An EPO titer in this optimized SF medium was 79% of that in IMDM supplemented with 5% dialyzed fetal bovine serum (dFBS). Furthermore, the in vitro and in vivo biological activities of EPO produced in the SF medium were comparable to those produced in the serum-supplemented medium. Taken together, the results obtained here show that a Plackett–Burman design facilitates the development of SF media for the production of EPO by suspension culture of rCHO cells. © 1999 Elsevier Science B.V. All rights reserved. Keywords: Plackett–Burman design; Serum-free medium; rCHO; Erythropoietin
* Corresponding author. Tel.: +82-42-869-2618; fax: +82-42-869-2610. E-mail address: [email protected] (G.M. Lee) 0168-1656/99/$ - see front matter © 1999 Elsevier Science B.V. All rights reserved. PII: S 0 1 6 8 - 1 6 5 6 ( 9 9 ) 0 0 0 0 4 - 8
86
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
1. Introduction Erythropoietin (EPO), a glycoprotein primarily produced by the kidney, is the principal regulator of red blood cell production (Fisher, 1983; Sytkowski, 1984). For the large-scale production of EPO, a gene coding for human EPO was cloned and the corresponding protein was expressed in Chinese hamster ovary (CHO) cells (Lin et al., 1985; Davis et al., 1987). For commercial production of EPO, recombinant CHO (rCHO) cells are frequently cultivated in roller bottles since they preferentially adhere to a surface. However, the use of roller bottles for large-scale cultivation requires intensive labor and often makes it difficult to control product quality. Alternatively, rCHO cells can be cultivated in suspension. In fact, the suspension culture-based manufacturing process has become the method of large-scale, commercial production of therapeutically important proteins from rCHO cells (Sinacore et al., 1996). Furthermore, it is desirable to use serum-free (SF) media in suspension culture because serum can cause problems in the subsequent processes (Glassy et al., 1988; Keen and Rapson, 1995). However, since there is no universal SF medium applicable to all cell lines, specific media suitable for each cell line need to be developed (Hayter et al., 1991; Keen and Rapson, 1995; Zang et al., 1995). The nutritional requirements of mammalian cells are so complex that extensive efforts have been made to identify suitable serum-substituting supplements (Castro et al., 1992). The classical approach of changing one medium component at a time is still being used but becomes impractical because it is time-consuming and has the risk of neglecting interactions among supplements (Castro et al., 1992; Freshney, 1994). The need for efficient methods for developing SF medium has led to adoption of a statistical experimental design (Ganne and Mignot, 1991; Castro et al., 1992). In this study, a Plackett – Burman statistical design, which provides an efficient way of screening a large number of medium supplements and identifying important ones, was used to develop an optimized SF medium for the production of EPO by suspension culture of rCHO cells. Furthermore,
the biological activity of EPO produced in this SF medium was examined.
2. Materials and methods
2.1. Cells, medium preparation and culture conditions The rCHO cells (LGE10-9-27) expressing high levels of human EPO were used in this study. These rCHO cells were developed by transfection of a vector containing the dihydrofolate reductase (dhfr) and human EPO genes into dhfr-deficient CHO cells (ATCC CRL 9096). Their EPO gene was coamplified with dhfr by stepwise increments in methotrexate (MTX, Sigma, St. Louis, MO) level up to 5 mM (Yoon et al., 1998). The cells were maintained using Iscove’s modified Dulbecco’s medium (IMDM, Gibco, Grand Island, NY) supplemented with 5% dialyzed fetal bovine serum (dFBS, Gibco) and 5 mM MTX. For the SF medium formulation, medium supplements, except for insulin, were dissolved in water before addition to the culture media. Insulin was dissolved in 1 M acetic acid. According to a Plackett–Burman matrix, 24 types of SF media were prepared for each set of experiments and filtered through a 0.2 mm membrane before use. MTX was not included in SF media. Exponentially growing cells in T-flasks containing IMDM with 5% dFBS and 5 mM MTX were used as an inoculum. After trypsinization, the cells were diluted in fresh media and centrifuged at 1200× g for 5 min at 4°C. After removing the supernatant, the cells were suspended in phosphate buffered saline (PBS). The cells were pelleted again by centrifugation and were resuspended in the medium under test. Cells were seeded at a density of 1.5× 105 cells ml − 1 to the spinner flasks (Bellco Glass, Vineland, NJ) containing 50 ml of prepared SF media. The experiments were carried out in triplicate. The spinners were stirred at 100 rpm and were incubated in a 5% CO2/air mixture, humidified at 37°C. After 48, 72 and 96 h cultivation, cell concentration and viability were determined by trypan blue dye exclusion. The spent media were aliquoted and kept frozen at −80°C.
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
87
Table 1 Plackett–Burman matrix of the experimental design Runs
Variable
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24
A + + + + + − + − + + − − + + − − + − + − − − − −
B + + + + − + − + + − − + + − − + − + − − − − + −
C + + + − + − + + − − + + − − + − + − − − − + + −
D + + − + − + + − − + + − − + − + − − − − + + + −
E + − + − + + − − + + − − + − + − − − − + + + + −
F − + − + + − − + + − − + − + − − − − + + + + + −
G + − + + − − + + − − + − + − − − − + + + + + − −
H − + + − − + + − − + − + − − − − + + + + + − + −
I + + − − + + − − + − + − − − − + + + + + − + − −
J + − − + + − − + − + − − − − + + + + + − + − + −
K − − + + − − + − + − − − − + + + + + − + − + + −
2.2. EPO assay The EPO concentration secreted into the medium was determined by a sandwich enzyme linked immunosorbent assay (ELISA) as described previously (Yoon et al., 1998).
2.3. Determination of biological acti6ities of EPO For purification of secreted EPO into the media, the culture supernatants were loaded on a reversed phase high performance liquid chromatography (HPLC) column (Hi-Pore RP-304 column, BIO-RAD, Hercules, CA) and were eluted with a linear gradient of ethanol. The activity of purified EPO was determined by in vitro and in vivo bioassay methods. An in vitro EPO assay that used the stimulatory effect of EPO on the incorporation of [3H]thymidine into DNA in cultured mouse spleen cells was done as described by Krystal (1983). An in vivo EPO
L − + + − − + − + − − − − + + + + + − + − + + − −
M + + − − + − + − − − − + + + + + − + − + + − − −
N + − − + − + − − − − + + + + + − + − + + − − + −
O − − + − + − − − − + + + + + − + − + + − − + + −
P − + − + − − − − + + + + + − + − + + − − + + − −
Q + − + − − − − + + + + + − + − + + − − + + − − −
R − + − − − − + + + + + − + − + + − − + + − − + −
S + − − − − + + + + + − + − + + − − + + − − + − −
T − − − − + + + + + − + − + + − − + + − − + − + −
W − − − + + + + + − + − + + − − + + − − + − + − −
X − − + + + + + − + − + + − − + + − − + − + − − −
Y − + + + + + − + − + + − − + + − − + − + − − − −
activity was evaluated by reticulocyte counts in peripheral blood of normal mice as described by Barbone et al. (1994).
2.4. Statistical methodology The Plackett–Burman design, which allows the investigation of up to N− 1 variables with N experiments, was applied to develop SF media for the suspension culture of rCHO cells. The basal medium used was IMDM with Fe(NO3)39H2O (2 mg l − 1), CuCl2 (0.0025 mg l − 1) and ZnSO47H2O (1 mg l − 1). The experiments were carried out according to the design matrix shown in Table 1. Each row represents one trial (culture medium) and each column represents a single variable (medium component). Two variables (I and P in Table 1) are dummy variables which are imaginary nutrients. Since no change is made to them, they are used to determine the standard error (S.E.). The (+ ) and (−) elements represent the
88
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
Table 2 Nutritional variables and their corresponding concentrations
A B C D E F G H I J K L M N O P Q R S T W X Y
Nutritional variables
Low level (mg 1−1)
High level (mg l−1)
Alanine Arginine Asparagine/aspartate Cystine Glutamine Glutamate Glycine Serine Dummy Methionine Phenylalanine/tyrosine Threonine/valine/isoleucine Leucine/tryptophan/lysine Proline/histidine Na2SeO3 Dummy Insulin Transferrin Hydrocortisone Ethanolamine Phosphatidylcholine Glutathione Pluronic F68
25 84 25/30 91 584 75 30 42
37.5 168 50/60 182 759 97.5 60 84
30 66/104 95/94/105 105/16/146 40/42 0.02
60 132/208 190/188/210 210/32/292 80/84 0.04
2.5 10 0 3 0 0 0
high and low levels of each variable present within each trial, respectively. Independent variables selected for analysis and their corresponding concentrations are summarized in Table 2. Medium components known to be essential for cell growth were selected as independent variables. They were insulin, transferrin, hydrocortisone, ethanolamine, phosphatidylcholine, glutathione, pluronic F68, sodium selenite and amino acids (all from Sigma). According to the previous work by Castro et al. (1992), amino acids involved in the major metabolic pathways (alanine, glutamine, glutamate, glycine and serine) were tested independently because the metabolism of one amino acid may be influenced by the levels of others. Arginine, cystine and methionine were also considered independently. Due to the limited number of variables, the remaining amino acids were tested in five subgroups considering their similarities in metabolic pathways. They were: (a) asparagine and aspartate; (b) phenylalanine and tyrosine; (c)
5 20 10 6 5 1 1
threonine, valine and isoleucine; (d) leucine, tryptophan and lysine; and (e) proline and histidine. A low concentration ( − ) of each nutrient was taken as that contained in the IMDM whereas a high concentration (+ ) was established at 1.3–2fold. Statistical analyses were made to identify medium components that had a positive effect on cell growth and/or EPO production. First, the effect of each variable upon measured response (EA) was determined by subtracting the average response of the low level (− ) from that of the high level (+ ) for each parameter (i.e. cell growth and EPO production). Next, the S.E. of EA was calculated as: (S.E.)2 = (sum of squares of the dummy effect)/ (number of dummy variables). Finally, the significance level of each variable effect was determined using t distribution associated with Student’s t- test: t-value= (EA)/(S.E.)
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
Confidence levels were accepted only to 80% level. Accordingly, it was possible to rank the medium components that significantly affect cell growth and/or EPO production, either positively or negatively.
3. Results In order to develop a SF medium for the production of EPO by suspension culture of rCHO, medium supplements important for cell growth and/or EPO production need to be identified. For an efficient test of a large number of medium supplements, a statistical approach based on a Plackett–Burman design was applied. Experiments were carried out according to the Plackett–Burman matrix shown in Table 1. A last row of (−) elements is a basic assembly referring to the basal medium composition for each experiment set. The basal medium was prepared by supplementing IMDM with Fe(NO3)39H2O, CuCl2 and ZnSO47H2O, salts that are generally included in SF medium formulations (Castro et al., 1992; Keen and Rapson, 1995). This medium does not contain any nucleoside precursors, and therefore it does not bypass dhfr selection. Insulin, transferrin, ethanolamine, sodium selenite, and other medium supplements were also added to the basal medium to determine their optimal concentrations. Thus, a total 28 different supplements were selected as variables, based on their growth promoting abilities reported in previous studies (Barnes and Sato, 1980; Glassy et al., 1988; Butler and Jenkins, 1989; Castro et al., 1992; Keen and Rapson, 1995). Since some amino acids were tested in sub-groups, 21 independent variables with two levels were tested in the 24 experiments. Their randomized order and the corresponding concentrations are given in Table 2. The parameters evaluated in statistical analyses were maximum viable cell concentration (cells ml − 1) and EPO concentration (mg ml − 1) in each culture. The rCHO cells were cultivated in 24 types of SF media. Their growth and EPO production were determined. After all sets of experiments were repeated three times, statistical analyses were
89
made. Table 3 presents the results of data analyses for the effect of each variable on each parameter, and its associated t-value and level of significance. Confidence levels were accepted only to 80%. The effects of dummy variables (I and P) were never as high as to allow of significance greater than 80%. These were all included in the calculation of the variance of the effects, as a measure of the uncertainty of the results. Both positive and negative variables were found in this analysis. Only positive variables which were referred to having stimulatory effects on cell growth and/or EPO production were considered for medium optimization. Negative variables were thought to have inhibitory effects on cell growth and/or antibody production at the high level. The components included in the positive group were glutamate, serine, methionine, phosphatidylcholine, hydrocortisone and pluronic F68. In order to formulate the SF medium, these components were added to IMDM at the corresponding high level shown in Table 2. For insulin and transferrin, the increase in their concentration did not show any significant effects on cell growth and antibody production. Thus, these and other components which were not included in the positive group were supplemented to the basal medium at the corresponding low level shown in Table 2. Finally, the SF medium for suspension culture of rCHO cells was developed (Table 4). In order to make a comparison between the optimized SF medium and IMDM supplemented with 5% dFBS in regard to cell growth and EPO production, rCHO cells were cultivated in both media (Fig. 1). The maximum viable cell concentration obtained in the SF medium was 7.1 × 105 cells ml − 1, which is only 54% of that in serumsupplemented medium. However, EPO titer in the SF medium was 17.9 mg ml − 1, which is approximately 79% of that in serum-supplemented medium. This is because EPO productivity per cell in the SF medium was higher. The EPO productivity per cell in the SF and serum-supplemented media was 9.62 and 6.88 mg 106 cells − 1 day − 1, respectively. Since the type of media may influence the biological activity of EPO, the in vitro and in vivo biological activities of EPOs produced in the SF
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
90
Table 3 Nutritional variables and their effects on viable cell concentration and EPO titera Variables
Alanine Arginine Asn/Asp Cystine Glutamine Glutamate Glycine Serine Dummy Methionine Phe/Tyr Thr/Val/Iso Leu/Try/Lys Pro/His Na2SeO3 Dummy Insulin Transferrin Hydrocortisone Ethanolamine Phosphatidylcholine Glutathione Pluronic F68 a
Viable cell concentration
EPO titer
Effect (cells ml−1)
t-Value
Significance (%)
Effect (mg ml−1)
t-Value
Significance (%)
2.08×103 3.54×104 4.83×104 −1.94×104 4.56×104 7.54×104 5.11×104 9.21×104 −2.63×104 7.99×104 −4.14×104 −6.92×103 −9.89×104 −6.56×104 −1.66×104 6.58×103 −8.33×101 2.09×104 5.91×104 3.19×104 5.94×104 −8.92×103 1.25×105
0.11 1.85 2.53 −1.01 2.38 3.94 2.67 4.81 −1.37 4.18 −2.16 −0.36 −5.17 −3.43 −0.87 0.34 −0.00 1.09 3.09 1.67 3.11 −0.47 6.55
7 68 76 50 74 84 77 87 60 85 72 22 88 82 45 21 0 53 80 66 80 28 90
1.85×10−1 2.53×10−2 −3.34×10−1 1.75×10−1 4.00×10−1 −2.06×10−1 −1.56×10−1 3.13×10−1 −1.53×10−1 2.11×10−1 −1.35×10−1 −3.04×10−1 −4.69×10−1 −2.50×10−1 9.46×10−3 1.64×10−1 2.32×10−1 2.66×10−1 4.21×10−1 −8.78×10−2 2.38×10−1 −1.13×10−1 3.26×10−1
1.17 0.16 −2.10 1.10 2.52 −1.30 −0.98 1.97 −0.97 1.33 −0.85 −1.92 −2.95 −1.58 0.06 1.03 1.46 1.68 2.65 −0.55 1.50 −0.71 2.06
55 0 72 53 76 58 49 70 49 59 45 69 79 64 4 51 62 66 77 32 63 39 71
Positive and negative variables are shown in bold and italic, respectively.
and serum-supplemented media were determined. As summarized in Table 5, the biological activitity of EPO produced in the SF medium was comparable to that in serum-supplemented medium.
4. Discussion The Plackett–Burman statistical design technique used in this study was found to be very efficient in the development of a SF medium for the production of EPO by suspension culture of rCHO cells. Medium supplements which had positive effects on cell growth and/or EPO production could be identified in only 24 tests. In determining the level of components necessary for SF medium formulations, this method allows evaluation of the probability of the observed effect purely by chance. Thus, it is possible to rank the variables that affect the parameters, either
positively or negatively. The level of confidence accepted for this analysis, which is a trade-off between the probability of meaningful variables and the effect of experimental errors, was 80%. The effects of dummy variables were never as high as to allow a level of significance greater than 80% (Table 3), showing validity of this analysis. Medium supplements which had positive effects on cell growth were glutamate, methionine, serine, hydrocortisone, phosphatidylcholine, and pluronic F68. Glutamate also arises as a result of the metabolism of glutamine. Thus, a low level of glutamate was expected to satisfy glutamate requirement of cells. However, it was not enough to do so. It might be because a significant portion of glutamate nitrogen was incorporated into aspartate and alanine via transaminase reactions (Ardawi and Newsholme, 1982; Hayter et al., 1991).
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
91
Table 4 Composition of an optimized SF medium IMDM supplemented with: Concentration (mg l−1)
Components Fe(NO3)39H2O CuCl2 ZnSO 7H O 4
2 0.0025 1 0.02 22.5 42 30 2.5 10 10 3 5 1
2
Na2SeO3 Glutamate Serine Methionine Insulin Transferrin Hydrocortisone Ethanolamine Phosphatidylcholine Pluronic F68
Methionine provides sulfur and methyl groups in metabolic reactions. Moreover, it is the first amino acid to be inserted in all eukaryotic polypeptide chains. Therefore, it is likely that requirement for methionine is significant in this rCHO cell line expressing EPO. Methionine depletion was also found in batch culture of some CHO cell lines (Kurano et al. 1990). In contrast to these results, an increase from 15 to 30 mg l − 1 in the amount of methionine was found to inhibit the growth of other CHO cell lines (Castro et al., 1992). This suggests that requirement for methionine significantly varies among CHO cell lines. Serine is known to be utilized extensively by CHO cells during batch culture (Hayter et al., 1991). Thus, serine may be a growth-limiting nutrient. In addition, glycine as a precursor of puriTable 5 In vitro and in vivo activities of erythropoietins (EPOs) produced in the serum-free (SF) and serum supplemented media Medium
In vitro activitya (×105 IU mg−1)
In vivo activityb (×104 IU mg−1)
SF Serum-supplemented
1.50 9 0.18 1.58 9 0.19
7.7590.53 8.06 90.67
a b
Average 9S.D., n =5. n= 10.
Fig. 1. Cell growth and erythropoietin (EPO) production during a batch culture of recombinant Chinese hamster ovary (rCHO) cells in SF medium () and serum-supplemented medium (). (A) Viable cell concentration; (B) EPO titer.
nes in nucleic acid synthesis is formed from serine and tetrahydrofolate. Since the cells were grown in the absence of MTX, their synthesis of tetrahydrofolate may be significant. The presence of sufficient levels of tetrahydrofolate can contribute to a high rate of serine metabolism in proliferating cells and an additional requirement for serine (Snell et al., 1987). Hydrocortisone is one of the growth hormones present in serum which have mitogenic effects. Its positive effects may be explained by the previous observation that it could promote cell proliferation (Freshney, 1994). Phosphatidylcholine is one of phospholipids which are major components of the cell membranes and are critical in maintaining the integrity and regulation of various membrane functions (Castro et al., 1996). Suspension culture may require a high level of phosphatidylcholine in order to induce changes in the fluidity of the membrane
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93
92
lipid bilayer as a result of changes in phospholipid fatty acyl composition. The changes in membrane fluidity may influence the shear sensitivity of cells in culture (Schmid et al., 1991). Pluronic F68, a non-ionic surfactant, also showed a positive effect on cell growth. The protective effect of pluronic F68 in adverse hydrodynamic environments has been reported (Papoutsakis, 1991). Suspension culture may require pluronic F68 for the protection of freely suspended cells against agitation damage. Based on the variables identified by the Plackett – Burman design, a SF medium for EPO production by suspension culture of rCHO cells (LGE10-9-27) has been formulated. Although the SF medium developed in this study includes expensive insulin and transferrin, their cost is less than 3% of the price of EPO on the market. In addition, by decreasing their concentration, the cost of SF medium can be reduced further. For the commercialization of a cell culture product, rapid development of production medium is important. This statistical design technique enabled the development of the SF medium rapidly in a relatively small number of experiments. While the EPO titer obtained in this SF medium was only about 79% of that in IMDM supplemented with 5% dFBS, the intrinsic benefits of using a SF medium, such as a facilitated down-stream purification and a reduced requirement for quality assurance tests, clearly outweigh the small decrease in EPO titer. Furthermore, the biological activity of EPO produced in this SF medium, one of the most important factors considered in development of SF media, was comparable to that in the serum-supplemented medium. In conclusion, a Plackett – Burman design technique enabled the efficient development of a SF medium for the production of EPO by suspension culture of rCHO cells (LGE10-9-27). This SF medium formulation will be a reference for developing a SF medium for suspension culture of other rCHO cell lines.
Acknowledgements The authors thank Ms H.S. Cheong for help
with determination of biological activities of EPO.
References Ardawi, M.S., Newsholme, E.A., 1982. Maximum activities of some enzymes of glycolysis, the tricarboxylic acid cycle and ketone-body and glutamine utilization pathways in lymphocytes of the rat. Biochem. J. 208, 743 – 748. Barbone, A.G., Aparicio, B., Anderson, D.W., Natarajan, J., Ritchie, D.M., 1994. Reticulocyte measurements as a bioassay for erythropoietin. J. Pharm. Biomed. Anal. 12, 515 – 522. Barnes, D., Sato, G.H., 1980. Methods for growth of cultured cells in serum-free medium. Anal. Biochem. 102, 255 – 270. Butler, M., Jenkins, H., 1989. Nutritional aspects of the growth of animal cells in culture. J. Biotechnol. 12, 97 – 110. Castro, P.M.L., Hayter, P.M., Ison, A.P., Bull, A.T., 1992. Application of a statistical design to the optimization of culture medium for recombinant interferon-gamma production by Chinese hamster ovary cells. Appl. Microbiol. Biotechnol. 38, 84 – 90. Castro, P.M.L., Ison, A.P., Hayter, P.M., Bull, A.T., 1996. CHO cell growth and recombinant interferon-g production: effects of BSA, pluronic and lipids. Cytotechnology 19, 27 – 36. Davis, J.M., Arakawa, T., Strickland, T.W., Yphantis, D.A., 1987. Characterization of recombinant human erythropoietin produced in Chinese hamster ovary cells. Biochemistry 26, 2633 – 2638. Fisher, J.W., 1983. Control of erythropoietin production. Proc. Soc. Exp. Biol. Med. 173, 289 – 305. Freshney, R.I., 1994. Culture of Animal Cells: A Manual of Basic Technique, 3rd edition. Wiley J. Liss, New York, pp. 85 – 98. Ganne, V., Mignot, G., 1991. Application of statistical design of experiments to the optimization of factor VIII expression by CHO cells. Cytotechnology 6, 233 – 240. Glassy, M.C., Tharakan, J.P., Chau, P.C., 1988. Serum-free media in hybridoma culture and monoclonal antibody production. Biotechnol. Bioeng. 32, 1015 – 1028. Hayter, P.M., Curling, E.M.A., Baines, A.J., Jenkins, N., Salmon, I., Strange, P.G., Bull, A.T., 1991. Chinese hamster ovary cell growth and interferon production kinetics in stirred batch culture. Appl. Microbiol. Biotechnol. 34, 559 – 564. Keen, M.J., Rapson, N.T., 1995. Development of a serum-free culture medium for the large scale production of recombinant protein from a Chinese hamster ovary cell line. Cytotechnology 17, 153 – 163. Krystal, G., 1983. A simple microassay for erythropoietin based on 3H-thymidine incorporation into spleen cells from phenylhydrazine treated mice. Exp. Hematol. 11, 649 – 660.
G.M. Lee et al. / Journal of Biotechnology 69 (1999) 85–93 Kurano, N., Leist, C., Messi, F., Kurano, S., Fiechter, A., 1990. Growth behavior of Chinese hamster ovary cells in a compact loop bioreactor. 2. Effects of medium components and waste products. J. Biotechnol. 15, 113–128. Lin, F.K., Suggs, S., Lin, C.H., Browne, J.K., Smalling, R., Egrie, J.C., Chen, K.K., Fox, G.M., Martin, F., Stabinsky, Z., Badrawi, S.M., Lai, P.H, Goldwasser, E., 1985. Cloning and expression of the human erythropoietin gene. Proc. Natl. Acad. Sci. USA 82, 7580–7584. Papoutsakis, E.T., 1991. Media additives for protecting freely suspended animal cells against agitation and aeration damage. Trends Biotechnol. 9, 316–324. Sinacore, M.S., Charlebois, T.S., Harrison, S., Brennan, S., Richards, T., Hamilton, M., Scott, S., Brodeur, S., Oakes, P., Leonard, M., Switzer, M., Anagnostopoulos, A., Foster, B., Harris, A., Jankowski, M., Bond, M., Martin, S., Adamson, S.R., 1996. CHO DUKX cell lineages preadapted to growth in serum-free suspension culture enable rapid development of cell culture processes for the manufacture of recombinant proteins. Biotechnol. Bioeng. 52, 518 – 528.
.
93
Schmid, G., Zilg, H., Eberhard, U., Johannsen, R., 1991. Effect of free fatty acids and phospholipids on growth of and production formation by recombinant baby hamster kidney (rBHK) and Chinese hamster ovary (rCHO) cells in culture. J. Biotechnol. 17, 155 – 167. Snell, K., Natsumeda, Y., Weber, G., 1987. The modulation of serine metabolism in hepatoma 3924A during different phases of cellular proliferation in culture. Biochem. J. 245, 609 – 612. Sytkowski, A.J., 1984. Erythropoietin: a prime regulator of red cell differentiation. Biomed. Pharm. 38, 369 – 371. Yoon, S.K., Ahn, Y.H., Kwon, I.C., Han, K., Song, J.Y., 1998. Influence of reducing agents on the secretion rate of recombinant erythropoietin from CHO cells. Biotechnol. Lett. 20, 101 – 104. Zang, M., Trautmann, H., Gandor, C., Messi, F., Asselbergs, F., Leist, C., Fiechter, A., Reiser, J., 1995. Production of recombinant proteins in Chinese hamster ovary cells using a protein-free cell culture medium. Bio/Technology 13, 389 – 392.