UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Estimation of mammalian biomass in high density cell culture using intracellular adenosine triphospate… Sonderhoff, Stefan Andrew 1995

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

Item Metadata

Download

Media
831-ubc_1995-0236.pdf [ 3.21MB ]
Metadata
JSON: 831-1.0086734.json
JSON-LD: 831-1.0086734-ld.json
RDF/XML (Pretty): 831-1.0086734-rdf.xml
RDF/JSON: 831-1.0086734-rdf.json
Turtle: 831-1.0086734-turtle.txt
N-Triples: 831-1.0086734-rdf-ntriples.txt
Original Record: 831-1.0086734-source.json
Full Text
831-1.0086734-fulltext.txt
Citation
831-1.0086734.ris

Full Text

ESTIMATION OF MAMMALIAN BIOMASS IN HIGH DENSITY CELL CULTURE USING INTRACELLULAR ADENOSINE TRIPHOSPHATE CONCENTRATION  By STEFAN ANDREW SONDERHOFF  B.Sc, The University of British Columbia, 1990  A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE  in  THE FACULTY OF GRADUATE STUDIES (Department of Microbiology and Immunology) We accept this thesis as conforming to the required standard  THE UNIVERSITY OF BRITISH COLUMBIA April 1995 © Stefan Andrew Sonderhoff, 1995  In  presenting  degree freely  at  the  available  copying  this  of  department publication  thesis  in  partial  fulfilment  University  of  British  Columbia,  for  this  thesis  or of  reference  by this  for  his  and  scholarly  or  thesis  study.  for  her  of I  I further  purposes  gain  shall  requirements  agree  that  agree  may  representatives.  financial  the  be  It not  is be  that  the  Library  MlCf?ohfo((^gu  The University of British C o l u m b i a Vancouver, Canada  Date  DE-6 (2/88)  Apr-,  I  /  9*>  Owd  an  granted  by  understood allowed  TjMVHaAr^A^U  advanced  shall  permission for  permission.  Department of  for  the that  without  make  it  extensive  head  of  copying my  my or  written  ABSTRACT  Analysis of cellular ATP in hybridoma and BHK-21 cell cultures was investigated as a method of determining viable biomass. Cell specific ATP content in hybridoma and BHK batch cultures varied from 8 to 2 fmol/viable cell. ATP levels and cell volumes were highest during exponential growth and decreased during the stationary and decline phases. Cell specific ATP correlated directly with the viable cell volume rather than the viable cell numbers.  In hybridoma cultures with between 90 and 35% viable cells, the average  intracellular ATP concentration of viable cells was 4.2 mM with less than 10% variation. Intracellular ATP decreased rapidly after cell death due to freeze-thaw. Within 20 min, the sample ATP concentration decreased by two orders of magnitude to approach background levels. There were no significant changes in the concentration of intracellular ATP during 1000 h of hybridoma continuous culture. During this time, hybridoma cells were exposed to pHs rangingfrom6.8 to 8.0 and dissolved oxygen concentrations ranging from 3 to 80% of air saturation. Analysis of cellular ATP concentration provided a more accurate estimate of viable biomass than LDH analysis or microscopic cell counts over the course of a 500 h high density perfusion culture of CHO cells.  ii  TABLE OF CONTENTS page Abstract  11  Table of Contents  iii  List of Tables  v  List of Figures  vi  Acknowledgments  x  1  INTRODUCTION  1  2  LITERATURE REVIEW  4  2.1  Direct Enumeration  5  2.1.1 Microscopic Counts 2.1.2 Flow Cytometry 2.1.3 Image Analysis 2.1.4 Electronic Counters 2.1.5 Packed Cell Volume 2.1.6 Dry Cell Weight Measurement of Cellular Components 2.2.1 Enzyme Activity 2.2.2 NAD(P)H 2.2.3 ATP Indirect Estimates of Biomass 2.3.1 Nuclear Magnetic Resonance 2.3.2 Metabolic Rates 2.3.3 Optical Density 2.3.4 Acoustic Resonance Densitometry 2.3.5 Conductivity 2.3.6 Electrochemical Methods  5 6 6 7 8 8 9 9 10 11 13 13 13 14 15 15 16  2.2  2.3  3  MATERIALS AND METHODS  17  3.1  17 17 17 18 18 18 19  Analytical Techniques 3.1.1 ATP Extraction 3.1.2 ATP Analysis 3.1.3 Cell Enumeration and Volume Measurements 3.1.4 Nuclei Counts 3.1.5 Flow Cytometry 3.1.6 Additional Analytical Methods iii  3.2  4  Tissue 3.2.1 3.2.2 3.2.3 3.2.4  Culture Batch Cultures of Hybridoma Cells Batch Cultures of BHK Cells Continuous Cultures of Hybridoma Cells High Density Perfusion Culture  20 20 20 21 22  EXPERIMENTAL RESULTS  27  4.1  27 27 28 29 29 33 34 36 36 37 39 40 42  4.2  4.3  4.4 4.5 4.6  4.7  ATP Extraction with TCA 4.1.1 Concentration of TCA 4.1.2 Dose Response Batch Growth Curves of Hybridoma Cells 4.2.1 Changes in Cell Specific ATP 4.2.2 Cell Volume Variations 4.2.3 Intracellular ATP Concentration Batch Growth Curves of BHK Cells 4.3.1 Replicate Culture System 4.3.2 Total Culture ATP 4.3.3 Cell Specific ATP 4.3.4 Intracellular ATP Concentration Stability of Intracellular ATP Following Cell Death Comparison of Different Methods for Detecting Changes in Cellular Viability Continuous Culture 4.6.1 Effect of Short Term Variations in pH 4.6.2 Effect of Long Term Decreased Medium pH 4.6.3 Effect of Long Term Increased Medium pH 4.6.4 Effect of Dissolved Oxygen Concentration on ATP Biomass Estimation During a High Density Perfused CHO Cell Culture  43 46 46 50 53 57 59  5  DISCUSSION  68  6  REFERENCES  71  iv  LIST OF TABLES  Page Table 1.  Assays of mammalian cell biomass.  v  4  LIST OF FIGURES  page Figure 1.  Exploded drawing of acoustic cell retention system.  23  Figure 2.  Schematic representation of the perfusion bioreactor system used.  25  Figure 3.  Total ATP recovery versus the concentration of TCA extractant used.  Figure 4.  28  Total ATP recovery and cell specific ATP versus the number of cells extracted.  Figure 5.  29  (a) Total, viable and nonviable cell concentration over the course Of a batch roller bottle hybridoma culture and (b) total ATP concentration over the course of batch roller bottle and spinner cultures.  Figure 6.  Cell specific ATP over the course of batch roller and spinner hybridoma cultures.  Figure 7.  32  Mean cell volume over the course of roller bottle and spinner batch cultures.  Figure 8.  31  33  Intracellular ATP concentration over the course of roller bottle and spinner batch cultures.  34  vi  Figure 9. Volume distribution of hybridoma cells before and after sorting for viability.  Figure 10. Viable BHK cells recovered from parallel cultures grown on the surface of six well plates.  Figure 11. ATP recovered from BHK cells grown in parallel cultures on the surface of six well plates.  Figure 12. Cell specific ATP of BHK cells grown in parallel cultures on the surface of six well plates.  Figure 13. Mean cell volume of BHK cells grown in parallel cultures on the surface of six well plates.  Figure 14. Intracellular ATP concentration of BHK cells grown in parallel cultures on the surface of six well plates.  Figure 15. Cell specific ATP over time after freezing and thawing a hybridoma cell suspension.  Figure 16. (a) Viable cell concentration by trypan blue exclusion, cell associated and supernatant associated ATP and (b) cell associated and supernatant associated LDH activity of hybridoma cells versus time after resuspension in PBS.  vii  Figure 17. (a) Culture pH and viability and (b) ATP and cell concentrations versus time during a continuous culture with short term pH 47  variations.  Figure 18. (a) Cell specific and intracellular ATP concentration and (b) culture pH versus time during a continuous culture with short term pH variations.  49  Figure 19. (a) Culture pH and viability and (b) cell specific and intracellular ATP versus time during a continuous culture with long term decreased pH.  50  Figure 20. (a) ATP and cell concentrations and (b) culture pH versus time during a continuous culture with long term decreased pH.  52  Figure 21. (a) Mean cell volume and (b) culture pH versus time during a continuous culture with long term decreased pH.  53  Figure 22. (a) Culture pH and viability and (b) ATP and cell concentrations versus time during a continuous culture with long term increased pH.  54  Figure 23. (a) Mean cell volume and (b) culture pH versus time during a continuous culture with long term increased pH.  55  Figure 24. (a) Cell specific and intracellular ATP concentration and (b) culture pH versus time during a continuous culture with long term increased pH.  56  Vlll  Figure 25. (a) Dissolved oxygen and viability and (b) ATP and cell concentrations versus time during a continuous culture.  58  Figure 26. (a) Cell specific ATP and intracellular ATP concentration and (b) dissolved oxygen versus time during a continuous culture.  59  Figure 27. Total and viable cell concentration and dilution rate over the course of a high density perfusion CHO cell culture.  62  Figure 28. Glucose concentration over the course of a high density perfusion CHO cell culture.  62  Figure 29. Osmolarity over the course of a high density perfusion CHO cell culture.  63  Figure 30. ATP and cell concentrations over the course of a high density perfusion CHO cell culture.  64  Figure 31. Cell associated LDH activity and cell concentration over the course of a high density perfusion CHO cell culture.  65  Figure 32. ATP concentration versus total cell concentration during a high density perfusion CHO cell culture.  66  Figure 33. (a) ATP recovered and (b) cell concentration versus culture productivity during a high density perfusion CHO cell culture. ix  67  ACKNOWLEDGMENTS  I would like to express my sincere thanks to my thesis supervisors Jamie Piret and Doug Kilburn for all their advice, unwavering support and patience, and for allowing me the freedom to go my own way to pursue new and interesting avenues of research. I would also like to express my appreciation to Jim Kronstad and Brett Finlay for serving on my supervisory committee. Special thanks also go to Phylis Pui for her support and invaluable experimental and editorial assistance.  Thanks also to Felix Trampler and John Poppleton for the  opportunities they provided me to follow my interests and further my career and to all my friends at the biotechnology lab who will be sorely missed. Lastly, though by no means least, I would like to thank my parents Geert and Diana Sonderhoflf for their love, support and pride in my accomplishments.  1  INTRODUCTION Mammalian cell culture is used to produce proteins with valuable therapeutic and  diagnostic applications (e.g. erythropoietin and tissue plasminogen activator). Though bacterial and yeast protein production is usually less expensive than mammalian cell protein production, many proteins require mammalian post-translational modifications for proper activity. Mammalian cell culture processes are characterized by low growth rates, cell densities and product concentrations.  Although there are a large number of cell  culture techniques available, most mammalian cell culture derived products are produced using relatively low cell density batch and fed batch suspension cultures.  Suspension  bioreactors have generally been favored by the biotechnology industry because they expose cells to homogeneous, controlled environments and can be readily sampled for quality control purposes. One factor that has complicated development of high cell density mammalian cell culture processes is the difficulty of accurately determining the viable biomass. Hemocytometer cell counts are widely used. Trypan blue or other vital stains allow the differentiation of viable and nonviable cells.  However, these methods of enumeration  provide only a rough estimate of the viable biomass because small changes in mean cell diameters result in large variations in cell volumes. The mean volume of hybridoma cells vary by more than a factor of 2 through batch suspension cultures (Frame and Hu, 1990; Ramirez and Mutharasan, 1990). In processes utilizing immobilized cells, it is even more difficult than in suspension cultures to accurately assay viable biomass. Crystal violet/citric acid has been used in microcarrier cultures to release and stain nuclei for hemocytometer enumeration (van Wezel, 1967). However, nuclei counting does not differentiate between the biomass of large and small cells, nor can it distinguish cells which have recently become nonviable. In high density cell culture processes where cells grow as clumps or in porous supports, it usually is not possible to recover representative samples of viable cells or nuclei. Other l  strategies to evaluate biomass have been based on measurements of cellular components such as DNA (Himmler et al., 1985), total protein (Lydersen et al., 1985) and dry cell weight (Kilburn and van Wezel, 1970; Frame and Hu, 1990). These methods do not distinguish between viable and nonviable cell biomass and hence overestimate the viable cell density in low viability cultures. In an attempt to assay only the viable fraction of cells, other workers have analyzed components which should be absent from nonviable cells.  Analysis of ATP  (Lundin et al., 1986), NAD(P)H (Siano and Mutharasan, 1991), esterases (Boag and Sefton, 1987) and LDH activity (Piret and Cooney, 1990) have been reported. However, the concentrations of these components may vary depending on the cell culture conditions. Ruddle and Rapola (1970), for example, showed that esterase and L D H concentrations in PK cells vary 2 to 8-fold between the exponential and plateau phases of growth. In addition, the half lives of most mammalian enzymes are on the order of days or weeks (Schimke, 1970).  Thus the enzymes trapped in nonviable cells or in the cell  immobilization matrix can cause an overestimation of bioreactor viable cell loading. Low molecular weight compounds such as ATP and NAD(P)H should be released or degraded rapidly when cells lyse. However, cellular ATP and NAD(P)H concentrations also may vary as a function of environmental conditions (Siano and Mutharasan, 1991). Average cell specific ATP levels have been reported to vary by a factor of 2 during batch cultures (Lundin et al., 1986). Lundin et al. suggested that these variations in cellular ATP content might result from changing cellular energy metabolism or cell volumes. Chapman et al. (1971) showed that intracellular ATP concentrations remained constant over 12 h cell cycle experiments in synchronously growing Chinese hamster fibroblasts. These results suggest that, at least over a narrow range of culture conditions, the ATP concentration was tightly regulated in mammalian cells. The following work investigates variations in cellular ATP content of BHK and hybridoma cells in batch and continuous cultures to examine the feasibility of using ATP analysis as an accurate measure of viable  2  biomass in culture systems. The methodology is then applied to an aggregated CHO perfusion culture.  3  2  LITERATURE REVIEW Increasing use of mammalian cell cultures for the production of proteins has  stimulated the development of methodologies to assay viable cell loading in bioreactors. Accurate measurement of viable biomass allows the calculation of cell specific growth rates and cell specific productivities and thus differentiates between effects which increase cell specific productivity and effects which increase cell growth.  In high cell density  perfusion bioreactors, it is often difficult or impossible to recover individual cells for manual or electronic enumeration. In such bioreactors the ability to discriminate between viable and nonviable biomass also becomes increasingly important, since viability often drops below 50% during culture. This review will discuss the relative advantages and shortcomings of a range of biomass assays, with a particular emphasis on immobilized cell bioreactor applications. Biomass analysis methods can be grouped into 3 categories: (1) direct enumeration of cells, (2) measurement of cellular components and (3) indirect estimates of biomass (Table 1). Table 1- Mammalian cell biomass assays  CELL  CELLULAR  INDIRECT  ENUMERATION  COMPONENTS  ESTIMATES  Viable Cell Counts  ATP, NADfP)H  Metabolic Rates  Nuclei Counts  Enzyme Activity  Turbidity  Electronic Counts  Nuclear Magnetic Resonance  Acoustic Resonance  Flow Cytometry  Conductivity  Image Analysis  Electrochemical  Packed Cell Volume Dry Cell Weight  4  2.1  Direct Enumeration  2.1.1  Microscopic Counts Manual cell counts are a rapid and generally accepted method for monitoring  mammalian biomass. These are usually performed in a hemocytometer on cells stained to differentiate between viable and nonviable cells. Trypan blue (De Luca, 1965), eosine blue (Hoskins et al., 1956) and erythrosin blue (Harris, 1966) are all commonly used to estimate cellular viability. These dyes measure the integrity of cellular membranes since they are excluded from the interior of viable cells by intact cellular membranes. Microscopic examination of a cell suspension is then used both to enumerate cells and distinguish stained nonviable cells from unstained viable cells. Though labor intensive, microscopic counts are commonly used off-line to monitor bioreactors from which representative cell suspensions can be obtained.  Though conceptually simple, the  distinction between viable and nonviable cells, cellular debris and particulates in the medium is somewhat subjective, and thus depends on the individual performing the task. When cells cannot be easily disaggregated or removed from solid supports, the cell number can be estimated using nuclei counts (van Wezel, 1967). A protocol for the removal and enumeration of cells attached to microcarriers was developed by van Wezel based on the technique of Sanford (Sanford et al., 1950). A solution of 0.1 M citric acid and 0.1% crystal violet is used to lyse attached cells and stain the released nuclei. Stained nuclei are then counted microscopically. This technique does not provide an estimate of culture viability or cell volume. It has often been assumed that in microcarrier cultures, nonviable cells detach from the support matrix and are removed with the culture supernatant (Croughan and Wang, 1989). Thus nuclei released from microcarriers should represent predominantly viable cells. This is, however, not a reasonable assumption in cases where multilayering of cells occurs or the porous nature of the support matrix entraps nonviable cells. Furthermore, only a portion of the attached cell nuclei may be  5  recovered depending on the support matrix on which the cells are attached (van Wezel, 1967).  2.1.2  Flow Cytometry Though rarely used for manual counts,fluoresceindiacetate and propidium iodide  stains are used routinely in flow cytometry to resolve viable and nonviable cells. Fluorescein diacetate (Ross et al., 1989) readily penetrates intact cell membranes, diffusing into the cytoplasm where it is converted intofluoresceinby cellular esterases. Fluorescein is then retained within viable cells by intact cell membranes such that viable cells are green when illuminated by 488 nm light. Propidium iodide (Ross et al., 1989) diffuses into nonviable cells and binds to nucleic acids causing the nucleus to fluoresce red when illuminated by 536 nm light. Propidium iodide is unable to penetrate intact membranes and thus does not stain viable cells. Flow cytometry can provide accurate estimates of cell number and viability. Measurements of the forward scatter of cells also provide estimates of cell volume.  Flow cytometry is however limited to the analysis of single cell  suspensions. In addition, the high capital and operating costs of this equipment prevent its use in routine monitoring of culture biomass.  2.1.3  Image Analysis Microscopic image analysis can be used to estimate the biomass of adherent cells.  Bjornsen (1986) described an image analysis system which, within 10 min, counted and estimated the volume of 250 bacterioplankton cells with a standard error of 5%. Micrographs of thin sections of collagen microcarriers were used to measure the proliferation of anchorage dependent CV1 and CHO cells (Foran et al., 1991). Truskey and Proulx (1990) used image analysis to estimate the volume of viable 3T3 cells attached to coated glass surfaces using phase contrast microscopy. A potential advantage of image analysis would be the ability to obtain consistent estimates of biomass based on 6  standardized criteria, thus eliminating much of the variability associated with manual determinations of biomass. With the exception of a few highly specialized applications, image analysis has not been used for the estimation of biomass in suspended cultures.  2.1.4  Electronic Counters First described by W. H. Coulter in 1953, electronic particle counters measure the  voltage pulse produced as particles suspended in a conducting solution pass through a small orifice positioned between two electrodes.  The electrical resistance of particles  (cells) is higher than that of the surrounding fluid. Thus the presence of a cell within the orifice increases the resistance between the electrodes in proportion to the volume displaced by the particle. Because a constant current is maintained across the orifice, this change in resistance is observed as a voltage pulse. Electronic particle counters are able to rapidly and accurately determine cell numbers and mean cell volumes of large numbers of cells. For accurate determination of cell number and volume by an electronic counter, samples must be single cell suspensions that are predominantly free of cell sized debris. Orifices used in electronic counters are also prone to clogging if cell clumps or larger debris are present in the suspension. Several workers have shown that the mean volume of a nonviable cell population is smaller than that of a viable population (Sen et al., 1989; Ramirez and Mutharasan, 1990). During periods of rapid growth, the larger volume of viable cells allows the nonviable cell population to be identified based on their distinct volume distribution (Sen et al., 1989). Though the viability of the culture can be estimated from the cell size distribution (Attallah and Johnson, 1981), the overlap of viable and nonviable cell volume distributions impairs the measurement of the viable cell volume when large numbers of nonviable cells are present (Sen et al., 1989). Cell volumes, as determined by electronic particle counters, vary by greater than 2fold over the course of batch cultures (Ruddle and Rapola, 1970; Attallah and Johnson,  7  1981; Frame and Hu, 1990; Ramirez and Mutharasan, 1990; Goebel et al., 1990). Since the total cell volume and dry cell weight are closely correlated (Frame and Hu, 1990), it follows that measurements of cell number alone will not always provide accurate estimates of biomass. The ability of electronic cell counters to measure both the number and volume of cells allows them to provide very accurate estimates of viable biomass, as long as single cell suspensions of high viability are used.  2.1.5  Packed Cell Volume Biomass estimates based on packed cell volumes (PCV) are rapid, easy and  frequently employed by industry. PCV is measured using special centrifuge tubes that allow the measurement of very small cell pellets. Though very simple to perform, this technique lacks the sensitivity of manual cell counts and does not distinguish between viable and nonviable cells. Estimation of biomass in cultures with low cell concentrations requires large sample volumes. This technique is also not applicable to cultures containing microcarriers or where representative suspension samples cannot be obtained.  2.1.6  Dry Cell Weight Dry cell weight assays are. commonly used to quantify biomass in bacterial and  yeast cultures. The widespread acceptance of dry weight as a measure of biomass and its ability take into account changes in cell size make it a standard with which to compare and validate other methods of microbial biomass measurement. The low cell concentrations of most mammalian cultures require that large samples be used for the determination of dry cell weight. Over the range of cell concentrations typical for a mammalian cell culture 5  6  (10 -10 cells/mL), providing 10 mg of dry cell mass would requirefrom30 to 1000 mL of the culture (Frame and Hu, 1990). This makes the routine use of dry weight as a means of quantifying biomass impractical. In addition, the dry cell weight includes both viable and nonviable cells. 8  2.2  Measurement of Cellular Components  2.2.2  Enzyme Activity Enzymes are cellular constituents that in many cases can be measured with a high  degree of sensitivity. The stability of intracellular enzymes ranges from hours to weeks (Schimke, 1970).  The level of enzyme expression may vary with time and culture  conditions (Tarentino et al., 1966; Fritz et al., 1969; Ruddle and Rapola, 1970; Jenkins et al., 1992). The activities of 11 commonly measured enzymes, monitored over a period of 7 days in a batch culture of PQX Bl/2 hybridoma cells, varied from as little as 16% for ketoglutarate dehydrogenase to greater than 10-fold for lactate dehydrogenase (LDH) (Jenkins et al., 1992). Even with these potential variations in the level of expression, LDH release is commonly used as a means of estimating cell death (Marc et al., 1991). Cell numbers in the extracapillary space of a hollowfiberbioreactor have been estimated using LDH, although this involved sacrificing the entire bioreactor (Piret and Cooney, 1990). Methyl thiazole tetrazolium (MTT) and 3,3,5-triphenyltetrazolium chloride (TTC) are reduced to highly colored insoluble formazan chromagens by mitochondrial dehydrogenases.  The production of formazans occurs in viable cells only and thus  estimates viable biomass only. This type of assay is sensitive to differences in the substrate concentration as well as variations in the level of cellular dehydrogenases (Hoskins et al., 1956; Al-Rubeai et al., 1990).  Varying levels of enzyme expression during growth,  different rates of enzyme release after cell death and different enzyme stabilities are all parameters that can vary with cell line and growth conditions, and thus should be determined on a case by case basis to maximize the accuracy of enzyme activity based estimates of viable biomass.  9  2.2.2  NAD(P)H Fluorescence of the reduced coenzymes NADH and NADPH has been used as a  non-invasive means of monitoring biomass in bacterial, yeast (Li et al., 1991), fungal and high density mammalian cell cultures (Siano and Mutharasan, 1991). When excited at 330-370 nm, reduced NAD(P)H fluoresces at 440-480 nm, while the oxidized form NAD(P) does not. Although both NADH and NADPHfluoresce,studies have indicated +  that the proportion offluorescencedue to NADPH, which is primarily involved in biosynthetic pathways, is negligible (Siano and Mutharasan, 1991). Other substances that fluoresce at this wavelength include the aromatic amino acids, pyridoxine and riboflavin. In yeast the proportion offluorescencedue solely to NAD(P)H ranged from 10-50% of the total culturefluorescence(Li et al., 1991). Because of the variations in the level of backgroundfluorescence,some workers select the wavelength which provides the best assay of viable biomass on a case by case basis. Changes in cell specific NAD(P)H concentration over the course of culture have been reported. NAD(P)H levels are sensitive to changes in dissolved oxygen, glucose concentration, pH and temperature (Aubin, 1979; Siano and Mutharasan, 1991). Hybridoma cells, for example, showed a rapid increase in intracellular NAD(P)H when they were exposed to oxygen limiting conditions (Siano and Mutharasan, 1991). Culture fluorescence can be readily used for on-line measurements, making possible the use of real timefluorescencebased control strategies.  Although several  fluorimeters are commercially available, NAD(P)Hfluorometryis limited to suspension cultures and situations where thefluorescenceof representative culture samples can be detected.  Sensitivity to changes in culture conditions and interference by a number of  media constituents prevent any direct correlation of culturefluorescencewith viable biomass in mammalian cell culture.  10  2.2.3 ATP Adenosine triphosphate (ATP) is the main currency of free energy in all cells, playing a crucial role in both catabolic and anabolic cellular metabolism. The concentration of ATP or the relative amount of ATP and its precursors ADP and AMP also regulate the rates of anabolic and catabolic activity. The  adenylate  energy  charge  (AEC),  calculated  from  the  equation  AEC = ([ATP]+0.5[ADP])/([ATP]+[ADP]+[AMP]), is used to describe the ratio of high energy phosphate bonds within the entire adenylate pool. The in vitro enzyme activity versus AEC response curves of several bacterial adenylate converting enzymes indicated the in vivo energy charge should be stable at 0.8-0.95. Response curves became nearly horizontal with AEC less than 0.5, and in some cases became antistabilizing (Atkinson, 1969). In growing E. coli AEC levels were above 0.8, while AEC levels below 0.5 were associated with cell death (Chapman et al., 1971). Measurements of mammalian cell AEC typically show even higher values. Murine hybridoma cells grown in perfusion culture maintained AEC levels between 0.95 and 0.97 during the entire cultivation (Ryll et al., 1991). Human lymphocytes also had similar AEC levels (de Korte et al., 1985). In some microorganisms AEC was more important to cell growth regulation than the absolute amount of ATP. Somlo (1970) described a Saccharomyces cerevisiae mutant that could survive, but not grow on ethanol. Although the absolute amount of ATP produced by this mutant was the same as in the wild type, the AEC was only 0.67 compared to 0.86 in the wild type, the lower level being associated with non-growing microbial cells. Within the AEC range typical for mammalian cells (i.e. >0.9), the AMP pool is much more sensitive to small changes in AEC than the ATP pool. A change in AEC from 0.92 to 0.8 could result from a 20% decrease in ATP and a 5 fold increase in AMP, and the ATP:AMP ratio could decrease from 45:1 to 8:1. This example demonstrates how  11  relatively small fluctuations in AEC might have profound regulatory effects (Dawes, 1986). Tight regulation of ATP producing and consuming metabolic processes maintains a relatively constant ATP concentration within the cell. Inhibition of a specific anabolic (ATP consuming) pathway results in a proportional decrease in catabolic (ATP producing) activity (Siems et al., 1984, 1986).  Similar results were reported for Ehrlich ascites  tumour cells (Miiller et al, 1986) and rat hepatocytes (Schneider et al., 1990). Depletion of intracellular ATP has the opposite effect resulting, for example, in a rapid decrease in protein synthesis (Freudenberg and Mager, 1971). Luciferase bioluminescence based assays of ATP are extremely sensitive, with a useful range of 10"6-10~1 1 M ATP. The high sensitivity of this assay allows the accurate determination of ATP in samples containing low cell numbers. ATP analysis has been used to estimate bacterial (Thore et al., 1975; Gikas and Livingston, 1993), yeast (Siro et al., 1982) and mammalian biomass (Hasenson et al., 1985; Lundin et al., 1986; Scheirer et al., 1987). A close correlation between the intracellular ATP concentration of bakers yeast and the yeast cell mass during exponential growth has been reported (Siro et al., 1982). Similarly, a close correlation (r=0.98) was shown between the number of LNCaP-r animal cells and the amount of ATP (Hasenson et al., 1985). In this case the cell specific ATP content was relatively constant at 20 fmol/cell during exponential growth, decreasing to around 15 fmol/cell during lag and stationary growth.  Such growth state dependent  changes in intracellular ATP were reported by Chapman et al. (1971) during a single cell cycle of synchronized Chinese hamster cells. The ATP pool increased from a low of 4 fmol/cell in early G l phase to 8 fmol/cell in G2. It was suggested that these changes in cell specific ATP likely result from changes in the cell volume. Changes in mean cell volume occur during batch growth of mammalian cells (Frame and Hu, 1990), suggesting  12  that if intracellular ATP remains proportional to cell volume rather than cell number, ATP would provide a better estimate of viable biomass.  2.3  Indirect Estimates of Biomass  2.3.1  Nuclear Magnetic Resonance Nuclear magnetic resonance (NMR) provides a non-invasive method of monitoring  23 the growth of animal cells. A technique using  Na NMR was developed to monitor the  growth of hybridomas within the extracapillary space of hollow fiber bioreactors (Mancuso et al., 1990). Exclusion of sodium ions from mammalian cells resulted in lower intracellular sodium concentration than in the medium. The total volume of cells within the bioreactor was calculated from the difference between the sodium concentration within the bioreactor and the medium. This method should assay only viable biomass since loss of membrane integrity rapidly results in a uniform distribution of sodium between the medium and the nonviable cells. Biomass estimations using this technique would also 23  accommodate changes in the mean cell volume.  Q  Na NMR is a relatively insensitive  technique, requiring concentrations of at least 10 cells/mL to obtain estimates of biomass within 1 h. Though expensive and relatively insensitive, ^'Na NMR may provide useful estimates of viable biomass in very high density immobilized systems such as hollow fiber bioreactors. 2.3.2  Metabolic Rates In the absence of more direct methods of biomass estimation, oxygen and glucose  uptake rates have frequently been used to estimate cell number within a bioreactor. Oxygen uptake rates were employed to estimate the biomass of numerous cell lines derived from human, monkey, hamster, chicken and insect sources that were immobilized on a ceramic matrix (Lydersen et al., 1985) or of FS-4 and ST cells grown on microcarriers (Fleischaker et al., 1981). The relationship between the oxygen uptake and  13  cell concentration has been studied for numerous hybridoma cell lines (Boraston et al., 1984; Wohlpart et al., 1990).  Oxygen uptake rates are influenced by a number of  environmental and growth related factors. The rates of oxygen utilization in hybridoma and ST cells were strongly affected by changes in media glucose concentrations (Frame and Hu, 1985; Siano and Mutharasan, 1991). Cell specific oxygen uptake rate is also influenced by cell density. An increase in hybridoma cell concentration from 10 to 10 cells/mL was accompanied by a 75% decrease in specific oxygen uptake (Wohlpart et al., 1990), likely due to decreased mean cell volume and growth rate as cell concentration increased. Glucose uptake rate has been shown to be regulated by its own concentration. The utilization of glucose by human diploid fibroblasts was 14-fold lower at 80 pM than at 5 mM (Zielke et al., 1984).  2.3.3  Optical Density Turbidity or optical density is commonly used to measure biomass in bacterial and  yeast cultures (Thatipamala et al., 1994). There has been little use of this method to estimate mammalian biomass. Optical density measurements are inherently useful only for biomass estimation of suspension cultures. On-line biomass sensors measuring absorbance (Konstantinov et al., 1992), light backscatter (Merten et al., 1987) and the combination of both (Cerex MD, USA and DKK, Tokyo, Japan) have been developed. Devices based on backscatter and absorbance provided a linear response over a limited concentration range, Ixl0 -2xl0 cells/mL in the case of backscatter (Merten et al., 1987). There has been 6  6  little published data describing the effectiveness of the two latter devices in mammalian culture, but in a recent review Konstantinov (1994) describes them as being potentially unreliable due to a complex mechanical sampling and debubbling systems.  14  Like measurements of forward scatter, measurements of optical density may partly take into account changes in cell volume. They do not, however, make any distinction between viable and nonviable cells, and thus are most useful when culture viability is high.  2.3.4  Acoustic Resonance Densitometry Based on the principle of mass damping of a vibrating resonant body, acoustic  resonance densitometry (ARD) has been used for the on-line monitoring of microbial (Blake-Coleman et al., 1986) and mammalian (Kilburn et al., 1989) biomass in batch, fedbatch and perfusion systems. ARD measurements are relatively insensitive, requiring a minimum of 1x10^ cells/mL for accurate measurements.  The described system was,  however, able to provide reproducible estimates of cell density up to 8x10  cells/mL in  suspension cultures of human lymphoma and hybridoma cell lines (Kilburn et al., 1989). Biomass estimates using ARD showed growth rate dependent variations. Because variations in growth rate would likely have been accompanied by changes in mean cell volume, it was suggested that ARD measurements represent estimates of total cell mass rather than cell number. ARD estimates of biomass are also unable to distinguish between viable and nonviable cells. Both viable and nonviable cells should contribute to changes in relative density, but because nonviable cells are both smaller and have an internal composition much closer to that of the surrounding medium, they should have a lower contribution to estimates of biomass.  Finally biomass estimates using ARD require a  relatively complex sampling system, which includes some means of obtaining samples of the cell suspension as well as samples of cell free medium.  2.3.5  Conductivity Blute et al. (1988) suggested the use of electrical conductivity measurements to  estimate the biomass within a high density hollow fiber bioreactor. As with cell volume measurements using electronic cell counters, the proportion of a known volume enclosed 15  by membranes can be calculated based on decreased conductivity compared to cell free medium. Although this method is relatively inaccurate, it has been reported effective at g  densities of 2x10 cells/mL. Estimates of biomass made using conductivity measurements likely include both viable and nonviable cells, but the volume fraction that is contributed by nonviable cells will be smaller. 2.3.6  Electrochemical Methods The relationship between the current generated between a platinum anode and  silver peroxide cathode and the cell concentration in a medium has been described by Sakato et al. (1981). Although the theoretical basis for this relationship is not entirely clear, a linear relationship between the cell concentration of L-1210 cells and the electrode current was observed (Sakato et al., 1981). A steam sterilizable probe employing this principle to estimate the biomass within a bioreactor has also been reported (Konstantinov et al., 1994). To date there have been few applications of this technology to cell culture, and there is little additional information available as to its limitations.  16  3  MATERIALS AND METHODS  3.1  Analytical Techniques  3.1.1  ATP Extraction ATP was extracted from suspensions of hybridoma, BHK and CHO cells by  adding an equal volume of freshly prepared extraction solution (5% trichloroacetic acid (TCA), 6 mM ethylenediaminetetra-acetic acid (EDTA)) immediately following sampling. The samples were then vortexed rapidly for 5 s and incubated for 15 min at room temperature. Extracted samples were stored at -70 °C until the ATP concentration was analyzed. To extract ATP directly from BHK-21 cells attached to the bottom of 6-well culture dishes, the medium was replaced with 2 mL of TCA extraction solution. After 15 min incubation at room temperature with occasional agitation, the extraction mixture was transferred to a microfuge tube and stored at -70 °C until analysis of the ATP concentration.  3.1.2  ATP Analysis ATP concentrations were measured using a bioluminescence assay kit (Bio-orbit)  in which light is generated as a byproduct of the ATP dependent conversion of luciferin to oxyluciferin by the enzyme firefly luciferase.  The amount of light produced by this  reaction is proportional to the amount of ATP in the sample or standard. The extracted samples were first diluted with tris acetate buffer (0.1M tris(hydroxymethyl)aminomethane with 2 mM EDTA, pH 7.75) in order to fall within the linear range of the assay (10"^ to 10"^ M ATP). A minimum 50 x dilution was used to avoid inhibition of the assay by TCA in the sample. The diluted samples (50 pL) were added to cuvettes containing 350 uL tris acetate buffer and 100 uL of monitoring reagent (luciferase, luciferin and bovine serum albumin in tris acetate buffer; LKB Wallac, Turku, Finland).  The luminescence was measured using an LKB Wallac 1250 Luminometer 17  attached to a potentiometric chart recorder (Goertz Metrawatt, Vienna, Austria). Following each determination, 10 uL aliquots of known ATP concentration were added to the cuvettes to serve as internal standards for the determination of the ATP concentration.  3.1.3  Cell Enumeration and Volume Measurements Cell numbers and volumes were measured using an Elzone 280PC electronic  particle counter with a 76 um diameter orifice (Particle Data Inc., Elmhurst IL). Calibration was performed with 10.2 and 20 um diameter latex beads (Particle Data). Samples were diluted to less than 2 x 10^ cells/mL with 0.2 um filtered, phosphate buffered saline (PBS). Triplicate counts of between 3000 and 7000 cells were averaged to estimate the total cell number. The arithmetic mean cell volume of each cell population was calculated from the analyzed size distribution. BHK-21 cells were pipetted several times with a 1 mL pipette to disaggregate cell clumps prior to analysis. The proportions of viable and nonviable cells were determined from hemocytometer counts with trypan blue staining (Sigma, St. Louis MO). The viable cell concentrations were calculated from the product of the total electronic particle count and the fraction of viable cells.  3.1.4  Nuclei Counts Nuclei counts were performed according to the method of van Wezel (1967).  One mL of cell suspension was centrifuged at 500g, the supernatant was removed and 1 mL of reagent was added (0.1% w/v crystal violet, 1% v/v Triton X-100 (BDH) and 0.1 M citric acid (BDH)). The cell suspension was then vortexed for 30 s at the maximum rate, and immediately counted using a hemocytometer.  3.1.5  Flow Cytometry The cells were concentrated by centrifugation at 200g for 10 min and resuspended  in the culture supernatant at 4 x 10^ viable cells/mL. Fluorescein diacetate (Sigma) was  18  added to the cell suspension at 0.4 ug/mL 3 min prior to sorting. Fluorescein diacetate is converted tofluoresceinby cellular esterases, and results in viable cellsfluorescinggreen. Nonviable cells do not fluoresce. Cells were sorted into two populations, fluorescing (viable) and non-fluorescing (nonviable), byflowcytometry (FACStar, Becton-Dickinson, Sunnyvale CA) at room temperature using PBS as the sheath fluid. Samples of the sorted populations were extracted with 5% TCA containing 6 mM EDTA, or counted and the volume distributions determined. The cells were exposed to PBS for less than 20 min prior to extraction of the nucleotides. Controls of unsorted cells were resuspended at room temperature in PBS for 20 min prior to extraction with TCA. These cells maintained the same mean volume and intracellular ATP concentration as cells isolated directly from the culture.  3.1.6  Additional Analytical Methods Media glucose concentration was monitored using a Glucose Analyzer 2  (Beckmann, Fullerton CA). pH was measured using a Model 168 pH/Blood Gas Analyzer (Ciba Corning, Richmond Hill ON). To determine cell and supernatant associated LDH activity, a sample of the culture was diluted to less than 4xl0 cells/mL in 0.1% saponin (J.T. Baker Chemical Co., 6  Phillipsburg NJ). 500 uL of the diluted sample was then further diluted 1:1 with 0.1% saponin solution. A second culture sample was centrifuged for 5 min at lOOOg to remove suspended cells, and 500 u.L of the supernatant was diluted 1:1 with 0.1% saponin. Samples were then frozen at -70 °C until analyzed. Assays of LDH activity in these samples were performed using an enzymatic LDH diagnostic kit (Sigma, St Louis MO).  19  3.2  Tissue Culture  3.2.1  Batch Cultures of Hybridoma Cells The hybridoma 2E11 cell line (Ziltner et al., 1988) produces a murine IgG  monoclonal antibody against IL-3.  It was grown in Dulbecco's Modified Eagle Medium  (DMEM) containing 4.5 g/L glucose, and supplemented with 10% fetal bovine serum (Gibco/BRL, Burlington ON). Cells were grown in 500 mL spinner flasks (Bellco Glass, Vineland NJ) operated at 40 RPM or in 850 cm^ roller bottles (Falcon, Becton Dickinson, Lincoln Park NJ) rotated at 0.5 RPM (Bellco).  Spinner flasks were incubated at 37 °C in a 5% C O 2 and 95%  humidity atmosphere.  Roller bottles were aerated with 5% C02/balance air after each  sampling and incubated at 37 °C.  3.2.2  Batch Cultures of BHK Cells A BHK-21 cell line (Funk et al., 1990) stably transfected to produce the N-  terminal human transferrin half molecule under the control of a mouse metallothionein promoter, was used as a model adherent cell line. Cells were maintained in 80 cm T2  flasks (Nunc, Rsokilde Denmark) containing DMEM with 5 or 10% fetal or newborn bovine serum (Gibco/BRL, Burlington.ON). T-flasks and culture dishes were incubated in a 5% C O 2 , 37°C and 95% humidity incubator (Forma, Marietta OH).  Cells were  detached from the surface of the culture dishes by 2-3 min treatment with 0.5 mL of 0.25% trypsin, 1 mM EDTA (Gibco/BRL, Burlington ON).  The trypsin was then  neutralized by the addition of 0.5 mL DMEM with 5% serum. Growth curves were performed using 6-well culture dishes (Nunc, Rsokilde Denmark). Experimental batch cultures were inoculated using BHK cells from a confluent T-flask which were diluted to 1.3 x 10^ cells/mL with fresh DMEM supplemented with 10% newborn calf serum. Three mL of the cell suspension was added to each of the wells in several 6-well plates. 20  Samples were generally taken on a daily basis. The medium was removed from 3 wells and a sample from each well was saved for later analysis of the glucose concentration. The cells attached to the bottom of each of the 3 wells were washed with PBS at room temperature and then removed using 0.5 mL of trypsin-EDTA solution. The trypsin was neutralized with 0.5 mL spent DMEM obtained from the same plate, and the cell suspension was repeatedly pipetted up and down until no cell clumps could be seen by microscopic examination of the suspension.  The cell number and mean cell volume of  each sample were analyzed using an electronic particle counter. Viability was determined by trypan blue exclusion. A sample of each suspension (0.5 mL) was extracted for ATP analysis. In three additional wells the ATP was directly extracted from the attached cells (Section 3.1.1).  3.2.3  Continuous Cultures of Hybridoma Cells Hybridoma 2E11 cells were grown in a glucose limited continuous stirred tank  bioreactor to investigate the effects of pH and oxygen on intracellular ATP. A 2 L water jacketed glass L H bioreactor (Slough, UK.) was operated with a 1 L working volume. Agitation was provided by a fully submerged marine type impeller and a Rushton surface impeller attached to the same shaft rotated at 100 RPM. Maintenance of pH, dissolved oxygen and temperature within the vessel was performed automatically by Anglicon 2000 Solo (TM) digital controllers. The pH within the bioreactor was monitored using an Ingold pH probe and maintained by the control of C O 2 addition to the head space of the bioreactor. Dissolved oxygen was monitored using an Ingold polarographic oxygen probe and controlled by adding a mixture of air and pure O2 to the head space. Nitrogen was added continuously to the head space to minimize overshooting of the set point. The output from each of the controllers was datalogged with an IBM-AT compatible computer running a BASIC program.  21  Cells were grown in DMEM with an initial concentration of 2 g/L glucose and 5% newborn calf serum (Gibco/BRL, Burlington ON). Medium was prepared 10 L at a time and maintained in a 4 °C water bath. Medium was fed into the bioreactor at dilution rates ranging from 0.2-0.4 day" using a Masterflex peristaltic pump. A second peristaltic pump 1  was operated continuously at a higherflowrate to remove medium from the bioreactor via a tube placed at the liquid surface to maintain a constant culture volume of approximately 1 L. Samples were removed aseptically from the bioreactor using a syringe. The lower section of the sample tube was clamped off and washed with 70% ethanol after each sample was taken. After sampling, the line below the clamp was filled with 70% ethanol. Above the clamp sterile air filled the sample line.  3.2.4  High Density Perfusion Culture A high density perfusion culture experiment was undertaken in collaboration with  Genentech Inc. (South San Francisco CA). The continuous perfusion system consisted of a 1.5 L working volume Applikon bioreactor (Applikon Dependable Instruments, Schedam, Netherlands) fitted with a novel 10 mL acoustic filter which aggregated cells in the outflow of the reactor and returned them to the culture. The filter consisted of a precisely constructed 10 mL volume acoustic resonator (15x15x45 mm) mounted vertically above a tube with a 16.5 mm inner diameter (Figure 1). The acousticfilterconsisted of a 4-layer composite resonator system composed of the bi-layer transducer, the medium and the reflector (Trampler et al., 1994). The transducer and reflector were sufficiently parallel (±0.05 mm tolerance) to fulfill the boundary conditions of standing resonance waves at the terminating resonator planes (Nowotny and Benes, 1987), and thus produce a high quality resonance field and minimal acoustically induced streaming. The composite transducer was constructed from a 15x25 mm and 1 mm thick PZT piezoelectric ceramic (Hoechst Ceramtech, Germany) mounted on a  22  2.7 mm borosilicate glass carrier. The PZT elements were electrically connected via the top and bottom surface and powered in the 2.1-2.17 MHz frequency range by a custom built automatically tuned high frequency power supply (UCCS-04, SonoSep Biotech Inc., Richmond B.C.).  To minimize convection within the chamber the transducer was  continuously cooled using a 10 L/min flow of room temperature air.  VCO Fitting  Reflector  Reflector Side Clamp Retaining Screw  Figure 1:  Exploded drawing of acoustic cell retention system.  Agitation of the bioreactor was provided by a marine impeller.  A partially  submerged rushton turbine enhanced surface aeration. Peristaltic pumps (Cole-Parmer,  23  Chicago IL) were used to add fresh medium as well as remove spent medium and product. Cell suspensions were drawnfromthe bioreactor through the inflow/settling tube and into the resonance chamber by a peristaltic pump (Watson-Mariow). Acoustically aggregated cells settled by gravity against the medium flow back to the bioreactor. The clarified medium was pumpedfromthe top of the acousticfilterinto the waste medium reservoir. Medium addition was controlled automatically through the use of a conductivity based level sensor. The bioreactor pH and oxygen were measured using Ingold probes (MettlerToledo, Urdorf, Switzerland) attached to an Applikon Bio Controller (Schedam, Netherlands). pH was controlled by the automatic addition of CO2 or base. Initially the base used was 0.3 M NaOH, but after 8 days of culture this was changed to 7.5% NaHC0 to prevent aggregation resulting from NaOH addition. The dissolved oxygen 3  concentration was initially regulated by controlling the head space O2 concentration, but after 5 days this was no longer sufficient and the culture was sparged with pure oxygen through a sintered stainless steel sparger (Applikon). A schematic representation of the experimental setup is shown in Figure 2.  24  Base for pH Adjustment Bioreactor Controller •  CO,  0  2  Air  >  "Waste Reservoir  " 20L  Gas Control Valves  Measured Variables  Feed Medium Reservoir  Liquid Level Dissolved Oxygen H Temperature  20L  P  Figure 2:  Schematic representation of the perfusion bioreactor system used.  These experiments used a CHO cell line producing y-interferon that was obtained from Genentech.  Cells were adapted for approximately 2 months in spinner flasks  (Bellco) to CHO-S-SFM medium (Gibco) prior to starting the perfusion culture. The CHO-S-SFM medium (Gibco) was supplied pre-sterilized in 10 L blood bags and transferred 20 L at a time into the perfusion culture medium reservoir. At this time 0.002% silicone antifoam C (Sigma) was added.  After 185 h (20 L medium), an  additional 2.5 g/L glucose (6 g/L total) (Sigma) and 3 mM glutamine (8.5 mM total) (Gibco) were added to the medium. After 286 h (40 L of medium), the amount of added glucose was further increased to 3.5 g/L (7 g/L total).  25  The culture was sampled daily. Cell concentration and viability were determined using dye exclusion assays and/or nuclei counts as required.  Intracellular ATP and  glucose concentrations and osmolality were also determined from the daily samples. Samples were also taken to allow measurement of LDH activity, but an error in the sample treatment protocol resulted in the loss of samples taken prior to 200 h of culture. The osmolarity amino acid composition and concentration of y-interferon in the culture supernatant were determined by Genentech.  26  4  EXPERIMENTAL RESULTS  4.1  ATP Extraction with TCA  4.1.1  Concentration of TCA Acids with strongly chiotropic anions such as TCA were previously shown to be  the most effective for extracting intracellular ATP (Lundin and Thore, 1975). A range of TCA concentrations was investigated to determine the optimum level for the extraction of ATP from hybridoma 2E11 cells (Figure 3). Multiple samples of hybridoma cells each containing approximately 3.5x10^ cells were extracted using 0.5, 1, 1.5, 2.5, or 5% w/v TCA. In all cases the EDTA concentration was maintained at 6 mM. TCA concentrations above 1% gave identical results within statistical error, while concentrations of 0.5% TCA either did not completely extract the ATP from the sample or did not completely neutralize ATPase activity (Figure 3). A TCA concentration of 2.5% was chosen for use in all subsequent extractions. This choice was consistent with Lundin (1986) who, after a series of extractions of LNCaP-r cells using concentrations of TCA between 0.5 and 10%, reported that the highest yield of ATP was obtained using 2.3% TCA. Lundin, however, reported a greater dependence of TCA concentration on the ATP recovery than seen in Figure 3. The cell specific ATP of approximately 4 fmol/cell obtained in this experiment was consistent with that obtained in subsequeniexperiments.  27  V  LY. CL  £ 0.400.200.00 -j 0  1—  ;  1  1  1  1  1  2  1  1  1  3  4  1  1  5  Percent TCA (w/v) Figure 3:  4.1.2  Total ATP recovered following extraction of 3.5xl0 concentrations. Error bars indicate one standard deviation.  5  cells using different T C A  Dose Response Samples containing from 0.4 to 8.6 million hybridoma cells were extracted with  1 mL of 2.5% TCA and the ATP recovery determined. Figure 4 shows that the ATP was proportional to the total number of cells extracted. The calculated cell specific ATP is also shown in Figure 4. The mean cell specific ATP of 4.7 fmol/cell had a coefficient of variation (C. V.) of less than 6% and was consistent with the range of values obtained in previous and subsequent experiments.  28  Cells Extracted (millions) Figure 4:  Total ATP recovered - • - and cell specific ATP - • - following extraction of different numbers of cells using 2.5% TCA. Error bars indicate one standard deviation.  4.2  Batch Growth Curves of Hybridoma Cells  4.2.1  Changes in Cell Specific ATP Hybridoma 2E11 is a robust, fast growing, anchorage-independent cell line. Initial  experiments to evaluate the usefulness of ATP as an indicator of biomass were performed using this cell line because of the relative ease of handling and growing suspension cells compared to anchorage-dependent cell lines.  Two methods of culture were used to  examine the possible effect of culture techniques on the intracellular ATP concentration. Initial growth curve experiments were performed using cultures grown in roller bottles,  29  while in later experiments the cells were grown in spinner flasks which more closely resembled the stirred tank bioreactors widely used in industry. The total, viable and nonviable cell concentrations from a batch roller bottle culture of 2E11 cells are shown in Figure 5a. The cells were inoculated at lxlO cells/mL 5  and attained a maximum specific growth rate of 0.03 h"V The cell numbers increased for 110 h, after which the viable numbers declined and the nonviable numbers increased rapidly.  The total cell numbers increased for approximately 40 h after the end of  exponential phase indicating a continued low rate of cell growth during the decline phase. The cell concentration and viability profiles of the spinner culture were similar to that of the roller bottle. The ATP concentration in batch cultures (Figure 5b) followed a trend similar to that of the viable cell concentration (Figure 5 a). For example, in the spinnerflaskthe ATP concentration increased during the exponential phase and peaked approximately 90 h after inoculation. The decline in ATP concentration from 90 to 110 h appeared inconsistent with the continued increase in viable cell number. Beyond 110 h, as the viable cell numbers declined, the ATP concentration also dropped. The cell specific ATP content based on the viable cell count (Figure 6) varied during the course of both roller bottle and spinner batch cultures. Cell specific ATP was highest during the exponential phase of growth and decreased by more than 2-fold in the decline phases of the batch cultures. The relatively low initial values of the cell specific ATP reflected that of the inoculum. Similar cell specific ATP results were obtained in 4 other batch experiments. Since whole culture samples were extracted to determine cell associated ATP, several samples of culture supernatant were analyzed to determine the background level of ATP. These culture supernatant samples contained less than 0.01 nmol ATP/mL, while the cell suspensions contained from 0.4 to 8 nmol/mL (Figure 5b). Thus >98% of the ATP measured in the culture samples was cell associated. 30  2 0 0  1 0 - ,  o E c c o c  <D O  c o O  0  (b)  Figure 5:  1 0 0  2 0 0  Time (h)  (a) Total - A - , viable -A- and nonviable - • - cell concentrations over the course of a batch hybridoma roller bottle culture, (b) Total ATP concentration over the course of spinner -0and roller - • - hybridoma batch cultures.  31  8-,  .8 o  64  E  < CD D_  cn  s  2-1  0  Figure 6:  I  50  ->  1—  100  150  ~200  Time (h)  Viable cell specific ATP over the course of spinner cultures. Error bars indicate one standard deviation.  and roller - • - hybridoma batch  The levels of cell specific ATP measured in the hybridoma 2E11 cell line were within the range of cellular ATP reported for other cell lines: HeLa, 9 fmol ATP/cell, mouse macrophages, 1 fmol ATP/cell (Thore et al., 1975), and guinea pig alveolar macrophages, 5 fmol ATP/cell (Cazin et al., 1990). Chinese hamster cells content varied from 4 to 8 fmol ATP/cell (Chapman et al., 1971). Variations in cell specific ATP have been reported in both bacterial (Miovic and Gibson, 1973; Knowles, 1977; Lundin, 1982) and mammalian cells (Chapman et al., 1971; Lundin et al., 1986).  32  4.2.2  Cell Volume Variations The average cell size of batch cultured hybridoma cells was the highest prior to  and during the exponential growth phase and decreased during the stationary and decline phases (Figure 7). Hybridoma cell volumes have previously been reported to range from 1,500 to 650 pm over the course of batch cultures (Frame and Hu, 1990; Ramirez and 3  Mutharasan, 1990). The cell volumes at the start of a batch were also influenced by the inoculum state.  When cells in mid-exponential phase were used to inoculate an  experimental batch culture (i.e. t = 40 h in Figure 7), the initial volume was larger than if late exponential phase cells (t = 80 h in Figure 7) were used as the inoculum.  1500  n  0H 0  Figure 7:  <  1  50  '  1  100 Time (h)  1  1  150  •  1  200  Variations in hybridoma 2E11 mean cell volume over the course of spinner - • - and roller - • - hybridoma batch cultures. Viabilities of less than 90% are indicated next to the data point.  33  4.2.3  Intracellular ATP Concentration The cell specific ATP profiles during batch cultures (Figure 6) were remarkably  similar to the mean cell volume profiles determined by electronic particle sizing (Figure 7). To investigate whether changes in the cell specific ATP were due to volume variations, the average intracellular ATP concentrations was calculated from the ratio of the cell specific ATP content and the mean cell volume.  Intracellular ATP concentration was nearly  constant throughout the course of batch cultures in either the spinner flask or roller bottle (Figure 8). The mean intracellular ATP concentration of 4.2 mM determined from the data of the two experiments had a C V . of only 14% despite the fact that the culture was subjected to a wide range of culture conditions, and that the calculation was dependent on 3 independent measurements (viable cell number, cell volume and ATP content).  10-.  8-  c  2-  0 -I  i 0  i  i 50  '  i 100  '  1  150  i 200  Time (h)  Figure 8:  Calculated intracellular ATP concentration over the course of spinner -•- and roller -ohybridoma batch cultures. Only the data from samples with greater than 90% viability are shown. Error bars indicate one standard deviation.  34  The average cell volume decreased in low viability cultures presumably due to the smaller size of nonviable cells. To demonstrate this, cells were sorted electronically into viable and nonviable populations using a fluorescence activated cell sorter with fluorescein diacetate (FDA) as a vital stain. The mean cell volume was then determined from the viable fraction.  The viability of the sorted samples prior to sorting is indicated on  Figure 7. The volume distribution of a culture sample at 48% viability before and after sorting for viability on the basis of FDA fluorescence is shown in Figure 9. The distinctive volume distribution of viable and nonviable cellfractionscan be clearly seen on this figure. The intracellular ATP concentrations of the 37 and 48% viable samples were calculated to be 3.8 and 4.1 mM respectively, close to the average (4.2 mM) of the greater than 90% viable cell samples from the spinner and roller cultures (Figure 8).  900-1  800-  Figure 9:  Volume distribution of hybridoma 2E11 cells before -A- and after - • - sorting for viable cells based on FDA fluorescence.  35  4.3  Batch Growth Curves of BHK Cells A number of batch growth experiments were performed using a BHK-21 fibroblast  cell line that has been stably transfected to express the N-terminal domain of human transferrin. BHK cells were chosen to investigate whether the ATP concentration within this anchorage dependent cell line would also remain constant during batch culture.  4.3.1  Replicate Culture System Parallel BHK cultures were grown in six well plates to allow multiple cell samples  to be taken during batch growth. At each time point three cultures were sacrificed. The C V . between the cell numbers in replicate samples averaged 5% and ranged from 2% to 14%. Cells were inoculated at 1x10^ cells/well in 3 mL of DMEM medium and attained a growth rate of 0.77 day'l during the exponential phase (Figure 10).  Cells became  confluent after 75 h growth (9x10 cells/cm ), but the cell number continued to increase at 4  2  a lower growth rate without any appreciable decrease in viability until the last sample was taken after 161 h of growth. With confluent cultures it was difficult to obtain the single cell suspensions required for accurate determinations of cell number and volume. After trypsinization, the cell suspension was repeatedly pipetted to break up cell aggregates in the suspension. The amount of shear required to obtain a single cell suspension continued to increase as the confluent culture began to form multiple layers of cells.  36  6-,  H—i—i—i—i—i—i—i—i—•—i—i—i—i—i— —i— —i  0  1  0  20  40  60  80  1  100 1 2 0 140 160 1 8 0  Time (h)  Figure 10:  4.3.2  Viable cell number of BHK cells grown in six well plates. Each point represents the average of three replicate wells. Error bars indicate one standard deviation.  Total Culture ATP In thefirstexperiment using BHK cells, ATP was extracted either directlyfromthe  cells attached to the bottom of the wells or from cells that had been trypsinized and  37  resuspended. The ATP recovery using these two extraction procedures are compared on Figure 11. Trypsinized samples consistently yielded a lower ATP recovery than directly extracted parallel cultures (Figure 11). The difference between the ATP recovery using these two extraction methods increased with culture time to a maximum of 20% at 161 h of culture time. The difference in ATP recovery by these two protocols at 161 h probably results from exposure of the resuspended cells to trypsin and shear forces. As the culture density increased, longer enzyme exposures of 5-7 min and higher shear forces were required to obtain single cell suspensions. reducing, the recovery of viable cells.  In  addition, the sensitivity of the cells to these stresses likely increased as culture conditions became less favorable for growth. 2.5-,  o.o  H—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i—i  0  Figure 11:  20 40 60  80 100 120 140 160 180 Time (h)  ATP recovered from cells grown in six well plates. ATP was extracted directly from attached cells - • - and from cells resuspended with trypsin Each point represents the average of three replicate wells. Error bars indicate one standard deviation.  38  4.3.3  Cell specific ATP The cell specific ATP (Figure 12) decreased over the course of culture in both the  directly extracted and trypsinized parallel cultures. In directly extracted cultures the cell specific ATP decreased by almost 50% while in trypsinized cultures the cell specific ATP decreased by 60%. The greater variation of trypsinized samples probably resulted from variations in the resuspension treatment of replicates. The cell specific ATP in parallel BHK batch cultures were similar both in trend and magnitude to that of hybridoma cells grown in batch culture (Figure 6). 8.0-,  T  Q> 2.0O 1.00.0-| • 1 0 20  r—I  40  1  1 60  i  1 80  1  1 1 1 1 I r—| 1 1 100 120 140 160 180  Time Figure 12:  Cell specific ATP from cells grown in six well plates. ATP was extracted directly from attached cells - • - and from cells resuspended with trypsin Each point represents the average of three replicate wells. Error bars indicate one standard deviation.  39  4.3.4  Intracellular ATP Concentration The average cell volume of BHK cells decreased from 2100 to 1500 urn (30%) 3  over the course of culture. (Figure 13). The decrease in the average cell volume (Figure 13) and the decrease in cell specific ATP (Figure 12) showed the same trend during batch cultures.  2200-1  I 1400 -J 1200 H  1000 I ' 0 20  Figure 13:  40  60  i  •  I  80 100 120 140 160 180 Time (h)  Mean cell volume of cells resuspended with trypsin. Each point represents the average of three replicate wells. Error bars indicate one standard deviation.  40  As with previous experiments, the change in cell specific ATP was largely a consequence of the change in the mean cell volume of the cells. The mean intracellular ATP concentration (3.25 mM) showed little variation during the course of the cultures (Figure 14), particularly when the cells were extracted in-situ. In this case the mean intracellular ATP concentration varied by a maximum of 20%, less than half of that calculated for cell specific ATP.  The intracellular ATP concentration of trypsinized  samples showed considerably more variation (20%) than the directly extracted samples.  4H  3H  ro M 2 CD O CO  1H  0  0  -r~ 25  ->—r  —f—  50  75  —i—>—i—~i — r  100  125  150  175  Time (h) Figure 14-  Intracellular ATP from cells grown in six well plates. ATP was extracted directly from attached cells - • - and from cells resuspended with trypsin -O-. Each point represents the average of three replicate wells. Error bars indicate one standard deviation.  41  The difference in ATP recovery obtained using these two sampling procedures increased with culture time, as would be expected if the technique for removing the cells from the surface of the dish were resulting in cellular damage. In subsequent BHK batch cultures, ATP was extracted directly from the cells without trypsinization.  4.4  Stability of Intracellular ATP Following Cell Death To demonstrate  that nonviable cells do not contribute significantly to  measurements of intracellular ATP, a population of nonviable cells obtained by cell sorting was analyzed for ATP. These cells contained less than 0.01 fmol ATP per nonviable cell. The stability of ATP following cell death was also investigated by freezing cells in liquid nitrogen. Following thawing, cell specific ATP in suspension samples decreased by two orders of magnitude, from 3.0 to 0.04 fmol/cell, in less than 20 min (Figure 15). This decrease in cell specific ATP was not accompanied by any increase in ATP concentration in the cell-free medium sampled at 22 min.  42  3.0 A  6  5  10  15  20  25  Time (min) Figure 15:  4.5  Cell specific ATP after freezing and thawing the cell suspension. The initial sample was taken prior to freezing.  Comparison of Different Methods for Detecting Changes in Cellular Viability An experiment was performed to compare the analysis of intracellular ATP with  several other standard techniques for detecting changes in cell viability. Hybridoma 2E11 cells were washed and resuspended in 37°C PBS in a spinner flask at approximately 7x10^ cells/mL. Depriving the cells of nutrient medium was expected to result in a gradual decrease in the viability of the cell suspension. The suspension was sampled at approximately 30 min intervals and the viability estimated using trypan blue staining, pellet  43  or supernatant associated lactate dehydrogenase (LDH) activity, and cell or supernatant ATP analysis. The results of this experiment are shown on Figure 16.  Figure 16:  (a) Viable cell concentration -A- by trypan blue exclusion, cell associated ATP and supernatant ATP - V - and (b) cell associated LDH activity - T - and supernatant LDH activity -V- after resuspension of hybridoma cells in PBS.  Cell viability as determined by trypan blue and the cell associated ATP decreased measurably after 6 hours. The rate of ATP decrease was approximately proportional to  44  the decline in viable cell number estimated using the trypan blue exclusion assay (Figure 16a).  Supernatant ATP concentration were two orders of magnitude less than ATP  concentration in the cells from an equivalent culture volume. This would be expected from the previously observed half life of ATP releasedfromlysed cells. Supernatant LDH activity is also used to estimate cell death particularly when individual cells cannot be recovered or otherwise enumerated (Piret and Cooney, 1990). After 6 h in PBS estimates of viability calculated from changes in cell associated LDH were considerably higher than estimates obtained either from the dye exclusion assay or from changes in the ATP concentration (Figure 16). While viability estimated by trypan blue exclusion decreased by almost 60% over the 9.5 h of this experiment, the level of cell associated LDH changed by only 25%. In addition the decrease in cell associated LDH activity was apparently not the result of its release from the cells as there was no increase in the supernatant LDH activity (Figure 16b).  Samples of the cell suspension taken  several days later, however, showed the bulk of formerly cell associated LDH to be present in the supernatant. The long retention of LDH within the cells probably resulted from the low shear environment in the spinner flask allowing cells to remain intact long after cell death had occurred. One would expect cell lysis to occur more rapidly in a harsher bioreactor environment.  The smaller size of the trypan blue molecule likely  allowed it to penetrate damaged cell membrane more rapidly than the much larger LDH molecules could exit. Even so delays in the release of the enzyme by the cells results in a tendency to over estimate cellular viability, particularly in cases where the culture viability changes quickly. In contrast ATP extracted from immobilized cells gives rapid indication of changes in cell viability.  4 5  4.6  Continuous Culture During batch cultures, the cell environment changes constantly. With hybridoma,  2E11, pH typically decreased from 7.6 to 6.8, glucose from 24 to 6 mmol/L, while lactate increased from 0 to 28 mmol/L and ammonium from 0 to 3 mmol/L.  This range of  changes in culture conditions did not significantly affect intracellular ATP concentration. However, in high density cultures, cells might be exposed to a more extreme range of culture conditions and these could potentially influence cellular ATP concentrations. To distinguish any possible changes in intracellular ATP resulting from variations in a single variable from those arising from changes in other culture conditions, several experiments were undertaken using a continuous culture system. The continuous culture system also allowed cells to be transiently exposed to extremes of pH and dissolved oxygen that might be encountered during high density cultures. A glucose limited chemostat culture was obtained by decreasing the glucose concentration in the DMEM feed medium from 4.5 g/L to 2 g/L and setting the dilution rate at 0.3 day"*.  Under these conditions a steady-state cell concentration around  10 cells/mL was obtained at a bioreactor glucose concentration of approximately 0.2 g/L. 6  The average cell volume was determined for each time point. Viable cell volume was estimated using software to eliminate the smaller peak of the bimodal volume distribution previously shown to represent nonviable cells and cellular debris by sorting experiments (Figure 9) and by Sen et al. (1989). Culture viability determined using this method correlated with the viability determined by trypan blue exclusion assays (regression coefficient of 0.99).  4.6.1  Effect of Short Term Variations in pH The medium pH was decreased from 7.4 to 7.0 for 18 h, then to 6.7 for 9 h after  which the pH was returned to 7.4 (Figure 17a). The viability of the culture decreased  46  rapidly during this experiment, dropping from 90 to 80% during the 27 h of exposure to lower pH. Even after the pH was returned to 7.4 the measured viability continued to decrease for at least an additional 20 h (Figure 17a).  Figure 17:  (a) Culture pH — and viability and (b) recovered ATP/mL - • - and cells/mL -A- versus time during a continuous culture of hybridoma cells. Error bars indicate one standard deviation.  47  ATP and cell concentration were closely correlated during growth at pH 7.4 (Figure 17b).  When the culture pH was decreased the amount of ATP also rapidly  decreased at a rate that was related to the magnitude of the decrease in pH (Figure 17b). The amount of ATP in the culture decreasedfrom6.4 to 5 nmol/mL during 20 h at pH 7.0 and to 3.5 nmol/mL after only 9 h at pH6.7.  During the same period the cell  concentration only decreased from 1 to 0.94 million/mL. Within 20 h of returning the pH to 7.4, the ATP concentration had returned to 5.6 nmol/mL (almost identical to its initial value if the decrease in culture viability is taken into account). The viable cell number continued to decrease during this period to a minimum value of approximately 0.8 million cells/mL. During growth at pH 7.4 cell specific ATP had a C V . of 7%, almost twice that of the intracellular ATP concentration (Figure 18a). Within 4 h of lowering the culture pH to 7.0 the cell specific ATP decreased by more than 15%, while intracellular ATP concentration decreased by less than 7%, only slightly more than the C V . at pH 7.4. Within 3 h of decreasing the pH to 6.7, both cell specific and intracellular ATP concentration had decreased by 14%.  Minimum cell specific and intracellular ATP  concentrations of 3.5 fmol/cell and 2.6 mM respectively were reached after 10 h at pH 6.7, a greater than 50% decrease in cell specific and 30% decrease in intracellular ATP compared to pH 7.4 levels. Within 19 h following the return to pH 7.4 both intracellular and cell specific ATP levels were within the normal concentration range. Intracellular and cell specific ATP were markedly influenced by the transient decrease to pH 6.7 (Figure 18), although the intracellular ATP concentration was less sensitive to decreased pH than was the cell specific ATP.  This result was somewhat  surprising since pHs of around 6.7 are not substantially lower than those obtained during batch cultures. The rapid decrease in viability was also unexpected. Previous experience with the 2E11 cell line had not indicated such extreme sensitivity to medium pH. It is  48  possible that the rapid shifts in pH during these experiments resulted in this sensitivity as they did not allow the cells time to adapt to altered conditions.  Figure 18:  (a) Cell specific ATP - • - and intracellular ATP concentration - • - and (b) culture pH versus time during a continuous culture of hybridoma cells. Error bars indicate one standard deviation.  49  4.6.2 Effect of Long Term Decreased Medium pH Cells in the CSTR were grown at decreased pH for longer periods to determine whether the cells would adapt to these conditions. During this experiment the cells were exposed to pH 7.1 for 168 h, and then pH 6.8 for 122 h (Figure 19a).  8.0  r  n  .-o 7.5  o'V  100 •80  A  .Q  r  £ 7.0 H  ro  60 > c  CD O  •40 a5 CL 6.5  H  r-20  1200 1400 Time (h) Figure 19:  (a) Culture pH — and viability -v- and (b) cell specific ATP - • - and intracellular ATP concentration - • - versus time during a continuous culture of hybridoma cells. Error bars indicate one standard deviation.  50  Exposure to pH 7.1 resulted in no consistent change in viability, average cell specific or intracellular ATP compared to pH 7.4 (Figure 19b). At pH 7.1 the average cell specific ATP was 5.8 (CV. 6%) compared to 6.4 fmol/cell (CV. 7%) at pH 7.4, and the intracellular ATP concentration was 4.1 compared to 3.9 mM. The sudden drop in cell volume observed at 1170 h (Figure 21) was likely due to a calibration error and is not considered to represent any cellular change. Further decreasing the culture pH from 7.1 to 6.8 profoundly affected the cell concentration and viability of the culture. During 122 h exposure to pH 6.8 the viable cell concentration decreased steadily from approximately 1 million to less than 0.3 million cells/mL (Figure 20). The viability of the culture also decreased from 85% to less than 60% during this period (Figure 19a). Returning the pH to 7.4 was accompanied by a rapid recovery of the viable cell population within the bioreactor. An increased growth rate was accompanied by a transient but dramatic increase in mean cell volume at 1400 h (Figure 21).  As the cell concentration stabilized, the mean cell volume returned to the  steady state pH 7.4 level. Exposure of the culture to pH 6.8 caused a transient decrease in the cell specific ATP (Figure 19b) that was similar to the changes observed during short term exposure to low pH (Figure 18a). This decrease and subsequent gradual increase in cell specific ATP generally mirrored the changes in cell volume. Intracellular ATP concentration showed substantially less variation than cell specific ATP, with most of the fluctuations falling within 10% of the mean (Figure 19b).  51  Figure 20:  (a) Recovered ATP/mL - • - and cells/mL -A- and (b) culture pH versus time during a continuous culture of hybridoma cells. Error bars indicate one standard deviation.  52  2200-1  ^ 2000-1 E u  1800.  £  1600'  E  o5 O  £ 1400' 1200 H n — i i  i—|—i—i—i—i—|—i—i—i  i  |—r  (a) 8.0-1  7.5x  7.0-  Q.  6.56.0  —*—i—i—i—i—i—i—i—i—i—i—i—i—r  1000  1200 1400 Time (h)  -,—i—|—r-  1600  (b) Figure 21:  4.6.3  (a) Mean cell volume and (b) culture pH versus time during a continuous culture of hybridoma cells.  The Effect of Long Term Increased Medium pH The effect of prolonged growth at elevated pH on cellular ATP was investigated  by decreasing the headspace C 0 concentration to increase the culture pH first to 7.7 for 2  143 h, and then to 8 for 92 h (Figure 22a). Viability, viable cell concentration and ATP in the culture decreased steadily during thefirst90 h at pH 7.7 (Figure 22 a,b). Changes in  53  culture ATP during this period approximated the changes in viable cell number. Mean cell volume increased steadily by approximately 20% during the first 90 h at pH 7.7 (Figure 23). Further increasing the pH to 8 accelerated the decrease in the viability of the culture. At the end of 92 h of exposure to pH 8, the viability had decreased from 80% to 36% (Figure 22a). Mean cell volume continued to increase by an additional 8% over the same period (Figure 23a). r  , {  Figure 22:  w  '  1600  1700 1800 Time (h)  1900  100  2000  (a) Culture pH — and viability and (b) recovered ATP/mL - • - and cells/mL -A- versus time during a continuous culture of hybridoma cells. Error bars indicate one standard deviation.  54  2000 -i  ^800-\ CD  E :> 1600'  © o c  CD  1400H •  —  1200  T  (a)  1  1  r  8.5-,  8.0  J  7.5  H  x  7.01600  (b) Figure 23:  T  T  1700  1800  1900  2000  Time (h) (a) Mean cell volume and (b) culture pH versus time during a continuous culture of hybridoma cells.  Intracellular ATP concentration was relatively unaffected by the exposure to increased pH (Figure 24), showing only a 10% C V . from a mean value of 3.5 mM. The lower absolute value of intracellular ATP concentration compared to earlier experiments was attributed to the use of a new ATP standard solution in the analysis.  55  8.5-j 8.0X Q.  7.57.0(b)  Figure 24:  1600  1700  1800  1900  2000  Time (h)  (a) Cell specific ATP - • - and intracellular ATP concentration - • - and (b) culture pH versus time during a continuous culture of hybridoma cells. Error bars indicate one standard deviation.  56  4.6.4  Effect of Dissolved Oxygen Concentration on ATP The effects of decreased dissolved oxygen on intracellular ATP concentration were  also investigated in the continuous culture system.  The culture was grown to  approximately 1.2 million cells/mL in a bioreactor continuously fed with DMEM containing 2 g/L glucose.  During all previous experiments the dissolved oxygen was  automatically maintained at approximately 80% of air saturation. Over the course of this experiment the dissolved oxygen was sequentially reduced first to 40%, then to 20% and finally to 5% of air saturation (Figure 25a). Samples were taken just prior to reducing the dissolved oxygen set point, immediately upon achieving the new dissolved oxygen concentration and several times thereafter until the set point was changed.  Viability  ranged from 80 to 90% throughout this experiment with the exception of one point taken after 149 hours with a measured viability of 74% (Figure 25a).  The cell concentration  varied only slightly from the average value of 1.2 million cells/mL (Figure 25b). Intracellular ATP concentration was virtually independent of dissolved oxygen concentration within the range studied (Figure 26a). Fluctuations in the intracellular ATP concentration were typical of those seen during normal growth at dissolved oxygen concentrations of 80% and at pH 7.4.  Cell specific ATP and intracellular ATP  concentration also showed only minimal and apparently random variations that generally coincided with variations in the measured cell number (Figure 4.26a). The C.V. around the average cell specific and intracellular ATP were only 9%.  57  Figure 25:  (a) Culture dissolved oxygen concentration and viability and (b) recovered ATP/mL - • and cells/mL -A- versus time during a continuous culture of hybridoma cells. Error bars indicate one standard deviation.  58  LO  V2  o  (a) c  100  CD D)  O  n  804  60H  > 40o «  20  Q 0  (b) Figure 26:  4.7  H -r 0  -  50  100  150  200  250  Time (h)  (a) Cell specific ATP - • - and intracellular ATP concentration - • - and (b) dissolved oxygen concentration versus time during a continuous culture of hybridoma cells. Error bars indicate one standard deviation.  Biomass Estimation During a High Density Perfused CHO Cell Culture The low growth rates, cell densities and product concentrations in conventional  batch and fed batch animal cell cultures often require the use of very large bioreactors (i.e. 10000 L) for commercial production. To increase the cell density and productivity, stirred suspension bioreactors can be operated with the continuous addition of fresh medium and  59  removal of spent medium containing the product.  A number of systems have been  developed to retain cells within the bioreactor. Spin-filter systems are widely commercially available. The rotation of cylindrical filters located inside the bioreactor helps reduce the fouling of the filter surface. Nonetheless, the usefulness of these systems is limited by progressive protein and cellular fouling of the filters. The time of productive operation of a spin-filter based perfusion bioreactor is limited both by filter fouling and the progressive accumulation of nonviable cells within the suspended phase of the bioreactor (Avgerinos et al., 1990; Esclade et al., 1991; Yabannavar et al., 1992). Scale-up of spin-filters has also been problematic due to the increasing difficulty of providing sufficient filter area to sustain adequate feed rates inside larger bioreactors.  Cross-flow and sedimentation systems are also limited by  fouling and scalability problems (Velez et al., 1989; Batt et al., 1990; Broise et al., 1992; Hiilscher et al., 1992). Centrifugation systems are not widely used, possibly because of the high capital cost ($40,000 to 100,000+ per unit), complicated operation and a perceived susceptibility to mechanical failure (Batt et al., 1990; Hiilscher et al., 1992). A novel acoustic resonance cell retention device (acoustic filter) was recently developed by SonoSep Biotech Inc. and the UBC Biotechnology Lab. This acoustic filter separates cells from medium using an acoustic resonance standing wave.  Forces  associated with the ultrasonic field retain and aggregate cells, which then sediment by gravity back into the bioreactor. Although this approach combines elements of filtration and sedimentation, there are several significant differences. Acousticfiltersperform cell separation using sound waves, thus the fouling problems associated with membrane based systems are eliminated. Because sedimentation is greatly enhanced by the aggregation of cells in the sound field, the need for inclined surfaces is also eliminated. The acoustic resonancefilteris also mechanically simple and contains no moving parts. An experiment was undertaken to test whether a prototype acoustic filter would effectively retain CHO cells during a long term perfusion culture. The CHO cell line used 60  in the experiment formed large cell aggregates during preliminary batch cultures (data not shown). There was no evidence to suggest that exposure to acoustic fields increased the formation of aggregates within the bioreactor during perfusion culture. The presence of such large numbers of cell aggregates within the culture makes the use of standard methods of biomass estimation difficult and, more importantly, inaccurate. To facilitate the estimation of viable biomass, intracellular ATP and LDH activity were measured during the culture. The culture was inoculated at a total cell concentration of 3.5x10)5 ii /mL ce  s  (Figure 27) and attained a maximum growth rate of approximately 0.7 day". Intermittent 1  perfusion was initiated during the second and third day after inoculation, and continuous perfusion performed after 66 h of culture (Figure 27).  A viable cell concentration, by  trypan blue exclusion, of around 10 million cells/mL was reached after 150 h of culture at a dilution rate of approximately 3.8 day" (5.6 L/day) with a residual glucose concentration 1  within the bioreactor of approximately 0.2 g/L (Figure 28).  After 166 h the feed was  supplemented with an additional 2.5 g/L of glucose and 3 mM glutamine.  At  approximately the same time it became obvious that the addition of NaOH to the culture to control pH resulted in increased formation of cell aggregates. To minimize aggregate formation, 0.8 M NaHC0 was subsequently used to control the culture pH. After 280 h 3  of culture the glucose had once more decreased to less than 0.3 g/L and the glucose in the feed medium was increased to 7 g/L.  The increased consumption of glucose and the  resultant increase in the rate of NaHC0 addition caused the medium osmolality to 3  increase from its initial value of 325 to a maximum of 468 mOsm/kg (Figure 29), a level which has been linked to decreased growth rates in hybridoma cells (Ozturk and Palsson, 1991). A malfunction in the level control system that occurred within hours of the second increase in the medium glucose transiently increased the glucose concentration within the bioreactor to 3 g/L. Once the normal feed rate was resumed, the glucose concentration within the bioreactor decreased rapidly to approximately 0.3 g/L.  6 1  r10  200  Time (h) Figure 27:  Total cell concentration determined using microscopic cell - A - or nuclei - • - counts, viable cell concentration -A- and dilution rate — over the course of a high density perfusion CHO cell culture.  level control failure  200  400  Time (h) Figure 28:  Glucose concentration over the course of a high density perfusion CHO cell culture.  62  480-j 460440§ 420-  1  tn  O  •  400-  £ 380E 360°  340320300  0  100  1  300  200  400  Time (h) Figure 29:  Osmolality over the course of a high density perfusion CHO cell culture.  Although the problem of aggregate formation in the bioreactor was less than in the preliminary spinner cultures, the size and number of cell aggregates increased steadily throughout the culture. After 150 h of culture, it was difficult to estimate the viable cell concentration using standard dye exclusion methods. Large variations in the measured viable cell concentration can be seen on Figure 27. After 250 h the dye exclusion assay was abandoned entirely in favor of counting stained nuclei.  The higher estimate of  bioreactor cell concentration obtained using nuclei counting (Figure 27) suggested that the large numbers of cell aggregates resulted in an underestimate of the actual cell population. During the period from 315 to 350 h of culture, the total cell population in the bioreactor 1  1  from nuclei counts decreased from 1.5x10 to 1.2x10 cells/mL, possibly due to the increase in osmolarity which reached its maximum at 315 h. The inconsistency of the measured cell concentration after 300 h of culture can be attributed to the difficulty of  63  obtaining representative samples of the culture. The inclusion in a sample of even one of the large cell aggregates present in the bioreactor at this time could have easily resulted in a measurable increase in cell concentration. The cell associated ATP (Figure 30), medium glucose (Figure 28) and osmolarity (Figure 29) were determined daily. LDH (Figure 31) was measured for all samples after 200 h. The increase in the recovered ATP approximated the measured increase in cell number until estimates of biomass determined using viable counts became unreliable after approximately 150 h. At this point both ATP and LDH (where available) likely provided a more accurate estimate of biomass than did manual cell counts.  200 Time (h) Figure 30:  ATP - • - and total cell concentration determined using microscopic cell - A - or nuclei counts over the course of a high density perfusion CHO cell culture.  64  3-  6-i  Time (h) Figure 31:  Cell associated LDH activity - • - and total cell concentration determined using microscopic cell -A- or nuclei - • - counts over the course of a high density perfusion CHO cell culture.  The ATP activity showed a transient increase in samples taken after 209 h of culture, while both ATP and LDH showed a substantial increase after 329 h (Figures 30 and 31). In both cases the increases immediately followed the addition of extra glucose to the medium (Figure 28). The 27% increase in ATP and the 50% increase in LDH activity which occurred after 329 h was also associated with the failure of the bioreactor's level control system. The sudden increase in the feed rate at this time resulted not only in a substantial increase in glucose concentration, but also a general improvement in culture conditions. Such an improvement in the culture conditions could result in an increased growth rate and an increase in the mean cell volume sufficient to account for the increase in cell specific ATP. Volume changes of more than 50% are typical of hybridoma batch cultures. Such a volume change combined with increased cellular biosynthetic activity could explain the 50% increase in LDH activity. Because of the high degree of clumping, it was not possible to determine if an increase in cell volume had occurred.  65  From 329 to 350 h the ATP concentration decreased to approximately the same average concentration as before the transient increase.  LDH activity also decreased  during this period, but only by 40% of the original increase. The failure of LDH activity to return to its pre-perturbation level may have been due to the stability of the enzyme which resulted in it being retained within viable and newly nonviable cells. There was a correlation between ATP concentration and total cell concentration (R=0.99) up to 185 h of culture after which it became difficult to obtain accurate cell counts (Figure 32). After 185 h there was a fairly consistent increase in the ATP per cell, implying that both trypan blue stained viable counts and nuclei counts underestimated the actual cell concentration.  Erratic increases and decreases in the measured cell  concentration after 185 h of culture also suggest large errors in the cell concentration estimated using manual counts.  80 -i  i 604 o E c  c  CD O O  O  20 H  Q.  1  0  1  50  1  1  1  1  100  150  r-  200  Total Cell Concentration (10 Cells/mL) 5  Figure 32:  ATP concentration versus total cell concentration before - • - and after -0- 185 h during a high density perfusion CHO cell culture.  66  The conclusion that ATP provided a more accurate estimate of biomass than cell counts was supported by the data on recombinant protein recovery. Production of yinterferon by the CHO cell culture was more closely correlated to measurements of the concentration of ATP in the culture than to the cell number as estimated using manual counts (Figure 33).  Figure 33:  (a) ATP recovered and (b) total cell concentration versus culture productivity over the course of a high density perfusion CHO cell culture.  67  5  DISCUSSION A method of determining biomass should fulfill several criteria. In addition to  being accurate and reproducible, it should differentiate between viable and nonviable cells, a characteristic that becomes increasingly important as the culture viability declines. Other criteria which make a method desirable include the ability to detect changes in cell volume. There are a wide variety of methods available by which viable biomass may be estimated.  With the exception of nuclei counts which may be used for immobilized  cultures, the need to recover representative cell samples restricts direct methods of biomass estimation primarily to single cell suspensions (van Wezel, 1967).  Cellular  components such as enzymes sometimes do not distinguish viable from recently nonviable cells, and the levels of expression have been shown to be sensitive both to growth state and culture conditions (Tarentino et al., 1966; Fritz et al., 1969; Ruddle and Rapola, 1970; Marc et al., 1991; Jenkins et al., 1992). Intracellular ATP was an accurate indicator of viable biomass over a wide variety of culture conditions. Although this work has investigated the effect of growth condition on intracellular ATP in only 3 cell lines, available literature indicates that ATP levels are remarkably insensitive to the effects of various agents and culture conditions (Chapman et al., 1971; Ryll et al., 1991). The results of this study indicate that the intracellular ATP concentration in viable 2E11 and BHK cells is remarkably stable over the course of batch culture. This is consistent with evidence that the activity of cellular enzymes responds rapidly to changes in adenylate concentrations, maintaining an equilibrium adenylate energy charge greater than 0.9 (Siems et al., 1984, 1986; Schneider et al., 1990). While continuous culture experiments indicated that rapid decreases in pH could transiently decrease intracellular ATP concentration, these pH shifts were also associated with dramatic decreases in culture viability. Prolonged exposure to these same low pH conditions did not change the intracellular ATP concentration significantly. Prolonged exposure to pHs up to 8 also caused substantial cell death, but did not significantly affect 68  the intracellular ATP concentration. Exposure to dissolved oxygen concentrations ranging from 5 to 80% of air saturation did not affect ATP concentrations. Under most circumstances, changes in the measured ATP reflected the changes in the viable cell number in the culture, and changes in the average cell specific ATP. Changes in average cell specific ATP were in turn largely attributable to changes in mean cell volume. With the exception of the few cases mentioned above, the intracellular ATP concentration was not significantly affected by culture conditions. The large number of enzymes with ATPase activity present in the cell resulted in a rapid decrease in ATP following cell death. As a result any change in viable biomass should be readily detectable by monitoring the ATP concentration, perhaps even before it would be detected by most conventional methods of biomass estimation. In addition the instability of ATP on cell death also ensures that nonviable cells will not contribute to estimates of culture biomass, and thus only the active biomass will be measured. ATP was of particular value in the estimation of biomass in high density cell cultures. While the formation of cell aggregates in the CHO cell culture prevented the estimation of viable biomass using standard cell counting, estimates of biomass using ATP extractedfromthe cells were possible throughout the culture. Biomass estimates based on ATP were also more closely correlated to the productivity of the culture than those based on microscopic cell counts. The results of this work indicate that cell associated ATP is an accurate measure of total viable cell volume (biomass) although not of cell concentration. The productivity of a given culture is dependent upon the amount of viable biomass rather than upon the actual number of cells present. Thus it follows that for the purposes of material balances or kinetic analysis, the biomass estimates provided by ATP analysis would be more useful than cell counts. This is particularly evident when one recognizes that the cell volume changes by more than 2-fold during a typical batch culture. In addition, since there is no requirement to recover individual cells, biomass estimation using ATP can be used to  69  monitor cell cultures containing cell aggregates, microcarriers or porous support matrices where it is currently very difficult, if not impossible, to obtain accurate estimates of biomass.  70  6  REFERENCES  1.  Al-Rubeai, M.A., Musgrave, SC., Lambe, C.A., Walker, A G . , Evans, N.H., Spier, R E . 1990. Methods for the estimation of the number and quality of animal cells immobilized in carbohydrate gels. Enzyme Mcrob. Technol. 12:459-463.  2.  Atkinson, 1969. Regulation of enzyme function. Annu. Rev. Mcrobiol. 23:47-68.  3.  Attallah, A M . , Johnson, R.P. 1981. A simple highly sensitive method for the determination of cell viability using an electronic particle analyzer, Coulter counter. J. Immunol. Meth. 41:155-162.  4.  Aubin, J.E. 1979. Autofluorescence of viable cultured mammalian cells. J. Histochem. Cytochem. 27(1): 36-43.  5.  Avgerinos, G.C, Drapeau, D., Socolow, J.S., Hsiao, K., Broeze, R.J. 1990. Spin filter system for high density cell culture: production of recombinant urinary type plasminogen activator in CHO cells. Bio/Technol. 8: 54-58.  6.  Batt, B.C., Davis, R H . Kompala, D.S. 1990. Inclined sedimentation for selective retention of viable hybridomas in a continuous suspension bioreactor. Biotechnol. Prog. 6: 458-464.  7.  Bjornsen, P.K. 1986. Automatic determination of bacterioplankton biomass by image analysis. Appl. Environ. Microbiol. 51(6): 1199-1204.  8.  Blake-Coleman, B.C., Clarke, D.J., Calder, M.R., Moody, S.C. 1986. Detennination of reactor biomass by acoustic resonance densitometry. Biotechnol. Bioeng. 28:12411249.  9.  Blute, T., Gilles, T., Dale, B. 1988. Biotechnol. Prog. 4:202-209.  10.  Boag, A.H., Sefton, M.V. 1987. Mcroencapsulation of humanfibroblastsin a waterinsoluble polyacrylate. Biotechnol. Bioeng. 30:954-962.  11.  Boraston, R , Thompson, P.W., Garland, S., Birch, J.R. 1984. Growth and oxygen requirements of antibody producing mouse hybridoma cells in suspension culture. Dev. Biol. Stand. 55:103-111.  12.  Broise, D., Noiseux, M., Massie, B., Lemieux, R. 1992. Hybridoma perfusion systems: a comparison study. Biotechnol. Bioeng. 40: 25-32.  71  13.  Cazin, M., Paluszezak, D., Bianchi, A., Cazin, J.C., Aerts, C , Voisin, C. 1990. Effects of anaerobiosis upon morphology and energy metabolism of alveolar macrophages cultured in gas phase. Eur. Respir. J. 3:1015-1022.  14.  Chapman, A.G., Fall, L., Atkinson, D.E. 1971. Adenylate energy charge in Escherichia coli during growth and starvation. J. Bact. 108(3): 1072-1086.  15.  Chapman, J.D., Webb, R.G., Borsa, J. 1971. ATP pool levels in synchronously growing Chinese hamster cells. J. Cell Biol. 49:229-233.  16.  Coulter, W.H. 1956. High speed automatic blood cell counter and cell size analyzer. Proc. Natl. El. Conf. 12:1034-1040.  17.  Croughan, M.S., Wang D.I.C. 1989. Growth and death in overagitated microcarrier cell cultures. Biotechnol. Bioeng. 33:731-744.  18.  Dawes, E . A 1986. Adenine nucleotides and their role in metabolism, p. 9-20 In: Microbial energetics. Blackie, London.  19.  de Korte, D., Haverkort, W.A., van Gennip, A.H., Roos, D. 1985 Nucleotide profiles of normal human blood cells determined by high performance liquid chromatography. Anal. Biochem. 147:197-209.  20.  De Luca, C. 1965. The use of trypsin for the determination of cellular viability. Exp. Cell Res. 40:186-188.  21.  Esclade, L.R.J., Carrel, S., Peringer, P. 1991. Influence of screen material on the fouling of spin filters. Biotechnol. Bioeng. 38: 159-168.  22.  Fleischaker, R.J., Weaver, J.C., Sinskey. 1981. Instrumentation for process control in cell culture. Adv. Appl. Microbiol. 27:137-167.  23.  Foran, D.J., Cahn, F., Eikenberry, E.F. 1991. Assessment of cell proliferation on porous microcarriers by means of image analysis. Analytical and Quantitative Cytology and Histology. 13(3): 215-222.  24.  Frame, K.K., Hu, W.-S. 1985. Oxygen uptake of mammalian cells in microcarrier culture-response to changes in glucose concentration. Biotechnol. Letters 7(3): 147152.  25.  Frame, K.K., Hu, W.-S. 1990. Cell volume measurement as an estimation of mammalian cell biomass. Biotechnol. Bioeng. 36:191-197.  72  26.  Freudenberg, H , Mager, J. 1971. Studies on the mechanism of the inhibition of protein synthesis induced by intracellular ATP depletion. Biochim. Biophys. Acta. 232:537555.  27.  Fritz, P.J., Vesell, E.S., White, E.L., Pruitt, K.M. 1969. The roles of synthesis and degradation in determining tissue concentrations of lactate dehydrogenase-5. P.N.A.S. 62:558-565.  28.  Funk, W.D., MacGillivray, R.T.A., Mason, A.B., Brown, S.A., Woodworm, R.C. 1990. Expression of the ammo-terminal half-molecule of human serum transferrin in cultured cells and characterization of the recombinant protein. Biochem. 29(6): 16541660.  29.  Gikas, P., Livingston, A.G. 1993. Use of ATP to characterize biomass viability in freely suspended and immobilized cell bioreactors. Biotechnol. Bioeng. 42:1337-1351.  30.  Goebel, N.K., Kuehn, R, Flickinger, M.C. 1990. Methods for determination of growth-rate-dependent changes in hybridoma volume, shape and surface structure during continuous recycle. Cytotechnol. 4:45-57.  31.  Harris, M. 1966. Criteria of viability in heat-treated cells. Exp. Cell Res. 44:658-661.  32.  Hasenson, M., Hartley-Asp, B., Kihlfors, C , Lundin, A., Gustafsson, J., Pousette, A. 1985. Effect of hormones on growth and ATP content of a human prostatic carcinoma cell line. LNCaP-r. Prostate. 7:183-194.  33.  Himmler, G , Palfi, G , Riiker, F., Katinger, H , Scheirer, W. 1985. A laboratory fermentor for agarose immobilized hybridomas to produce monoclonal antibodies. Devel. Biol. Standard. 60:291-296.  34.  Hoskins, J.M., Meynell, G.G, Sanders, F.K. 1956. A comparison of methods for estimating the viable count of a suspension of tumour cells. Exp. Cell Res. 11:297-305.  35.  Hulscher, M., Scheibler, U., Onken, U. 1992. Selective recycle of viable animal cells by coupling airlift reactor and cell settler. Biotechnol Bioeng. 39: 442-446.  36.  Jenkins, H.A, Fitzpatrick, L., Butler, M. 1992. Energy metabolism of murine hybridomas. p. 243-247 In: Production of biologicalsfromanimal cells in culture. R E . Spier, J.B. Griffiths and C. Macdonald (eds.), Butterworth-Heinemann, Oxford.  37.  Kilburn, D.G., Fitzpatrick, P., Blake-Coleman, B.C., Clarke, D.J., Griffiths, J.B. 1989. On-line monitoring of cell mass in mammalian cell cultures by acoustic densitometry. Biotechnol. Bioeng. 33:1379-1384.  73  38.  Kilburn, D.G., van Wezel, A.L. 1970. The effect of growth rate in continuous-flow cultures on the replication of rubella virus in BHK cells. J. Gen. Virol. 9:1-7.  39.  Knowles, C.J. 1977. Microbial regulation by adenine nucleotide pools, p. 241-283 In: Microbial Energetics. BA. Haddock and W.A Hamilton (eds), Cambridge University Press, Cambridge.  40.  Konstantinov, K.B., Chuppa, S., Sajan, E., Yoon, S., Golini, F. 1994. Real-time biomass-concentration monitoring in animal-cell cultures. Trends in Biotechnol. 12(8):324-333.  41.  Konstantinov K.B., Pambayun, R., Matanguihan, R., Yoshida, T., Perusich, C , Hu, W.-S. 1992. On line monitoring of hybridoma cell growth using a laser turbidity sensor. Biotechnol. Bioeng. 40:1337-1342.  42.  Konstantinov, K.B., Yoon, S. 1994. A novel method for continuous real-time monitoring of the oxygen uptake rate (OUR) in bioreactors. American Chemical Society Spring Meeting, San Diego, CA, USA.  43.  Li, J.-K., Asali, E.C., Humphrey, A.E. 1991. Monitoring cell concentration and activity by multiple excitation fluorometry. Biotechnol. Prog. 7:21-27.  44.  Lundin, A. 1982. Analytical applications of bioluminescence: thefireflysystem, p. 4374 In: Clinical and Biochemical Luminescence. L. J. Kricka and T. J. Carter, (eds.), Marcel Dekker, New York.  45.  Lundin, A , Hasenson, M., Persson, J., Pousette, A. 1986. Estimation of biomass in growing cell lines by adenosine triphosphate assay. Meth. Errzymol. 133:27-42.  46.  Lundin, A., Thore, A 1975. Comparison of methods for extraction of bacterial adenine nucleotides determined byfireflyassay. Appl. Microbiol. 30(5):713-721.  47.  Lydersen, D.K., Pugh, G.G., Paris, M.S., Sharma, B P , Noll, L.A. 1985. Ceramic matrix for large scale animal cell culture. Bio/Technol. 3:63-67.  48.  Mancuso, A , Fernandez, E.J., Blanch, H.W., Clark, DS. 1990. A nuclear magnetic resonance technique for determining hybridoma cell concentration in hollow fiber bioreactors. Bio/Technol. 8:1282-1285.  49.  Marc, A , Wagner, A., Martial, A , Goergen, J.L., Engasser, J.M., Geaugey, V., Pinton, H. 1991. Potential and pitfalls of using LDH release for the evaluation of animal cell death kinetics, p. 569-575 In: Production of biologicalsfromanimal cells in culture. R.E. Spier, J.B. Griffiths and B. Meignier (eds.), Butterworth-Heinemann, Oxford.  74  50.  Merten, O., Palfi, G., Stahli, J., Steiner, J. 1987. Application of sensors for the determination of physical and biochemical parameters in the fermentation of animal cells. Develop. Biol. Stand. 66:357-360.  51.  Miovic, M.L., Gibson, J. 1973. Nucleotide pools and adenylate energy charge in balanced and unbalanced growth of Chromatium. J. Bact. 114:86-95.  52.  Muller, M., Siems, W., Buttgereit, F., Dumdey, R, Rapoport, S.M. 1986. Quantification of ATP-producing and consuming processes of Ehrlich ascites tumour cells. Eur. J. Biochem. 161:701-705.  53.  Nowotny, H., Benes, E. 1987. General one-dimensional treatment of layered piezoelectric resonators with two electrodes. J. Acoust. Soc. Am. 82:513-521.  54.  Ozturk, S.S., Palsson, B.O. 1991. Effect of medium osmolality on hybridoma growth, metabolism, and antibody production. Biotechnol. Bioeng. 37:989-993.  55.  Piret, J.M., Cooney, C L . 1990. Mammalian cell and protein distributions in ultrafiltration hollowfiberbioreactors. Biotechnol. Bioeng. 36:902-910.  56.  Ramirez, O.T., Mutharasan, R. 1990. Cell cycle- and growth phase-dependent variations in size distribution, antibody productivity, and oxygen demand in hybridoma cultures. Biotechnol. Bioeng. 36:839-848.  57.  Ross, D.D., Joneckis, C C , Ordonez, J.V., Sisk, A M . , Wu, R.K., Hamburger, A.W., Nora, R E . 1989. Estimation of cell survival by flow cytometric quantification of fluorescein diacetate / propidium iodide viable cell number. Cane. Res. 49:3776-3782.  58.  Ruddle, F.H., Rapola, J. 1970. Changes in lactate dehydrogenase and esterase-specific activities, isozymic patterns,' and cellular distribution during the growth cycle of PK cells in vitro. Exp. Cell Res. 59:399-412.  59.  Ryll, T., Jager, V., Wagner, R. 1991. Intracellular concentration of ATP and other nucleotides during continuous cultuivation of hybridoma cells, p. 236-243 In: Production of biologicalsfromanimal cells in culture. R.E. Spier, J.B. Griffiths and B. Meignier (eds.), Butterworth-Heinemann, Oxford.  60.  Sakato, K., Tanaka, H , Samejima, H., Karube, I., Suzuki, S. 1981. Electrochemical measurements of cell populations. Ann. NY. Acad. Sci. 369:321-334.  61.  Sanford, K.K., Earle, W.R., Evans, V.J., Waltz, H.K., Shannon, J.E. 1950. The measurement of proliferation in tissue cultures by enumeration of cell nuclei. J. Nat. Cane. Inst. 11:773-795.  75  62.  Scheirer, W., Varecka, R, Bachmayer, H. 1987. A convenient system for the serial quantitation of viable cells measuring luminescence of ATP. Develop. Biol. Standard. 66:349-355.  63.  Schimke, R. T. 1970. Regulation of protein degradation in mammalian tissues, p. 177227 In: Mammalian protein metabolism. H.N. Munro (ed), vol. 4. Academic Press, London.  64.  Schneider, W., Siems, W., Grune, T. 1990. Balancing of energy-consuming processes of rat hepatocytes. Cell Biochem. Funct. 8:227-232.  65.  Sen, S., Srienc, F., Hu, W.-S. 1989. Distinct volume distribution of viable and nonviable hybridoma cells: aflowcytometry study. Cytotechnol. 2:85-94.  66.  Siano, S.A., Mutharasan, R. 1991. NADHfluorescenceand oxygen uptake responses of hybridoma cultures to substrate pulse and step changes. Biotechnol. Bioeng. 37:141159.  67.  Siems, W., Dubiel, W., Dumdey, R., Muller, M., Rapoport, S.M. 1984. Accounting for the ATP consuming processes in rabbit reticulocytes. Eur. J. Biochem. 139:101-107.  68.  Siems, W., Muller, M., Dubiel, W., Dumdey, R., Rapoport, S. 1986. ATP production and consumption of rabbit reticulocytes increase in an ammo-acid-enriched medium. Biomed. Biochim. Acta. 45:585-591.  69.  Siro, M.-R., Romar, H , Lovgren, T. 1982. Continuousflowmethod for extraction and bioluminescence assay of ATP. Appl. Microbiol. Biotechnol. 15:258-264.  70.  Somlo, M. 1970. Respiration linked ATP formation by an "oxidative phosphorylation mutant" of yeast. Arch. Biochem. Biophys. 36:122-133.  71.  Tarentino, A L . , Richert, D.A., Westerfeld, W.W. 1966. The concurrent induction of hepatic a-glycerophosphate dehydrogenase and malate dehydrogenase by thyroid hormone. Biochim. Biophys. Acta. 124:295-309.  72.  Thatipamala, R., Rohani, S., Hill, G. 1994. On-line monitoring of biomass concentration based on a new equation using a spectrophotometer-process control applications. J. Biotechnol. 38:33-42.  73.  Thore, A , Ansehn, S., Lundin, A , Bergman, S. 1975. Detection of bacteriuria by luciferase assay of adenosine triphosphate. J. Clin. Microbiol. 1:1-8.  74.  Trampler, F., Sonderhoff, S.A, Pui, P.W.S., Kilburn, D.G., Piret, J.M. 1994. Acoustic cellfilterfor high density perfusion culture of hybridoma cells. Bio/Technol. 12:281284.  76  75.  Truskey, G.A., Proulx, T.L. 1990. Quantitation of cell area on glass and fibronectincoated surfaces by digital image analysis. Biotechnol. Prog. 6:513-519.  76.  van Wezel, A.L. 1967. Growth of cell-strains and primary cells on micro-carriers in homogeneous culture. Nature 216:64-65.  77.  Velez, D., Miller, L., Macmillan, J.D. 1989. Use of tangential flow nitration in perfusion propagation of hybridoma cells for production of monoclonal antibodies. Biotechnol. Bioeng. 33: 938-940.  78.  Wohlpart, D., Kirwan, D., Gainer, J. 1990. Effects of cell density and glucose and glutamine levels on the respiration rates of hybridoma cells. Biotechnol. Bioeng. 36:630-635.  79.  Yabannavar, V.M., Singh, V., Connelly, N.V. 1992. Mammalian cell retention in a spinfilter perfusion bioreactor. Biotechnol. Bioeng. 40:925-933.  80.  Zielke, H.R., Zielke, C.L., Ozand, P.T. 1984. Glutamine: a major energy source for cultured mammalian cells. Fed. Proc. 43:121-125.  81.  Ziltner, H.J., Clark-Lewis, I., Fasekas, B., Groth, S., Orban, P C , Hood, L.E., Kent, S.B.H., Schrader, J.W. 1988. Monoclonal antipeptide antibodies recognize JL-3 and neutralize its bioactivity in vivo. J. Immunol. 4:1182-1187.  77  

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.831.1-0086734/manifest

Comment

Related Items