Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Kinetics study of hybridoma growth and antibody production Henry, Olivier 2000

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2000-0077.pdf [ 4.03MB ]
Metadata
JSON: 831-1.0058764.json
JSON-LD: 831-1.0058764-ld.json
RDF/XML (Pretty): 831-1.0058764-rdf.xml
RDF/JSON: 831-1.0058764-rdf.json
Turtle: 831-1.0058764-turtle.txt
N-Triples: 831-1.0058764-rdf-ntriples.txt
Original Record: 831-1.0058764-source.json
Full Text
831-1.0058764-fulltext.txt
Citation
831-1.0058764.ris

Full Text

KINETICS STUDY OF HYBREDOMA GROWTH AND ANTIBODY PRODUCTION by Olivier Henry B.Eng., Ecole Polytechnique, 1997 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE FACULTY OF GRADUATE STUDIES Department of Chemical and Biological Engineering and The Biotechnology Laboratory We accept this thesis as conforming to the required standard The University of British Columbia February 2000 © Olivier Henry, 2000 In presenting t h i s thesis i n p a r t i a l f u l f i l m e n t of the requirements for an advanced degree at the Uni v e r s i t y of B r i t i s h Columbia, I agree that the Library s h a l l make i t f r e e l y a v a i l a b l e for reference and study. I further agree that permission for extensive copying of t h i s thesis for s c h o l a r l y purposes may be granted by the head of my department or by his or her representatives. I t i s understood that copying or p u b l i c a t i o n of t h i s thesis for f i n a n c i a l gain s h a l l not be allowed without my written permission. Department of C.HEHIrfil B^}Qlj[lMk[IVG The Un i v e r s i t y of B r i t i s h Columbia Vancouver, Canada ABSTRACT Monoclonal Antibodies (MAb) produced by hybridomas have an expanding market for use in diagnostic assays, affinity separations and therapeutic applications. The objectives of this thesis were to study the kinetics of growth and antibody production of a hybridoma cell line in five different modes of operation and establish to what extent these kinetics are comparable from one mode to another. Hybridoma cells producing monoclonal antibodies were grown in batch, fed-batch, semi-continuous, continuous and perfusion cultures. The kinetics were compared in terms of growth rate, nutrient consumption rate and antibody formation rate. Batch cultures reached a maximum cell concentration of around 3x106 cells/mL and antibody concentration of 100 |j,g/mL. Glutamine was likely the substrate limiting the batch cultures since its depletion coincided with the end of exponential growth phase. No other amino acid was consumed completely and the levels of toxic metabolites (23 mM lactate; 2.6 mM ammonium) were much lower than those reported to affect growth or MAb production. A simple unstructured model was developed based on the batch data during the exponential phase. Simulations were performed and compared with experimental results obtained in semi-continuous and continuous cultures at dilution rates varying from 0.016 to 0.036 h"1. Cell yields from glucose in batch exponential phase were approximately 30% less than in the continuous processes. The cell specific glucose and glutamine uptake rates increased at greater dilution rates in both continuous and daily fed semi-continuous cultures. Comparing the cultures on the basis of specific dilution rates, perfusion results were in good agreement ii with semi-continuous and continuous data in terms of glutamine and glucose uptake rates. Although the specific antibody production rate only varied within a narrow range (0.7 to 2 pg/cell.h), the data suggest that the formation of antibody was growth-associated. It is concluded that modeling and analysis of semi-continuous data can describe continuous processes sufficiently well to be a useful tool for perfusion system optimization, while providing the relative simplicity of batch operation. In fed-batch, through the daily addition of a concentrate feed, cultures yielded antibody concentrations two times greater (200 (Xg/mL) compared to batch. This increase in final antibody concentration was closely related to the increase in the viable cell index. iii TABLE OF CONTENTS ABSTRACT ii TABLE OF CONTENTS.... IV LIST OF TABLES VI LIST OF FIGURES VIH LIST OF SYMBOLS AND UNITS XI ACKNOWLEDGMENTS XII 1 INTRODUCTION 1 1.1 Monoclonal antibody 1 1.2 Alternative reactor operation modes 2 1.3 Thesis objectives. 4 1.4 Thesis outline.. 5 2 LITERATURE REVIEW 6 2.1 Cell Metabolism 6 2.1.1 Glucose 6 2.1.2 Glutamine 6 2.1.3 Ammonium 7 2.1.4 Lactate 8 2.1.5 Serum 9 2.2 Semi-continuous cultures. 10 2.3 Continuous cultures 10 2.4 Fed-batch cultures 13 2.5 Perfusion 15 2.6 Culture mode comparison 18 3 MATERIAL AND METHODS 26 3.1 Cell line and medium 26 3.2 Analytical Techniques 30 3.2.1 Cell concentration, viability and sample analysis 30 3.2.2 Cell volume 30 3.2.3 Monoclonal antibody assay 30 3.2.4 Glutamine assay 31 3.2.5 Amino acids analysis 32 iv 4 Batch Modeling 34 4.1 Correlations for the growth rate 35 4.2 Correlations for glucose and glutamine uptake rates 39 4.3 Correlation for the production rate. 41 5 RESULTS AND DISCUSSION 43 5.1 Cell Volume. 43 5.2 Inoculum state 46 5.2.1 Phase of inoculum 46 5.2.2 Inoculum size 48 5.3 Batch cultures 49 5.3.1 Batch results 49 5.3.2 Batch model parameters 53 5.3.3 Batch simulations 53 5.4 Semi-Continuous cultures. 55 5.4.1 Semi-continuous results 55 5.4.2 Simulation results 60 5.5 Continuous cultures 62 5.5.1 Continuous results 62 5.5.2 Simulation results 66 5.6 Perfusion cultures 69 5.6.1 Perfusion results 69 5.7 Fed-batch cultures 73 5.8 Comparison of kinetic rates 77 5.8.1 Specific glucose uptake rate 78 5.8.2 Specific glutamine uptake rate 80 5.8.3 Specific monoclonal antibody production rate 82 6 CONCLUSIONS & RECOMMENDATIONS 85 7 REFERENCES 87 v LIST OF TABLES Table 2-1: Semi-continuous cultures of hybridoma cells 10 Table 2-2: Continuous cultures of hybridoma cells 12 Table 2-3: Fed-batch production of MAb and improvement over batch culture 14 Table 2-4: Perfusion cultures and improvement over batch culture 17 Table 2-5: Comparison between reactor modes of operation. Sections are left blank when no data are available 19 Table 3-1 Composition of Dulbecco's Modified Eagle Medium 26 Table 3-2: Concentrate formulation used for the fed-batch cultures 29 Table 4-1: Correlations for the growth rate 37 Table 4-2: correlations for the glucose uptake rate 40 Table 4-3: Correlations for the glutamine uptake rate 40 Table 4-4: Correlations for the antibody production rate 42 Table 5-1: Effect of inoculum age. Errors for antibody concentrations are standard deviations based on duplicate measurements. Error on cell concentration measurements is 10% 46 Table 5-2: Inoculum size effect. Errors for antibody concentrations are standard deviations based on duplicate measurements. Error on cell concentration measurements is 10% 48 Table 5-3: Amino acids consumption/production rates in batch culture. Negative values indicate metabolite consumption, whereas positive values indicate metabolite production 52 Table 5-4: Parameters value assumed for the simulations. The errors are standard deviations 53 Table 5-5: Semi-continuous culture results. The concentrations measured prior to media exchange are given 57 Table 5-6: Amino acid concentrations (mM) in semi-continuous culture. The asparagine and glycine peaks were merged on the spectrum due to poor resolution and so the combined concentrations were estimated 58 Table 5-7: Amino acid consumption/production rates (pmol/cell.d) for semi-continuous culture. Negative values indicate metabolite consumption, whereas positive values indicate metabolite production. 59 Table 5-8: Continuous culture results. Concentrations are average from steady-state measurements 63 v i Table 5-9: Amino acid concentrations (mM) in continuous culture. When two peaks were merged on the spectrum due to poor resolution, the combined concentrations were estimated 65 Table 5-10: Amino acids consumption/production rates (pmol/cell.d) in continuous culture. Negative values indicate metabolite consumption, whereas positive values indicate metabolite production 66 Table 5-11: Perfusion culture results. Cells were grown in spinners with a working volume of 200 mL 71 Table 5-12: Amino acid concentration (mM) in perfusion culture. When two peaks were merged on the spectrum due to poor resolution, the combined concentrations were estimated. 72 Table 5-13: Amino acid consumption/production rates (pmol/cell.d) in perfusion culture. Negative values indicate metabolite consumption, whereas positive values indicate metabolite production 72 Table 5-14: Fed-batch results 75 Table 5-15: Amino acid concentrations (mM) in fed-batch culture. When two peaks were merged on the spectrum due to poor resolution, the combined concentrations were estimated. 76 vii LIST OF FIGURES Figure 1-1: Bioreactor operation modes. (A) Batch (B) Fed-batch (C) Continuous (D) Semi-continuous (E) Perfusion. Solid lines indicate a continuous flow and dashed lines correspond to semi-continuous flow 3 Figure 2-1: Simplified schematic diagram of hybridoma cell metabolic pathways 7 Figure 3-1: Schematic diagram of the apparatus used for the perfusion culture of hybridoma cells 28 Figure 5-1: Mean cell volume (open circles) and viability (filled triangles) profiles in batch culture. A indicates the early exponential phase, B the exponential phase, C the stationary phase and D the decline phase. Error bars for cell volumes are standard deviations from duplicate counts. A conservative estimate of error was applied to viability data (20 %) 44 Figure 5-2: Volume specific (filled circles) and cell specific (open circles) glucose uptake rates (GUR) profiles in batch culture. Errors were calculated assuming a 10% error on cell concentration measurements and 0.5 mM error on glucose measurements 45 Figure 5-3: Effect of the inoculum phase on cell growth. Cultures were started with inoculum taken from a batch at 27 (filled triangles), 53 (open circles), 66 (filled squares) and 79 h (open diamonds). A conservative estimate of error (10%) was assumed for cell concentration measurements 47 Figure 5-4: Cell concentration (filled circles) and cell viability (open triangles) profiles in batch culture. Errors were calculated assuming a conservative estimate of error (10%) on cell concentration measurements and 20 % in viability measurements 49 Figure 5-5: Glucose (filled circles) and lactate (open diamonds) concentrations in batch culture. The measurement errors for glucose and lactate are 0.5 mM 50 Figure 5-6: Glutamine (open diamonds) and ammonium (filled circles) concentration profiles in batch culture. The measurement error for ammonium is 0.5 mM 51 Figure 5-7: Monoclonal antibody concentration (filled circles) profile for batch culture. Error bars are standard deviations from two separate assays 51 Figure 5-8: Simulation of cell concentration profile. Circles represent experimental points from a batch culture and the model prediction is given by the line. A conservative estimate of error (10%) was assumed for cell concentration measurements 54 Figure 5-9: Simulation of glucose and glutamine concentration profiles. Circles represent experimental data points from a batch culture and the model prediction is given by the line. The measurement error for glucose concentrations is 0.5 mM 54 viii Figure 5-10: Simulation of monoclonal antibody profile. Circles represent experimental data points from a batch culture and the model prediction is given by the line. Error bars are standard deviations from duplicate measurements 55 Figure 5-11: Simulation of cell concentration profile in semi-continuous mode (D = 0.019 h"1). Filled circles represent experimental data points prior to medium exchange. These points correspond to the top part of the saw-tooth profile predicted by the model (solid line) 61 Figure 5-12: Simulation of glucose concentration profile in semi-continuous mode (D = 0.019 h"1). Filled circles represent experimental data points prior to medium exchange. These points correspond to the bottom part of the saw-tooth profile predicted by the model (solid line) 61 Figure 5-13: Simulation of antibody profile in semi-continuous culture (D = 0.019 h"1). Filled circles represent experimental data points prior to media exchange. These points correspond to the top part of the saw-tooth profile predicted by the model (solid line). Error bars are standard deviations from duplicate measurements 62 Figure 5-14: Simulation of cell concentration profile in continuous culture (D = 0.028 h"1). The model predictions (solid line) are compared with data from two separate experiments (filled circles and triangles) run under the same operating conditions 68 Figure 5-15: Simulation of glucose concentration profile in continuous culture (D = 0.028 h"1). The model predictions (solid line) are compared with data from two separate experiments (filled circles and triangles) run under the same operating conditions 68 Figure 5-16: Simulation of antibody profile in continuous culture (D = 0.028 h"1). The model predictions (solid line) are compared with data from two separate experiments (filled circles and triangles) run under the same operating conditions 69 Figure 5-17: Viable cell and antibody concentration profiles for a perfusion culture. Open circles represent the viable cell concentrations and filled triangles the antibody concentrations. The average feed rate was 195 ml/d 70 Figure 5-18: Cell concentration (filled circles) and viability (filled triangles) profiles in fed-batch culture. The starting glucose concentration was 5 mM 74 Figure 5-19: Antibody concentration (filled circles) profile in fed-batch culture. Error bars are standard deviations. The starting glucose concentration was 5 mM 74 Figure 5-20: Glucose (filled circles) and lactate (filled triangles) concentration profiles in fed-batch culture. Arrows indicate when concentrate additions were made 75 Figure 5-21: Antibody concentration as a function of the viable index for batch (open symbol) and three fed-batch (filled symbols) cultures. All cultures exhibited ix a similar average specific production rate given by the slope of the solid line 77 Figure 5-22: Specific glucose uptake rate as a function of growth rate for different culture modes 78 Figure 5-22: Specific glucose uptake rate as a function of growth rates for batch (open circles), semi-continuous (filled circles), continuous (open triangles) and fed-batch (cross) cultures. The solid line represents the batch model prediction 78 Figure 5-23: Specific glucose uptake rate as a function of specific dilution rate for continuous processes 80 Figure 5-23: Specific glucose uptake rate as a function of specific dilution rate for semi-continuous (filled circles), continuous (open triangles) and perfusion (filled squares) cultures 80 Figure 5-24: Specific glutamine uptake rate as a function of growth rate for different culture modes 81 Figure 5-24: Specific glutamine uptake rate as a function of growth rate for batch (open circles), semi-continuous (filled circles) and continuous (open triangles) 81 Figure 5-25: Specific glutamine uptake rate as a function of specific dilution rate 82 Figure 5-25: Specific glutamine uptake rate as a function of specific dilution rate for semi-continuous (filled circles), continuous (open triangles) and perfusion (filled squares) cultures. The solid line represents the batch model prediction 82 Figure 5-26: Specific antibody production rate as a function of growth rate 83 Figure 5-26: Specific antibody production rate as a function of growth rate for batch (open circles), semi-continuous (filled circles) and continuous (open triangles) cultures. In the batch model, a constant production rate of 1.5 pg/cell.h was assumed 83 Figure 5-27: Specific antibody production rate as a function of specific dilution rate 84 Figure 5-27: Specific antibody production rate as a function of specific dilution rate for semi-continuous (filled circles), continuous (open triangles) and perfusion (filled squares) cultures 84 LIST O F SYMBOLS AND UNITS Amm - Ammonium concentration (mM) D = Dilution rate (h'1) Gin = Glutamine concentration (mM) Glu = Glucose concentration (mM) Lac = Lactate concentration (mM) m = maintenance term (mmol/cell.h) P = Monoclonal antibody concentration (ug/mL) q? = Monoclonal antibody specific production rate (ng/cell.h) qs = Substrate specific consumption rate (mmol/cell.h) S = Substrate concentration (mM) t = Time (h) X = Viable cell concentration (cells/mL) Yx/s =• Yield of cells on glucose (cell/mmol) Greek symbols: a = Death rate (h'1) P = Constant for growth-associated formation of monoclonal antibody (ug/cell) y = Constant for growth-unassociated formation of monoclonal antibody (ug/cell. h) /i = Growth rate (h"1) xi ACKNOWLEDGMENTS I would like to express sincere gratitude to my two supervisors Dr. Ezra Kwok and Dr. James Piret for the guidance and support they have provided throughout the pursuit of this project. The materials I was exposed to through their engaging courses really helped me define my specific interests. I am also indebted to all my friends and co-workers of the Biotech Lab for their encouragement, inspiration, support and for enduring me whistling all the time. It has been a remarkable place to study and a unique environment to do research. I would like to thank the two research assistants, Faiz and Chris for their invaluable help and for teaching me some of the analytical techniques so important to this work. I am also grateful for the precious advice of Jason and all the helpful discussions we have had were a great source of inspiration. Many thanks to a summer student, Patricia, who did some of the experiments presented in this work and whose friendship and dynamism were greatly appreciated. Also, special thanks go to my friend Amanda whose constant support was as much appreciated as all the cookies she baked for me. Finally, my gratitude go to Yi-ta for sharing not only contaminations, but also ideas and a lot of his time to patiently take care of my training. x i i 1 I N T R O D U C T I O N 1.1 Monoclonal antibody Monoclonal antibodies (MAb) are proteins that bind to particular molecules with a high degree of specificity. They appear in the blood serum in response to foreign macromolecules (referred to as antigens). Antibodies are now important biological products that are gaining widespread use in many technical and medical fields. Because of their binding specificity, antibodies provide tools for molecular biology, chemical assays, biological separation processes and for the diagnosis and treatment of some diseases such as cancer, autoimmune disorders and infectious diseases. They are also increasingly applied to other fields such as in vivo imaging and biosensors. Currently, one-third of all biotechnology products in development are MAbs (Kling, 1999). Hybridomas are cells resulting from the fusion of two mammalian lymphoid cells. One is an antibody producing cell with a finite lifetime. The other is an "immortal", cancerous lymphocyte that normally does not produce antibodies. The resulting hybridoma combines the principal characteristics of the two parents cells: it produces a single, specific type of antibody and can live indefinitely in culture. Many of the experimental MAb therapeutic strategies require high in vivo dosages, from 0.5 to more than 5 mg/kg (Aulitzky et al., 1991). Concentrations of MAb obtained in vitro in traditional batch reactors are usually on the order of microgram-per-milliliter compared to 1 rnilligram-per-milliliter levels in mouse ascites fluid. Ascites production of MAb is still widely used, but is subject to criticism on the basis of both technical and humane criteria (McArdle, 1997), adding to the need to develop high-level expression systems for MAb. The increasing demand for MAb has stimulated efforts to maximize bioreactor cell density and MAb productivity, ameliorate media formulations and improve process optimization. 1.2 Alternative reactor operation modes Batch culture (Figure 1-1A) is the traditional production technology for animal cells. The reactor operation is technically simple and relatively short in duration, therefore lowering the risk of contamination while easing implementation, scale-up and validation. Although widely used, the batch mode is likely not the optimal process for MAb production. Many strategies have been developed to overcome the early nutrient depletion of the batch mode. In fed-batch cultures (Figure 1-1B), nutrients are added to extend the time of antibody production and, in some cases, increase the cell density. Most feeding strategies are designed to avoid nutrient depletion and/or to reduce lactate and ammonium formation by controlling glucose and/or glutamine concentrations at low levels (Omasa et al., 1992; Kurokawa et al., 1994; Ljunggren and Haggstrom, 1994; Schwabe et al., 1999). The major advantages of fed-batch operation over batch or even continuous operation are that products accumulate to higher concentrations, thus providing higher yields on medium and easing purification of the product. 2 Although more complex to operate, chemostat or continuous cultures (Figure 1-1C) provide homogeneous and constant metabolite concentrations at steady state and therefore are a preferred tool for carrying out research on metabolism, growth and antibody production kinetics. However, the cell and product concentrations are typically low, making chemostats less attractive for production. A. Batch B. Fed-Batch C. Continuous fresh • • • fresh medium spent medium D. Semi-Continuous fresh • medium T • -• y spent medium E. Perfusion retention device fresh ^ i ^^%SY+ cell-free spent medium 4-• • -^ -spent medium Figure 1-1: Bioreactor operation modes. (A) Batch (B) Fed-batch (C) Continuous (D) Semi-continuous (E) Perfusion. Solid lines indicate a continuous flow and dashed lines correspond to semi-continuous flow. 3 In semi-continuous cultures (Figure 1-1D), a portion of the reactor volume is harvested at regular intervals and replaced by an equal quantity of fresh medium, approximating continuous operation while retaining the simplicity of the batch mode. It also reduces the down-time of a bioreactor (between batch cleaning and sterilization) and eliminates the need for inoculum after the startup. However, this mode of operation still suffers from many of the disadvantages of batch culture such as time-variant conditions and low product concentrations. In a perfusion system (Figure 1-1E), the cells are retained in the reactor while medium is added and spent medium is removed either continuously or semi-continuously. Higher cell densities yield order of magnitude greater volumetric productivities. However, the difficulties of cell retention have slowed implementation of large scale perfusion systems. 1.3 Thesis objectives Fed-batch and perfusion systems are desirable for their high product concentrations and high volumetric productivities, respectively. However, the cost and development time associated with these processes slow their optimization. On the other hand, batch and semi-continuous cultures, because of their simplicity, provide cell kinetics in a relatively short time. It has not been clearly demonstrated how much the intrinsic differences between these modes of operation prevent the transposition of the cell kinetics measurements from one mode to another. Thus, the first objective of this study was to determine the kinetics of hybridoma cell growth and antibody production in batch, fed-batch, semi-continuous, continuous and 4 perfusion cultures. Then, by comparing the metabolic and production rates with respect to growth rate and cell specific dilution rate, the second objective was to determine to what extent the results readily obtained in the batch and semi-continuous modes could predict the performance of continuous and fed-batch processes. 1.4 Thesis outline To meet these objectives, the thesis is organized as follows. Chapter 2 presents a literature review of hybridoma growth and antibody production in the various culture modes. In Chapter 3, the materials and methods are described. The development of a simple kinetic model is presented in Chapter 4. Cell culture results are discussed in Chapter 5 along with a comparison of the different modes. Finally, Chapter 6 provides conclusions and suggestions for future work. 5 2 L I T E R A T U R E R E V I E W 2.1 Cell Metabolism Mammalian cells are used to produce complex proteins difficult to produce in microbial systems. However, the physiology of mammalian cell is more complicated. Figure 2-1 depicts the main metabolic pathways of hybridomas. The importance of some of the components shown in this figure will be discussed in the following sections. 2.1.1 Glucose Glucose is metabolized by glycolysis from which the carbon flux can take one of five major paths: the pentose phosphate pathway, the formation of lipids or lactate, the formation of amino acids or the tricarboxylic acid cycle. In mammalian cells, the rate of glycolysis is much faster than the maximum rate of utilization of glycolytic intermediates, therefore most glucose is converted to lactate (Batt and Kompala, 1989). 2.1.2 Glutamine Glutamine is an essential amino acid and the primary source of nitrogen for cultured mammalian cells (Zeilke et al., 1984). It also provides an additional carbon and energy source. Glutamine can be used directly in the synthesis of proteins, be converted into glutamate or alanine, or enter the TCA cycle. It also contributes to the formation of the major intracellular building blocks: amino acids, nucleotides and lipids. 6 Amino Acids Glucose Figure 2-1: Simplified schematic diagram of hybridoma cell metabolic pathways. 2.1.3 Ammonium Ammonium is a waste product of glutamine metabolism in mammalian cell cultures. It is mainly produced by the deamination of glutamine to form glutamate. The concentration of 7 ammonium in cell culture is influenced by the mode of reactor operation, glutamine concentrations, and cellular activity. Typical concentrations of ammonium in batch mode range from 2 to 5 mM, but can be higher especially in fed-batch cultures. The influence of ammonium concentration on hybridoma cell growth and monoclonal antibody production has been reported for both batch and continuous culture. Ammonium concentrations in the range of 2 to 10 mM were observed to inhibit cell growth by 50% (Ozturk et al., 1992), probably because of changes in the intracellular pH. Glacken (1987) reported a decrease in antibody productivity at elevated ammonium levels, but other studies have reported cell specific antibody productivity not influenced by ammonium concentration (McQueen and Bailey, 1990; Ozturk et al., 1992). These discrepancies are likely due to individual cell line differences. 2.1.4 Lactate Lactate is mainly produced from glucose metabolism (product of glycolysis) but can also be produced in small amounts from glutamine metabolism (Zeilke et al., 1984). The lactate concentration is a function of the glucose concentration, the mode of operation and the cellular activity. Lactate can inhibit cell growth by medium acidification but inhibition was also observed at constant pH (Ozturk et al., 1992). Most studies have reported that lactate inhibition becomes significant at concentrations greater than 40 mM (Ozturk et al., 1992). Ozturk et al. (1992) found no effect of the lactate on the specific MAb production, whereas Glacken (1987) observed a decrease in specific MAb formation. 8 2.1.5 Serum Mammalian cell growth is principally regulated not only by nutrient levels but also by growth factors which interact with great specificity with cell surface receptors. These growth factors are responsible for the serum requirements of mammalian cells. Serum also supplies hormones, binding and transport proteins and trace nutrients. It has been reported that serum increases the specific growth rate while acting as a shear-protective agent and thus decreasing the death rate (Ozturk and Palsson, 1991). This result is in agreement with many other studies (Sanfeliu et al., 1996; Zeng et al., 1998). Ozturk et al. (1991) also showed that the specific MAb production rates were independent of the serum concentration for two different cell lines. However, Gaertner et al. (1993) showed that raising the concentration of serum resulted in higher rates of antibody production. But in their study, specific MAb productivity exhibited a linear dependence on serum concentration, so they came to the conclusion that there was probably a growth factor present at a limiting concentration. Despite the aforementioned benefits, the use of serum has major drawbacks: it is expensive, it has a poorly defined complex composition, it increases the risk of contamination from viruses and mycoplasma and makes the downstream processing more complicated due to high protein concentrations. Thus, a great deal of effort has been expended in developing cost-efficient serum-free media (Glacken, 1987; Suzuki and Ollis, 1990; Hiller et al., 1991). 9 2.2 Semi-continuous cultures Compared to continuous and batch experiments, very little has been published regarding the semi-continuous culture of hybridoma cells (Table 2-1). Ramirez and Mutharasan (1990) grew cells at four semi-continuous dilution rates in T-flasks. Both the specific glucose uptake rate and the yield of lactate from glucose increased with dilution rate. The specific MAb production rate was greater at lower dilution rates (Ramirez and Mutharasan, 1990; Leno et al., 1992). In contrast, Reuveny et al. (1986) using cultures that were fed twice daily found no relationship between the cell specific production rate and the dilution rate. Decreasing antibody concentrations with increasing dilution rates have been reported (Reuveny et al., 1986; Leno et al., 1992). All these studies reported higher metabolic activity at higher dilution rates (Reuveny et al., 1986; Ramirez and Mutharasan, 1990; Leno et al., 1992). Table 2-1: Semi-continuous cultures of hybridoma cells. Dilution rate (h1) Cell concentration (cells/mL) Antibody concentration (mg/L) Specific production rate (pg/cell.h) Reuveny et al. (1986) 0.002-0.017 0.9 - 2.4xl06 45 - 290 0.44 - 0.67 Ramirez and Mutharasan (1990) 0.007 - 0.042 2-3.4xl0 6 n/a 0.05 - 0.20 Leno et al. (1992) 0.018-0.055 1 - 2.6 xlO6 2-13 0.02 - 0.27 2.3 Continuous cultures Unlike batch and semi-continuous modes, in which the cells are exposed to a changing environment, continuous operation allows maintenance of constant conditions in the bioreactor. For this reason, it provides a better method of determining various kinetic 10 parameters, and many studies of hybridoma cells have been carried out in continuous bioreactors (Table 2-2). Some have reported non-growth associated antibody production rates (Hiller et al., 1991; Portner and Schafer, 1996). In contrast, maximum antibody production and specific productivity were obtained for other hybridomas at an intermediate dilution rate (Ray et al., 1989; Follstad et al., 1999). Robinson and Memmert (1991) have studied the production of a chimeric antibody and the specific rate of MAb production was strictly growth associated (linear function of the growth rate). Unlike all the aforementioned studies, Frame and Hu (1991) and Martens et al. (1993) have shown that, for their respective cell lines, the productivity of antibody was described by a mixed-type model with a non-growth-associated term and a negative-growth associated term. Exponentially decreasing antibody concentrations and specific productivities with increasing dilution rates were also reported (Miller et al., 1988). Most studies are in agreement regarding the increase of glucose uptake rate with respect to growth rate (Miller et al., 1988; Ray et al., 1989; Frame and Hu, 1991; Hiller et al., 1991; Robinson and Memmert, 1991). In some cases, the uptake rates increased strongly and ceased to be linear at high growth rates (Frame and Hu, 1991; Robinson and Memmert, 1991). Increasing yields of lactate on glucose with increasing dilution rates and relatively constant yields of ammonium on glutamine have been reported (Miller et al., 1988; Hiller et al., 1991). In contrast, Ray et al. (1989) observed lower yields of lactate on glucose at high dilution rates. 11 Table 2-2: Continuous cultures of hybridoma cells. Dilution rate (h1) Cell concentration (cells/mL) Antibody concentration (mg/L) Specific production rate (pg/cell.h) Low et al. (1987) 0.01-0.048 0.4- 1.4 xlO6 1-7 0.05-0.1 Miller et al. (1988) 0.012-0.055 0.7-2.2xl06 5-80 0.3 - 0.6 Follstad et al. (1999) 0.01 -0.04 0.5-1.5xl06 n/a 0.9-2.4 Ray etal. (1989) 0.012-0.039 0.8-2.2xl06 60 - 230 1.2-2.5 Hiller et al. (1991) 0.017-0.054 0.2- 1.4 xlO6 8-48 0.45 - 1.1 Robinson and Memmert (1991) 0.009 - 0.043 0.5-0.9xl06 45 0.58-3.2 Frame and Hu (1991) 0.01-0.05 0.65-1.5 xlO6 30-250 n/a Martens et al. (1993) 0.01-0.058 1.9-2.2xl06 n/a 0.1 - 1.5 Some discrepancies regarding the effect of the dilution rate on the cell concentration also exist in the literature. Cell concentrations relatively unaffected by the dilution rate were reported (Robinson and Memmert, 1991; Martens et al., 1993). In other studies, the viable cell concentrations went through a maximum with respect to the dilution rate (Miller et al., 1988; Hiller et al., 1991). Moreover, Follstad et al. (1999) noted that, at the same dilution rate, two cultures could be obtained which exhibited similar growth rates and viabilities, but drastically different cell concentrations. Taken together, these results show how hybridomas are unpredictable from one cell line to another and, for process optimization, there is a need to determine these dependencies each time. 12 2.4 Fed-batch cultures To extend culture lifetime, one or several limiting nutrients can be supplemented at higher concentrations to obtain medium fortification. Despite such batch mode optimization (Jo et al., 1990), at high concentrations, some medium components can have inhibitory effects on growth and/or production. Therefore, a more advanced strategy has been to feed the culture in a fed-batch mode. Concentrate and control strategies have been developed to avoid nutrient depletion or waste accumulation. Viable cell concentrations as high as 1.2xl010 cells/mL (Zhou et al., 1997) and final antibody concentrations up to 2.4 g/L (Xie and Wang, 1996) have been reported. Most concentrate formulations are based on analysis of consumption rates obtained from batch data. In many cases, concentrates are fed daily without any feedback control mechanism. Table 2-3 shows some of the different strategies used for the fed-batch production of MAb. Since the final antibody concentration in a culture is essentially a function of the cell longevity and MAb secretion rate, fed-batch process strategies aim at maximizing these two parameters in order to maximize the final antibody yield. The feeding protocols vary from one study to another, but the vast majority aim at keeping the concentration of glucose and glutamine low to reduce the formation of ammonium and lactate (Omasa et al., 1992; Kurokawa et al., 1994; Ljunggren and Haggstrom, 1994; Schwabe et al., 1999). By controlling both the glucose and glutamine concentrations at low levels, Kurokawa (1994) increased the specific antibody production rate by 2.6-fold. Using a similar strategy, 13 Ljunggren et al. (1994) did not improve upon the results obtained in batch, probably because the effects of ammonium and lactate are cell line dependent. Both studies reported significantly reduced specific consumption rates of glucose and glutamine in fed-batch. Table 2-3: Fed-batch production of MAb and improvement over batch culture. Reference Strategy Cell and MAb concentrations Increase over batch Glacken et al. (1986) Controlled addition of glucose and glutamine 1.2 xlO6 cells/mL Similar Omasa et al (1992) Glutamine addition n/a n/a Fike et al. (1993) Continuous addition of a nutrient supplement 3.8 xlO6 cells/mL 132 mg/L 1.3-fold 4-fold Bibila et al (1994) Concentrated medium 2x106 cells/mL 1000 mg/L 2-fold 7-fold Bushell etal (1994) Amino acids addition 1.4xl06 cells/mL 36 mg/L 1.4-fold 2.8-fold Kurokawa et al. (1994) Glucose/Glutamine control 4.1xl06 cells/mL 172 mg/L n/a n/a Ljunggren et al. (1994) Glutamine limited fed-batch Glucose limited fed-batch Glu + Gin limited fed-batch lxlO6 cells/mL 40 mg/L 1.2xl06 cells/mL 64 mg/L 1.4xl06cells/mL 92 mg/L Reduction Reduction Reduction Reduction Similar Similar Xie and Wang (1994) Stoichiometric nutrients requirements 6.3xl06 cells/mL 900 mg/L 2.7-fold 17-fold Zhou etal. (1995) On-line OUR* measurements and nutrient consumption analysis 1.4xl07 cells/mL 58 mg/L 7-fold 7-fold Xie and Wang (1996) Stoichiometric nutrient requirements 1.7xl07 cells/mL 2400 mg/L 7-fold 46-fold Zhou et al. (1997) Feeding strategy based on the integral of viable cell 6.6x106 cells/mL 2700 mg/L n/a n/a Schwabe et al. (1999) Glutamine/Glucose addition based on OUR* measurements 3xl06 cells/mL 50 mg/L 1.5-fold 2-fold * OUR: Oxygen uptake rate 14 Based on amino acid analysis, Schwabe et al. (1999) noted that their batch cultures were limited by the depletion of essential amino acids such as glutamine, methionine, isoleucine and leucine. Omasa et al. (1992) reported a 1.5-fold increase in specific antibody productivity following the addition of glutamine. Although the maximum specific production rates were the same in batch and fed-batch, Xie and Wang (1994) noted an average specific production rate two times greater in fed-batch compared to batch. It was suggested that this increase was the result of decreased lactate concentration in the culture. Bushell et al. (1994) tested two concentrates, one based on consumption rates calculated during the exponential phase and the other based on the stationary phase (production phase). The formulation designed from the exponential phase resulted in a higher MAb concentration while the stationary formulation did not lead to any improvement over batch culture. The "viable index", which is obtained by integrating the viable cell concentrations over the culture time, has been used to quantify the increase of cell growth in fed-batch. Many authors have correlated increases in the viable index with increases in the final antibody concentration (Luan et al., 1987; Xie and Wang, 1996; Zhou et al., 1997). 2.5 Perfusion Simple configurations such as batch and continuous cultures lead to low cell concentrations and low productivities. In the perfusion mode, cells are retained at a relatively high concentration inside the reactor. Many retention devices or reactor configurations have been employed: sedimentation chambers, membrane units, centrifuges, hollow fiber bioreactors (HFBR), ceramic cartridges, immobilized cell systems, spin-filters and acoustic separators; all 15 with their intrinsic advantages and disadvantages. Table 2-4 shows a list of several studies performed at high cell densities. The heterogeneous systems for which cell concentration measurements are usually not available (e.g. HFBRs) are not included in this list. The cell densities achieved in perfusion reactors (up to 107 cells/mL) are typically an order of magnitude higher than those reached in batch. The specific glucose uptake rate has been reported to decrease as cell density increased (Banik and Heath, 1995), but the consumption rate was found to be more a function of the specific growth rate as in low-density cultures (Hiller et al., 1993; Banik and Heath, 1995). Both the yield of lactate on glucose and the yield of ammonium on glutamine were lower at lesser specific growth rates (Hiller et al., 1993; Banik and Heath, 1995). The MAb specific production rate increased with increasing cell density and was relatively independent of the specific growth rate (Hiller et al., 1993; Banik and Heath, 1995). 16 u > *> X J M w c •o 2 •—i m .q ^ H H O t - H 00 T 3 2 00 cd rt 00 II ed ed ~ C ~ C ed <2 • VO cd ed ^ c CJ u 2 ^ CN O o OJ 3 O *d J D w > O +J c E CJ > o ex 6 •a c CQ co CJ 3 o c o '55 CN I e %-> e I H OJ •«-» Sc c OD OJ -*—' 53 •o a> a , ed co CJ •c •a .S m ^ H •*—< ^ H CO cd o •c •o .S ">> r- E o ^ *CJ ^ E VO ^ OJ -+-» 53 I H OJ <C "o X co c a I H E 3 5 cj o E x cj I 1 CJ CJ u c 4J 3 VO oo OJ OJ 3-3 oo ON OJ N OJ > CN ON ON 1 3 O OJ CN ON ON OJ OJ 00 c OJ CN ON ON OJ •a O C/3 o 1 W -a c cd C OJ co co cd i-l i l w , o in vo t~- vo 2 "a-" "a X CN I - H 1 - 1 s> ON X o CN cn 2 E x ° vo CN oj 00 cd O 60 C '•5 CJ -4-» C C i c C/3 I H OJ ^ H «c I H CJ C C *o B 53 CO 3 O u < 4J «c co 3 2 o a. o CJ 00 a •c c OJ U CN ON ON OJ c o co c x> ON ON OJ I H CJ ON ON 78 CJ I H H «/-> ON ON •3 CQ vo ON ON OJ C o co i i o 17 2.6 Culture mode comparison As described in the previous sections, hybridoma growth and MAb production in batch, fed-batch, semi-continuous and perfusion cultures have been studied extensively. However, because most studies are done using different cell lines with significantly different kinetics, it is hard to detect the differences related solely to the mode of cultivation. Fewer studies have compared different modes using the same cell line and most of these comparisons are based only on volumetric productivity or yield on medium (Table 2-5). These parameters are highly dependent on the operating conditions chosen and should be examined with caution before general conclusions can be drawn. Similar volumetric productivities in batch, semi-continuous and continuous reactors were reported (Reuveny et al., 1986; Kurkela et al., 1993), but much higher values were obtained in perfusion cultures (Reuveny et al., 1986; Bartley and MacLeod, 1992; Griffiths et al., 1992; Schmid et al., 1992). Very few studies show direct comparisons of the kinetics of growth and MAb production between different culture modes (Table 2-5). To study the relationship between growth and shear sensitivity, Petersen et al. (1990) compared batch and fed-batch cultures. Even though all the cultures had similar growth rates, considerable variations were observed in the metabolite yields. Good agreement between specific rates obtained in continuous and batch experiments have also been reported (Griffiths et al., 1992; Schmid et al., 1992; Southwick and Mannix, 1992). 18 — a 1 o e H S *- w in t w oo H N n ON |C 8 4) 1 -53 O cj CS VO VO cs o o | ? l , VO cs ?; CO VO o +i in ON 4> '—-1^  •« S M o r-~ <n °° m co t-~ cs cs m 2 e o o o o ^ o\ M 2 M M r H ^ I— § "3 'Is I > C w O O O O X X X X ON ON r-o - H ' CS r^ ' '•8 e o > Vi 3 O g J3 J3 J3 CS TT 00 VO O O O - H O O C/3 O O O O Cfl 3 O c 3 oisnjj ntin oisnjj 'BR o OJ ra U PH 2 oil 2 '£ '•S 1-9 T3 OJ HH JO <u PH 44 B 1(2 OJ ^ >> vg C 00 OJ ON 3 ^ OJ p4 13 « ON </> 00 S -a OJ c w /-^ OJ ON *5 ON PH w 19 l l f l © O B CN cs cn cs VO vo CO cs 00 VO cs uo o cs uo cs wo I uo cs CO O N o uo VO VO 4> 15 •» •5 .2 « I y o cs CO oo od o cs o cs - " s i |5 a s o < I 3 B SJ o B O 1= .2 ^ ° 2 s . O B s 1.1 8 g l> g ^ U X 00 © co o x X cs X o uo o "x cs E i.2 ts -o >• o CO O c 3 0IS ontin 0IS ontin erfu PH 3* O uo g o o °53 o tg N P H >—• O ^ > ' 5 5 uo cS II £ 8 1 co 3 O 3 o U T D OJ e TD o <u T 3 C CJ O X ) ' 5 5 -a ^ J OJ «4H PH * 4> 2 OJ '3 o^  cd O N -a ^ T 3 £ * O N PQ 2 C cs O N OJ co 8 cs O N O N 20 1 s«s C N •g .2 « C N tN 00 C N co C N C O C N '•a ^ O B e i i i l O X vo X X ro X VO X 00 c > o U o PH W CJ a o o '55 C U 3) ^ VO I ? O a -1 ? CO 3 O o u B T 3 a OJ c oo OJ CD °5 5 oo O C N X > ON £ ccj -»-» OJ '3 OJ I C N ON ON ccj $ T D ^ '3 C N S O N •8 2 C/3 s_ CJ 3 s s 8^ 00 ON O N cB v»» 21 l-c I E ® cu S s l l S a. 00 "3-CO +1 ro +1 ON CO VO vd CN +1 o ro S S c o I E 1 to v. .2 " £ =L cd 3 CS 00 o b ON o CN 00 '•8 s I t= cu o cd 3 1 3 . IT) I—H CO b i b i i i ON o ro b b 00 O CN i r i IT) >—1 +1 +1 +1 3±1 ro ir> CN 3±1 r»' o ro ro CN IT) CN VO o CN CN CN 00 CN IT) ro e s o cu '-S cj es cu .t » S3 U o e o o o CO X ON b X IT) CN l l es > i3 CO CO 3 3 / x > X pinne O O •c •c pinne J ank) tinu mat CO 3 mat C/D G c i O I tch ( O o i irred o o 1 rami( ntinu ramie Bai Sei +-» C/D Sei (ce Co (ce CO 3 O o U CU ca o "o K S +-> CD PQ w CN a • s i PQ CO 3 O 3 _ •S CN 1 1 •I 41 E CO cj C/D O S .2 '•3 C cj o U -CJ a e 2 ro ON CJ I-I * ' 3 M-H -a i s 3 13 OJ Tt cj ON OJ S "cd 22 In perfusion mode, slight increases in MAb concentration compared to continuous cultures at the same dilution rate have been reported (Bartley and MacLeod, 1992; Schmid et al., 1992), but the specific production rates were similar (Schmid et al., 1992). In contrast, Griffiths et al. (1992) observed a 2-fold increase in the specific productivity in perfusion compared to chemostat or batch cultures. It was suggested that MAb denaturation by proteases is decreased because the time antibodies are in contact with the cells and medium is reduced. Lower production rates in perfusion than in batch were reported by Krieger-Lassen (1992), but the retention device used might be the reason for this unusual result. Reuveny et al. (1986) studied the effect of four different culture modes on growth and MAb production. Although greater cell and antibody concentrations were achieved in semi-continuous cultures compared to batch, the specific production rates were similar regardless of the metabolic activity of the cells. In those experiments, lactate and ammonium concentrations were found to be relatively constant and independent of the dilution rate. Thus, the presence of these inhibitors cannot explain the variation in cell densities observed. During the perfusion culture, at steady state, the MAb specific production rate was higher than those obtained in batch or semi-continuous cultures. The authors suggested that the cell may be secreting one or several factors which stimulate antibody productivity and, because of the high cell density, these factors are available at stimulatory levels in perfusion culture. However, the higher dilution rates of perfusion cultures should negate such an effect. 23 Al-Rubeai et al. (1992) observed a 2-fold increase in MAb specific productivity in perfusion culture compared to batch and continuous modes. This increase was related to a decrease in cell viability and intracellular antibody content. In their batch experiment, the level of intracellular MAb was shown to have no effect on the specific productivity during the growth phase, but a great influence during the death phase. Based on this observation, the authors suggested that the higher production rates of perfusion cultures were due to the increase in death rate of cells storing large amounts of antibody. Goergen et al. (1992) compared the kinetics of growth, consumption of glucose and glutamine and the production of lactate, ammonium and antibodies in batch and continuous cultures. They observed a 40% lower glucose consumption rate and a 70% higher MAb production rate in continuous culture. However, in latter cultures, cells were exposed to glucose limitations and they were not able to maintain a true steady state in terms of cell and substrate concentrations. It is possible that, under these conditions of stress, changes occurred in the metabolism and physiology of the cells resulting in an increased antibody production. Merten et al. (1994) compared semi-continuous and continuous cultures. However, two different media were used making their analysis questionable. Two batch experiments were carried out using both media to give a basis of comparison. It should also be pointed out that the continuous experiments were performed in a continuous stirred tank reactor whereas the semi-continuous experiments were done in roller bottles. Significantly higher cell densities were obtained in the semi-continuous mode. Although the batch results showed some 24 discrepancies due to the different media, the authors suggested that the mode used was the main reason for the differences observed. According to them, the continuous medium change leads to constant but permanently higher levels of toxic substances in the case of the continuous cultures. The MAb specific productivity was 10-times greater in the continuous culture. The authors have related this result to both the different media and mode of cultivation used. Taken together, these results show that the cell kinetics may be highly variable depending on the cultivation mode. In the context of process development and optimization, the impact of potential differences between each mode, added to the differences among cell lines, needs to be addressed. 25 3 M A T E R I A L A N D M E T H O D S 3.1 Cell line and medium The hybridoma cell line used in this study was TFL-P9. secreting rat anti-mouse IgGi monoclonal antibody and was provided by the Terry Fox Laboratory (Vancouver, BC). The medium used to support cell growth was Dulbecco's Modified Eagle Medium (DMEM 12800 Gibco, BRL, Burlington, ON) supplemented with 5% Fetal Bovine Serum (Sigma). The composition of the medium used is presented in Table 3-1. Table 3-1 Composition of Dulbecco's Modified Eagle Medium. Component Concentration Component Concentration (mg/L) (mg/L) Inorganic salts: Amino acids CaCl2 (anhyd.) 200 L-Arginine« HC1 84 Fe(N0 3 ) 2 «9H 2 0 0.10 L-Cystine« HC1 63 KC1 400 L-Glutamine 584 MgS0 4 (anhyd.) 97.67 Glycine 30 NaCl 6400 L-Histidine HC1«H20 42 NaHC0 3 3700 L-Isoleucine 105 NaH 2 P0 4 « H 2 0 125 L-Leucine 105 Other components: L-Lysine«HCl 146 D-Glucose 4500 L-Methionine- 30 Phenol Red 15 L-Phenylalanine 66 Sodium Pyruvate 110 L-Serine 42 Vitamins L-Threonine 95 D-Ca pantothenate 4 L-Tryptophan 16 Choline Chloride 4 L-Tyrosine*2Na»2H20 104 Folic Acid 4 L-Valine 94 i-Inositol 7.2 Niacinamide 4 Riboflavin 0.4 Thiamine HC1 4 Pyridoxine HC1 4 26 After thawing, cells were expanded and subcultured in T-75 and T-150 flasks, maintained at 37°C in a humid atmosphere of 5% CO2 in an incubator (Forma, Marietta,OH). Experiments were carried out in 250 mL spinner flasks (Bellco Glass, Vineland, NJ). For some cell lines, there are many reports indicating a loss of MAb producing ability when the subculture time increases. For that reason, cells were never maintained more than a month before inoculation. For continuous experiments, a multi-channel pump (Watson Marlow 205S) was used to deliver the small flow rates (50-150 mL/day) required. The retention device used for the perfusion cultures was an acoustic filter BioSep 10L (PSI systems, Coquitlam, BC). Figure 3-1 shows the schematic for the system used in these experiments. As suggested by the user manual, silicon tubing (Cole-Parmer) was used for all the lines. The entire system was made autoclavable for sterilization, after which the portable system was inoculated in a laminar flow hood. A sampling port allowed samples of 5 mL to be aseptically taken on a daily basis. For fed-batch experiments, to avoid the long and challenging task of identifying the limiting substrates, concentrated complete media have been used successfully (Bibila et al., 1994). The lack of a reliable structured model hinders the optimization of important operation parameters such as the frequency and the mode of feeding. For convenience and simplicity, daily additions of concentrate were performed. Concentrate optimization requires significant 27 28 effort and it was not the goal of this study to maximize antibody production, but rather to show the similarities and/or differences between the fed-batch and batch modes. The daily concentrate volume addition was calculated based on the assumption of a constant glucose uptake rate obtained from batch data. Table 3-2 presents the concentrate formulation. Essentially, the various salts, glutamine and glucose were fed at five times their normal concentrations in DMEM. NaCl was not supplemented so that the osmolality would not become too high. Enzymatic soy hydrolysate was used to provide the amino acids. As mentioned earlier, lactate has been shown to have negative effects on growth and production in hybridoma culture. With the addition of glucose in the culture, higher lactate levels might be expected which could possibly affect growth and/or production. In order to determine if lactate effects were significant, two medium formulations were tested for the start of the culture: a low glucose medium (5 mM) and a high glucose medium (25 mM). Table 3-2: Concentrate formulation used for the fed-batch cultures. Component Concentration Glucose 120 mM Glutamine * 3 g/L Peptone (enzymatic hydrolysate)* 5 g/L CaCl2 (anhyd.) lg/L Fe(N0 3 ) 2 «9H 2 0 0.0005 g/L KC1 2 g/L MgS0 4 (anhyd.) 0.49 g/L NaH2PCV H 2 0 0.63 g/L Phenol red 0.075 g/L MEM vitamin 20% (V/V) solution (Gibco)** * in HC1 solution **5% (V/V) in NaOH solution 29 3.2 Analytical Techniques 3.2.1 Cell concentration, viability and sample analysis The total and viable cell concentrations were counted via a haemocytometer. The viable cell population was distinguished from dead cells by the trypan blue exclusion method (1:1 mixture of cell culture and trypan blue in normal saline). The remainder of the cell-containing sample was centrifuged and the supernatant was frozen in aliquots at -20°C for later analysis. Glucose, pH, lactate and urea nitrogen were measured using a clinical NOVA analyzer. 3.2.2 Cell volume Cell numbers and volumes were measured by an Elzone 280 PC electronic particle counter (Particle Data Inc., Elmhurst, EL). The system was calibrated with 10 and 20 u.m diameter latex beads. Samples were diluted in phosphate-buffered saline (PBS). The arithmetic mean cell volume of the populations of cells were calculated from the analyzed size distributions. Duplicate counts were done for each measurement. 3.2.3 Monoclonal antibody assay Antibody concentration was determined by a sandwich ELISA assay in 96-well microtiter plates (Immulon 1, flat bottom, Dynatech Labs). Goat anti-mouse IgG (ICN Biomedicals) was diluted in phosphate buffer and adsorbed to the wells which were then stored overnight at 4°C. The wells were washed 3 times with phosphate-buffered saline and blocked with bovine serum albumin for 2 h. Various dilutions of samples with the proper concentrations of 30 standard antibody (purified mouse IgG from StemCell technologies) were added to the plates and incubated for 1.5 h at room temperature. Subsequently, the plates were washed 3 times and a solution of phosphatase conjugated with rabbit anti-rat IgG (ICN Biomedicals) was added to each well. After 1.5 h of incubation at room temperature, the wells were washed 3 times and a 1 mg/mL p-nitrophenyl-phosphate (PNPP, Sigma) solution in 10% diethanolamine buffer (5 mL diethanolamine, 36 mL distilled water, 5 mg MgCb, 10 mg NaN3, pH adjusted to 9.8 with HCl) was added. The optical density at 405 nm of each well was recorded by an ELISA reader (Vmax Kinetic microplate reader, Molecular Devices). A calibration curve for the standard antibody was prepared for each plate. The linear region of this curve was used to calculate the concentration of MAb in the culture supernatant. The range of best linear response for the conditions described was between 0.01 and 0.2 u.g/mL 3.2.4 Glutamine assay Glutamine concentrations were determined rapidly by a colorimetric glutaminase enzyme assay in microtitre plates. Derivatization of samples with phenylisothiocyanate (PITC) and HPLC analysis was required for quantification of other amino acids or confirmation of glutamine concentrations obtained with the plate assay. Glutamine standards with concentrations of 0.03 to 1 mM were prepared. Samples were diluted in water to fall within the range of standard concentrations. A 96-well microtitre plate assay was separated in two sections. Standards and samples were added in duplicate at a volume of 70 uL/well to each of the two halves of the plate in an identical pattern. A glutaminase solution (108 uL glutaminase, 2160 uL acetate buffer, 1080 uL hydroxylamine and 1155 u.L distilled water) was added to the first half of the plate (50 uL/well). In the second half of the plate, 50 31 uL/well of a solution (2160 uL acetate buffer, 1080 uL hydroxylamine and 1265 uL of water) were added. The plate was then incubated at 37°C for 75 min. Then, a solution (containing 1440 uL (^ -nicotinamide adenine dinucleotide, 144 |iL adenosine 5'-diphosphate, 90 uL glutamate dehydrogenase, 9.6 mL of Tris/Hydrazine buffer and 3.1 mL distilled water) was added at a volume of 120 uL /well and the plate was incubated for 30 min at 37°C. The absorbance of each well at 340 nm was measured using a T H E R M O m a x plate reader. The increase in absorbance in the first half of the plate was due to the presence of both glutamine and glutamate in the sample. The second half did not contain glutaminase and the absorbance increase was due to the presence of glutamate. Thus, to obtain glutamine concentrations, the absorbance for each well in the second half of the plate was subtracted from the corresponding well in the first half. 3.2.5 Amino acids analysis The Pico-Tag system developed by Waters (Millipore) was used to determine the concentrations. The method consists of deproteination of samples, followed by the formation of a phenylthiocarbamyl (PTC) derivative of the amino acids via reaction with phenylisothiocyanate (PITC). The derivatized amino acids undergo liquid chromatography separation on a bonded phase column, followed by ultraviolet detection at 254 nm. Deproteinization of samples was accomplished with Millipore Ultra Free cartridges (10,000 NMWCO). The supernatant was vacuumed dry, then 25 uL of redry solution (40% methanol, 40% distilled water, 20% triethylamine) was added and the samples were vacuum dried again. Finally, 20 uL of derivatizing solution (70% methanol, 10% distilled water, 10% 32 triethylamine, 10% phenylisothiocyanate) were added. The initial drying step removes solvents and volatile components such as HCl. The redry step neutralizes any residual acid that may cling to the glass tube, and the coupling step makes the PITC (phenylthiocarbamyl) derivatives that are actually analyzed. After derivatization, the samples were either stored dry at -20°C (stable for 3 weeks) or 100 uL of diluent was added (95% 5 mM sodium phosphate dibasic, titrated to pH 7.5 with 10% phosphoric acid and 5% acetonitrile). The samples were then transferred to autoinjector vials sealed with Teflon-lined closures and analyzed within 24 h. A Shimadzu HPLC was used with a Waters Sentry guard column and a Waters 30 cm Nova Pak C18 reverse phase HPLC column. 10 p i of sample were injected into the column for analysis. The elution buffers used for the HPLC run were as follows: eluent A: 70 mM sodium acetate, pH 6.5 and 2.5% acetonitrile; eluent B: 45% acetonitrile, 40% H 2 0 and 15% MeOH. L-norleucine was used as an internal standard. It should be pointed out that Cystine/cysteine cannot be accurately quantified with this method. 33 4 Batch Modeling Optimization of the performance and productivity of hybridoma cell cultivation systems and bioreactor control depend on an understanding of the cell metabolism and antibody production. Therefore, models describing the kinetics of cell growth and antibody production are needed. Such models can be broadly divided into two categories: structured and unstructured models. Unstructured models quantify cell mass as a single component whereas structured models break the cell mass into distinct subcomponents. Structured models may be attractive due to their general applicability, but such models are difficult to build due to unknown and complicated mechanisms. They also require a large number of parameters which can be difficult to identify or estimate. So far, most of the attempts to model hybridoma growth and antibody production have used unstructured models with only a few exceptions (Batt and Kompala, 1989; Suzuki and Ollis, 1990; Bibila and Flickinger, 1991) Also, because the kinetics can vary greatly from one cell line to another, most of the models found in the literature cannot be generally applied (Portner and Schafer, 1996). In the present work, the approach followed was to develop a simple unstructured model based on the kinetics obtained from batch data. This provided a suitable basis for the simulation of hybridoma growth and MAb production in continuous cultures. 34 Mass balances around a batch reactor yield the following equations for the concentrations of viable cells (X), substrates (S), and products (P): dX = yX-aX (4.1) dt dS_ dt = -qsX (4.2) (4.3) where u is the specific growth rate (h"1) and a the death rate (h') of the cells, qs the specific substrate uptake rate (mmol/cell.h) and qP the specific production rate (ug/cell.h). This model structure is valid for any biological system grown in batch mode. The rates ju, cc, qs and qp are dependent on growth conditions and some correlations are discussed in the next sections. 4.1 Correlations for the growth rate A number of factors have been reported to affect the growth of animal cells in culture, including physiochemical parameters (e.g. temperature, pH, osmolarity), levels of nutrients (glucose, glutamine), growth factors (insulin) and the formation of toxic metabolites (ammonium, lactic acid). Unstructured Monod-type models have been widely applied to animal cell systems to describe the growth rates of animal cells as functions of substrate concentrations. Glucose and glutamine are the main carbon and nitrogen sources and can be analyzed easily. Therefore kinetic growth models are often based on the concentrations of these compounds (Table 4-1). In some cases, models have been extended to include 35 limitations due to inhibitory metabolites (Omasa et al., 1992; Ozturk et al., 1992; Kurokawa et al., 1994; Pelletier et al., 1994). The following general form from the hybridoma literature has been proposed by Miller et al. (1988): max Glu KGlu+Glu Gin KG[n+Gln KLac KLac+Lac. K Amm K Amm •Amm (4.4) Glu, Gin, Lac and Amm denote the glucose, glutamine, lactate and ammonium concentrations (mM), respectively, while the K{ are the corresponding saturation constants (mM). Most correlations found in the literature are simplifications of equation (4.4). It is also remarkable that, although each study was done on a different cell line, most of the kinetic constants in Table 4-1 fall within a relatively narrow range. The serum has a major influence on the growth rate and was included in the Glacken et al. (1988) model. A deviation from the Monod-type kinetics was observed in the data of Frame and Hu (1991) at very low growth rates. They modified the Monod model to include a threshold substrate concentration. Other recent studies raised the possibility that the transition of the cells from growth to death might be due to an unknown inhibiting factor (Lee et al., 1995; Sanfeliu et al., 1996; Zeng et al., 1998). 36 Table 4-1: Correlations for the growth rate Reference Correlation Glacken et al. (1988) MmaxSerUmGl" / W = 0.061 h"1 KGln = 0.15 mM Kser=\.6% =12000 mM 2 KAmm = 45 mM 2 M - -[Serum + K^jGln+KsJ Miller et al. (1988) ( Glu Y Gin Y Y Kj^, \ = 0.0625 h"1 Kaiu = 2 mM KGi„ = 0.44 mM £ L f l C =103 mM KAmm=\2 mM m\Kail + Glu\K0b, + Gln^K^ +Lac + Amm) Frame and Hu (1991) P = 0 S<S, umin + ( M m a x ~ M m i n ) { S ~ S , ) S>S, Ks ^ = 0.013 h"1 / W = 0.053 h"1 S, = 0.03 mM Ks = 0.025 mM Omasa et al. (1992) / V = Vmax V KLOC +Lac) //max = n/a KLac = 9.4 mM de Tremblay et al. (1992) f Glu ^ \KGlu +Glu) ' Gin \ ^Gin+Gln) / W = 0.047 h"1 Kaiu = 1 mM Koin = 0.3 mM Ozturk et al. (1992) Ka,+[Arf "•Amm fio = 0.037 h"1 KAmm = 24 mM 2 Gaertner et al. (1993) 0 .043 S - 0 0 7 B B: base medium concentration (empirical correlation) Pelletier et al. (1994) ' Gin Y KLac Koin+GlnXK^+Lac^ \( r \ A-Amm {.KAmm +Ammy / W = 0.065 h"1 KGin : 0.3 mM KLOC : 55 mM KAmm : 5 mM Kurokawa et al. (1994) Glu KGlu + Glu + Lac / W = 0.033 h"1 Kaiu = 0.3 mM KLOC =14 mM Jeong et al. (1995) Gin KGl„+Gln = 0.0334 h"1 KGin = 0.089 mM Portner et al. (1996) Gin KGln+Gln /An« = 0.036 h"1 Koin = 0.06 mM Zengetal. (1998) (, XV Glu ) V P){KGlu+Glu) f Gin \ KKGln +Gln^ / U * = 0.072 h"1 <*= 0.009 L/109cells.h KG,U = 0.29 mM Koi„ = 0.09 mM 37 In the present work, the following correlation was used: \ X«>J (4.5) where X„ denotes the maximum cell concentration or as it is also referred to "the carrying capacity" of the system. Once integrated, this correlation yields the logistic equation (an s-shape curve) often encountered in biological processes. The use of this correlation requires the a priori knowledge of the maximum cell concentration which must be determined experimentally. But, unlike most of the aforementioned correlations, using the logistic equation circumvents the need to identify the limiting or inhibiting factor. Because of a lack of reliable measurements of nonviable cell concentrations, it is difficult to obtain a death rate. Moreover, death rate only becomes significant during the later stage of a culture, normally avoided in a bioreactor. For these reasons, the death rate was neglected from the model (Equation 4.1). Implications of this assumptions will be further discussed later (Section 5.3.3). 38 4.2 Correlations for glucose and glutamine uptake rates In mammalian cell culture, the glucose uptake rate is often described as a function of the growth rate (Table 4-2). The relationship between the two is typically assumed to follow the maintenance-energy model: qs=-^- + m (4.6) with YX/s denoting the yield coefficient of cells on substrate (cell/mmol) and m the maintenance term (mmol/cell.h). In some cases, the maintenance energy model was found to be valid only at low growth rates. Hiller et al. (1993), Robinson et al. (1991) and Harigae et al. (1994) all found that the specific glucose uptake rate increased strongly and ceased to be linear at higher growth rates. The correlation proposed by Frame and Hu (1991) takes into account that observed non-linearity. Saturation-type models have also been proposed in which the uptake is expressed as a function of the substrate concentration (de Tremblay et al., 1992; Gaertner and Dhurjati, 1993; Zeng et al., 1998). Like glucose, glutamine is often assumed to follow the maintenance-energy model (Table 4-3). 39 Table 4-2: correlations for the glucose uptake rate. Reference Correlation Yx/s (108 cells mmol"1) m (IO"10 mmol cell"1 h-1) Miller et al. (1988) M 2.0 0.5 Hiller et al. (1991) -J—- + m YXIS 1.02 Linardos et al. (1991) 5.93 1.96 Harigae et al. (1994) 5.88 0.51 Kurokawa et al. (1994) 0.25 0.27 Frame and Hu (1991) _ J P_ + « /'min [Jxis Ypis\ UPIS YXIS. Anin (h'1) = 0.013 B(mg Ab/mg cells) = 0.027 a(mg Ab/mg cells.h) = 0.0067 Yx/s (mg cells/ mg glucose) = 0.36 Yp/s = (mg Ab/ mg glucose) = 0.408 de Tremblay et al. (1992) YX/S K g + S Tx/s (108 cells mmol"1) = 1.09 mg (10"10 mmol cell"1 h"1) =0.78 £g(10- 1 0 mmolceU-1)= 19 Gaertner et al. (1993) aS b + S a(10-10 mmol ceir1h"1) = 2.7 b (mM) = 0.34 Zeng et al. (1998) u. aS —— + m + YXIS Xvb + S Yx/s (108 cells mmol"1) = 1.71 m (IO-10 mmolceU^h"1) = 0.86 a (IO-10 mmol celf1 h"1) = 0.94 i(10'10mmolceir1) = 2.9 Table 4-3: Correlations for the glutamine uptake rate. Reference Correlation Tx>y (10 8 cells mmoll) in (10 1 0 mmol cell1 h"1) Miller et al. (1988) M + m Yx/s 6.3 Hiller et al. (1991) 16.9 1.56 Linardos et al. (1991) 6.30 0.29 de Tremblay et al. (1992) 3.80 Harigae et al. (1994) 3.44 0.08 Pelletier et al. (1994) 2.63 Zeng et al. (1998) u aS -J-— + m + YXIS X v h + YX/S(10* cells mmol"1) = 9.5 m (IO"10 mmol cell"1 h"1) = 0.093 a (IO"10 mmol cell"1 h_1) = 0.33 Z>(10'10mmolcell"1) = 0.89 40 Maintenance (iri) is often assumed to have a negligible effect on the consumption term (Hiller et al., 1991). Thus, in the present study, the following model structure was used to describe both glucose and glutamine consumption rates: (4-7) *x/s 4.3 Correlation for the production rate The productivity of antibody is usually expressed as a function of the specific growth rate with a non-growth-associated term (f) and a growth-associated term (JJ). Depending on the cell line used, the growth-associated term can be positive, negative or neglected. qp = y+8M (4.8) Linardos et al. (1991) instead expressed the production rate as a function of the death rates, a. It can be important to consider the instability of MAb productivity, especially for long term cultures (Morrison et al., 1974; Frame and Hu, 1990; Zeng, 1996). The model proposed by Zeng et al. (1996) takes into account the productivity loss (Table 4-4). 41 Table 4-4: Correlations for the antibody production rate. Reference Correlation r (mg/1010cells.h) P (mg/109cells) Frame and Hu (1991) PH+Y 0.00672* -0.027** Hiller et al. (1993) 7.1 0 Pelletier et al. (1994) 0 18 Portner et al. (1996) 0.17 0 Linardos et al. (1991) Aa + B A,B: constants a. death rate Zeng et al. (1996) qP=(r + Ba)(B + e-AAt) A,B: constants At: period of time after time zero. a(mg/109cells.h)= 1.54 Gaertner et al. (1993) (B - 0.07)(9.65 + 5.2L + 296) A2 5(1+ —)(£ +57) 48 B: base medium concentration L: lactate concentration (pg/cell.h) S: serum concentration (pg/cell.h) A: ammonium concentration (mM) de Tremblay et al. (1992) a M Ku+ju a(mg 10"8 cells d_1) = 2.6 Ku(d'1) = 0.001 *(mg Ab/mg cells.h) **(mg Ab/mg cells) In this work, the growth-associated term for the antibody correlation (B in Equation 4.8) was neglected since many studies have reported constant production at different growth rates (Hiller et al., 1993; Portner and Schafer, 1996). Thus, the following equation was used for the specific productivity: qP = 7 (4-9) In summary, with the model structure assumed, the following parameters needed to be identified from batch experiments: the maximum growth rate (//max), the maximum cell density (Xmax), the yields of cells on glucose and on glutamine (Yx/s) and the MAb specific productivity (f). 42 5 RESULTS AND DISCUSSION In this chapter, the kinetics of hybridoma growth and antibody production in five different culture modes are presented and discussed. First, the effects of cell volume variations and inoculum state on the rates are assessed. Second, based on batch data, a kinetic model is developed and compared to experimental data from semi-continuous and continuous cultures. Then, fed-batch and perfusions cultures results are presented and, finally, the rates of nutrient uptake and MAb formation are compared for all modes of cultivation. 5.7 Cell Volume Because this work is primarily concerned with cell specific rates, it is important to consider how changes in cell might affect these rates. In the batch cultures, the hybridoma mean cell volumes remained roughly constant at 1100 um3 during the early exponential phase of cell growth (Figure 5-1), then the volumes decreased during the late exponential, stationary and decline phases. Since only one peak was observed in the cell volume distribution, it was not possible to discriminate between the volume of viable cells and the volume of nonviable cells. As shown on Figure 5-1, the decline in mean cell volume correlated closely with the decline in cell viability. Figure 5-2 shows both the cell and volume specific glucose uptake rates for a batch experiment. These rates are obtained by differentiating the glucose concentrations with respect to time, and dividing by the viable cell concentrations or cell volumes. During the 43 exponential phase, the volume specific uptake remained relatively constant whereas the cell specific uptake rates decreased more rapidly. This is an indication that at least some of the variation observed in cell specific uptake rate was the result of cell volume decreases resulting in decreased nutrient requirements. The effects of volume changes on the metabolic rates can be expected to be more pronounced in late batch and fed-batch cultures, whereas the more homogenous conditions found in continuous systems should yield more consistent cell volumes. During the course of a batch experiment, the culture medium osmotic strength increases (Ozturk and Palsson, 1991) mainly due to the conversion of substrate to multiple products (e.g. Glucose —> 2 lactate). Thus, the hybridoma cell volume measurements performed in 300 mOsM PBS may have overestimated the culture cell volumes (Figure 5-1) at later stages of batch cultures. 1600 Time (h) Figure 5-1: Mean cell volume (open circles) and viability (filled triangles) profiles in batch culture. A indicates the early exponential phase, B the exponential phase, C the stationary phase and D the decline phase. Error bars for cell volumes are standard deviations from duplicate counts. A conservative estimate of error was applied to viability data (20 %). 44 Time (h) Figure 5-2: Volume specific (filled circles) and cell specific (open circles) glucose uptake rates (GUR) profiles in batch culture. Errors were calculated assuming a 10% error on cell concentration measurements and 0.5 mM error on glucose measurements. Sonderhoff et al. (1992) also observed that mean cell volumes reach maximum values during exponential growth and decrease during the stationary and decline phases. In perfusion experiments, Hiller et al. (1993) found average steady-state cell diameters were larger at higher growth rates. However, for growth rates ranging from 0.08 to 1.2 h"1 the diameter only varied from 13.5 to 13.9 um. Ramirez et al. (1990) obtained a similar batch profile and they used the cell volume as an indicator of cell cycle stage changes. They have also observed that the specific oxygen uptake rate closely followed the changes in cell volume. 45 5.2 Inoculum state To determine how critical the state of the inoculum is for obtaining reproducible results, the effects of inoculum "phase" and concentration have been examined. 5.2.1 Phase of inoculum Cells taken at 27, 53, 66 and 79 h (i.e. early exponential phase, mid-exponential phase, late exponential phase and stationary phase) of a batch culture were centrifuged and used as inocula for four batch cultures. Batch cultures inoculated by later phase inocula had longer lag phase (Figure 5-3) while the maximum cell and antibody concentrations decreased (Table 5-1). However, the maximum growth rates were relatively unaffected by the phase of the inoculum. Table 5-1: Effect of inoculum age. Errors for antibody concentrations are standard deviations — «« • ~ — • „ io irw. Inoculum state Maximum viable cell concentration Maximum growth rate Final MAb concentration Early exponential phase (t = 27 h) 2.9 x 106 cells/mL 0.057 h"1 110 ± 17 (ig /mL Mid exponential phase (t = 53 h) 2.8 x IO6 cells/mL 0.058 h"1 96 ± 14 (Ig/mL Late exponential phase (t = 66 h) 2.6 x 106 cells/mL 0.054 h"1 81 ± 12 (Ig/mL Stationary phase (t = 79 h) 2.4 x 106 cells/mL 0.054 h"1 77 ± 12 (Ig/mL Martial et al. (1991) also reported that increasing the age of the inoculum resulted in a longer lag phase, a lower maximum specific growth rate and a reduced maximal cell density. However, antibody concentrations did not vary significantly suggesting that a higher specific production rate was obtained when inoculating with stationary or decline phase cells. 46 1.0E+07 1 .OE+04 -I 1 1 1 0 50 100 150 Time (h) Figure 5-3: Effect of the inoculum phase on cell growth. Cultures were started with inoculum taken from a batch at 27 (filled triangles), 53 (open circles), 66 (filled squares) and 79 h (open diamonds). A conservative estimate of error (10%) was assumed for cell concentration measurements. 47 5.2.2 Inoculum size Cultures were started at 1.5 and 2x10s cells/mL to investigate the effect of the inoculum size on growth and production. Lower inoculum concentrations were also tested but resulted in a prolonged lag (105 cells/mL) or culture failure (<5xl04 cells/mL). At 1.5 and 2xl05 cells/mL there were no significant effects on maximum growth rates or cell concentrations (Table 5-2). The slight increase in final antibody concentrations is related to the total number of viable cells, as will be discussed later in section 5.7. Table 5-2: Inoculum size effect. Errors for antibody concentrations are standard deviations based on duplicate measurements. Error on cell concentration measurements is 10%-Inoculum concentration (cells/mL) Maximum cell concentration (cells/mL) Maximum growth rate (h1) Final MAb concentration 1.5 x 105 2.9 x 10b 0.061 89 ± 13 Ug/mL 2x 10s ' 2.9 xlO 6 0.059 110 ± 15 pig/mL Dutton et al. (1999) reported that varying the initial cell concentration from 1 to 3x103 cells/mL did not significantly affect the maximum antibody production or the maximum cell concentration. Since the goal of this study was ultimately to compare the kinetics of different culture modes, the cultures were all initiated with 2 x 105 cells/mL of mid-exponential phase cells in order to reduce potential inoculum variations effects. 48 5.3 Batch cultures 5.3.1 Batch results Figure 5-4 shows the cell concentrations and viability for a typical batch experiment. Since no lag phase was observed, the cells grew exponentially from the start. The viability was maintained at over 90% throughout the exponential phase. A maximum cell concentration of 2.4 xlO6 cells/mL was reached. The cell concentration then started a decline phase where the cell viability decreased below 60%. 1.0E+07 Time (h) Figure 5-4: Cell concentration (filled circles) and cell viability (open triangles) profiles in batch culture. Errors were calculated assuming a conservative estimate of error (10%) on cell concentration measurements and 20 % in viability measurements. The exponential phase was characterized by rapid glucose uptake until the stationary phase (Figure 5-5). Glucose was not limiting this culture as the concentrations leveled off at around 49 6.5 mM. The lactate profile was almost the mirror image of the glucose concentrations. During the exponential phase, the average yield of lactate on glucose was 1.35 ± 0.33 mol/mol. Unlike glucose, the glutamine (Figure 5-6) was depleted before the end of the culture and the point at which the glutamine concentration reached zero coincides with the time at which growth stopped. Thus, glutamine was probably the limiting substrate for this culture, although many other factors could have limited cell growth, such as the accumulation of a toxic metabolite or the depletion of a serum component. The average yield of ammonium on glutamine was 0.42 ± 0.21 mol/mol during the exponential phase. Both lactate (Figure 5-5) and ammonium (Figure 5-6) final titers were below the values reported in the literature to be growth inhibiting (Glacken, 1987; Ozturk et al., 1992). The pH profile (data not shown) decreased throughout the culture but remained within the range in which there should be no detrimental effect (7.4 to 6.8). Time (h) Figure 5-5: Glucose (filled circles) and lactate (open diamonds) concentrations in batch culture. The measurement errors for glucose and lactate are 0.5 mM. 50 50 100 Time (h) 150 Figure 5-6: Glutamine (open diamonds) and ammonium (filled circles) concentration profiles in batch culture. The measurement error for ammonium is 0.5 mM. The MAb concentration (Figure 5-7) initially increased steadily and finally reached a maximum value of around 100 ug/mL which remained constant until the end of the culture. Unlike in reports for other hybridoma cell lines, the antibody production by the TFL-P9 took place mainly during the growth phase. o '*-> CO k_ "c 0) O CZ o o >, T3 O .O 140 120 100 80 60 40 20 — i — 50 100 150 Time (h) Figure 5-7: Monoclonal antibody concentration (filled circles) profile for batch culture. Error bars are standard deviations from two separate assays. 51 In order to further investigate the possible limitation by a medium component, an amino acid analysis was performed. Table 5-3 presents the results obtained for 18 amino acids. The analysis revealed that, except glutamine, none of the other amino acids were completely depleted during the culture. It also showed that glutamate, glycine, proline and alanine were produced whereas the concentrations of histidine, methionine and tryptophan declined little in the culture. Table 5-3: Amino acids consumption/production rates in batch culture. Negative values indicate metabolite consumption, whereas positive values indical e metabolite production. Amino Acid Initial concentration (mM) Final concentration (mM) Production/consumption rate (pmol/cells.d) Alanine 0.00 1.42 0.761 Arginine 0.43 0.28 -0.131 Asparagine 0.00 0.00 0.000 Aspartate 0 0 0.000 Cystine not measured not measured n/a Glutamate 0.07 0.68 0.220 Glycine 0.44 0.99 0.021 Histidine 0.20 0.17 -0.049 Isoleucine 0.68 0.17 -0.275 Leucine 0.83 0.23 -0.256 Lysine not measured not measured n/a Methionine 0.31 0.28 -0.063 Phenylalanine 0.40 0.27 -0.084 Proline 0.00 0.28 0.104 Serine 0.41 0.17 -0.102 Threonine 0.83 0.71 -0.227 Tryptophan 0.09 0.08 -0.003 Tyrosine 0.38 0.30 -0.110 Valine 0.86 0.41 -0.257 52 5.3.2 Batch model parameters Table 5-4 presents the values assumed for the different parameters of the batch model presented in section 4. They were obtained by doing a non-linear regression on data points from four batch cultures, considering only the exponential phase. Table 5-4: Parameters value assumed for the simulations. The errors are standard deviations. Parameter Maximum cell concentration (Xmax) Value (2.9±0.3) xlO6 cells/mL Maximum growth rate (/w) (0.063±0.012) h"1 Yield of biomass on glucose (YX/s) (1.7± 0.5)xl08 cells/mmol Yield of biomass on glutamine (YX/s) (5.7± 2.1)xl08 cells/mmol Cell specific production rate (qP) (1.5±0.4)pg/cell.h 5.3.3 Batch simulations Using Equations 4.1 to 4.3, simulations of cell growth, glucose consumption and antibody production were performed. The results are presented in Figures 5-8 to 5-10 together with the experimental data from a batch culture. There was a good agreement up to the time when the stationary phase began. Past this point, the effect of cell death, not taken into account in the model, becomes significant (Figure 5-8). During the exponential phase, the good glucose and glutamine predictions (Figure 5-9) confirm the validity of assuming constant yields of cells on substrates. However, towards the end of the culture, it becomes evident that a growth-associated MAb production rate model is needed to extend the range of the prediction. 53 1.0E+07 cu o c o co 5 1.0E+06 o o "CD o o .O co > 1.0BO5 50 100 Time(h) Figure 5-8: Simulation of cell concentration profile. Circles represent experimental points from a batch culture and the model prediction is given by the line. A conservative estimate of error (10%) was assumed for cell concentration measurements. 25 | 20 c o 15 cu o a o o 0J CO o o CD 10 o f \ • \ --\ * \ . • • • --\° \ ° o -1 —O-O— 50 100 Time (h) 4- 3.00 4.00 c o CO L. —^' c 2.00 § o o CD C E CO -*—' 1.00 0.00 150 Figure 5-9: Simulation of glucose and glutamine concentration profiles. Circles represent experimental data points from a batch culture and the model prediction is given by the line. The measurement error for glucose concentrations is 0.5 mM. 54 200 50 100 150 Time (h) Figure 5-10: Simulation of monoclonal antibody profile. Circles represent experimental data points from a batch culture and the model prediction is given by the line. Error bars are standard deviations from duplicate measurements. 5.4 Semi-Continuous cultures 5.4.1 Semi-continuous results Semi-continuous cultures were also performed by harvesting then replacing a fixed volume on a daily basis. During the semi-continuous operation, the conditions in the reactors varied within a certain range. Different pseudo-steady state conditions were obtained by varying the volume exchanged. The "average dilution rate" is given by: DavJh-1) avg Volume changed per day (Volume of reactor)( 24 h) (5.1) 55 Semi-continuous operation was typically initiated after the cells were cultivated in batch mode for around 2 days. Each set of results presented in Table 5-5 was obtained from a separate bioreactor run. Despite some scatter, the general trends observed were increases in both specific consumption rates with increasing dilution rates (the probability that a constant relationship better describes the data is p < 0.05 for glutamine uptake, and p < 0.0004 for glucose uptake). However, no clear dependence on the dilution rate was observed for the MAb production rate and for the cell, antibody, lactate, ammonium or glucose concentrations. The highest dilution rate (0.035 h"1), which corresponds to a replacement of 83 % of the reactor volume, was characterized by both low cell and antibody concentrations as a result of the wash-out phenomenon. As in the batch operation, glutamine appears to have been limiting for cultures at dilution rates lower than 0.021 h"1, although the scatter in the data does not allow clear observations of any resulting metabolic shift. Since no trends were observed for the ammonium and lactate concentrations, their average yields appeared to be independent of the growth rate. 56 UT3 o 00 o fN o o cu ee fN O e o u > 0\ © o 00 1—1 © o o o o o © © CN u-> CN CN CN 00 VO o CN CN e o 5 £ O CU u u "S o cn CN O CN 00 cn Ov vo e o e cu e o u _ "O ^ © M « 1 Os cn CN CN 00 CN Ov 00 00 00° CN 00 CN © e cu w e o u CU o 00 o Ov VO cn o co "5 co *«3 cO co © o c o • PM 2 c cu u B o u E cs * H s 5 cn d Ov d cn d ov d d d cn d CN 00 o 2 Da W S =3 CU u cn -J? u 2 2 S O 3j vo VO Ov Ov d 1 CO CO *i3 CO 13 ea cu cs •«-> s JS CU CA •E *cu 5 E O 3\ IT) Ov oo CN 00 Ov o cs VO d VO ob CN i—i CN O CN CN CN CN VO d CN Ov O CN OV c es •a •M c cu w c O cj cu « e CN CN cn CN CN VO CN VO CN cn CN cn CN CN C E-2 3 « u § c E 3 E c VO VO CN IT) CN c n Ov vo oo cn r l -cn O CN Ov wo II' £ E o « •a o 1 a s: 57 As in the batch mode, most of the amino acids were consumed (Table 5-6) but glutamine was the only component completely depleted, especially at low dilution rates. Glutamate, alanine, proline and methionine were produced whereas histidine and tryptophan concentrations remained more or less constant. These results are quite similar to those obtained from the batch experiments. Most consumption and production rates (Table 5-7) increased with increasing dilution rates but the values were generally lower than those encountered in batch. Table 5-6: Amino acid concentrations (mM) in semi-continuous culture. The asparagine and glycine peaks were merged on the spectrum due to poor resolution and so the combined concentrations were estimated. Amino Acid Concentration in D M E M Average dilution rate 0.01 h 1 0.021 h 1 0.028 h 1 0.035 h 1 Alanine 0 1.61 1.41 1.21 0.67 Arginine 0.38 0.14 0.13 0.20 0.33 Asp + Gly 0.38 0.39 0.27 0.30 not measured Asparagine 0 not measured not measured not measured 0.09 Aspartate 0 0.00 0.00 0.00 0.00 Cystine 0.23 not measured not measured not measured not measured Glutamate 0 0.40 0.25 0.33 0.19 Glycine 0.38 not measured not measured not measured 0.32 Histidine 0.19 0.13 0.16 0.25 0.19 Isoleucine 0.76 0.14 0.18 0.23 0.21 Leucine 0.76 0.05 0.28 0.35 0.59 Methionine 0.19 0.58 0.81 1.01 0.53 Proline 0 0.19 0.19 0.17 0.20 Serine 0.38 0.31 0.24 0.29 0.21 Threonine 0.76 0.52 0.56 0.64 0.52 Tyrosine 0.44 0.21 0.22 0.25 0.21 Valine 0.76 0.27 0.40 0.45 0.40 Phenylalanine 0.38 0.22 0.22 0.23 0.21 Tryptophan 0.076 0.05 0.05 0.09 0.05 Lysine 0.76 0.42 0.41 0.42 0.37 58 Table 5-7: Amino acid consumption/production rates (pmol/cell.d) for semi-continuous culture. Negative values indicate metabolite consumption, whereas positive values indicate Amino Acid Average dilution rate 0.01 h 1 0.021 h 1 0.028 h 1 0.035 h 1 Alanine 0.185 0.278 0.383 0.409 Arginine -0.028 -0.050 -0.056 -0.028 Asp + Gly 0.002 -0.022 -0.025 n/a Asparagine n/a n/a n/a 0.054 Aspartate 0.000 0.000 0.000 0.000 Cystine n/a n/a n/a n/a Glutamate 0.046 0.050 0.105 0.116 Glycine n/a n/a n/a -0.037 Histidine -0.007 -0.005 0.020 -0.002 Isoleucine -0.072 -0.115 -0.170 -0.335 Leucine -0.082 -0.095 -0.130 -0.106 Lysine -0.039 -0.069 -0.108 -0.238 Methionine 0.044 0.123 0.259 0.209 Phenylalanine -0.019 -0.032 -0.048 -0.105 Proline 0.022 0.037 0.054 0.123 Serine -0.008 -0.029 -0.027 -0.106 Threonine -0.027 -0.040 -0.037 -0.146 Tryptophan -0.003 -0.004 0.003 -0.014 Tyrosine -0.026 -0.043 -0.059 -0.139 Valine -0.057 -0.071 -0.099 -0.220 59 5.4.2 Simulation results Figures 5-11 to 5-13 show the simulation results obtained when applying the batch model (Equations 4.1 - 4.3) to the semi-continuous culture data. None of the parameters were adjusted to get a better fit since the purpose was to highlight the similarities and differences between batch and semi-continuous operations. At each harvest time, two concentrations are shown corresponding to the values before and after fresh medium was added. For clarity, only the experimental cell and glucose concentrations prior to medium exchange are shown. Thus, the predictions should be compared with the top and the bottom portions of the saw-tooth profiles for cells and glucose, respectively. Despite some scatter in the cell concentration data (Figure 5-11), on average the model provides a relatively good description of cell growth. However, some discrepancies appear in the glucose concentration profile (Figure 5-12). The model most often predicts a higher pseudo-steady-state glucose concentration. This indicated that the specific consumption rate of glucose was greater in semi-continuous than batch culture. This will be further demonstrated in section 5.8.1. The variability of the cell and glucose data may be attributable to variations in the fresh medium additions (differences in the pH, temperature and even composition since some key components are degraded over time). The antibody pseudo-steady-state concentration predicted by the model is significantly higher than the experimental results (Figure 5-13). Such discrepancy may partly be explained by the great error (26%) associated with the constant specific productivity obtained from batch data. 60 1.0E+07 1.0E+05 200 400 Time(h) 600 800 Figure 5-11: Simulation of cell concentration profile in semi-continuous mode (D = 0.019 h"1). Filled circles represent experimental data points prior to medium exchange. These points correspond to the top part of the saw-tooth profile predicted by the model (solid line). 200 400 Time(h) 600 800 Figure 5-12: Simulation of glucose concentration profile in semi-continuous mode (D = 0.019 h"1). Filled circles represent experimental data points prior to medium exchange. These points correspond to the bottom part of the saw-tooth profile predicted by the model (solid line). 61 165 200 400 600 800 Time(h) Figure 5-13: Simulation of antibody profile in semi-continuous culture (D = 0.019 h"1). Filled circles represent experimental data points prior to media exchange. These points correspond to the top part of the saw-tooth profile predicted by the model (solid line). Error bars are standard deviations from duplicate measurements. 5.5 Continuous cultures 5.5.1 Continuous results In order to obtain true steady-state data, cultures with continuous feeding and medium removal were performed (Table 5-8). Feeding of fresh medium was usually started after 2 days when the glucose concentration was around 15 mM, insuring that the cells remained in the exponential phase. 62 CO H - l es u a B 00 t N O 00 CN o © o «N o i-H © VO 1—I o © o X cs o X O N CN O "x o *x O N O i - H X CN o X VO CN X 00 CN °H-» 2 e eu . « CN CN CN CN VO VO I T ) VO 00 es fc. a C U u e o C J >> O M •s s £ "©3D < 3 r o od Ov O N 00 O N CN O N 00 b e © es u c cu ej e o C J C U cn O C J i l l cd cd cd VO o b cd o es fc. e O J C J C © C J C U e e CS r o vd oo O N ON r o od co vd oo s o H - l es fc. ** e «u C J C o cj cu tJ £ es CN CN CN b oo CN CN CN CN CN CN CN "3-CN 1—H b o CN b c C cj E s •< CJ cj es fc. es £ -8 S o 3 S 0 cd ed cd *c3 cd o in CN 00 o ON r o ON ON o oo r o b <-! eu +J es fc. eu es C U , s . CJ c : S cs H-l S 3 CJ "© s C J CJ H - l es fc. e o • mm H - l C J 3 © CN oo o r o CN CN 00 r o vo IT) O N ir> ej cs eu C A © C J 63 No clear dependence on the dilution rate was observed for both cell concentrations and glucose consumption rates. Moreover, three distinct experiments run under the same conditions (0.028 h"1) had considerable differences in the steady-state cell concentrations. For similar growth rates, steady states exhibiting different cell densities and metabolic rates were also reported by Follstad et al. (1999). The antibody concentrations were significantly higher at low dilution rates (p < 0.01), while the cell specific production rates decreased at low dilution rates (p < 0.06). Both the glutamine concentrations and consumption rates increased with increasing dilution rates, but more data would be needed to assess the significance level of these trends. The glucose concentrations were relatively constant over the range of dilution rates studied. Only the highest dilution rate had a significant increase due to washout. At low dilution rates, glutamine was almost depleted which is consistent with the semi-continuous results. The yields of lactate on glucose were generally higher at greater dilution rates (p < 0.03). Such decreases in the efficiency of glucose utilization have also been reported by Miller et al. (1988). Over the limited data range available, the yield of ammonium on glutamine appears to be relatively constant and independent of the growth rate. The steady-state ammonium concentrations were slightly lower at high dilution rates (p < 0.0004) but no such trend was evident for the lactate concentrations. 64 Similarly to the semi-continuous results, increased production or consumption rates of amino acids were obtained at higher dilution rates and only glutamine appears to be limiting (Tables 5-9 and 5-10). Glutamate and alanine, which are formed from glutamine metabolism, were produced whereas glutamine, isoleucine and leucine were consumed at the highest rate. These results are similar to those reported by Hiller et al. (1991) who, in a serum free continuous culture, have observed that glutamine, isoleucine, leucine and valine were consumed at the highest rates and were almost depleted at low dilution rates. Table 5-9: Amino acid concentrations (mM) in continuous culture. When two peaks were merged on the spectrum due to poor resolution, the combined concentrations were estimated. Amino Acid Concentration in D M E M Average dilution rate 0.016 h 1 0.02 h 1 0.028 h 1 Alanine 0 1.19 1.327 1.08 Arginine + Proline 0.38 0.16 0.19 0.21 Asparagine + Glycine 0.38 0.26 0.28 0.23 Aspartate 0 0.00 0.00 0.00 Cystine 0.23 not measured not measured not measured Glutamate 0 0.23 0.24 0.22 Histidine + Threonine 0.95 0.30 0.31 0.38 Isoleucine 0.76 0.14 0.19 0.42 Leucine 0.76 0.05 0.09 0.24 Lysine 0.76 0.39 0.42 0.52 Methionine 0.19 1.08 0.74 1.21 Phenylalanine 0.38 0.14 0.15 0.19 Serine 0.38 0.24 0.28 0.26 Tryptophan 0.076 not measured not measured not measured Tyrosine 0.437 0.19 0.20 0.23 Valine 0.76 0.29 0.33 0.52 65 Table 5-10: Amino acids consumption/production rates (pmol/cell.d) in continuous culture. Negative values indicate metabolite consumption, whereas positive values indicate metabolite production. Amino Acid Average dilution rate 0.016 b"1 0.02 h"1 0.028 h"1 Alanine (mM) 0.153 0.275 0.404 Arginine + Proline (mM) -0.028 -0.040 -0.062 Asparagine + Glycine (mM) -0.015 -0.021 -0.055 Aspartate (mM) 0.0 0.0 0.0 Cystine (mM) n/a n/a n/a Glutamate (mM) 0.029 0.050 0.082 Histidine + Threonine (mM) -0.084 -0.133 -0.213 Isoleucine (mM) -0.079 -0.119 -0.127 Leucine (mM) -0.091 -0.141 -0.193 Lysine (mM) -0.048 -0.071 -0.090 Methionine (mM) 0.115 0.114 0.383 Phenylalanine (mM) -0.031 -0.047 -0.070 Serine (mM) -0.018 -0.021 -0.044 Tryptophan (mM) n/a n/a n/a Tyrosine (mM) -0.032 -0.049 -0.077 Valine (mM) -0.061 -0.090 -0.090 5.5.2 Simulation results Using mass balances, the batch model (Equations 4.1-4.3) can be readily extended to continuous culture as follows: dX dt = fjX-DX dS dt = -qsX + D(Sin-S) dP_ dt = qPX - DP (5.2) (5.3) (5.4) where D is the dilution rate defined as the ratio of the volumetric feed flow rate over the volume of the reactor and Sm is the feed medium substrate concentration. 66 Figures 5-14 to 5-16 show the simulation results obtained when applying the model Equations 5.2 to 5.4, using the batch parameters. The simulations are compared to two data sets from experiments run at the same operating conditions. The cell concentration prediction was in agreement with the experimental data (Figure 5-14). However, as can be seen in Figure 5-15, the batch model consistently predicted a higher steady-state glucose concentration then was obtained experimentally. This offset was similar to that found in the semi-continuous culture analysis. The product concentration profile illustrates the inherent variation common to most biological processes. Although the steady-state cell and glucose concentrations overlapped for the two experiments, one of the runs produced slightly more antibody. As is evident in the glucose profile, the transient data for the two cultures differed at the beginning and so it is possible that these differences affected the remainder of the culture. However, it should also be pointed out that an error associated with the antibody measurement could explain the apparent discrepancy. In any case, the assumption of a constant specific production rate provided a reasonably good description of the average data. 67 1 .OE+07 co o ' c o co c o o c o o a3 o .a CO > 1 .OE+06 1.0E+05 100 —Simulation A Experimental data set #1 • Experimental data set #2 200 Time(h) 300 400 Figure 5-14: Simulation of cell concentration profile in continuous culture (D = 0.028 h"1). The model predictions (solid line) are compared with data from two separate experiments (filled circles and triangles) run under the same operating conditions. o CO k. -*—' c cu o c o o CD CO o o 3 o 24 20 16 12 8 1 V I 16 -^ W-1 t S 19 • Experimental data set #1 A Experimental data set #2 —Simulation A * * * 4 t * t * A A A * a 100 200 Time(h) 300 400 Figure 5-15: Simulation of glucose concentration profile in continuous culture (D = 0.028 h"1). The model predictions (solid line) are compared with data from two separate experiments (filled circles and triangles) run under the same operating conditions. 68 185 • Experimental data set #1 A Experimental data set #2 —Simulation 0 50 100 150 200 250 300 350 400 Time(h) Figure 5-16: Simulation of antibody profile in continuous culture (D = 0.028 h"1). The model predictions (solid line) are compared with data from two separate experiments (filled circles and triangles) run under the same operating conditions. 5.6 Perfusion cultures 5.6.1 Perfusion results In order to investigate how high cell density systems compare with batch and continuous cultures, three perfusion experiments were performed. The reactor maintained cell concentrations more than 2 times greater than the maximum encountered in batch culture. The cell and antibody concentration profiles from one run are shown in Figure 5-17. 69 1 .OE+07 E 8.0E+06 <u e •B 6.0E+06 O.OE+00 -o- Viable cell concentration A Antibody concentration 100 200 Time (h) 300 180 150 120 + 90 60 30 400 3 c o o o e o o >» TJ O .Q Figure 5-17: Viable cell and antibody concentration profiles for a perfusion culture. Open circles represent the viable cell concentrations and filled triangles the antibody concentrations. The average feed rate was 195 ml/d. In comparison of the three perfusion runs (Table 5-11), the cell concentrations, product concentrations, production rate and glutamine uptake rate all increased with the amount of medium fed per day. However, the limited number of experimental points available does not permit an assessment of the significance level of these trends. The specific glucose uptake rates, as well as the ammonium and lactate concentrations were quite similar for the three operating conditions tested. No clear trend could be observed for the yield of lactate on glucose and the yield of ammonium on glutamine, but the values obtained fall in the range of those encountered in other culture modes. 70 Table 5-11: Perfusion culture results. Cells were grown in spinners with a working volume of 200 mL. Average feed rate 140 mL/d 195 mL/d 245 mL/d Cell concentration (cells/mL) 5xl06 6.7x106 7x106 Antibody concentration (Hg/mL) 91 120 126 Glucose concentration (mM) 10.7 6.7 10.7 Glutamine concentration (mM) 1.69 0.53 0.62 Glucose uptake rate (pmol/cells.h) 0.08 0.11 0.11 Glutamine uptake rate (pmol/cells.h) 0.46 0.64 0.80 Specific production rate (pg/cells.h) 0.50 0.68 1.19 Average lactate concentration (mM) 21.6 18.0 19.0 Average ammonium concentration (mM) 2.1 2.2 2.3 Average yield of lactate on glucose 1.67 0.94 1.39 Average yield of ammonium on glutamine 0.68 0.98 0.57 None of the amino acid appears to be limiting for the perfusion cultures (Table 5-12). Glutamate and alanine were produced (Table 5-13) as previously observed. Unlike the other culture modes, the three perfusion runs had significant residual glutamine concentrations. Since the reactor relied only on surface aeration, oxygen may have been the factor limiting cell growth. 71 Table 5-12: Amino acid concentration (mM) in perfusion culture. When two peaks were merged on the spectrum due to poor resolution, the combined concentrations were estimated. Amino Acid Concentration in DMEM Average feed rate 140 mL/d 190 mL/d 250 mL/d Cystine 0.22 not measured not measured not measured Aspartate 0.00 0.00 0.00 0.00 Glutamate 0.00 0.26 0.24 0.17 Serine 0.38 0.39 0.29 0.27 Asparagine + Glycine 0.38 0.41 0.34 0.33 Histidine + Threonine 0.95 0.62 0.44 0.41 Alanine 0.00 1.27 1.50 1.24 Arginine + Proline 0.38 0.40 0.45 0.34 Tyrosine 0.44 0.39 0.27 0.26 Valine + Methionine 0.95 0.92 0.63 0.60 Isoleucine 0.76 0.71 0.51 0.48 Leucine 0.76 0.46 0.34 0.29 Phenylalanine 0.38 0.31 0.22 0.20 Tryptophan 0.08 not measured not measured not measured Lysine 0.76 0.77 0.58 0.52 Table 5-13: Amino acid consumption/production rates (pmol/cell.d) in perfusion culture. Negative values indicate metabolite consumption, whereas positive values indicate metabolite production. Amino Acid Average feed rat e 140 mL/d 190 mL/d 250 mL/d Cystine n/a n/a n/a Aspartate 0.000 0.000 0.000 Glutamate 0.057 0.046 0.043 Serine 0.003 -0.017 -0.028 Asparagine + Glycine 0.006 -0.007 -0.013 Histidine + Threonine -0.073 -0.100 -0.135 Alanine 0.279 0.293 0.310 Arginine + Proline 0.004 0.014 -0.011 Tyrosine -0.011 -0.032 -0.045 Valine + Methionine -0.006 -0.063 -0.088 Isoleucine -0.011 -0.049 -0.070 Leucine -0.067 -0.083 -0.118 Phenylalanine -0.014 -0.031 -0.045 Tryptophan n/a n/a n/a Lysine 0.003 -0.035 -0.059 72 5.7 Fed-batch cultures Extension of culture time was achieved by daily additions of a concentrated feed medium. The volumes of concentrate added were calculated based on the assumption of a constant glucose uptake rate obtained from batch. The maximum cell concentrations (Figure 5-18) were similar to the ones obtained in batch but, as expected, viability was maintained for a 2-fold longer period of time in fed-batch. This resulted in 2-fold higher final antibody concentrations (Figure 5-19). The glucose concentrations increased throughout the process (Figure 5-20). Since the feed was based on an average consumption rate measured in batch, this suggests that the uptake rates were lower in fed-batch. To study the effect of potential inhibition by lactate, two starting media were tested: one with an initial glucose concentration of 25 mM and an other of 5 mM. The use of a low glucose medium did not show any improvement both in terms of cell and antibody concentrations. These results are summarized in Table 5-14. The table suggests that although the final levels of lactate were slightly higher in the high-glucose medium, it did not reach a critical concentration at which it would influence the kinetics of growth and/or production of the cells (Ozturk et al., 1992). 73 1.0E+08T cu o X> ra > 1.0E+05 -Viable cell concentration -Cell viability + 50% 30% 10% -10% 20 40 60 80 100 120 140 Time (h) Figure 5-18: Cell concentration (filled circles) and viability (filled triangles) profiles in fed-batch culture. The starting glucose concentration was 5 mM. 300 E 250 § 2 0 0 ro I 150 o c £ 100 -D o A •5 50 > > ( T 4 • T • 0.E+00 5.E+07 1.E+08 2.E+08 Time (h) 2.E+08 3.E+08 3.E+08 Figure 5-19: Antibody concentration (filled circles) profile in fed-batch culture. Error bars are standard deviations. The starting glucose concentration was 5 mM. 74 60 80 Time (h) • Glucose - Lactate 24 2 0 ? E 16 I 12 I o c o u 0) to ts 8 140 Figure 5-20: Glucose (filled circles) and lactate (filled triangles) concentration profiles in fed-batch culture. Arrows indicate when concentrate additions were made. Higher ammonium concentrations than those normally encountered in batch were reached (Table 5-14). It is possible that these high levels were responsible for the transition of the cells from growth to death (Ozturk et al., 1992). Table 5-14: Fed-batch results. Batch Fed-batch (high glucose) Fed-batch (low glucose) Viable index (cells.h/mL) 1.46x10s 2.30x10s 2.94x10s Final antibody (u.g/mL) 107 184 220 Final ammonium concentration (mM) 3.2 8.7 4.7 Final lactate Concentration (mM) 23.1 24.7 22.4 75 Table 5-15 presents the concentrations of various amino acids at different culture times. Since enzymatic soy hydrolysate was used in the concentrate, the consumption/production rates were not calculated. Despite the addition of concentrate to the culture at regular intervals, glutamine levels remained quite low in the second half of the culture. There were also high uptake rates of leucine and isoleucine (Table 5-15), which is consistent with previous observations. Table 5-15: Amino acid concentrations (mM) in fed-batch culture. When two peaks were merged on the spectrum due to poor resolution, the combined concentrations were estimated. Concentration after 10 h Concentration after 43 h Concentration after 77 h Concentration after 118 h Cystine n/a n/a n/a n/a Aspartate 0.000 0.000 0.021 0.049 Glutamate 0.148 0.266 0.659 0.761 Serine 0.448 0.150 0.380 0.255 Asparagine+Glycine 0.308 0.235 0.499 0.819 Glutamine 3.209 0.335 0.081 0.092 Histidine+Threonine 0.554 0.293 0.449 0.399 Alanine 0.367 0.594 1.895 2.750 Arginine+Proline 0.245 0.149 0.409 0.629 Tyrosine 0.383 0.173 0.252 0.254 Methionine+Valine 0.823 0.407 0.665 0.483 Isoleucine 0.723 0.205 0.246 0.153 Leucine 0.539 0.129 0.128 0.093 Phenylalanine 0.375 0.161 0.432 0.209 Tryptophan n/a n/a n/a n/a Lysine 0.854 0.338 0.483 0.486 Further work is needed before the batch model could describe cell growth in fed-batch. The approach would require the identification of the limiting nutrient or the inhibiting component 76 to adequately model cell growth. However, to predict the final antibody production, it appears sufficient to assume a constant average specific production rate given by: (5.5) As shown in Table 5-14 and Figure 5-21, a 2-fold increase in the viable index resulted in a 2-fold increase in the final antibody concentration. 300 0.0E+00 1.0E+08 2.0E+08 3.0E+08 Viable Index (cells.h/mL) Figure 5-21: Antibody concentration as a function of the viable index for batch (open symbol) and three fed-batch (filled symbols) cultures. All cultures exhibited a similar average specific production rate given by the slope of the solid line. 5.8 Comparison of kinetic rates In order to compare the kinetics of growth and antibody production between the different modes, the specific rates of nutrient uptake and metabolite production are represented as 77 functions of the specific growth rate. This approach is used to test the hypothesis that, despite the variability in results often encountered, parameters estimated from batch or semi-continuous data allow the predictive modeling of other modes. 5.8.1 Specific glucose uptake rate In Figure 5-22, the batch model was obtained assuming a constant yield of cells on glucose based on data taken during the exponential phase. The beginning of a batch culture is typically characterized by high glucose uptake rates. These rates tend to decrease during the exponential phase and decrease almost linearly with respect to growth rates as the cells enter the stationary phase. 0 . 8 Figure 5-22: Specific glucose uptake rate as a function of growth rates for batch (open circles), semi-continuous (filled circles), continuous (open triangles) and fed-batch (cross) cultures. The solid line represents the batch model prediction. 78 The semi-continuous and continuous rates (Figure 5-22) show a clear difference as all the rates were consistently greater than those observed in batch for the same growth rates (not taking into account the initial points from batch culture). This illustrates the offset observed in the glucose simulated profiles for continuous and semi-continuous cultures presented in previous sections. Many factors could contribute to these differences. First of all, the ever-changing conditions in batch could have a detrimental effect on the cells as opposed to the steady-state or pseudo-steady-state conditions obtained in continuous or semi-continuous processes. It could also be the result of the rapid uptake and depletion of a growth factor in batch that the addition of fresh medium in the continuous processes would prevent. When developing the model, the maintenance term was assumed to be zero, but a regression of the continuous data yields an intercept of 0.01 pmol/cell.h. Also included on Figure 5-22 are the fed-batch uptake rates. However, these data are quite erratic and do not support clear conclusions. Since no precise value of the growth rate was obtained for the perfusion cultures, a new independent variable was introduced, the "specific dilution rate" obtained by dividing the dilution rate by the average cell concentration. This variable allowed us to plot together the specific glucose uptake rates of all the continuous processes. As shown in Figure 5-23, considering the errors associated with the data points, there is a relatively good agreement between all the different continuous processes and it illustrates how parameters estimated 79 from semi-continuous and continuous data could be used in a model to predict perfusion results. 0.4 0.E+00 1.E-08 2.E-08 3.E-08 Cell specific dilution rate (mL/cell.h) Figure 5-23: Specific glucose uptake rate as a function of specific dilution rate for semi-continuous (filled circles), continuous (open triangles) and perfusion (filled squares) cultures. 5.8.2 Specific glutamine uptake rate Figure 5-24 shows the specific glutamine uptake rate as a function of the growth rate. Again, there is a high metabolic activity at the beginning of the batch culture followed by a significant decrease as the cells enter the exponential phase. To fill in the curve, more batch points would be needed at lower growth rates, but the rapid transition from exponential to stationary phase made this difficult to obtain. Figure 5-22 shows remarkable agreement between the continuous and semi-continuous data over the range of dilution rates studied. It 80 is evident that assuming a constant yield of cells on glutamine based on data from batch would give a reasonably good description of the continuous data. The data also suggest that the maintenance term for glutamine is negligible. Figure 5-25 shows that the specific glutamine uptake rates in perfusion cultures are also in good agreement with the continuous and semi-continuous results. = 5 ^ ca IK 4-£ "33 I o % E 3. a. o o Qi CL 00 O Batch • Semi-Continuous A Continuous Batch model Growth rate (h ) Figure 5-24: Specific glutamine uptake rate as a function of growth rate for batch (open circles), semi-continuous (filled circles) and continuous (open triangles). 81 CD 2 cu ro D. 3 2.4 2.0 sz 1.6 I £ 1.2 | | i t 3 0.8 O o CD O -oo —s— > ^ I • ' » I T • Semi-Continuous A Continuous • Perfusion 0.4 0.0 0.E+00 1.E-08 2.E-08 3.E-08 Cell specific dilution rate (mL/cell.h) Figure 5-25: Specific glutamine uptake rate as a function of specific dilution rate for semi-continuous (filled circles), continuous (open triangles) and perfusion (filled squares) cultures. The solid line represents the batch model prediction. 5.8.3 Specific monoclonal antibody production rate Figures 5-26 and 5-27 present the specific MAb production rate as a function of the growth rate and the specific dilution rate respectively. The scatter and the large errors associated with the experimental values do not support clear conclusions. Most of the data fall in the range of 1 to 1.5 pg/cell.h, but the results tend to suggest a growth-association for the production rate and this apparent trend was observed in all continuous modes. The error in antibody determination can be as high as 20% and the error associated with the rates even greater. According to Phillips et al. (1991), most of the data reported in the literature should be treated with some caution since different interpretations can be made when considering 82 the errors involved. Many studies have reported increased MAb production at low growth rates (Reuveny et al., 1986; Miller et al., 1988; Martens et al., 1993), but it is not a universal finding. There are many contradictory results on MAb production kinetics in the literature, even when they are obtained in continuous cultures. Most probably, these differences may be attributed to the different cell lines used. In many cases, the specific productivity was found to be independent of the growth rate, or, like in the present work, growth associated (Low et al., 1987; Robinson and Memmert, 1991; Gaertner and Dhurjati, 1993; Harigae and Matsumura, 1994). Low et al. (1987) have suggested that, at low specific growth rates associated with long residence time, the lower specific MAb production rates were due to the degradation or inactivation of MAb. Inhibition of specific antibody production rate at high antibody concentration could also explain the observed profile, but it has never been reported in the hybridoma literature. tu c o 4.0 S 3.0 o 3 T J 2 £~ o_ = n o 2 <=» :2 Q. -•—' — c CO o 'co CO C L w o-Batch • Semi-Continuous A Continuous 0.00 0.01 0.02 0.03 0.04 Growth rate (h"1) 0.05 0.06 Figure 5-26: Specific antibody production rate as a function of growth rate for batch (open circles), semi-continuous (filled circles) and continuous (open triangles) cultures. In the batch model, a constant production rate of 1.5 pg/cell.h was assumed. 83 c o 4.0 * 3.0 o •a o •a o $2.0 Q. ca o it o CO Q. CO 1.0 0.0 • Semi-Continuous A Continuous • Perfusion 0.E+00 A. 1.E-08 2.E-08 Cell specific dilution rate (mL/cell.h) Figure 5-27: Specific antibody production rate as a function of specific dilution rate for semi-continuous (filled circles), continuous (open triangles) and perfusion (filled squares) cultures. 84 6 CONCLUSIONS & RECOMMENDATIONS A simple model was developed based solely on batch data taken during the exponential phase. When applied to continuous or semi-continuous data, an offset was noticed in the glucose steady-state concentration. The apparent glutamine limitation observed in batch and some continuous cultures did not result in any significant metabolic shift. Continuous and perfusion cultures were adequately described by semi-continuous data (Figure 5-23 and 5-25). In terms of antibody formation, the results suggest a positive association between the specific production rate and the growth rate (Figure 5-26). Considering the errors involved, the constant value assumed in the batch model gives a reasonable approximation but the predictions could be further improved by including a growth related term obtained from semi-continuous data. Taken together, these results show the value of the semi-continuous mode as a way to approximate continuous operation while keeping the relative simplicity of batch culture (i.e. spinners). These advantages make semi-continuous culture a useful tool for preliminary optimization of complicated systems such as perfusion culture, with the potential to greatly facilitate and accelerate process development. The volumetric productivity was greatest in perfusion systems, while the highest concentrations were obtained in fed-batch processes. Fed-batch increased the viability index and the antibody production compared to batch (Figure 5-21), while retaining the relative simplicity and reliability of batch stirred tank operation. Cell growth in fed-batch could not be 85 described by the batch model, but the antibody concentration can be predicted from batch data assuming an average production rate. Modeling cell growth in fed-batch would require the identification of the limiting substrate or the inhibiting component. It was demonstrated that some of the variation observed in the kinetic rates is attributable to changes in the cell volume (Figure 5-2). These rates may also be influenced by the state of the inoculum (Tables 5-1 and 5-2). The cell specific death rate was neglected in the analysis, but it may play a significant role in continuous cultures operated at very low dilution rates. It is suggested that the effect of cell death be investigated and characterized to determine its influence on the various kinetic parameters. It has been emphasized many times in this work that different cell lines may exhibit totally different kinetics. In order to assess whether the conclusions reached in this study are unique to the cell line employed, the same approach should be repeated using another cell line. 86 7 REFERENCES Al-Rubeai "Specific Monoclonal Antibody Productivity and the Cell Cycle Comparisons of Batch, Continuous and Perfusions Cultures", Cvtotechnol.. 9, 85-97, 1992. Aulitzky, W., Schulz, T., Tilg, H., Niederwieser, D., Larcher, K., Ostberg, L., Scriba, M., Martindale, J., Stern, A., Grass, P., Mach, M., Dierich, M. and Huber, C. "Human Monoclonal Antibodies Neutralizing Cytomegalovirus (CMV) for Prophylaxis of CMV Disease: Report of Phase I, Trial in Bone Marrow Transplant Recepients", J. Infect Pis.. 163, 1344-1347, 1991. Banik, G. and Heath, C. A. "Hybridoma Growth and Antibody Production as a Function of Cell Density and Specific Growth Rate in Perfusion Culture", Biotechnol. Bioeng., 48, 289-300, 1995. Bartley, A. and MacLeod, A. J., "A Comparative Study of Monoclonal Antibody Yield Using Batch, Continuous or Perfusion Suspension Culture Techniques". In Animal Cell Technology: Developments, Processes & Products, R. E. Spier, J. B. Griffiths and C. Macdonald, Butterworth-Heinemann Ltd, Oxford, U.K., 376-378, 1992. Batt, D. C. and Kompala, D. S. "A Structure Kinetic Modeling Framework for the Dynamics of Hybridoma Growth and Monoclonal Antibody Production in Continuous culture", Biotechnol. Bioeng.. 34, 515-531, 1989. Bibila, T. and Flickinger, M. C. "A Structured Model for Monoclonal Antibody Synthesis in Exponentially Growing and Stationary Phase Hybridoma Cells", Biotechnol. Bioeng.. 37, 210-226, 1991. Bibila, T. A., Ranucci, C. S., Glazomitsky, K., Buckland, B. C. and Aunins, J. G. "Monoclonal Antibody Process Development Using Medium Concentrates", Biotechnol. Prog.. 10, 87-96, 1994. Blasey, H. D., Isch, C. and Bernard, A. R., "Production of a Monoclonal Antibody with a Serum Dependant Hybridoma in Different Reactor Systems". In Animal Cell Technology: Products of Today, Prospects for Tomorrow, R. E. Spier, J. B. Griffiths and W. Berthold, Butterworth-Heinemann Ltd, Oxford, U.K., 329-330, 1994. Bushell, M. E., Bell, S. L., Scott, M. F., Spier, R. E., Wardell, J. N. and Sanders, P. G. "Enhancement of Monoclonal Antibody Yield by Hybridoma Fed-Batch Culture, Resulting in Extended Maintenance of Viable Cell Population", Biotechnol. Bioeng.. 44, 1099-1106, 1994. de Tremblay, M., Perrier, M., Chavarie, C. and Archambault, J. "Optimization of Fed-Batch Culture of Hybridoma Cells using Dynamic Programming: Single and Multi Feed case", Bioprocess Eng.. 7, 229-234, 1992. 87 Dutton, R. L., Scharer, J. M. and Moo-Young, M. "Hybridoma Growth and Productivity: Effects of Conditioned Medium and Inoculum Size", Cvtotechnol.. 29, 1-10, 1999. Fenge, C , Fraune, E. and Schugerl, K., "Perfusion Bioreactor Performance at Different Cell Bleed Rates". In Animal Cell Technology: Developments, Processes & Products, R. E. Spier, J. B. Griffiths and C. Macdonald, Butterworth-Heinemann Ltd, Oxford, U.K., 365-372, 1992. Fike, R., Kubiak, J., Price, P. and Jayme, D. "Feeding Strategies for Enhanced Hybridoma Productivity: Automated Concentrate Supplementation", BioPharm. October , 49-54, 1993. Follstad, B. D., Balcarcel, R. R., Stephanopoulos, G. and Wang, D. I. "Metabolic Flux Analysis of Hybridoma Continuous Culture Steady State Multiplicity", Biotechnol. Bioeng.. 63, 675-683, 1999. Frame, K. K. and Hu, W. S. "The Loss of Antibody Productivity in Continuous Culture of Hybridoma Cells", Biotechnol. Bioeng.. 35, 469-476, 1990. Frame, K. K. and Hu, W. S. "Kinetic Study of Hybridoma Cell Growth in Continuous Culture: II. Behavior of Producers and Comparison to Nonproducers.", Biotechnol. Bioeng.. 38, 1020-1028, 1991. Gaertner, J. G. and Dhurjati, P. "Fractional Factorial Study of Hybridoma Behavior. 1. Kinetics of Growth and Antibody Production.". Biotechnol. Prog.. 1993, 309-316, 1993. Glacken, M. W. "Reduction of Waste Product Excretion via Nutrient Control: Possible Strategies for Maximizing Product and Cell yields on Serum in Cultures of Hybridoma Cells", Biotechnol. Bioeng.. 28, 1376-1389, 1986. Glacken, M. W., "Development of Mathematical Descriptions of Mammalian Cell culture Kinetics for the Optimization of Fed-Batch Bioreactors". Ph.D. Thesis, Massachusetts Institute of Technology, Cambridge, MA., 1987. Glacken, M. W., Adema, E. and Sinskey, A. J. "Mathematical Descriptions of hybridoma Culture kinetics: I. Initial Metabolic Rates", Biotechnol. Bioeng.. 32, 491-506, 1988. Goergen, J. L., Engasser, M. and Engasser, J. M. "Comparison of Specific Rates of Hybridoma Growth and metabolism in batch and continuous cultures", Cvtotechnol.. 10, 147-155, 1992. Griffiths, J. B., Looby, D. and Racher, A. J. "Maximisation of Perfusion Systems and Process Comparison with Batch-Type Cultures", Cvtotechnol.. 9, 3-9, 1992. Harigae, M. and Matsumura, M. "Kinetic Study on HBs-MAb Production in Continuous Cultivation", J. Biotechnol.. 34, 227-235, 1994. 88 Hiller, G. W., Aeschlimann, A. D., Clarck, D. S. and Blanch, H. W. "A Kinetic Analysis of Hybridoma Growth and Metabolism in Continuous Suspension Culture on Serum-Free Medium", Biotech. Bioeng.. 38, 733-741, 1991. Hiller, G. W., Clark, D. S. and Blanch, H. W. "Cell Retention - Chemostat Studies of Hybridoma Cells - Analysis of hybridoma Growth and Metabolism in Continuous Suspension Culture on Serum-Free Medium", Biotechnol. Bioeng.. 42, 185-195, 1993. Himmelfarb, P., Thayer, P. S. and Martin, H. E. "Spin Filter culture: The Propagation of Mammalian Cells in Suspension", Science. 164, 555-557, 1969. Jeong, Y.-H. and Wang, S. S. "Role of Glutamine in Hybridoma cell Culture: Effects on Cell Growth, Antibody Production, and Cell Metabolism", Enz. Microb. Technol.. 17, 47-55, 1995. Jo, E . - C , Park, H.-J., Park, J.-M. and Kim, K.-H. "Balanced Nutrient Fortification Enables High-Density Hybridoma Cell Culture in Batch Culture", Biotechnol. Bioeng.. 36, 717-722, 1990: Johnson, M., Lanthier, S., Massie, B., Lefebvre, G. and Kamen, A. "Use of the Centritech Lab Centrifuge for Perfusion Culture of Hybridoma Cells in Protein-Free Medium", Biotechnol. Prog.. 12, 855-864, 1996. Kling, J. "Restoring magic to the bullets", Modern Drug Discovery. March/April, 33-45, 1999. Krieger-Lassen, K. M. and Emborg, C , "Perfusion Systems for Hybridoma Cell Based on Sedimentation in Chambers and Erlenmeyer Flasks". In Animal Cell Technology: Developments, Processes & Products, R. E. Spier, J. B. Griffiths and C. Macdonald, Butterworth-Heinemann Ltd, Oxford, U.K., 351-353, 1992. Kurkela, R., Fraune, E. and Vihko, P. "Pilot-Scale Production of Murine Monoclonal Antibodies in Agitated, Ceramic-Matrix or Hollow-Fiber Cell Culture Systems", BioTechniq.. 15, 674-683, 1993. Kurokawa, H., Park, Y. S., Lijima, S. and Kobayashi, T. "Growth Characteristics in Fed-Batch Culture of Hybridoma Cells with Control of Glucose and Glutamine Concentrations", Biotechnol. Bioeng.. 44, 95-103, 1994. Lee, Y.-K., Yap, P.-K. and Teoh, A.-P. "Correlation Between Steady-State Cell Concentration and Cell Death of Hybridoma Cultures in Chemostat", Biotechnol. bioeng.. 45, 18-26, 1995. Leno, M., Merten, O.-W. and Hache, J. "Kinetic Analysis of Hybridoma Growth and Monoclonal Antibody Production in Semi-Continuous Culture", Biotechnol. Bioeng.. 39, 596-606, 1992. 89 Linardos, T. I., Kalogerakis, N. and Behie, L. A. "The Effect of Specific Growth Rate and Death Rate on Monoclonal Antibody Production in Hybridoma Chemostat Cultures", Can. J. Chem. Eng.. 69, 429-438, 1991. Ljunggren, J. and Haggstrom, L. "Catabolic Control of Hybridoma Cells by glucose and Glutamine Limited Fed Batch Cultures", Biotechnol. Bioeng.. 44, 808-818, 1994. Low, K. S., Harbour, C. and Barford, J. P. "A Study of Hybridoma Cell Growth and Antibody Production Kinetics in Continuous culture", Biotechnol. Techniq.. 1, 239-244, 1987. Luan, Y., Mutharasan, R. and Magee, W. "Strategies to Extend Longevity of Hybridomas in Culture and Promote Yield of Monoclonal Antibodies", Biotechnol. Lett.. 9, 691-696, 1987. Martens, D. E., de Gooijer, C. D., van der Velden-de Groot, C. A. M., Beuvery, E. C. and Tramper, J. "Effect of Dilution Rate on Growth, Productivity, Cell Cycle and Size, and Shear Sensitivity of a Hybridoma Cell in a Continuous Culture", Biotechnol. Bioeng.. 41, 429-439, 1993. Martial, A., Dardenne, M., Engasser, J. M. and Marc, A. "Influence of Inoculum Age on Hybridoma Culture Kinetics". Cvtotechnol.. 5, 165-171, 1991. Mazars, C , Provenchere, S., Pouget, L., Pingeon, B. and Henno, P., "Comparison of Different Cell Culture Systems for the Production of Monoclonal Antibodies". In Advances in Animal Cell Biology and Technology for Bioprocesses, R. E. Spier, J. B. Griffiths and B. Meignier, Butterworth-Heinemann Ltd, Oxford, U.K., 350-354, 1989. McArdle, J. "Alternatives to Ascites Production of Monoclonal Antibodies", Animal Welfare Information Center Newsletter. 8, 1997. McQueen, A. and Bailey, J. E. "Effect of Ammonium ion and Extracellular pH on Hybridoma Cell Metabolism and Antibody Production", Biotechnol. Bioeng.. 35, 1067, 1990. Miller, W. M., Blanch, H. W. and Wilke, C. R. "A Kinetic Analysis of Hybridoma Growth and Metabolism in Batch and Continuous Culture: Effect of Nutrient Concentration, Dilution Rate, and pH", Biotechnol. Bioeng.. 32, 947-965, 1988. Morrison, S. L., Baumal, R., Birshtein, B. K., Kuehl, W. M., Preud'homme, J. L., Frank, L., Jasek, T. and Scharff, M. D., In Cellular Selection and regulation in the Immune Response, G. M. Edelman, Raven Press, New York, 233-244, 1974. Omasa, T., Higashiyama, K., Shioya, S. and Suga, K. "Effects of Lactate Concentration on Hybridoma Culture in Lactate-Controlled Fed-Batch Operation", Biotechnol. Bioeng.. 39, 555-564, 1992. 90 Omasa, T., Ishimoto, M., Higashiyama, K., Shioya, S. and Suga, K. "The Enhancement of Specific Antibody Production Rate in Glucose and Glutamine Controlled Fed-batch Culture.", Cvtotechnol.. 8, 75-84, 1992. Ozturk, S. S. and Palsson, B. O. "Effect of Medium Osmolarity on Hybridoma Growth, Metabolism, and Antibody Production", Biotechnol. Bioeng.. 37, 989-993, 1991. Ozturk, S. S. and Palsson, B. O. "Examination of Serum and Bovine Serum Albumin as Shear Protective Agents in Agitated Cultures of Hybridoma Cells", J. of Biotechnol.. 18, 13-28, 1991. Ozturk, S. S., Riley, M. R. and Palsson, B. O. "Effects of Ammonia and Lactate on Hybridoma Growth, Metabolism, and Antibody Production", Biotechnol. Bioeng.. 39, 418-431, 1992. Pelletier, F., Fonteix, C , Lourenco da Silva, A., Marc, A. and Engasser, J.-M. "Software Sensors for the Monitoring of Perfusion Cultures: Evaluation of the Hybridoma Density and the Medium Composition from Glucose Concentration Measurements", Cvtotechnol.. 15, 291-299, 1994. Petersen, J. F., Mclntire, L. and Papoutsakis, E. T. "Shear Sensitivity of Hybridoma Cells in Batch, Fed-Batch and Continuous Cultures", Biotechnol. Prog.. 6, 114-120, 1990. Philips, P. J., Marquis, C. P., Barford, J. P. and Harbour, C. "An Analysis of Some Batch and Continuous Kinetic Data of Specific Monoclonal Antibody Production from Hybridomas", Cvtotechnol.. 6, 189-195, 1991. Portner, R. and Schafer, T. "Modelling Hybridoma Cell Growth and Metabolism - a Comparison of Selected Models and Data", J. of Biotechnol.. 49, 119-135, 1996. Ramirez, 0. T. and Mutharasan, R. "Cell Cycle and Growth Phase Dependent Variations in Size Distribution, Antibody Productivity, and Oxygen Demand in Hybridoma Cultures", Biotechnol. Bioeng.. 36, 839-848, 1990. Ray, N. G., Karkare, S. B. and Runstadler, P. W. J. "Cultivation of Hybridoma Cells in continuous cultures: Kinetics of Growth and Product Formation", Biotechnol. Bioeng.. 33, 724-730, 1989. Reuveny, S., Velez, D., Miller, L. and Macmillan, J. D. "Comparison of Cell Propagation Methods for their Effect on Monoclonal Antibody Yield in Fermentors", J. Immunol. Methods.. 86, 61-69, 1986. Robinson, D. K. and Memmert, K. W. "Kinetics of Recombinant Immunoglobulin Production by Mammalian Cells in Continuous Culture", Biotechnol. Bioeng.. 38, 972-976, 1991. 91 Robinson, D. K., Widmer, J. and Memmert, K. "Effect of Specific Growth Rates on Productivity in Continuous Open and Partial Cell Retention Animal Cell Bioreactors", I Biotechnol.. 22, 41-50, 1992. Sanfeliu, A., Cairo, J. J., Casas, C , Sola, C. and Godia, F. "Analysis of Nutritional Factors and Physical Conditions Affecting Growth and Monoclonal Antibody Production of The Hybridoma KB-26.5 Cell Line", Biotechnol. Prog.. 12, 209-216, 1996. Schmid, G., Wilke, C. R. and Blanch, H. W. "Continuous Hybridoma Suspension Cultures With and Without Cell Retention: Kinetics of Growth, Metabolism and Product Formation", J. of Biotechnol.. 22, 31-40, 1992. Schwabe, J. 0., Portner, R. and Markl, H. "Improving an On-line Feeding Strategy for Fed-batch Cultures of Hybridoma Cells by Dialysis and Nutrient-Split Feeding", Bioproc. Eng.. 20, 475-484, 1999. Shintani, Y., Kohno, Y.-L, Sawada, H. and Kitano, K. "Comparison of Culture Methods for Human-Human Hybridomas Secreting anti-HBsAg Human Monoclonal Antibodies", Cvtotechnol.. 6. 197-208. 1991. Sonderhoff, S. A., Kilburn, D. G. and Piret, J. P. "Analysis of Mammalian Viable Cell Biomass Based on cellular ATP", Biotechnol. Bioeng.. 39, 859-864, 1992. Southwick, W. A. and Mannix, C. J., "Comparison of Growth and Antibody Production of a Hybridoma Cell Line Grov/n in Static Flask and Suspension Culture". In Animal Cell Technology: Developments, Processes & Products, R. E. Spier, J. B. Griffiths and C. Macdonald, Butterworth-Heinemann Ltd, Oxford, U.K., 342-344, 1992. Suzuki, E. and Ollis, D. F. "Enhanced Antibody Production at Slowed Growth Rates: Experimental Demonstration and a Simple Structured Model", biotechnol. Prog.. 6, 231-236, 1990. Trampler, F., Sonderhoff, S. A., Phylis, W. S. P., Kilburn, D. G. and Piret, J. M. "Acoustic Cell Filter for High Density Perfusion Culture of Hybridoma Cells", Bio/Technology. 12, 281-284, 1994. Velez, D., Reuveny, S., Miller, L. and Macmillan, J. D. "Effect of Feeding Rate on Monoclonal antibody Production in a Modified Perfusion-Fed Fermentor", J. Immunol. Methods. 102, 275-278, 1987. Xie, L. and Wang, D. "Fed-Batch Cultivation of animal Cells Using Different Medium Design Concepts and Feeding Strategies", Biotechnol. Bioeng.. 43, 1175-1189, 1994. Xie, L. and Wang, D. I. C. "High Cell Density and High Monoclonal Antibody Production Through Medium Design and Rational Control in a Bioreactor", Biotechnol. Bioeng.. 51, 725-729, 1996. 92 Zeilke, H. R., Zeilke, C. L. and Ozand, P. T. "Glutamine: A Major Energy Source for Cultured Mammalian Cells", Federation Proc. 43, 121-125, 1984. Zeng, A.-P. "Mathematical Modeling and Analysis of Monoclonal Antibody Production by Hybridoma Cells", Biotechnol. Bioeng.. 50, 238-247, 1996. Zeng, A. P., Deckwer, W. D. and Hu, W. S. "Determinants and Rate Laws of Growth and Death of Hybridoma Cells in Continuous Culture", Biotechnol. Bioeng.. 57, 642-654, 1998. Zhou, W., Chen, C - C , Buckland, B. and Aunins, J. "Fed-Batch Culture of Recombinant NSO Myeloma Cells with High Monoclonal Antibody Production", Biotechnol. Bioeng.. 55, 783-792, 1997. Zhou, W., Rehm, J. and Hu, W. "High viable cell Concentration Fed-Batch Cultures of Hybridoma Cells Through On-Line Nutrient Feeding", Biotechnol. Bioeng.. 46, 579-587, 1995. 93 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0058764/manifest

Comment

Related Items