Growth, metabolic, and antibody production kinetics of hybridoma cell

Biotechnol. Pmg. 1001, 7, 481-494

48 1

Growth, Metabolic, and Antibody Production Kinetics of Hybridoma Cell Culture: 2. Effects of Serum Concentration, Dissolved Oxygen Concentration, and Medium pH in a Batch Reactor Sadettin S. Ozturk' and Bernhard 0. Palsson Cellular Biotechnology Laboratory, Department of Chemical Engineering, University of Michigan, Ann Arbor, Michigan 48109

The effects of serum, dissolved oxygen (DO) concentration, and medium pH on hybridoma cell physiology were examined in a controlled batch bioreactor using a murine hybridoma cell line (167.4G5.3). The effect of serum was also studied for a second murine hybridoma cell line (S3H5/72bA). Cell growth, viability, cell density, carbohydrate and amino acid metabolism, respiration and energy production rates, and antibody production rates were studied. Cell growth was enhanced a n d cell death was decreased by increasing the serum level. The growth rates followed a Monodtype model with serum being the limiting component. Specific glucose, glutamine, and oxygen uptake rates and specific lactate and ammonia production rates did not change with serum concentrations. Amino acid metabolism was slightly influenced by the serum level. Cell growth rates were not influenced by DO between 20% and 80% air saturation, while the specific death rates were lowest a t 20-50 % air saturation. Glucose and glutamine uptake rates increased at DO above 10 % and below 5 % air saturation. Cell growth rate was optimal a t pH 7.2. Glucose and glutamine uptake rates, as well as lactate and ammonia production rates, increased above pH 7.2. Metabolic rates for glutamine and ammonia were also higher below pH 7.2. The consumption or production rates of amino acids followed the glutamine consumption very closely. Cell-specific oxygen uptake rate was insensitive to the levels of serum, DO, and pH. Theoretical calculations based on experimentally determined uptake rates indicated that the ATP production rates did not change significantly with serum and DO while it increased continually with increasing pH. The oxidative phosphorylation accounted for about 60% of total energy production. This contribution, however, increased a t low pH values t o 76%. The specific antibody production rate was not growth associated and was independent of serum and DO concentrations and medium pH above 7.20. A 2-fold increase in specific antibody production rates was observed a t pH values below 7.2. Higher concentrations of antibody were obtained a t high serum levels, between 20 % and 40% DO, and a t p H 7.20 due to higher viable cell numbers obtained.

1. Introduction Monoclonal antibodies (MAbs)derived from hybridoma cell cultures are already used extensively in diagnostic assays, and increasing interest has been shown in their potential therapeutic uses and in their application to affinity production systems and to in vivo imaging. The best large-scale production method for MAbs is in vitro cultivation in bioreactors by techniques similar to those used for microbial fermentations. A cost-effective production of MAbs largely depends on understanding the effech of the bioreactor process variables on the physiology of the cells. Several studies have shown that serum level in the medium is very important in determining the physiology of hybridoma cells (Low and Harbour, 1985; Velez et al., 1986; Reuveny et al., 1987; Glacken et al., 1988; Renard et al., 1988; Dalili and Ollis, 1989; Heath et al., 1989; MacMichael, 1989a; Ozturk and Palsson, 1990b, 1991b). In these studies, it was found that serum level directly affects the growth rate. Similarly, the specific monoclonal antibody production rate was reported to be influenced (Dalili and Ollis, 1989; Heath et al., 1989). Serum level

* To whom correspondenceshould be addressed at Vera Corp., 6 Etna Rd., Lebanon,NH 03766. 8756-7938/91/3007-0481$02.50/0

has been reported to change cell metabolism and the amount of antibodyproduced (Daliliand Ollis,1989;Heath et al., 1989; Ozturk and Palsson, 1990b). Effects of oxygen on hybridoma cell growth were studied extensively in the past (Miller et al., 1987; Mizrahi, 1984; Reuveny et al., 1987; Phillips et al., 1987; MacMichael, 1989b; Ozturk and Palsson, 1990d). Cell growth is known to be depressed at low DO. High DO concentrations have also been reported to depress the cell growth (Miller et al., 1987). Oxidative damage was reported for other mammalian cell growth at DO concentrations more than 80 % air saturation (Kilburn et al., 1969; Rueckert and Mueller, 1960; Suleiman and Stevens, 1987). Although a DO concentration range between 30 % and 60 7% air saturation is accepted as optimum for cell growth, the sensitivity of each cell line seems to be different (Reuckert and Mueller, 1960,Boraston et al., 1984;Balin et al., 1976). Dissolved oxygen concentration can influence cell metabolism, mainly the pathways that involve redox potential (Balin et al., 1976;Kilburn et al., 1969;Self et al., 1968; Brosemer and Rutter 1961; Miller et al., 1987). The DO level was reported to alter the antibody production from hybridoma cells. The amount of antibodyproduced was higher around 50% DO (Miller et al., 1987; Mizrahi, 1984;Phillips et al., 1987).

@ 1991 American Chemical Society and American Institute of Chemical Engineers

402

Medium pH is another key parameter in cell culture as it affects cell viability, cell growth, and metabolic activity. The optimal pH for cell growth varies with cell type but is usually in the range of 7.2-7.6 (Eagle, 1973;Rubin, 1971; Thomas, 1986; Harbour et al., 1989). The growth of hybridoma cells has been reported to show similar response to pH (Miller et al., 1988; MacMichael, 1989b; Maiorella et al., 1989). Cell metabolism is influenced significantly by medium pH. Common to many cells is an increase in the rate of glycolysis at elevated pH (Barton, 1971; MacMichael, 1989b; Maiorella et al., 1989; Miller et al., 1988; Rubin, 1971). Similar increase has been observed in glutamine consumption (Miller et al., 1988). Lactate yield from glucose was reported to increase a t high pH values (Barton, 1971;Birch et al., 1980;Leist et al., 1986). Miller et al. (1988) observed that ammonia yield from glutamine decreased complementary to the decrease in lactate yield. Medium pH was reported to have a very cell-line-specific effect on the oxygen consumption rate (Danes, 1963). Contradictory results have appeared in the literature for the effects of pH on antibody production rate. Reuveny et al. (1987) reported for VI1 H-8 hybridoma cell line that pH in the range of 6.8-7.2 did not affect antibody production. Wergeland et al. (1987) found that pH of 7.3 is optimal for antibody production for a murine hybridoma line. Miller et al. (1988) found that the specific antibody production rate is higher a t 6.8 and 7.7 than a t pH 7.1-7.4 for cell line AB2-143.2. Maiorella et al. (1989) observed an increase in antibody production rate under acidic conditions. Design, scale-up, and control of bioreactors for the production of MAbs require quantitative description of hybridoma cell growth and MAb production kinetics. The regulation of cell metabolism by the process variables also needs to be understood for developing more efficient medium formulations, determining reactor operating conditions, and developing process control schemes. We have performed a comprehensive kinetic study on hybridoma cell cultures using a model hybridoma cell line. The effects of important environmental and process variables on hybridoma cell growth, metabolism, and antibody production have been investigated. We have elsewhere reported the effects of initial cell density (Ozturk and Palsson, 1990b),ammoniaand lactate (Ozturk et al., 19911, and osmolarity (Ozturk and Palsson, 1991~)on these parameters. Effects of dissolved oxygen were studied in a continuous reactor (Ozturk and Palsson, 1990d). In this series we have investigated the effects of serum, dissolved oxygen, and medium pH in a batch bioreactor under controlled pH and dissolved oxygen conditions. For each variable, cell growth and death, metabolic characteristics of the cell, respiration activity, energy metabolism, and antibody production rates are quantified.

2. Materials and Methods 2.1. Cell Lines, Medium, and Culture Maintenance. Murine hybridoma cell line (167.4G5.3) was provided by Dr. Latham Claflin from the Medical Center at The University of Michigan. The antibody produced by this cell line is an IgG1, directed against phosphorylcholine (PC) (Briles et al., 1984). Hybridoma cells were made by fusion of BALB/c spleen cells with the nonsecreting plasmacytoma fusion line P3X63-Ag8.653. Antibody was generated from mice immunized with PC-keyhole limpet hemocyanin (KLH). The second murine hybridoma cell line (S3H5/y2bA2) was provided by Dr. Mark Kaminski from The University of Michigan Medical Center. The antibody produced by this cell line is a y2b isotype and is specific to the anti-idiotype on the surface of 38C13 lymphoma

Biotechnol. Prog., 1991, Vol. 7, No. 6

cells (Kaminski et al., 1987). Cells were propagated in T-flasks (Bellco Glass, Inc., Vineland, NJ) using Iscove's modified Dulbecco's medium (IMDM, Gibco Laboratories, Green Island, NY) containing 5 5% fetal bovine serum (FBS, Gibco). The media was supplemented by 100 units/ mL potassium penicillin G and 100 kg/mL streptomycin sulfate (Sigma Chemical Co., St. Louis, MO). Cells were kept at 37 "C in a humidified incubator under 5% COz. 2.2. Bioreactor Experiments. A 1.5-L Celligen bioreactor described in the preceding paper was used to investigate cellular kinetics during batch cultivation ( 0 2 turk and Palsson, 1991d). IMDM supplemented with 100 units/mL potassium penicillin G and 100 pg/mL streptomycin sulfate was used as medium. The pH and dissolved oxygen (DO) concentration were kept constant at the set points using a microprocessor controller throughout the run. No acid or base addition was made as the pH could be controlled at the set point by varying the COZ content of the gas phase. Agitation was provided by screenimpeller system (New Brunswick) at 80 rpm. Temperature was controlled a t 37 "C. Cells were inoculated typically at an initial concentration of 5 X lo4 cells/mL. Effects o f Serum. Murine hybridoma cell lines 167.4G5.3 and S3H5/y2bA were cultivated in the Celligen bioreactor using IMDM supplemented with different amounts of FBS to cover a concentration range of 1.2510% (v/v). Medium pH and dissolved oxygen (DO) concentrations were kept constant at 7.2 and at 20% air saturation, respectively. Effectso fDissolved Oxygen. The murine hybridoma cell line 167.4G5.3 was cultivated in the Celligen bioreactor using IMDM containing 5% FBS. The pH was kept constant at 7.2 throughout the culture by adjusting the COZconcentration in the head space. The DO was kept constant throughout each run at the set value by adjusting the 02 concentration in the head space. Effects of Medium p H . The murine hybridoma cell line 167.4G5.3 was cultivated in the Celligen bioreactor using IMDM containing 5% FBS. Dissolved oxygen concentration was kept controlled at 50% air saturation. The pH was kept constant throughout each run at the set value by adjusting the COZconcentration in the head space. 2.3. Analytical Techniques and Determination of Kinetic Constants. The techniques for sample analysis and for determination of kinetic constants were described in the preceding paper (Ozturk and Palsson, 1991d). 3. Results 3.1. Effects of Serum. Concentration time profiles are presented for cell, antibody, glucose, glutamine, lactate, and ammonia in the supplementary material. Effects of Serum on Cell Growth. Figure 1A shows maximum viable cell concentrations obtained in the bioreactor. For both cell lines used, higher maximum cell concentrations were obtained at higher serum concentrations. Increasing the serum concentration from 1.257% to 10% FBS increased the maximum cell concentration 2-fold and the increase in maximum cell concentration obtained was similar for both cell lines used. However, S3H5/y2bA cells always had higher maximum cell concentrations compared to the 167.405.3 cells. The net growth rates for the two cell lines were different, as discussed below. Cell viability decreases throughout batch cultivation and the decrease is accelerated during the decline phase [e.g., Ozturk and Palsson (1991d)l. At any time during batch culture, the drop in cell viability was more pronounced for the cultures containing low serum (data not

Biotechml. hog.., lQQl, Vol. 7, No. 6

483

E 3,

"

1.21

1

1.1

.

O.MI

.

.

.

--

0.01 0.00 0

.

.

;

4

6

& -.-

10

FBS, v/v

0

FBS, % v/v

.

,

1

1 0 ' 1 2

1.21

2.1

10

1

FBS, %v/v 0.Wl

,

;

*

.

,

8

Ib

I2

1

*\

o.ml/S3H1C A ".*OO

,

,

4

.

6

FBS, %v/v

Figure 1. (A) Maximum viable cell concentrations obtained in batches with different serum concentrations for the cell lines used. (B)Overall viabilities obtained from maximum total and viable cell counts. (C) Experimental growth rates for two cell lines used and the prediction of growth rates by the Monod model. The specific growth rates were fitted by a Monod-type rate equation as explained in the text. Experimental data for 167.405.3 cells (open circles) and for S3H5/y2bA cells (closed circles) and model prediction (solidline). (D) Experimentaldeath rates in the exponential phase for 167.4G5.3cells (open circIes) and for S3H5/y2bA cells (closed circles).

shown). Serum, therefore, helps to maintain the culture viability. In Figure 1B we present the ratio of maximum viable cell concentration to maximum total cells obtained. This fraction gives the overall viability of the culture determined by the relative growth and death rates. Thus, a healthier culture with high growth and low death rates gives higher values. The overall viability was higher at higher serum concentration for both cell lines. Cell line S3H5/72bA resulted in higher fractions than 167.4G5.3 cells. Again this was due to lower death rates for S3H5/ 72bA cells. Specific growth and death rates calculated during the exponential phase are presented in Figure 1C. It was observed that serum increases the specific growth rate in a linear fashion at low concentrations, and this effect gradually decreases until finally a constant growth rate is obtained. A Monod-type saturation model with serum as the limiting component can be used to represent the effects of serum on the growth rate:

where F m a is maximum growth rate and K, is the Monod constant. This equation states that serum is the limiting element determining the growth rate, and it has been used by previous investigators (Glacken et al., 1988; Dalili and Ollis, 1989). Figure 1C shows the experimental data for the two cell lines used and the model prediction using kinetic parameters of p m a = 0.053 h-l and Km = 1.39% serum (v/v). Equation 1 represented the experimental data well for both the cell lines with identical numerical values of model parameters. The death rates were calculated in the exponentialphase and are presented in Figure 1D. The specific death rates decreased by increasing the serum concentration in the culture. The death rates for S3H5/y2bA cells were lower at any given serum concentration than those of 167.4G5.3 cells. The lower death rate is the reason for the higher

viabilities obtained (Figure 1B) for S3H5/72bA cells. The changes in the death rate with serum concentration were qualitatively similar for both cell lines. Effects of Serum on Cell Metabolism. Metabolic rates and metabolic yields are summarized in Tables I and I1 for 167.4G5.3 and S3H5/72bA cells, respectively. The specific uptake and production rates were evaluated in the exponential phase, as described earlier (Ozturk and Palsson, 1991d). Although the cell growth rates were influenced by serum concentration, the metabolic rates and yield coefficients did not show any significant changes (Tables I and 11). We have shown that the specific metabolic rate decreases following the exponential phase (Ozturk and Palsson, 1991d). It is important to note that cells maintained the same metabolic yield coefficients through the decline phase. The transient rates during the course of batch experiments, including both exponential and death phases, are shown in Figure 2 for 167.4G5.3cells grown at different serum concentrations. Figure 2A shows that the lactate production rate and glucose consumption rate are correlated with an overall yield coefficient of 1.60 mol/mol. This graph shows that although the rates decrease during the time course of batch culture, the yield is essentially the same. This yield coefficient was also independent of serum concentration, as all the data points from different serum-containing cultures lie on the same line. Unlike glucose, glutamine was completely depleted a t all serum levels and the cell growth ceased after depletion of glutamine. The cultures were nutritionally limited by glutamine. The specific uptake rates for glutamine in the exponential phase are presented in Tables I and I1for the cell lines used. The glutamine uptake rates were about 5-6 times lower than the glucose uptake rates for both cell lines. Glutamine is also consumed by chemical decomposition. The chemicaldecompositionaccounted for about 45% of the total depletion of glutamine [see Ozturk and Palsson (1990a)l. The concentration of serum did not influence the specific uptake rate of glutamine and the two cell lines used did not show significant differences in glutamine uptake rates (Tables I and 11). The production rates of ammonia in the exponential phase are also presented in Tables I and 11. Specific ammonia production rates were about 60% of the glutamine consumption. The production rate of ammonia and the yield coefficient of ammonia from glutamine ( Y N ~ + Jwere G ~also ) independent of serum concentration. These parameters also appeared to be independent of the cell line used. In Figure 2B all the transient rates for glutamine consumption and ammonia production are presented for 167.4G5.3 cells. This graph includes the results from different serum concentrations. The metabolic rates were observed to change during batch cultivation [e.g., Ozturk and Palsson (1991d)l. However, the ratio of ammonia production rate to glutamine utilization, i.e., the ammonia yield from glutamine, waa invariant. When these two rates were plotted against one another, a good correlation was obtained with an overall yield coefficient of 0.62 mol/mol, which is close to the yield coefficient observed during the exponential phase. As shown in Figure 2C, glucose and glutamine uptakes were correlated with each other, emphasizing the invariance of cell metabolism in the batch culture. Consequently,the metabolic fate of key substrates appears to change neither with time nor with serum concentration. Effect o f Serum on Amino Acid Metabolism. The specific consumption and production rates of amino acids

404


Table 1. SDecific Metabolic Rates at Different Serum Concentrations for 167.4635.3 Cells. glucose lactate glutamine ammonia FBS, utilization production utilization production YLac/Glu, rate rate rate mol/mol 7" (v/v) rate 1.25 2.5 5 10

0.225 0.205 0.214 0.241

0.381 0.330 0.379 0.422

0.046 0.045 0.046 0.040

0.028 0.028 0.031 0.024

1.69 1.62 1.77 1.75

mol/mol

oxygen utilization rate

0.61 0.62 0.58 0.60

0.083 0.080 0.078 0.088

YNH,+/GI~,

The rates are in micromoles per IO6 cells per hour. The decomposition of glutamine was considered in these calculations and the actual rates (due to cellular activity) for glutamine and ammonia are reported. The pH-dependent decomposition constants of glutamine were taken from Ozturk and Palsson (1990a). (I

Table 11. Specific Metabolic Rates at Different Serum Concentrations for S3H5/?2bA Cells. glucose lactate glutamine ammonia FBS, utilization production utilization production (v/v) rate rate rate rate 1.25 2.5 5 10

0.346 0.315 0.316 0.292

0.557 0.517 0.530 0.492

0.042 0.049 0.048 0.044

0.025 0.029 0.030 0.027

YLec/Glur

YNH,+/Ghr

1.62 1.64 1.68 1.68

0.61 0.58 0.62 0.61

mol/mol

mol/mol

The rates are in micromoles per lo6 cells per hour. The decomposition of glutamine was considered in these calculations and the actual rates (due to cellular activity) for glutamine and ammonia are reported. The pH-dependent decomposition constants of glutamine were taken from Ozturk and Palsson (1990a). (I

are presented in Tables I11 and IV for 167.4G5.3 cells and S3H5/y2bA cells, respectively. The amino acid consumption/production rates were similar for the two cell lines. Alanine, glutamate, and glycine were produced by both cells and all other amino acids were consumed. The only difference between the two cell lines was in serine metabolism. Serine was produced by 167.4G5.3cells but consumed by S3H5/72bA cells. The glycine production rate was higher in S3H5/y2bA cells. This difference suggests that all serine is converted to glycine by S3H5/ 72bA cells. The consumption and production rates of amino acids were 1order of magnitude lower than glutamine utilization rates, with alanine being the only exception. The alanine production rate was about 70-80% of the glutamine consumption rate (Tables I11 and IV). The production of glutamate, serine, and glycine by hybridoma cells has been reported in the literature (Adamson et al., 1987;Miller et al., 1989). Glutamate is produced probably by the glutaminase reaction from glutamine. Serine and glycine are produced by transamination of glycolytic intermediates with glutamate (Ozturk and Palsson, 1991d). The consumption of amino acids was most pronounced for branched-chain amino acids such as leucine, isoleucine, and valine. The consumption of tyrosine, lysine, and threonine was also significant. In general, all amino acids except alanine, glutamate, serine, and glycine were used for cell growth. Metabolic rates were influenced by serum concentration for some amino acids, whereas they remained relatively constant for the others. A decrease in production rates of serine and glycine was observed for 167.4G5.3 cells at high serum concentration. The glycine production rate also decreased for S3H5/y2bA cells a t high serum levels (Table IV). Consumption rates of valine, isoleucine, leucine, tyrosine, arginine, aspartate, and histidine also decreased for this cell line (Table IV). Effectso f Serum on Antibody Synthesis. The total antibody concentrations obtained a t the end of the batch culture and individual specific antibody productivities in different serum concentrations are presented in Table V and in Figure 3A for the two cell lines used. The antibody concentration accumulated in batch mode was influenced by the serum concentration used. Specific antibody productivities were determined by the integral method (Oz-

turk and Palsson, 1991d). Figure 3B shows that a plot of antibody concentration against the integral of viable cells gives a straight line, which demonstrates a constant specific antibody production rate. This figure showsall data points from both the exponential and decline phases of growth. All the data points fall on the same line for a given serum concentration, and we can conclude that cells secreted antibody a t the same rate in the decline phase as they did during the growth phase. The data from all the serum concentrations followed the same straight line, and thus the specific antibody production rate is also independent of the serum concentration. This result is consistent with the results obtained in spinner flasks (Ozturk and Palsson, 1991b). Intracellular antibody content of the cells was determined by flow cytometric analysis in the exponential phase (Ozturk and Palsson, 1991d). The intensity of green fluorescence from GAM-FITC was used as a measure of antibody content of the cells. We have shown above that the specific antibody production rate was independent of serum concentration as measured from the secreted amount of antibody in the medium. Figure 4 shows that serum also did not change the intracellular antibody content of the cells. Effect of Serum on Cell Respiration and Energy Metabolism. Oxygen uptake rates were studied for 167.4G5.3 cells. The specific uptake rates of oxygen in different serum concentrations are presented in Table I. A relatively constant oxygen uptake rate was obtained with a value 0.080 f 0.008 pmo1/(106 cells-h). The production rate of ATP was estimated using the lactate production and oxygen consumption rates as discussed in the preceding paper (Ozturk and Palsson, 1991d). Figure 5A shows the ATP production rates for 167.4G5.3 cells in different serum concentrations. The estimated ATP production rates were independent of the serum concentration used, and a relatively constant value of 0.85 f 0.05 pmol/(106 cells-h) was obtained. The contributions of glycolysis and oxidative phosphorylation were calculated from lactate production and oxygen consumption rates. In all serum concentrations used, glycolysis contributed to the ATP production by 40%, with the balance provided by oxidative phosphorylation. 3.2. Effects of Dissolved Oxygen. Concentration

Biotechnd. Rog., 1991, Vol. 7, No. 6

485

Table 111. Amino Acid Consumption Rates for 167.4G6.3 Cells.

0 0. u.00

0.05

0.10

0.15

0.20

q-Clu, pmo1/10A6cell-hr

0.25

0.30

M

u.00

0.01

0.02

0.03

0.04

I 0.05

q_Gln, pmo1/10A6cell-hr 0.3

1

q-Cln, pmol/10A6cell-hr

F i g u r e 2. Relationship among metabolic rates in batch cultivation of 167.4G5.3cells. The transient rates a t different batch times are plotted against another. (A) Lactate and glucose uptake rates in different serum concentrations throughout the batch run. These rates correlate to each other with an overall yield coefficient of 1.60. (B)Ammonia and glutamine uptake rates. These rates correlate with each other with an overall yield coefficient of 0.62. (C)Glucose and glutamine uptake rates are also correlated with each other, showing that these metabolites areutilized at aconstant ratio. Legend 10%serum (open circles), 5% serum (closedcircles),2.5% serum (opensquares),and1.25% serum (closed squares).

profiles for cell number, metabolites and antibody are presented in the supplementary material. Effects of DO on Cell Growth. The specific growth and death rates are summarized in Figure 6A. The specific growth rates were affected by DO over the range studied. The cell growth rates decreased below 10% and above 8070 DO. Cell growth rate did not change significantly between 20% and 80% DO. The variation in death rates was more pronounced. A broad minimum between 20 % and 50% DO was found, with sharp increases below and above this range. In Figure 6B,the maximum viable and total cell concentrations are presented as a function of dissolved oxygen concentrations. Viable cell concentrations showed insignificant variation between 5 % and 80% DO. EffectsofDO on Cell Metabolism. Cells metabolized both glucose and glutamine. Glucose was not a limiting nutrient as there was always excess glucose in all the cultures. Lactate was produced in all cultures. Glucose uptake and lactate production rates are presented in Table VI and in Figure 7 as a function of DO. Both glucose

amino acid Aspartate Glutamate Asparagine Serine Glutamine Histidine Glycine Threonine Arginine Alanine Tyrosine Methionine Valine Phenylalanine Isoleucine Leucine Lysine

1.25 0.86 [2.65] 0.65 [4.70] 45.90 1.24 [3.29] 1.46 2.512 [32.99] 2.67 1.03 4.04 0.70 6.43 6.91 2.21

2.5 1.52 [3.17] 0.83 [3.01] 44.70 1.12 [ 1.861 1.79 3.01 [31.56] 2.71 1.09 4.98 0.80 7.08 7.13 3.25

5 1.47 [3.43] 0.85 [2.26] 45.80 0.94 [0.78] 1.36 2.71 [33.63] 2.65 1.10 4.45 0.85 5.90 6.77 2.65

10 1.47 [3.32] 0.82 [1.07] 39.60 1.06 [0.48] 1.47 2.63 [30.05] 2.23 1.21 4.72 0.54 4.34 5.41 2.99

The rates are in nanomoles per 106 cells per hour. Square brackets indicate that the amino acid is being produced. Table IV. Amino Acid Consumption Rates for S3HBly2bA Cells.

FBS,% (v/v) amino acid 1.25 2.5 5 10 Aspartate 1.11 1.38 1.61 2.09 Glutamate [8.25] [3.04] [3.24] [3.65] Asparagine 2.60 2.22 1.78 1.74 4.73 Serine 5.69 4.77 7.40 42.11 Glutamine 48.20 43.60 48.80 Histidine 1.88 1.96 3.09 3.04 Glycine [7.03] [5.77] [5.27] [3.50] Threonine 8.31 6.11 4.94 2.42 4.52 4.05 Arginine 3.51 3.74 Alanine [30.50] [30.58] [31.63] [30.89] Tyrosine 4.76 5.45 3.83 3.32 Methionine 1.97 2.30 1.53 1.65 Valine 8.25 5.39 4.99 10.57 Phenylalanine 1.41 0.85 0.45 0.75 Isoleucine 8.39 11.63 6.16 4.77 Leucine 9.93 14.06 7.71 6.64 Lysine 4.81 4.43 4.92 3.78 The rates are in nanomoles per 106 cells per hour. Square brackets indicate that the amino acid is being produced. Table V. Antibody Concentrations and Antibody Production Rates in Different Serum Concentrations 167.4G5.3cells S3H5/y2bA cells

FBS, % (v/v)

1.25 2.5 5 10

final IgGl concn, mg/L 12.5 21 28 30

IgGi production rate, pg/(cell.h) 0.225 0.219 0.227 0.225

final IgGzb concn, mg/L 20 23 30 33

IgG2b production rate, pg/(cell.h) 0.173 0.177 0.157 0.168

utilization and lactate production rates increased above and below 10% DO (Table VI). The increases in glucose and lactate metabolic rates were parallel to each other. The rates at 1% and 100?’ 6 DO were 40-50% higher as compared to the rates at 10% DO. The yield coefficient of lactate from glucose increased from 1.54 (mol/mol) at 100% DO to a value of 2.1 a t 1%DO. Glutamine was consumed and it was the limiting nutrient for all the cultures except the one carried out at 1% DO. The glutamine utilization rates are presented in Table VI. Glutamine also decomposes chemically in the medium, and the necessary corrections in evaluating the uptake rates are described in Ozturk and Palsson (1990a).

Biotechnol. h g . , 1991, Vol. 7, No. 6 1.50

N:r Oxidative Phosphorylation

(A)

W Glycolysis

1.25

20

p

10

0

2

0 167.4653 cells 0 S3H5 cells

4

6

8

1

0

1

2

1.25

FBS, 5% vlv

5

2.5

10

FBS, %v/v Nutrients

Serum

Energy

Y-

O.ue+O

5.0e+7

1.0e+8

1.5ec8

2.0e+8

Maintenance

Growth

Integral of viable celldml-hr

Figure 3. (A) Effects of serum on the antibody (IgG1 for

167.405.3 cells and IgGzb for S3H5/72bA cells) concentrations accumulated in terminal batch experimentsfor the two cell lines used. (B)Antibodyproduction kinetics in the culture of 167.4G5.3 cells in different serum concentrations for 167.4G5.3 cells. Antibody (IgG1)concentrationswere plotted againstthe integral of viable cells to obtain the antibody production rates, qmb, as indicated in the text.

Figure 5. (A) Energy metabolismfor 167.4G5.3cells in different serum concentrations. ATP production rates were calculated

from lactate production (glycolysis contribution) and oxygen consumption (oxidativephosphorylation contribution)rates. (B) Effect of serum on the distribution of energy to maintenance and growth. I 0.010

0.05

I

,

4

5%

3

-

i

s

0.006

26 5

i

5 . 0.004 n 0.01' 0

"

"

20

40

"

60

"

80

"

LOO

"0.002 120

DO, % Air saturation

log(Intracel1ular IgG)

Figure 4. Intracellular antibody content of the 167.4G5.3 cells measured in a flow cytometer. Antibody content of the cells in different serum concentrationswas the same.

The influence of DO on the glutamine uptake rate showed a trend similar to that obtained for the glucose uptake rates (Table VI). There was an increase in glutamine utilization rate above 10% DO and the uptake rate increased by 40% a t 100% DO. The production rates of ammonia followed a similar pattern; a minimum a t 10 % DO with increases above and below 10% DO (Table VI and Figure 6). Effects of DO on Amino Acid Metabolism. Amino acid metabolism was studied for 17 amino acids over the range of DO used. The specific amino acid consumption and production rates are summarized in Table VII. Glutamate, serine, glycine, and alanine were produced while all other amino acids were consumed a t all DO values (Table VII). Glutamine consumption and alanine production exhibited the highest rates. The production rates of alanine were comparable to the glutamine consumption rates. The consumption and production rates of all other

1

5

10

20

50

80

100


Figure 6. (A) Specific cell growth and death rates and (B) cell

concentrationsobtained in batch culture at different dissolved oxygen concentrations. amino acids were about an order of magnitude lower. Branched amino acids such as isoleucine, leucine, and valine were consumed at significant rates. Effectso f DO on Antibody Synthesis. The antibody concentrations obtained were influenced by DO (Figure 8A), exhibiting a maximum between 20% and 50% DO. About 33 mg/L of IgGl was obtained in this range, while cells in low and high DO produced about 50% and 70% less antibody, respectively. The location of the maximum in antibody concentration is directly related to the cell

407

Blotechnol. Prog., 1991, Vol. 7, No. 6

Table VI. Metabolic Rates at Different Dissolved Oxvnen Concentrations Exbressed as Percent Air Saturation. glucose lactate glutamine ammonia oxygen dissolved utilization production utilization production consumption YLac/Glu, oxygen, 70 rate rate rate rate rate mol/mol 1 0.29 0.61 0.057 0.036 n.d. 2.10 ~

~

5 10 20 50 80 100

0.20 0.18 0.22 0.25 0.28 0.27

~

0.38 0.31 0.38 0.39 0.43 0.42

~~

0.043 0.045 0.048 0.056 0.064 0.068

~

0.025 0.024 0.028 0.031 0.039 0.043

~~~

0.078 0.085 0.080 0.086 0.085 0.086

1.89 1.74 1.68 1.60 1.54 1.54

YN&+/Gh,

mol/mol 0.62 0.58 0.53 0.58 0.55 0.61 0.63

a The rates are in micromoles per 106 cells per hour. The decomposition of glutamine was considered in these calculations and the actual rates (due to cellular activity) for glutamine and ammonia are reported.

0.1' 0

"

20

"

40

"

60

"

80

" LOO

DO, % Air saturation 2.4 I

.

'

I

11.2

0 NH4ICln

0 Lsricl" 0 . 4 1 . 0

, 20

.

I

40

.

60

I 80

.

I

0.2 1ou


Figure 7. (A) Glucose and glutamine uptake rates and (B)yield coefficients of lactate and ammonia at different dissolved oxygen concentrations.

concentration. However, the maxima for antibody and cell concentrations were not the same. Cell growth was optimal around 50% air saturation while optimal antibody production was located at lower dissolved oxygen concentrations, around 30 % dissolved oxygen. This was due to the changes in cell viabilities. Except at 1% DO, cells were more viable at low DO; consequently, viable cell concentration is higher near the end of the culture, resulting in more integrated antibody production. Antibody production was non growth associated in all the cultures. Antibody production continued throughout the exponential and decline phases. The plot of antibody concentration against the integral of viable cells, described in Ozturk and Palsson (1991d), yielded straight lines with a constant slope (Figure 8B). All the data points obtained, including exponential phase and decline phase, followed the same line, showing (1)that cells had the same specific antibody production rate in both phases of growth and (2) that the specific production rate was independent of DO. Effects o f DO on Cell Respiration and Energy Metabolism. Oxygen consumption rates were not affected by DO as presented in Table VI. The calculated ATP production rates are presented in Figure 9A as a function of oxygen concentration. Cells showed similar ATP production rates at all oxygen concentrations. The specific ATP production rate is constant around 0.85 pmol/(l06

cells-h). Cells obtained only about 40% of their energy from glycolysis, and the rest is provided by glutamine oxidation. The NADH yield from glutamine was calculated as explained in the preceding paper (Ozturk and Palsson, 1991d). Figure 9B shows that although about 3.7 mol of NADH is produced/mol of glutamine below 50% air saturation, this yield decreases to 2.5 at 100% DO. 3.3. Effects of Medium pH. Concentration profiles for cell number, metabolites, and antibody are presented in the supplementary material. EffectsofpHonCell Growth. Cellgrowth wasgreatly influenced by the medium pH. The specific growth and death rates are presented in Figure 10A as a function of medium pH. Maximal specific growth rates were obtained at about pH 7.2; minimum specific death rates were obtained around the same pH. The curve of cell growth rate versus pH was not symmetric and a drop in the growth rate was more pronounced at low pH values than at high pH values. The maximum viable and total cell concentrations are presented in Figure 10B. These values showed again maxima at 7.2, due to the maximum growth rate and minimum death rate at this pH. Effects of p H on Cell Metabolism. The specific uptake rates of glucose and production rates of lactate are summarized in Table VIII. The glucose uptake rate increased 1order of magnitude by increasing the pH value from 6.90 to 7.65 (Table VI11 and Figure 11). The specific lactate production also increased at elevated pH. The yield coefficient of lactate from glucose increased slightly at high pH values (Table VIII). The specific glutamine uptake rate increased with pH above pH 7.2 (Table VI11 and Figure 11). Raising the pH from 7.2 to 7.65 resulted in more than a 2-fold increase in glutamine utilization rate, which is comparable to the corresponding increase in glucose uptake rate. Glutamine uptake rate also increased at acidic pH values but to a lesser extent than at alkaline pH values. Ammonia production was also affected by the pH (Table VIII). The change in ammonia production rate with pH was similar to that of glutamine. The yield coefficient of ammonia from glutamine showed a slight decrease with pH. The relative uptake rates of glucose and glutamine changed with pH. The ratio of glucose to glutamine uptake rates was around 6 at high pH values and decreased to 1 at acidic pH values. This change was mostly due to the sharp decrease in the specific glucose uptake rate with low pH, and also glutamine uptake rates increased below the optimal growth pH. Oxygen consumption rate was not influenced by the pH. A relatively constant uptake rate of oxygen was obtained (Table VIII). Effect of p H on Amino Acid Metabolism. Table IX summarizes the influence of medium pH on amino acid metabolism. Glutamate, serine, glycine, and alanine were produced while all other amino acids were consumed at

4aa

Blotechnol. h g . , 1991, Vol. 7, No. 6

Table VII. Consumption and Production Rates of Amino Acids as a Function of Oxygen Concentration in Batch Culture under Different Oxygen Concentrations. dissolved oxygen, 96 amino acid Aspartate Glutamate Asparagine

5 0.37 [2.88]

1

1.56 [3.63] 1.41 [2.56] 57.30 1.02 [2.04] 3.50 4.98 [31.56] 4.23 1.58 6.31 2.55 8.19 8.46 5.17

Serine

Glutamine Histidine Glycine

Threonine Arginine Alanine

Tyrosine Methionine Valine

Phenylalanine Isoleucine Leucine Lysine

10 0.36 [2.82] 0.83 [1.71] 45.20 0.80 [ 1.001 1.29 2.14 [ 30.661 2.29 1.04 5.06 0.93 6.37 6.18 2.88

1.12

[LO91 44.30 0.73 [0.87] 2.49 3.15 [29.01] 3.13 1.34 4.83 1.94 6.38 5.96 2.73

20 0.75 [3.42] 0.90 [2.46] 47.40 0.80 [0.94] 1.29 2.40 [33.77] 2.77

50 0.96 13.631 1.23 [3.23] 54.20 0.96 [ 1.571

80 0.59 [4.16] 1.68 [4.08] 63.12 1.30 [1.87] 2.92 4.15 [42.09] 3.51 1.67 6.00 2.45 10.14 10.94 4.17

2.28

3.61 [35.95] 3.68 1.39 5.53 2.07 8.39 8.30 4.01

1.11

4.60 0.85 6.85 7.56 3.03

100 1.05 [4.62] 2.19 [4.52] 67.21 1.52 [2.37] 2.74 4.03 [45.15] 4.16 1.80 6.94 2.39 1.059 10.24 5.08

The values in square brackets are production rates. The rates are in nanomoles per 106 cells per hour.

1

1.50 [

-

4JL

1.25

-a

."0

20

60

40

80

100

120

I

W

Oxidative Phosphorylation Glycolysis

10

5


50

20

100

80


0

O Q ' O.ue+O

' 5.0e+7

'

'

l.Oec8

'

' 1.5e+8

.

'

2.0e+8

Integral of viable cells/ml-hr

Figure 8. (A) Effectsof DO on the amount of antibody produced in batch mode and (B)evaluation of antibody production rates at different oxygen levels. Experimental data obtained at different DO levels are plotted against the integral of viable cells. all pH values. The specific alanine production increased at high pH values. Glutamate production rates were also higher at high pH values. Glycine production also increased, whereas serine production showed a minimum at pH 7.2. The consumption rates of valine, leucine, and isoleucine were significant, and these rates were also increased by pH. Effects o f p H on Antibody Synthesis. Antibody production continued into the decline phase of growth for all the pH values studied. Antibody production proceeded at the same rate during both cell growth and decline. The antibody concentrations and the specific antibody production rates were influenced by the pH (Figure 12). Antibody concentrations obtained in batches showed a maximum at pH 7.2, the optimum pH for cell growth. The specific antibody production rates were higher under acidic conditions. At pH 6.9 the antibody production rate was about 2 times higher than at pH 7.2. Specific antibody production rates did not change above pH 7.2.

2

0

20

40

60

80

100

120


Figure 9. Energy metabolism under different oxygen concentrations. (A) ATP production rates as a function of oxygen concentration. (B)NADH yield from glutamine.

Effects o f pH on Cell Respiration and Energy Metabolism. Medium pH did not alter the cell specific oxygen uptake rates, as shown in Table VIII. The estimated ATP production rates are presented in Figure 13A as a function of pH. The ATP production increased with pH. This was due to the increased glycolytic activity as measured by glucose uptake rate. Oxidative phosphorylation contributed to the ATP production at a relatively constant rate as qo, was relatively independent of pH. Then the percentage of this contribution decreased with pH. At pH 6.9, cells obtained about 76% of the energy from oxidative phosphorylation, whereas a t pH 7.65 this contribution decreased to 25 % . The contribution of glutamine to energy metabolism is also considered. Assuming that the oxidative phosphorylation is used to oxidize only NADH derived from the glutamine oxidation, we calculate the apparent NADH yield of glutamine. Figure 13B shows the result of these calculations. The NADH yield coefficient of glutamine


489

1

3

'1.0

b.8

'1.2

1.4

1.6

7.8

PH

viable

Ei total

6.9

1.05

1.2

1.35

1.5

1.65

PH Figure 10. (A) Specific cell growth and death rates calculated in the exponential phase of growth and (B) cell concentrations obtained in batch culture and different pH values.

showed a maximum value of 3.5 at around pH 7.2. At this value of NADH yield we would obtain about 11 mol of ATP/glutamine consumed, assuming one NADH yields 3 ATP. This number shows that glutamine is incompletely oxidized. 4. Discussion 4.1. Effects of Serum. Serum is an important component in cell culture medium, supporting and promoting cell growth. Our experimental data have shown that serum increases the specific growth rate while decreasing the death rate. The influence of serum on specific growth rate was modeled using Monod-type saturation kinetics. The specific growth rates for two different cell lines could be described by this model with the same parameters, i.e., K,and pmax. The Monod-type kinetics for the serum effect on growth rate was also used successfully by other investigators (Glacken et al., 1988; Dalili and Ollis, 1989). Glacken et al. (1988) observed that for the cell line they used, the parameter K,is sensitive to initial cell concentration. However, for both cell lines we have not observed the effects of initial cell density on the growth rate [for 167.4G5.3cells see Ozturk and Palsson (1990b)l for serum-containing batches when inoculum density was varied between lo3 and 105 cells/mL. We have, however, observed that the parameter K,decreases when the cells are exposed to low serum concentration for more than 3 months. This "adaptation" process for the cells and the consequences of long-term exposure to serum were discussed elsewhere (Ozturk and Palsson, 199Oc, 1991a). For both cell lines used, serum decreased the specific death rates. This shear protective effect of serum was observed by other researchers (Kunas and Papoutsakis, 1989;McQueen et al., 1989). This effect of serum was also reported in spinner flasks (Ozturk and Palsson, 1991b) for 167.4G5.3 cells. Qualitatively, the decrease in death rate by serum was similar for both cell lines, but the specific death rates were cell line dependent. The S3H5fy2bAcells had a lower death rate than 167.4G5.3.

.

The increase in growth rates resulting from the use of higher serum concentration did not influence cell-specific metabolic rates. The nutrients were consumed and products formed to different extents because of differences in cell concentrations created by serum. Low and Harbour (1985) reported that the serum level did not influence the glucose utilization and lactate production rates. The results of Dalili and Ollis (1989) were, however, different. These authors found an increase in glucose uptake rates at low serum concentrations. One drawback in these experiments was that they were carried out in T-flasks with no pH and oxygen control. Many such differences are attributed to cell line variations. Both cell lines used in this study, however, showed no influence of serum on the metabolic rates. Further, the cells showed similar metabolic behavior, with only minor differences. Glucose uptake and lactate production rates were, for instance, higher in S3H5/y2bA cells. However, the yield coefficient of lactate from glucose ( Y L ~ ~ /was G ~ very " ) similar and serum did not influence this yield coefficient for both cell lines (Tables I and 11). The specific antibody production rates were independent of the serum concentration for both cell lines. Then the increase in antibody concentration by serum was basically due to the increased viable cell concentration at high serum levels. It is also important to note that the IgGzb-producing cell line (S3H5/y2bA) had lower specific antibody productivity. However, as these cells had higher apparent growth rates (similar growth but lower death rates) than the 167.4G5.3 cells, they grew into higher cell concentrations and yielded higher antibody concentrations. Previous investigators have also suggested that serum affects antibody production (Dalili and Ollis, 1989;Heath et al., 1989). An increase in antibody production rate by increasing serum concentration was reported (Dalili and Ollis, 1989),but the opposite trend was reported by Heath et al. (1989). Glacken et al. (1988) reported a constant antibody production rate for growth rates higher than 0.02 h-' and a decrease at growth rates below 0.02 h-l. One drawback of these studies was that they were not carried out with controlled pH and dissolved oxygen levels. We have observed that pH has a significant effect on the specific antibody production rate. Several studies have shown an increase in antibody production at low growth rates (Miller et al., 1988; Suzuki and Ollis, 1989). It has been suggested that the specific antibody production rate is higher in the GO phase of a cell cycle (Suzuki and Ollis, 1989). If this is so, then we would expect to find an increase in specific antibody production at the low growth rates obtained at low serum concentrations. Our data do not show such an effect of growth rate on the antibody production rate when the specific growth rate is changed between 0.025 and 0.045 h-l. Our results are in agreement with the Data of Glacken et al. (1988), who showed a constant antibody production when the growth rate is higher than 0.02 h-l. When we analyzed the cell cycle in different serum concentrations, we found only a 20% increase in the fraction of GO + G1 phase (data not shown). This difference in cell cycle apparently did not reflect on the antibody synthesis rate. Energy was provided by glycolysis and glutamine oxidation. The total ATP production rate, estimated at 0.85 pmol/ (lo6cells-h), was not affected by serum. At low serum concentration, the cells obtained the same energy but grew more slowly compared to growth in high serum. The apparent ATP growth yield for the hybridoma cell

Blotechnol. Prog., 1991, Vol. 7, No. 6

490

Table VIII. Metabolic Rates at Different pH Values. glucose lactate glutamine utilization production utilization PH rate rate rate 6.90 7.05 7.20 7.35 7.50 7.65

0.15 0.28 0.41 0.63 1.07 1.43

0.09 0.17 0.24 0.38 0.60 0.76

0.065 0.052 0.046 0.069 0.102 0.123

ammonia production rate

oxygen consumption rate

mol/mol

YLsc/Glut

YNH,+/GLn,

0.047 0.038 0.029 0.046 0.066 0.071

0.087 0.082 0.083 0.082 0.084 0.083

1.68 1.67 1.69 1.64 1.79 1.88

0.73 0.63 0.66 0.65 0.58

mol/mol 0.72

4 The rates are in micromoles per 106 cells per hour. The decomposition of glutamine was considered in these calculations and the actual rates (due to cellular activity) for glutamine and ammonia are reported. The pH-dependent decomposition constants of glutamine were taken from Ozturk and Palsson (1990d).

PH 2 . 0 , .

,

.

,

.

I

,

I

.

,

1.0

0.9

:

E

1.6

I

1.0 6.8

7.0

7.2

7.4

7.6

I 0.5

7.1

PH

Figure 11. (A) Glucose and glutamine utilization rates at different pH values. (B) Lactate yield coefficient from glucose and ammonia yield coefficient from glutamine. considered here was calculated as Y'ATP= QATP/P = (2.95) X 1O1O cells/mol. This yield is comparable to those reported for bacterial cells and mouse LS cells (Kilburn et al., 1969). In most mammalian cells, including hybridomas, oxidative phosphorylation is operated by glutamine as a little glucose enters to the TCA cycle. Cells obtained about 40% of energy grom glycolysis and 60% from oxidative phosphorylation. This distribution did not change with serum concentration or with the differences in growth rates obtained at different serum concentration. Constant energy metabolism at different serum concentrations needs to be explained. Serum changes the growth rate without influencing the cell metabolism and energy production. One possible explanation for this behavior can be stated as follows. Cells utilize more nutrients than needed for growth and maintenance and somehow dissipate the excess generated. Serum components thus control the coupling of metabolic resources to cell growth and proliferation, as schematically indicated in Figure 5B. Further, growth-associated costs appear to be small compared to the total energy produced from nutrient utilization. Our experimental data from chemostat culture (Ozturk and Palsson, 1990d)indicate that the per cell ATP production rates were about 40% lower than those measured under the experimental conditions considered here. This comparison suggests that the energy

that is actually needed for growth can be lower than the ATP production rates of 0.85 f 0.05 pmol/(106 ce1ls.h) estimated here. 4.2. Effects of Dissolved Oxygen. Cell growth rate diminished at DO levels below lo%, probably due to the limited availability of oxygen for cellular functions and only the culture at 1%DO appeared to be oxygen limited. Cell growth was inhibited by 20% at 100% DO, most likely due to oxidative damage. Oxidative damage has been reported for other mammalian cells (Kilburn et al., 1969a; Rueckert and Mueller, 1960; Suleiman and Stevens, 1987). The growth rate thus showed a plateau between 20% and 80% DO. Our data is consistent with the data by Balin et al. (1976), who reported that WI-38 cells had similar growth rates between 16% and 85% air saturation. The death rates were higher at low (below 10% ) and high (above 50%) DO. The reason for the dependency of death rate on DO is not clear. Both viable and total cells showed a maximum at 50% DO because of higher growth and lower death rates at this DO concentration. Glucose and lactate metabolic rates increased at DO levels above 107%. The increase in glucose uptake a t higher oxygen concentrations was also observed for WI-38 cells (Balin et al., 1976). Note that cell growth rate did not increase at high oxygen concentrations. Hence the increased consumption rate of glucose is not used for cell mass. Extra energy obtained from glucose is possibly used to repair the oxidative damage or simply dissipated. Lactate production was in parallel to glucose consumption. More lactate was produced at oxygen concentrationsabove 10% air saturation. Although lactate yield from glucose dropped, the production of lactate was always significant even at very high DO levels. These data indicate that most of the pyruvate is converted to lactate or transaminated to alanine and these processes do not depend on the availability of oxygen. Both glucose uptake and lactate production rates increased at 1% DO. The increase in glucose uptake rates at low oxygen concentrations is known for microbial and mammalian cells (Pasteur effect). Since oxidative phosphorylation is limited, cells use more glucose for energy production through glycolysis, and almost all glucose is converted to lactate to give a theoretical maximum yield coefficient of 2. Lactate can be produced from glutamine (Zielke et al., 1980). The lactate yield coefficient at 1% DO was greater than 2, indicating possible production of lactate from glutamine in our cells. The metabolism of glutamine and ammonia was also influenced by oxygen concentration. Both glutamine consumption and ammonia production were minimal at 10% DO. Ammonia production was proportional to glutamine utilization (Table VI). There was a minimum ammonia production at 107% oxygen concentration, and the rates of ammonia production increased at oxygen concentrations below and above 10% air saturation. The


49 1

Table IX. Amino Acid Consumption and Production Rates at Different pH Values. PH amino acid

6.9 1.06 [4.64] 1.40 [4.73] 65.30 1.80 [1.44] 3.40 1.58 [22.98] 3.59 1.351 3.01 1.85 6.75 7.13 2.78

Aspartate Glutamate Asparagine Serine Glutamine Histidine Glycine Threonine Arginine Alanine

Tyrosine

Methionine Valine Phenylalanine Isoleucine Leucine Lysine

7.05 0.89 [4.47] 1.33 [2.70] 55.30 0.62 [1.02] 1.60 3.73 [31.57] 3.22 1.257 4.94 1.53 6.64 6.48 3.87

7.2 0.76 [3.42] 1.25 [2.43] 46.30 0.68 [0.95] 1.30 3.00 [32.67] 2.32 1.100 5.25 1.26 6.92 6.72 4.09

7.35 0.87 [9.03] 2.29 [1.39] 65.30 1.19 [1.29] 1.97 3.37 [45.74] 3.93 2.001 6.63 1.45 8.89 9.69 4.35

7.5 0.87 [14.57] 2.26 [4.12] 101.20 0.95 12.201 3.04 3.71 [48.58] 4.84 2.189 6.96 1.99 10.21 10.68 4.46

7.65 0.96 [29.15] 2.74 [4.03] 123.10 1.37 [4.88] 4.03 5.34 [46.97] 6.42 2.352 7.58 3.90 11.43 11.65 5.91

The rates are in nanomoles per 1@ cells per hour. Square brackets indicate that the amino acid is being produced. Oxidative phosphorylation Glycolysis

10.4

c

0 b.8

1.0

1.2

1.4

1.6

6.9

1.8

PH Figure 12. Antibody titers (open circles) and antibody production rates (closed circles) at different pH values. Specific antibody productivity was higher at low pH values. However, cultures resulted in maximum antibody at pH 7.2 due to high cell concentration. yield of ammonia from glutamine changes little with oxygen concentration and remained around 0.6 mol/mol. Important effects of DO were observed on amino acid metabolism. In the preceding paper, we have outlined the metabolic pathways of amino acids [Figure 8 in Ozturk and Palsson (1991d)l. The production or consumption of amino acids often accompanies a change in redox potential. Oxygen can alter the redox potential and, therefore, can influence the amino acid metabolism. Experimental results showed that this was the case (Table VII). There was also an influence of glutamine utilization on metabolic rates of amino acids. These rates followed the same trend as glutamine. When amino acid metabolic rates were normalized against glutamine consumption,the effects of oxygen were less pronounced. Only alanine production showed a significant change relative to glutamine at low DO. The yield coefficient of alanine from glutamine, Y ~ l + l ~was , 0.58 mol/mol at 1%DO, while it was about 0.72 at other oxygen concentrations! [The yield coefficients for amino acids were calculated from the metabolic rates, Le., Y A A / G=~qAA/qGh.] Alanine formation depends on the availability of pyruvate for the transaminase reaction [Figure 8 in Ozturk and Palsson (1991d)l. At low oxygen levels, glucose is converted more and more to lactate, leading to a lower concentration of pyruvate. Serine and glycine yield coefficients showed an increase with DO above 5 % air saturation. These amino acids are dependent on glycolytic flux. The increase in

1.2

7.05

1.35

1.5

1.65

PH 0.8

I

\s

n i

I \

0.6

-

0.5

-

0.4

-

0.3

-

0.2L b.8

I 4.0

'

"

7.0

A

"

"

1.2

7.4

"

1.6

" 1.8

PH

Figure 13. Energy metabolism under different pH values. (A) ATP production rates as a function of pH. (B)Fraction of energy produced by oxidative phosphorylation and NADH yield from glutamine. Energy obtained from glutamine oxidation seems to have an optimum around pH 7.2. glucose uptake rates with DO explains elevated yields of serine and glycine. Specific oxygen uptake rate was not influenced by DO between 5% and 100% air saturation. Fleischaker and Sinskey (1981) and Miller et al. (1987) reported similar data. If the oxygen uptake rate followssaturation kinetics, the Km value must be much lower than 5% DO. Miller et al. (187) reported a value of K m = 0.6% air saturation. Energy production rates were evaluated theoretically. ATP production rate increased with DO only very slightly between 5 % and 100% DO. The relative contribution of glycolysis and oxidative phosphorylation also seemed to be less sensitive to DO levels in this range. About 4 mol of NADH was produced/mol of glutamine below 50% air

492

saturation. If 3 mol of ATP is obtained for the generation of each NADH, then 12 mol of ATP is produced/mol of glutamine. Complete oxidation of glutamine to COZresults in 21 mol of ATP. Hence, under our experimental conditions, glutamine oxidation was always incomplete. The NADH yield of glutamine decreases at high DO levels mainly because of increased glycolytic activity. As cells obtain more energy from glycolysis,they use less glutamine oxidation. Such regulations in energy metabolism have been reported in the literature (Zielke et al., 1984). Maximum antibody concentration was obtained at 30 % DO while the maximum in cell concentration was at 50% DO. However, these maxima in antibody and cell concentrations were very flat. Miller et al. (1987) and Reuveny et al. (1987) also reported different maxima for cell growth and antibody production. The specific antibody production rate was not influenced by the DO. These data indicate that cells are producing antibody at a constant rate independent of the environmental changes created by different oxygen levels. Thus, for cost-effective production, one should focus on maximizing the viable cell numbers. 4.3. Effects of Medium pH. Specific growth rates exhibited a maximum around pH 7.2 in the batch mode. This maximum in the cell growth rate is in the pH range reported by Eagle (1973) for a variety of cell lines. Maiorella et al. (1989) reported the same pH optimum for the growth of a trioma cell line. For AB2-143.2 hybridoma cell, Miller et al., (1988) observed lower cell concentrations at pH 6.8 and 7.7. MacMichael (1989b) reported higher cell growth rates between pH 7.2 and 7.6. The metabolic rates were higher at high pH values. The increase in glucose and lactate metabolism at high pH values is a well-known phenomenon for tumor cells (Barton, 1971; MacMichael, 1989a,b; Maiorella et al., 1989; Miller et al., 1988;Rubin, 1971). One possible explanation lies in the increase of the activity of glycolytic enzymes. It is known that hexokinase has an optimal pH of 8 (Eigenbrodt et al., 1985). Another explanation is that increased pH alters the membrane potential, changing the glucose transport rate through the membrane (Wilhelm et al., 1971). The lactate production rate increased because of higher glycolytic activity. Hybridoma cells convert most of the pyruvate to lactate for reasons that are not fully understood, as described previously (Ozturk and Palsson, 1991d). The yield coefficient of lactate from glucose increased at elevated pH values. This may suggest a self-regulation to keep internal pH constant. The greater amount of lactate formed allows the cells to survive under alkaline conditions. Similarly, less lactate is produced at low pH. A similar decrease in lactate production at low pH was observed for various cell lines (Barton, 1971; Birch et al., 1980; Leist et al., 1986; Miller et al., 1988). Glutamine consumption rate showed a minimum at pH 7.2. The glutamine uptake rate was higher at low pH where glucose utilization was minimal. Glutamine was used preferentially at low pH, and the ratio of glucose uptake to glutamine uptake rate decreased constantly. Glutamine appears to substitute for glucose under acidic conditions. Glutamine uptake rate increased at high pH values, possibly due to increased activity of glutaminase, which is known to have a pH optimum around 8 for tumor cells (McKeehan, 1986). An increase in glutamate yield from glutamine supports this hypothesis. Glutamate yield from glutamine increased from 0.07 at pH 6.95 to 0.23 at pH 7.65. The ammonia production rate followed a trend similar to glutamine. However, ammonia yield decreased


at elevated pH values (Figure 11). A similar result was observed by Miller et al. (1988). Amino acid metabolism followed the same trend as glutamine. However, alanine yield from glutamine (Y&/cln = Q M ~ / Q C had ~ ~ ) a maximum around 7.2, which is the optimal cell growth pH value. Alanine production depends on the availability of pyruvate and glutamate as it is produced by transamination of these substrates. A t low pH both pyruvate and glutamate levels were low, and less alanine was produced. At pH values greater than 7.2, the alanine yield coefficient also decreased. This decrease may be related to limited pyruvate availability from glucose. At high pH values, more lactate was produced from pyruvate. Glutamate yield increased because of increased activity in the glutaminase reaction, as discussed above. The yield of serine from glutamine (Ys,,/G, = qser/ Q G ~decreased ) from 0.072 to 0.030 mol/mol by changing the pH from 6.9 to 7.65. Serine is produced from pyruvate by a transaminase reaction. The decrease in serine yield is again limited by availability of pyruvate. The consumption rates of valine, leucine, isoleucine, and lysine were influenced by pH in a way similar to glutamine. However, the yield coefficients for these amino acids were also pH dependent. They all showed a maximum at pH 7.2. Oxygen uptake rate was not influenced by pH. Danes et al. (1963) have reported that the oxygen uptake rates of different cells have different responses to pH. Miller et al. (1988) observed that, for hybridoma cell line AB2143.2, pH did not influence the specific uptake rate of oxygen. Antibody production was optimal at pH 7.2 mainly because of higher cell concentrations. However,there was also the influence of pH on specific antibody productivity. The specific antibody production rate showed a 2-fold increase under acidic conditions (Figure 12). Miller et al. (1988)and Maiorellaet al. (1989)observedsimilar increases in specific antibody production. Miller et al. (1988) reported an increase also under alkaline conditions. The increase in specific antibody productivity was not enough to compensate for the decrease in cell number under acidic conditions and could not be utilized for better production in the batch mode. The antibody production is non growth associated for this cell line. Then, both cell growth and antibody production can be optimized separately by changing the pH. Optimum pH (here pH 7.2) for growth is needed at the beginning of a batch culture to reach high cell densities faster. Then the pH can be decreased for a higher antibody production rate. Estimated ATP production rates increased with high pH. This increase was due to the increase in glycolytic flux. A 3-fold increase in ATP production, however, did not lead to a higher growth rate. We have seen that the energy metabolism is not related to cell growth when altered by serum. Cells seem to dissipate energy for unknown reasons. At elevated pH, cells do not use the extra energy generated from glycolysis and glutamine oxidation. Oxidative phosphorylation contributed to ATP production about 76% a t pH 6.90. This contribution decreased gradually at elevated pH values due to increased glycolytic flux. As the lactate was the primary end product from glucose, we could attribute the energy production from oxidative phosphorylation to glutamine oxidation. Under these conditions we observed that glutamine provided about 1.5-3.5 NADH depending on the pH. A maximum at pH 7.2 for the NADH yield corresponds to pH optimum for cell growth, suggesting that cell growth

493

Biotechnol. FYOQ..,1991, Vol. 7, No. 6

may be related to, among other factors, the efficiency of glutamine oxidation.

5. Conclusions We have investigated the effects of serum concentration, the dissolved oxygen tension, and medium pH on hybridoma culture kinetics in a controlled batch bioreactor. Together with other data presented elsewhere (Ozturk and Palsson, 1990a,b,d, 1991b,c; Ozturk et al., 1991), this work presents an integrated and comprehensive effort in characterization of hybridoma cell growth, metabolism, and antibody production. We have studied a broad spectrum of processvariables and investigated the response of hybridoma cells in terms of growth, viability, metabolic rates for glucose, lactate, glutamine, ammonia, oxygen, and amino acids, respiration activities and energy metabolism, and antibody production rates. Several conclusions can be drawn from the present study: (1)Cell growth is enhanced by serum, which acts as a growth-promoting and a shear-protective agent. A pH of 7.20 is optimal for the cell line used in this study. Cell growth is not sensitive to DO levels between 20% and 80% air saturation. Cell growth is depressed at very low (1% ) and very high (100%)DO levels in batch mode. (2) Metabolic rates are not dependent on serum concentration for both cell lines used. Increasing the DO levels increases metabolic activities. Cells are metabolically more active when they are limited by oxygen. Elevating pH increases glucose consumption rate significantly, whereas glutamine consumption shows a minimum at pH 7.20. Medium pH alters the metabolic yield coefficients. Glucose and glutamine consumption rates proceed at a relatively constant ratio. However,cells prefer more glucose at high pH values and under oxygen limitation. (3) Amino acid metabolism follows the same pattern as glutamine and glucose. The metabolic rates of amino acids are 1 order of magnitude lower than glutamine consumption with the exception of alanine. Alanine production rate relative to glutamine is affected by DO and pH. The medium pH alters the relative rate of glutamate, serine, and branched amino acids. (4) Oxygen uptake rate is independent of serum concentrations, DO levels, and medium pH for the cell line used. (5) Cells obtain the same amount of metabolic energy at different serum and DO levels. However, increase in pH results enhances the ATP production. Cells obtain about 55 % of the energy from oxidative phosphorylation. This percentage increases at low pH values. (6) Antibody production is not growth associated. The specific antibody production rate is independent of serum concentrations and DO levels. Only acidic pH yields a higher antibody production rate. The amount of MAb produced ina batch mode depends mostly on the viable cell concentrations.

Acknowledgment This work was supported by National Science Foundation Grant EET-8712756. We thank Drs. J. Latham Claflin and Mark S. Kaminski for providing the hybridoma cell lines used in this study. A part of this work was first presented at the AIChE annual meeting in San Francisco, CA, November 1989. Supplementary Material Available: Ten figures showing more experimental data presented in graphical form as described in the text (10 pages). Ordering information is given on any current masthead page.

Literature Cited Adamson, S. R.; Behie, L. A.; Gaucher, G. M.; Lesser, B. H. Metabolism of Hybridoma Cells in Suspension Culture:

Evaluation of Three Commercially Available Media. In Commercial Production of Monoclonal Antibodies: A Guide for Scale-up; Seaver, S. S., Ed.; Marcel Dekker, Inc.: New York, 1987. Balin, A. K.; Goodman, D. B. P.; Rasmussen, H.; Cristofalo, V. J. The Effect of Oxygen Tension on the Growth and Metabolism of WI-38 Cells. J. Cell. Physiol. 1976,89, 235-250. Barton, M. E. Effect of pH on the Growth Cycle of HeLa Cells in Batch Suspension Culture Without Oxygen Control. Biotechnol. Bioeng. 1971, 13, 471-492. Birch, J. R.; Edwards, D. J. The Effect of pH on the Growth and Carbohydrate Metabolism of a Lymphoblastoid Cell Line. Dev. Biol. Stand. 1980, 46, 59-63. Boraston, R.;Thompson, P. W.; Garland, S.; Birch, J. R. Growth and Oxygen Requirements of Antibody Producing Mouse Hybridoma Cells in Suspension Culture. Dev. Biol. Stand. 1984, 55,103-111.

Briles, D. E.; Forman, C.; Hudak, S.; Claflin, J. L. The Effects of Idiotype on the Ability of IgGl anti-phosphorylcholine Antibodies to Protect Mice from Fatal Infection with Streptococcuspneumoniae. Eur. J.Immunol. 1984,14,1027-1030. Brosemer, R. W.; Rutter, W. J. The Effect of Oxygen Tension on the Growth and Metabolism of a Mammalian Cell. Erp. Cell Res. 1961, 25, 101-113. Dalili, M.; Ollis, D. F. Transient Kinetics of Hybridoma Growth and Monoclonal Antibody Production in Serum-Limited Cultures. Biotechnol. Bioeng. 1989, 33, 984-990. Danes, B. S.; Broadfoot, M. M.; Paul, J. A Comparative Study of Respiratory Metabolism in Cultured Mammalian Cell Strains. Exp. Cell Res. 1963, 30, 369-378. Daskal, I.; Sinclair, R. Influence of Growth Rate on Protein Synthesis and Distribution of Ribosomes in HeLa Cells. I n Vitro 1975, 11, 338-346. Eagle, H. The Effect of Environmental pH on the Growth of Normal and Malignant Cells. J. Cell. Physiol. 1973,82, 1-8. Eigenbrodt, E.; Fister, P.; Reinacher, M. New Perspectives on Carbohydrate Metabolism in Tumor Cells. In Regulation of Carbohydrate Metabolism; Beitner, R., Ed.; CRC Press: Boca Raton, FL, 1985; Vol. 11, pp 141-179. Fleischaker, R. J.,Jr.; Sinskey, A. J. Oxygen Demand and Supply in Cell Culture. Eur. J.Appl. Microbiol. Biotechnol. 1981,12, 193-197.

Glacken, M. W.; Adema, E.; Sinskey, A. J. Mathematical Descriptions of Hybridoma Culture Kinetics: I. Initial Metabolic Rates. Biotechnol. Bioeng. 1988, 32, 491-506. Harbour, C.; Low, K. S.;Marquis, C. P.;Barford, J. P. pH Control Options for Hybridoma Cultures. Biotechnol. Tech. 1989,3, 73-78.

Heath, C. A.; Dilwith, R.; Belfort, G. Methods for Increasing Antibody Production in Suspension and Entrapped Cell Cultures: Biochemical and Flow Cytometric Analysis as a Function of Serum Content. J. Biotechnol. 1989,15,71-89. Kaminski, M. S.; Kitamura, K.; Maloney, D. G.; Levy, R. Idiotype Vaccination Against Murine B Cell Lymphoma: Inhibition of Tumor Immunity by Free Idiotype Protein. J. Immunol. 1987,138, 1289-1296. Kilburn, D. G.; Lilly, M. D.; Webb, F. C. The Energetics of Mammalian Cell Growth. J. Cell Sci. 1968,4, 645-654. Kilburn, D. G.; Lilly, M. D.; Webb, F. C. The Effect of Dissolved Oxygen Partial Pressure on the Growth and Metabolism of Mouse LS Cells. J. Cell Sci. 1969, 4, 25-37. Kunas, T. K.; Papoutaakis, E. T. Increasing Serum Concentrations Decrease Cell Death and Allow Growth of Hybridoma Cells at Higher Agitation Rates. Biotechnol. Lett. 1989, 11, 525530.

Leist, C.; Meyer, H.-P.; Fiechter, A. Process Control During the Suspension Culture of a Human Melanoma Cell Line in a Mechanically Stirred Loop Bioreactor. J. Biotechnol. 1986, 4,235-246.

Low, K.; Harbour, C. Growth Kinetics of Hybridoma Cells: (1) The Effects of Varying Foetal Calf Serum Levels. Deu. Biol. Stand. 1985,60, 17-24. MacMichael, G. J. The Use of Continuous Culture to Enhance MonoclonalAntibody Production. Hybridoma 1989a,8,117126.

494

MacMichael, G. J. The Effects of pH and Oxygen on the Growth, Monoclonal Antibody Production, and Metabolism of a Mouse Hybridoma. Am. Biotechnol. Lab. 198913, 7,44-47. Maiorella, B.; Harwarth, B.; Shauger, A.; Inlow, D. Optimized pH Control for Antibody Production. Paper presented at the 198th National Meeting of the American Chemical Society, Miami, FL, 1989. McQueen, A.; Meilhoc, E.; Bailey, J. E. BiotechnoL Lett. 1989, 9,831-836. McKeehan, W. L. Glutaminolysis in Animal Cells. In Carbohydrate Metabolism in Cultured Cells; Morgan, M. J., Ed.; Plenum Press: New York, 1986; pp 111-150. Miller, W. M.; Blanch, H. W.; Wilke,C. R. The Effects of Dissolved Oxygen Concentration on Hybridoma Growth and Metabolism in Continuous Culture. J. Cell. Physiol. 1987, 132, 524-530. Miller, W. M.; Blanch, H. W.; Wilke, C. R. Kinetic Analysis of Hybridoma Growth and Metabolism in Batch and Continuous Culture: Effect of Nutrient Concentrations, Dilution Rate, and pH. Biotechnol. Bioeng. 1988,32,947-965. Mizrahi, A. Oxygen in Human Lymphoblastoid Cell Line Cultures and Effect of Polymers in Agitated and Aerated Cultures. Deu. Biol. Stand. 1984, 55, 93-102. Ozturk, S. S.; Palsson, B. 0. Chemical Decomposition of Glutamine: Effect of Media Type, pH, and Serum Concentration. Biotechnol. Prog. 1990a, 6, 121-128. Ozturk, S. S.; Palsson, B. 0. Effects of Initial Cell Density on Hybridoma Cell Growth, Metabolism, and Antibody Production Rates. J. Biotechnol. 1990b, 16, 259-278. Ozturk, S. S.; Palsson, B. 0.Loss of Antibody Production During Long-term Cultivation of a HybridomaCell Line in Low-serum and Serum-free Media. Hybridoma 199Oc, 9, 167-175. Ozturk, S. S.; Palsson, B. 0. Effects of Dissolved Oxygen on Hybridoma Growth, Metabolism, and Antibody Production in Continuous Culture. Biotechnol. Prog. 1990d, 6, 437-446. Ozturk, S. S.; Palsson, B. 0. Physiological Changes During the Adaptation of Hybridoma Cells to Low Serum and Serumfree Medium. Biotechnol. Bioeng. 1991a, 37, 35-46. Ozturk, S. S.; Palsson, B. 0. Examination of Serum and Bovine Serum Albumin as Shear Protective Agents in Agitated Cultures of Hybridoma Cells. J. Biotechnol. 1991b, 18, 1328. Ozturk, S. S.;Palsson, B. 0. Effects of Osmolarity on Hybridoma Cell Growth, Metabolism, and Antibody Production Rates. Biotechnol. Bioeng. 1991c, 37, 989-993. Ozturk, S. S.; Palsson, B. 0. Metabolic Growth, and Antibody Production Kinetics of Hybridoma Cell Culture: 1. Analysis of Data from Controlled Batch Reactors. Biotechnol. Prog. 1991d, preceding paper in this issue.


Ozturk, S. S.; Riley, M. R.; Palsson, B. 0. Effects of Ammonia and Lactate on Hybridoma Cell Growth, Metabolism, and Antibody Production Rates. Biotechnol. Bioeng. 1991,in press. Phillips, H. A.; Schrer, J. M.; Bols, N. C.; Moo-Young, M. Effect of Oxygen on Antibody Productivity in Hybridoma Culture. Biotechnol. Lett. 1987, 9, 745. Renard, J. M.; Spagnoli, R.; Mazier, C.; Salley, M. F.; Mandine, E. Evidence That Monoclonal Antibody Production Kinetics is Related to the Integral of Viable Cells Curve in Batch Systems. Biotechnol. Lett. 1988, 10, 91-96. Reuveny, S.; Velez, D.; Macmillan, J. D.; Miller, L. Factors Affecting Monoclonal Antibody Production in Culture. Deu. Biol. Stand. 1987, 66, 169-175. Rubin, H. J. Cell. Biol. 1971, 51, 686-702. Rueckert, R. R.; Mueller, G. C. Effect of Oxygen on HeLa Cell Growth. Cancer Res. 1960,20,944-949. Self, D. A.; Kilburn, D. G.; Lilly, M. D. The Influence of Dissolved Oxygen Partial Pressure on the Level of Various Enzymes in Mouse LS Cells. Biotechnol. Bioeng. 1968, 10, 815-828. Suleiman, S. A.; Stevens, J. B. The Effect of Oxygen Tension on Rat Hepatocytes in Short-term Culture. In Vitro Cell. Dev. Biol. 1987, 23, 332-338. Suzuki, E.; Ollis, D. F. Cell Cycle Model for Antibody Production Kinetics. BiotechnoL Bioeng. 1989, 34, 1398-1402. Thomas, J. N. Nutrients, Oxygen, and pH. In Mammalian Cell Technology; Thilly, W. G., Ed.; Butterworths: Boston, 1986; pp 199-221. Velez, D.; Reuveny, S.; Miller, L.; Macmillan, J. D. Kinetics of Monoclonal Antibody Production in Low Serum Growth Medium. J. Immunol. Methods 1986,86, 45-52. Wergeland, I.; Wallberg, C.; Ericsson, R. Hybridoma Culture for the Continuous Production of Monoclonal Antibodies: a Fractionalized Factorial Design Experiment for Process Optimization. In Modern Approaches to Animal Cell Technology; Spier, R. E., Griffiths, J. B., Eds.; Butterworths: Boston, 1987; p 795. Wilhelm, G.; Schulz.; Hofmann, E. pH-Dependence of Aerobic Glycolysis in Ehrlich Ascites Tumour Cells. FEBSLett. 1971, 17, 158-162. Zielke, H. R.; Sumbilla, C. M.; Sevdalian, D. A.; Hawkins, R. L.; Ozand, P. T. Lactate: A Major Product of Glutamine Metabolism by Human Diploid Fibroblasts. J. Cell. Physiol. 1980,104,433-441. Zielke, H. R.; Zielke, C. L.; Ozand, P. T. Glutamine: a Major Energy Source for Cultured Mammalian Cells. Fed. Proc. 1984, 43, 121-125. Accepted September 30, 1991. Registry No. DO, 7782-44-7.

Growth, metabolic, and antibody production kinetics of hybridoma cell

Recommend Documents