OPTIMIZATION OF ENVIRONMENT FOR HIGH DENSITY VERO CELL CULTURE

J. Cell Sci. 81, 65-103 (1986) 65Printed in Great Britain © The Company of Biologists Limited 1986

OPTIMIZATION OF ENVIRONMENT FOR HIGHDENSITY VERO CELL CULTURE: EFFECT OFDISSOLVED OXYGEN AND NUTRIENT SUPPLY ONCELL GROWTH AND CHANGES IN METABOLITES

ARA T. NAHAPETIAN*, JAMES N. THOMASjAND WILLIAM G. THILLYJBiotechnology Group, Room El 8-666, Department of Applied Biological Sciences,Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA

SUMMARYThis study was initiated for optimization of the environment of a technologically useful

mammalian cell line for high density production. Cultures of Vero cells on microcarriers wereperfused with 100%, 50%, 25% and 12-5% modified L15 media (galactose was replaced with10 mM-fructose, with 4 mM-glutamine and 5 % foetal bovine serum) in phosphate-buffered saline ateither 4 or 8 vol. day"1. Cell growth, pH, dissolved oxygen, and changes in the metabolites, lactateto pyruvate and lactate to ammonia indices, demonstrated that under the conditions used in thepresent study, perfusion of cultures with 50% L15 medium in PBS at 8 vol. day"1 provided theoptimum microenvironment for Vero cell growth. The highest cell density in the perfused cultureswas 3xlO7 cells ml"1, which at these conditions was ten times higher than the maximum celldensity (3xlO6 cells ml"1) obtained in a batch culture. Nutrient supply and conditioning factorswere the most probable growth-limiting factors in cultures that were perfused with 12-5% and25 % LI5 media, while multilayering, limitation of available oxygen, and accumulation of meta-bolic endproducts in the cellular microenvironment were the most probable causes of a density-dependent inhibition of cell growth observed under the optimized and overfed (supply of 100%L15 medium at the rate of 8 vol. day"1) culture conditions. Under the optimized environmentalcondition, the major source of energy was probably glutamine during the first week. However,significant utilization of fructose became evident at higher cell densities during the second week,when lactate production dramatically declined and reached an almost undetectable level, whilerespiration progressively assumed the predominant role in energy production. It is postulated that'available' oxygen in the multicell-layered microenvironment of the optimized cultures was higherthan in the overfed culture due to the greater utilization rate of oxygen for oxidation of excessnutrients in the overfed culture.

INTRODUCTION

The first comprehensive study of specific environmental factors such as nutrientrequirement for cell growth and survival in culture was reported nearly 30 years ago(Eagle, 1955). Most media used for cell culture at present, which are basically the

•Present address: E. I. duPont de Nemours and Co., Wilmington, DE 19898, USA.•(•Present address: Genentech, Inc., 460 Point San Bruno Blvd, South San Francisco, CA

94080, USA.J Author for correspondence.

Key words: perfusion, Vero, cell culture, microcarriers, nutrients.

66 A. T. Nahapetian, J. N. Thomas and W. G. Thilly

same as or a modification of the original formulation reported by Eagle, containglucose as the main source of carbohydrate.

In a number of studies the use of glucose as the main energy source has beenquestioned. Rapid glucose depletion of culture media was reported in a number ofinvestigations (Graff & McCarty, 1957; Westfall, Evans, Shannon & Earle, 1953;Himmelfarb, Thayer & Martin, 1969). The actual glucose requirement of cells inculture for either proliferation or maintenance was suggested to be much less thanthe levels cells could consume when the monosaccharide was supplied at relativelyhigh (5-11 mM) concentrations (Graff et al. 1965). Imamura, Crespi, Thilly &Brunengraber (1982) demonstrated that when the glucose in Dulbecco's modifiedEagle's medium (DMEM) was replaced with fructose, in terms of cell productionboth carbohydrates seemed to be equally effective for Madin-Darby canine kidney(MDCK) cells. They noted that the depletion rate of fructose from culture mediumwas much slower than that for glucose. Substitution of fructose for glucose in culturemedia was recommended (Imamura et al. 1982). Other studies have suggested thatcarbohydrates were essential for cells in culture because they were required forsynthesis of nucleic acids (Griffiths, 1972; Zielke, Zielke & Ozand, 1984; McGowen,Russell & Bucher, 1984), but not for their role as a source of energy.

Energy requirements of mammalian cells can also be met by degradation of aminoacids (Thilly, Bamgrover & Thomas, 1982; Knop et al. 1984). The relative signifi-cance of glutamine was indicated in the initial formulation of Eagle's basal mediumfor in vitro cell growth and maintenance; the recommended glutamine concentration(2 mM) in the medium was ten times the concentration of the next highest essentialamino acid on the list. Moreover, glutamate alone or supplemented with ammoniaand ATP could not replace glutamine for cell growth and survival (Eagle, 1955). Itwas suggested that glutamine had an essential metabolic role, although the role wasnot clearly defined. The significance of glutamine was confirmed in other studies(McCarty, 1962; Kruse, Miedema & Carter, 1967; Kruse, Keen & Whittle, 1970),and it was proposed that the amino acid was a major source of energy for culturedmammalian cells (Kovacevic & Morris, 1972; Zielke et al. 1978, 1984; Tildon &Roeder, 1984; Brand, Williams & Weidemann, 1984).

Cell growth and survival in culture can be affected by accumulation of metabolicend products. They could either inhibit (Rubin, 1966; Burk, 1966; Stoker & Rubin,1967; Thilly et al. 1982; Butler, Imamura, Thomas & Thilly, 1983) or stimulate(Rubin, 1966; Horng & McLimans, 1975) cell proliferation. Accumulation of lactate(Graff & McCarty, 1957; Ehrlich, Stewart & Klein, 1978; Imamura et al. 1982;Kashiwagura, Wilson & Erecinska, 1984; Hue & Bartrons, 1984), ammonia(Imamura e/ a/. 1982; Butler et al. 1983), and glutamate (Kruse et al. 1967; Butleret al. 1983) were reported by a number of investigators. It was suggested that, sincelactate was known to be one of the main sources of non-volatile acid equivalents inmammalian cell cultures, the excessive rise in concentration of the metabolite was themain cause of decrease in culture pH, which could inhibit cell growth (Imamuraet al. 1982). In contrast, others have shown that high concentrations of lactate inculture media can stimulate hepatocyte DNA replication (McGowan et al. 1984).

Optimized environment for Vero cell culture 67

Ammonia is known to be toxic at concentrations higher than 2 mM in culture media(Visek, Kolodny & Gross, 1972; Holley, Armour & Baldwin, 1978; Thilly et al.1982). Concentrations of ammonia were reported to rise above the toxic level in bothbatch (Imamura.ef al. 1982) and perfused (Butler et al. 1983) cultures.

Rubin (1966) reported production of heat-labile, large molecule(s) (termed con-ditioning factor(s)) by chick embryo cells in vitro that could enhance cell growth.Similarly, Horng & McLimans (1975) observed a dramatic increase in calf anteriorpituitary cell proliferation when microcarriers were treated with a conditionedmedium before cell inoculation. However, Levine, Wang & Thilly (1979) could notconfirm the growth-enhancing effect by similar treatment of the beads.

Cell growth and survival are highly sensitive to environmental changes in oxygentension. Oxygen concentrations in the medium lower than 0-5mil"1 (2% atmos-pheric oxygen tension) could lead to cessation of cell growth (Graff & McCarty,1957). Excess dissolved oxygen in the medium (more than 23mil"1 under 95%atmospheric pressure oxygen) was also reported to be toxic for cells in culture(Burrows, 1924; Krebs, 1950; Jones & Bonting, 1956). Using mathematical models,Stevens (1965) showed that for maintenance of a monolayer of rabbit liver cells,medium overlay thickness in a static culture should not exceed 0*34mm. Thesignificance of oxygen availability for animal cells in culture was confirmed in anumber of investigations (McLimans, Blumenson & Tunnah, 1968; Himmelfarbetal. 1969; Richter, Sanford & Evans, 1972; Knazek, Kohler & Gullino, 1972;Jensen, Wallach & Lin, 1974; Werrlein & Glinos, 1974; Butler et al. 1983; Knopet al. 1984; Kashiwagura et al. 1984). Moreover, oscillatory concentration gradientsof oxygen were detected in the overlay fluid above the attached cells. It was suggestedthat the respiratory system of a large number of cells was alternately and syn-chronously turned on and off in response to environmental changes, due to localexhaustion and diffusion of essential nutrients and metabolites (Werrlein & Glinos,1974).

Limitation of surface for attachment of anchorage-dependent cells also caninfluence growth of mammalian cells in culture. Abercrombie & Ambrose (1962)demonstrated cessation of migration and growth of cells in batch culture when theentire surface available for attachment became confluent with a cell monolayer. Theterm 'contact inhibition' was used to describe this phenomenon. In perfusedcultures, however, multilayering was observed together with a density-dependentretardation of cell proliferation (Kruse, Myhr, Johnson & White, 1963; Kruse &Miedema, 1965; Kruse etal. 1967, 1970; Kruse, Whittle & Miedema, 1969;Knazek, Gullino, Kohler & Dedrick, 1972; Knazek, Kohler & Gullino, 1974; Butleret al. 1983). Limitation of attachment surface was alleviated by the discovery (vanWezel, 1967) and development of microcarriers (Levine, Wong, Wang & Thilly,1977; Levine et al. 1979), which opened the path for growth of anchorage-dependentcells in suspension culture.

At present, mammalian cells are grown and maintained in both batch and perfusedsystems. In most cases, however, batch methods are preferred because of theirsimplicity, ease of operation and the availability of material. Tissue culture was


considered to be a valuable tool for biological and metabolic studies because it couldprovide the means for studying homogeneous cell populations under defined con-ditions resembling those present in vivo (Graff & McCarty, 1957). Batch cultures donot fulfil this requirement due to wide fluctuations in the environmental components(Graff & McCarty, 1957; Kruse e< al. 1963; Knazek ef al. 1972; Jensen et al. 1974;Gebhardt & Mecke, 1979; Thilly et al. 1982; Barbehennef al. 1984) by providing analternating environment of 'feasting and fasting' for cells in culture (Graff &McCarty, 1957). Excess and deficiency of nutrients and metabolic endproducts thatwould result in non-physiological extremes in culture pH, osmolarity, and gastensions {pQz andpCOi) can exert detrimental effects on cell growth and survival.Perfused systems were developed to minimize the adverse fluctuations in the en-vironmental components (Graff & McCarty, 1957; Kruse et al. 1963, 1967, 1969,1970; Graff et al. 1965; Kruse & Miedema, 1965; Miedema & Kruse, 1965;Rose, 1967; Schleicher & Weiss, 1968; Himmelfarb et al. 1969; Knazek et al.1972; Knazek, 1974; Chick, Like & Lauris, 1975; Sun & Macmorine, 1976; Fike,Glick & Burns, 1977; Knazek, Lippman & Chopra, 1977; Rutzky, Tomita,Calenoof & Kahan, 1977; Ehrlich et al. 1978; Quarles, Morris & Leibovitz, 1978;Gebhardt & Mecke, 1979; Rutzky, Tomita, Calenoff & Kahan, 1979; Quarles,Morris & Leibovitz, 1980; Smith & Vale, 1980; Butler et al. 1983; Strand, Quarles &McConnell, I984a,b; Knop et al. 1984).

A major shortcoming in almost all perfused anchorage-dependent cell cultures wasthe lack of an accurate monitoring procedure for cell enumeration during theexperimental period. Cell growth was estimated either indirectly by measuringglucose utilization (Schleicher & Weiss, 1968) or by daily termination of singleperfused cultures (Kruse et al. 1967). Frequent sampling of single homogeneouscell populations for direct observation, biochemical testing or enumeration wasmade possible by the development and later improvement and application of micro-carriers (van Wezel, 1967; Horng & McLimans, 1975; Levine et al. 1977, 1979;Thilly et al. 1982; Butler et al. 1983). Moreover, there are limited data on changesin metabolic activity of anchorage-dependent mammalian cells at different cell den-sities. Changes in concentrations of media glucose, lactic acid (Kruse & Miedema,1965; Ehrlich et al. 1978), amino acids, ammonia (Kruse et al. 1967; Butler et al.1983), ATP, and Pi levels (Knop et al. 1984) were presented only in some of theinvestigations.

The present study was initiated for production of a technologically useful mam-malian cell line (Vero) at high densities. For optimization of the culture environ-ment, the effects of nutrient supply and removal of spent medium were studied bymonitoring cell growth. Changes in medium pH, dissolved oxygen (DO), andconcentrations of glucose, fructose, pyruvate, glutamine, lactate, ammonia andglutamate were monitored to determine their relative significance for optimization ofthe environment for Vero cell growth and maintenance. Data are presented for onlythe optimized and overfed perfused culture conditions. They are discussed, togetherwith those obtained for batch cultures with and without cells; first, in order tocompare cell growth in the two systems and, second, to differentiate between the


changes in the culture media that were due to either metabolic activity of the cells orenzymic and spontaneous degradation of media components under the experimentalconditions. We present evidence that: (1) a Vero cell density of 3x 107 cells ml"1 canbe achieved in an optimized environment; (2) nutrient supply and synthesis ofgrowth-stimulating endogenous metabolites (conditioning factors) are the mostprobable growth-limiting factors in underfed cultures, while depletion of oxygen andaccumulation of bicarbonate, lactate, ammonia, and other unknown metabolites inthe cellular microenvironment, are the most probable causes of a density-dependentinhibition of cell growth observed under overfed culture conditions; (3) underoptimized environmental conditions, available oxygen in the cellular micro-environment and accumulation of unknown metabolites are probably the limitingenvironmental factors; (4) there is a change in metabolic activity of Vero cells as theyproliferate and increase in density; and (5) the major source of energy is probablyglutamine during the first week, while significant utilization of fructose becomesevident at higher cell densities during the second week when respirationprogressively assumes the predominant role in energy production.

MATERIALS AND METHODS

MediaA bicarbonate-free medium, L15 (Flow Laboratories, McLean, VA), originally developed by

Leibovitz (1963), was modified (Barngrover, Thomas & Thilly, 1985) by replacing galactose withlOmM-fructose and supplementing with 4mM-L-glutamine (Sigma Chemical Co., St Louis, MO)and 5% foetal bovine serum (Flow Laboratories). The basal medium ( l x ) was diluted withphosphate-buffered saline (PBS) for preparation of 50% (0-50X), 25% (0-25X) and 12-5%(0-125X) basal experimental media.

CellsVero cells (American Type Culture Collection), a cell line isolated from African green monkey

kidney cells, were provided by Flow Laboratories. The cells were used after tests for mycoplasma(Russell, Newman & Williamson, 1975) proved to be negative. Stock Vero cells were grown andmaintained in 490 cm2 polystyrene roller bottles (Corning Glass Works, Corning, NY) with 100 mlof IX basal medium. The bottles were rolled at one vessel rev. min"1 on a cell production rollerapparatus (Bellco Glass, Inc., Vineland, NJ) at 37°C in a walk-in incubator (Harris EnvironmentalSystems, Winchester, MA). The cells were either subcultured or used for studies at the day ofconfluence (4-6 days following inoculation). After removal of the spent medium from the rollerbottles, the cells were washed twice with 30-ml samples of PBS. They were then harvested bytreatment with 3 ml solution of 0-1% trypsin (Flow Laboratories) and 0-02% EDTA (AldrichChemical Co., Inc., Milwaukee, WI) in PBS. After trypsinization for 3 min at 37°C, the cells weredetached by manually hitting the sides of the roller bottles. They were then suspended in about100ml l x basal medium, transferred into a 250ml centrifuge tube, and, after centrifugation at1000 rev. min"1 for 3 min, resuspended in 100 ml of the basal medium.

Perfuston apparatus.The perfusion set-up, which was a modification of the system reported by Butler et al. (1983), is

illustrated in Fig. 1. It consisted essentially of four basic components: (1) two reservoirs, 1000 and2000 ml Erlenmeyer flasks, for fresh and spent media, respectively; (2) a reactor vessel, 500 ml,three-necked, round-bottomed distilling flask (Corning Glass Works), equipped with a dissolvedoxygen probe (LSL Biolafitte, Princeton, NJ); inlets for fresh medium (used also for sampling)and air/oxygen gas mixture and outlets for spent medium and the gas mixture; (3) peristaltic

70 A. T. Nahapetian, J. N. Thomas and W. G. ThillyAir+ O2 mixture

Glasswoolfilter

Mediumreservoir

Amplifier

pH meter

oco

Wastereservoir

Fig. 1. Microcarrier perfusion apparatus for Vero cell culture.

pumps (Multistatic, Buchler Instruments, Inc., Fort Lee, NJ); and (4) silicon rubber tubing(0-2cm inner diam.; 0-5cm outer diam.), which was used together with a non-toxic silicone glue(RTV 108 translucent silicone rubber, General Electric Co., Waterford, NY) for connecting thetwo flasks to the reactor vessel (culture vessel). The flasks were equipped with glass-wool filters,which prevented destruction of the vessels due to gas expansion during autoclaving. In addition,the filters provided free gas exchange between inside and outside environments of the flasks understerile conditions. The specific design and dimensions of the glass suction tube for spent medium(10cm length and 2-5cm inner diam.) prevented removal of microcarriers with the outflowingspent medium at the highest flow-rate (800 ml day"1) used in the present study. All glasscomponents were siliconized (Prosil-28, PRC Research Chemicals, Inc., Gainesville, FL) andrinsed thoroughly with distilled water before use in order to prevent bead adhesion to the glasssurfaces. Once the perfusion set-up was completed, about 100 ml of distilled water was introducedinto the reactor vessel for both protection of the dissolved oxygen probe membrane duringsterilization and complete destruction of microorganisms. After the entire system was wetted withdistilled water, it was autoclaved at 121CC for 90min. The set-up was then cooled down to roomtemperature, and the reactor vessel liquid was removed for preparation of perfused culture.

Perfused culturePerfused cultures were prepared by minor modification of the procedure described by Butler

et al. (1983). A sample (100ml) of sterile microcarrier suspension in PBS (Superbeads, suppliedby Flow Laboratories) at 20mgml~' bead concentration was transferred into a 250 ml centrifugetube. Once the beads settled, the clear supernatant was removed, and the beads were washed twice


with 50 ml of experimental medium. The bead suspension, which provided approximately 11 300cm2 of surface area (equivalent to 23 490-cm2 roller bottles), was then transferred quantitatively tothe reactor vessel. The suspension volume in the flask was made up to 92 ml with the respectiveexperimental medium. The perfusion system was then transferred into the warm room (37°C), andthe mixture was stirred at 60-90 rev. min"1 by means of a rotating magnetic base (Bellco Glass,Vineland, NJ) overnight for microcarrier environment equilibration. The following day, afterdetermination of initial dissolved oxygen in the mixture, the culture preparation was completed byaddition of 5-8 ml Vero cell suspension. Initial cell inoculum provided between 0-7X108 and1-26X108 cells, which corresponded to about 6-11 cells/bead. Samples of lml and 2ml of theculture were transferred into two 15-ml centrifuge tubes for determination of initial cell number,pH and metabolites. During the study, the perfused system was stirred and maintained at 37°Cexcept at the time of sampling, when it was transferred into a laminar flow hood and the operationwas performed at room temperature. Perfusion of the culture at either 400 (4 vol.) or 800 (8 vol.)ml day"1 rate was started 24 h following the inoculation. The adaptation period was necessary forcomplete attachment of cells to the carriers. Free cells were removed, while those attached to thebeads remained in the reactor vessel, due to the special design of the spent medium separatorremoval system. The dissolved oxygen in the culture vessel was monitored throughout theexperimental period. When the reading reached a level below 2% of its original value, a mixture ofhumidified air and oxygen gas was introduced into the culture vessel before cell inoculation. Thegas flow was regulated manually until a DO reading close to 50 % of its original value was obtained.Daily, or every other day, samples of the culture (lml) or medium without cells (2ml) wereobtained for monitoring cell growth and changes in culture pH and metabolites during the study.Similarly, 2-3 drops of the culture were transferred into a multiwell plate (Flow Laboratories) forvisual observation. After determination of pH, the samples without cells were frozen and stored at—20°C for later analysis.

Batch cultureBatch cultures were prepared according to procedures described previously (Levine et al. 1977,

1979). A 25 ml sample of the Superbead suspension, which provided about 2800 cm2 of surface areafor cell attachment, was transferred into a 50 ml centrifuge tube. The beads were washed twice in asiliconized 250 ml spinner flask (Wilbur Scientific, Boston, MA) equipped with sampling andsterile gas-exchange ports. Volume in the flask was made up to 98 ml with the respectiveexperimental medium. The mixture of bead and medium in the spinner flask was maintained underthe conditions described earlier for the perfused culture, either with or without the addition of cellsuspension after the overnight equilibration period. The cell suspension provided between1-75X107 and 2-52X107 cells, which corresponded to a cell/bead ratio of between 6 and 8,respectively. Batch cultures were maintained under the same conditions as described earlier for theperfused system. Similarly, they were also monitored for cell growth and changes in medium pHand metabolites.

Cell enumerationBefore inoculation, viable cell counts were determined by the Trypan Blue dye-exclusion

method (Patterson, 1979). Equal samples (0-5 ml) of the cell suspension and Trypan Blue reagent(0-2 % Trypan Blue in PBS) were mixed in a 15 ml tube, and cell number was determined using ahaemacytometer. Also, following the inoculation, the number of attached cells on the microcarrierswas determined by a procedure developed by Sanford et al. (1951) and modified by others formicrocarrier cultures (van Wezel, 1967; Levine et al. 1977). A sample (1 ml) of the microcarrierculture was transferred into a 15 ml centrifuge tube. The sample was then centrifuged at1000 rev. min"1 for 3 min. This step was necessary only for determination of initial cell number,when the cells were not attached to the beads. For subsequent samples, after the first 24 h after thesampling, the tubes were left standing for 3 min. The latter procedure was adequate for separation.The supernatant was then removed, and the microcarrier culture was washed twice with 2 ml ofPBS for complete removal of free cells. After recording the bead volume, which was about 0-1 or0-4 ml for samples obtained from batch or perfused cultures, respectively, the mixture was made upto 1-4 ml with a solution of 0-1 % Crystal Violet and 0-1 M-citrate in water. The mixture was leftstanding at either 37°C or 25°C for 2 or 48 h, respectively, for releasing and staining the nuclei.

72 A. T. Nahapetian, jf. N. Thomas and W. G. Thilly

After the incubation period, the mixture was drawn in and out of Pasteur pipettes a number oftimes before application to a haemacytometer for enumeration. The beads could not enter into thehaemacytometer, so they did not interfere with the counting process. For the same reason, and alsobecause the volume occupied by the beads was not available for distribution of the nuclei, beadvolume was subtracted from the total volume of the mixture for calculation of the number of nucleiin the original sample. The correction was especially critical at 20 g I"1 bead concentration, whenthey occupied close to 50% of total culture volume.

Visual examinationTwo to three drops of the microcarrier culture were transferred into a multiwell plate, and the

cells were stained with 5 drops of the Crystal Violet reagent for 3 min. After washing three timeswith 2-ml portions of PBS, the stained cells on the microcarriers were examined with a microscope.

pH and dissolved oxygenThe pH of the experimental samples was determined with a Corning pH meter (model 125,

Corning Glass Works). After replacing the pH electrode with the dissolved oxygen probe, the samemeter was used with an amplifier constructed at the Massachusetts Institute of Technology formonitoring DO in the perfused cultures.

MetabolitesThe frozen samples were thawed and mixed with an equal volume of 6% (v/v) solution of

perchloric acid in water. After cooling in an ice bath, the deproteinized samples were neutralizedby 2M-potassium hydrogen carbonate solution in water. They were then centrifuged at2000 rev. min"1 for 5 min in a refrigerated centrifuge (model DPR-6000, International EquipmentCo., Needham Heights, MA). Glucose, fructose (Bernt & Bergmeyer, 1981a,b), pyruvate, lactate(Czok & Lamprecht, 1981), glutamine (Lund, 1981), glutamate (Bernt & Bergmeyer, 1981a,6)and ammonia (Kun & Kearney, 1981) concentrations in the supernatant were determined bystandard enzymic methods. The net change in quantity of a metabolite in the perfused systemexpressed in pmol cell" day" was calculated as follows: from the data on culture medium volume(100 ml) and metabolite concentration on two subsequent days, the daily change in total quantity ofa metabolite in the reactor vessel was calculated. Then, from the data on concentration of themetabolite in fresh and spent media and volume of spent medium in the waste reservoir, total dailychange in quantity of the metabolite, excluding the change in the reactor vessel, was calculated.The algebraic sum of the two changes was then divided by the average total cell number for the twodays. For a similar calculation in a batch culture, the daily change in total quantity of a metabolitein culture medium was divided by the mean total cell number for the two days. Positive andnegative values represented net production and utilization, respectively.

RESULTS

Cell number

Cell number in the perfused cultures either remained at a relatively constant levelor declined sharply after the start of perfusion with 25% or 12-5% L15 media,respectively (Fig. 2). In contrast, the cells in simultaneously run batch culturesmultiplied for at least 3 days.

Perfusion of the cultures with 50% and 100% L15 media, however, resulted insignificant increases in cell number until the last day of the study (Fig. 3). Vero cellgrowth was apparently higher at the higher perfusion rate (800 ml day" ). During thefirst 6 days, there were no significant differences between growth rates of 1X and0-5x perfused cultures. However, the growth rate was apparently higher when thecultures were perfused with 50% L15 medium during the second week. Visual


100

o

X

i8.IE

I

1 2 3 4Time (days)

Fig. 2. Growth curves of Vero cells in perfused ( ) and batch ( ) cultures.Perfusion of the cultures (100 ml) began 24 h after cell inoculation at either 4 (A) or 8(B) vol. day"1. Microcamer concentration in perfused and batch cultures was 20 and5gl~ ' , respectively; experimental media were: 0-250X (#) , modified L15 (contained4mM-glutamine and lOmM-fructose, instead of galactose) + 5 % foetal bovine serumdiluted 1:3 (v/v) with PBS; 0-125X (O), modified L15+S % foetal bovine serum diluted1:7 (v/v) with PBS. Each circle represents a single value.

examination of microcarrier cultures in the latter group showed healthy cells thatremained firmly attached on the bead surfaces. Some cell detachment was observedin cultures with the more concentrated (100%) basal medium by the 16th day. Thehighest cell density in the perfused cultures under the most favourable nutrientconcentration (50%) and perfusion rate (800 ml day"1) was 3X107 cells ml"1. At thehighest perfusion rate, growth curves of the cultures had a clear biphasic pattern(Fig. 3). In general, Vero cells proliferated with a 24 h doubling time until day 6,when the growth rate was considerably diminished (96h doubling time), remainingat the low level during the remaining experimental period.

There was a significant increase in cell number in batch cultures that were runsimultaneously with the 50% and 100% basal media (Fig. 3). The highest celldensity under optimum batch culture medium concentration (100%) was 3xlO6


cells ml"1, which was reached by day 6. Following day 6, however, attached cellnumbers steadily declined until the last day of the study.

pH

Medium pH in the reactor vessel (Fig. 4) was lower than that collected in the wastereservoir. The difference was especially apparent for the IX perfused system. Ingeneral, pH of the perfused cultures, after an initial drop, was maintained between6-9 and 7-1 during the entire period of the study. In batch cultures (Fig. 5), after aninitial drop, pH was maintained at neutrality until day 6, when it began to risesteadily and reached 7-6-7-8 by the 16th day. A slight but detectable rise in mediumpH was observed in the spinner flasks without cells (Fig. 5).

Dissolved oxygen

Relative dissolved oxygen (Fig. 6) dropped sharply and reached zero by days 3and 4, respectively, in IX and 0-5 X perfused cultures. When it declined to valueslower than 2 % of saturation in any one of the perfused flasks, the flow-rate of the airand oxygen mixture was increased for all perfused cultures until the DO was raised to50—90% of full saturation. Dissolved oxygen remained at relatively higher levels in0-5 X than in 1X perfused cultures, especially during the first 6 days of the study.

Metabolites

During the first 24 h before perfusion, there were sharp declines in concentrationsof glucose, fructose, pyruvate and glutamine (Fig. 7), with concomitant increases inconcentrations of lactate, ammonia and glutamate (Fig. 8). Although there wereapparent decreases in glutamine and increases in ammonia (Fig. 9) and pyruvate,medium concentrations of the remaining metabolites did not change significantly inthe control batch spinner flasks without cells, which were run simultaneously. Thedata illustrate cyclic changes in concentrations of glucose, fructose and glutamine inthe l x perfused culture.

After the initial sharp decline, glucose concentrations (Fig. 7) in both perfusedand batch 0-5X cultures were maintained at very low levels (0—50 nmol ml"1) for theremaining period of the study. In the IX batch culture, after the initial decline andan apparent increase on day 2, concentration of glucose decreased to the low levelsobserved in the 0-5X cultures. Glucose concentration in l x was higher than 0-5Xperfused cultures, especially on days 3, 4 and the last 4 days of the study. Followingperfusion, fructose concentration (Fig. 7) in both 0 5 X and l x perfused culturesincreased for 1 and 2 days, respectively, and reached slightly higher than their initial

Fig. 3. Growth curves of Vero cells in perfused ( ) and batch ( ) cultures.Perfusion of the cultures (100 ml) began 24 h after cell inoculation at either 4 (A) or 8(B) vol. day"1. Microcarrier concentration in perfused and batch cultures was 20 and5gl~', respectively; experimental media were: 1-0X (#), undiluted modified L15(contained 4 mM-glutamine and lOmM-fructose, instead of galactose) +5% foetal bovineserum; 0-5X (O), modified LlS+5% foetal bovine serum diluted 1:1 with PBS. Eachcircle represents a single value.


1000

o

E

&

J{3C

u

oXi

8.

3C

3

uuu

100

10

r

r

•J

1

/

iif

/9

1

1

f \\\1

1 1

\

1

b\\\\\\\

s1 V J .

1 1 1

B •

-

"*— -• •

i i i

2 4 6 8 10 12 14 16

Time (days)

Fig. 3


values. Between days 3 and 6, fructose concentration declined in both l x and 05 Xperfused cultures. The rate of decrease was apparently greater in the former,however. Following day 6, fructose concentration in 0*5X perfused cultures dim-inished steadily until it reached the lowest level (2500 nmol ml"1) by day 12. After aslight increase in the last 4 days, glucose concentration was maintained at about3000 nmol ml"1. Following day 6, fructose concentration in lx perfused culturesfluctuated between 7000 and 8500 nmol ml"1 during the remaining period of thestudy. In batch IX and 0-5x cultures, following a slow decline for 2 and 3 days,respectively, fructose concentrations declined at a higher rate until day 6. Followingday 6, concentration of the monosaccharide in the 05X batch culture was main-tained at about 1000 nmol ml"1 until the last day. In contrast, concentration offructose in IX batch cultures continued to decline during the same period.

Concentration of pyruvate (Fig. 7) in both IX and 0-5 X perfused culturesincreased between days 1 and 2 and then declined to about 200 nmol ml"1 by day 6.The latter concentration was maintained in l x , while it declined further in 0-5Xperfused cultures to 100 nmol ml"1 during the remaining experimental period. In thebatch cultures, after a sharp drop, pyruvate concentration remained at a low level(50 nmol ml"1) between days 7 and 16.

Glutamine concentration (Fig. 7) in both lx and 0-5x perfused cultures rosesharply between days 1 and 2, due to the initiation of perfusion. The changes inconcentration appeared to be cyclic between days 2 and 16, especially for the IXperfused culture. After a slow decline between days 2 and 3, glutamine concentrationdecreased sharply on days 3 and 4. The decline was followed by an increase betweendays 4 and 10, then a return to the day 4 value by the 14th day of the experiment. Inthe 0-5 X perfused cultures, after a slow decline between days 2 and 4, concentrationof the metabolite fluctuated between 250 and 1000 nmol ml"1 between days 4 and 16.Concentration of glutamine in 0-5 X and IX batch cultures decreased and reachednearly zero by days 6 and 10, respectively.

Lactate concentration (Fig. 8) in the 1X perfused culture declined during the firstday following the start of perfusion. Concentration increased sharply between days 2and 6 to a maximum of 4500 nmol ml"1, which was followed by a decline to a finalvalue of 3000 nmol ml"1 by day 12. Concentration in 0*5 X perfused cultures wassignificantly lower. After initiation of perfusion with 0-5 X L15 medium, lactateconcentration rose relatively slowly until day 5, when it reached a maximum of1500 nmol ml"1, maintained for 5 days. Following day 10, the concentration de-clined and reached its initial level in fresh 50% L15 medium towards the end of thestudy. Concentration of lactate in lx batch culture rose, and by day 3 reached amaximum (4000 nmol ml"1), which was maintained until day 6. Following day 6,

Fig. 4. Changes of pH in perfused cultures (A) and in waste reservoir media (B).Perfusion of the cultures (100ml) began 24h after cell inoculation at 8 vol. day"1.Microcarrier concentration was 20gl~' ; experimental media were: 1-0X ( • ) , undilutedmodified L15 (contained 4 mM-glutamine and 10 mM-fructose, instead of galactose) + 5 %foetal bovine serum; 0-5X (O), modified L15+5 % foetal bovine serum diluted 1:1 withPBS. Each circle represents a single value.


Q.

6 8 10

Time (days)12 14 16

6 8 10Time (days)

12 14 16

Fig. 4


after a sharp decline, the metabolite concentration reached a level that was hardlydetectable and remained there until the end of the study. After a rise to1500 nmol ml"1, which was maintained until day 5, concentration of lactate in 0-5xbatch culture declined and remained close to the initial value in 50% L15 mediumuntil the last day of the study.

Ammonia concentration (Fig. 8) in both IX and 0-5 X perfused cultures declinedto 500 nmol ml"1 24 h after the start of perfusion. After a lag period, ammoniaconcentrations increased and reached maximum levels on days 3 and 6, respectively,for IX (lSOOnmolml"1) and 0-5X (nOOnmolml"1) perfused cultures. Followingday 6, ammonia concentration in the perfused cultures declined until day 10, thenrose and returned to the day 6 value by the 12th day. The apparent cyclic pattern ofthe changes continued throughout the experimental period. In the IX and 0-5 Xbatch cultures, however, there was a steady increase in concentration of themetabolite until days 6 and 7, respectively. After a 1-day lag period, ammoniaconcentration increased until the last day of the study, reaching 4700 and2500 nmol ml"1, respectively, in l x and 0-5X batch cultures.

The cyclic pattern of changes in concentration of glutamate (Fig. 8) in theperfused cultures was similar but more pronounced than that described above forammonia. Following a decrease on day 1 and a lag period between days 2 and 3,concentration of the amino acid increased and reached maximum levels by days 5and 6, respectively, in l x (SOOnmolmP1) and 0-5X (300-500nmolml"1-) perfusedcultures. Concentration of the metabolite fluctuated between 200 and 600 nmol ml"1

during the remaining period of the study. In IX and 0-5 X batch cultures, however,there were steady increases in concentrations of glutamate until days 4 and 5,respectively. Following day 4, glutamate concentration in the l x batch culture waselevated in an apparent cyclic manner until it reached 800 nmol ml"1 by day 16.Similarly, concentration of the amino acid in the 0-5X culture increased, althoughwith less-apparent fluctuations, and reached about 600 nmol ml"1 during the sameperiod.

Net utilization and production of all the metabolites are illustrated in Figs 10and 11, respectively. Utilization of glucose in 1X was apparently higher than in 0 5 Xperfused cultures, especially 24 h before and after the start of perfusion. Moreover,utilization was relatively complete (Fig. 7) and, consequently, was much higher incultures with a smaller number of cells (Fig. 10). There was a net utilization offructose in all cultures at all times except during the 5 days following the start ofperfusion, when there were positive, negative or zero net changes in quantity of themetabolite. In contrast to glucose, with fructose there were no apparent differencesin utilization among the cultures on day 1. Between days 1 and 6, however,utilization of the monosaccharide in l x tended to be more than in 0-5X cultures.

Fig. 5. Changes of pH in batch cultures (A) and in batch spinner flasks without cells (B).Microcarrier concentration was 5gl~'; experimental media were: 1-0X (#), undilutedmodified L15 (contained 4mM-glutamine and lOmM-fructose, instead of galactose) +5 %foetal bovine serum; 0-SX (O), modified L15 + 5 % foetal bovine serum diluted 1:1 withPBS. Each circle represents a single value.


CL

80

7-5

7-0

6-5

-

_

\^\

l i

1 1

1

I

I

1

1 1

y//

i i

i 1

A

-

s

^ • '

1

_

- 0

-

i

2 4 6 8 10 12 14 16Time (days)

6 8 10 12Time (days)

Fig. 5

14 16


The difference was less during the period following day 6, when utilization offructose was maintained between 0-5 and 1 pmolcelP1 day"1.

Glutamine utilization in IX was apparently higher than in 0-5 X cultures 1 daybefore and after initiation of perfusion. However, during the remaining period of thestudy, utilization of the amino acid in IX was either equal to or less than in 0-5 Xperfused cultures. In contrast, utilization of pyruvate in IX was either equal to orless than in 0-5 X perfused cultures during the initial 2 days of the study. The greaternet decreases in pyruvate in the more diluted cultures lasted until day 6. During theperiod following day 6, however, utilization of the metabolite in IX was higher orequal to 0-5X perfused cultures.

Production rates of lactate, ammonia and glutamate (Fig. 11) declined followingthe initiation of perfusion and reached their lowest level by the end of the secondweek. The rate of lactate production in lx was apparently higher than in 05Xcultures during the entire period of the study. The difference was more apparentby the 16th day, when the net change of the metabolite was near 0 and 0-5pmolcelP1 day"1, respectively, in 0-5X and IX perfused cultures. Ammonia pro-duction rate in 1X was higher than in 0-5 X cultures during the first 24 h of the study.Following perfusion, however, ammonia production in l x was less than in 0-5xperfused cultures until day 6, when the difference became less-apparent and re-mained so until the last day. The changes in production rates of glutamate weresimilar to those that were described for ammonia. However, the difference betweenthe production rates of the metabolite in IX and 0-5X cultures were less-apparentthan those observed for ammonia.

Changes in lactate/pyruvate ratio (Fig. 12) were not apparently different betweenIX and 05 X perfused cultures during the first 3 days of the study. However, afterthe third day, the ratio in IX (15-25) was higher than in 0-5X (5-10) perfusedcultures. In IX batch culture, after a rise to a non-physiological level (45) by day 7,the ratio dropped to its initial value by the 16th day. The lactate/pyruvate ratio in0-5x batch cultures, however, was maintained between 8 and 17 between days 2 and16. The cyclic pattern of changes in the lactate/ammonia ratio (Fig. 13) was almostidentical to that observed for concentration of glutamine (Fig. 7).

DISCUSSION

Earlier investigations suggested that mammalian cell growth and survival inculture could be affected by availability of nutrients (Eagle, 1955; Graff & McCarty,1957), growth-promoting endogenous metabolites synthesized by the cells in culture(Rubin, 1966; Horng & McLimans, 1975), accumulation of endproducts (Rubin,1966; Burk, 1966; Stoker & Rubin, 1967; Holley et al. 1978; Thilly et al. 1982;Butler et al. 1983), changes in oxygen tension (Graff & McCarty, 1957; Stevens,1965; McLimans et al. 1968; Himmelfarb et al. 1969; Richter et al. 1972; Knazeket al. 1972; Jensen et al. 1974; Werrlein & Glinos, 1974; Butler et al. 1983; Knopet al. 1984; Kashiwagura et al. 1984), and limitation of surface for attachment ofanchorage-dependent cells (Abercrombie & Ambrose, 1962; Kruse et al. 1963;


6 8 10Time (days)

16

6 8 10Time (days)

12 14 16

Fig. 6. Changes in dissolved oxygen in perfused cultures. Perfusion of the cultures(100 ml) began 24 h after cell inoculation at 8 vol. day"1. Microcarrier concentrationwas 20gl~ ' ; experimental media were: (A), undiluted modified L15 (contained4mM-glutamine + lOmM-fructose, instead of galactose) + S % foetal bovine serum;B, modified L15+5% foetal bovine serum diluted 1:1 with PBS. ( • ) Dissolved oxygenwas raised by manually increasing oxygen supply; (O) daily mean value.


6 8 10Time (days)

12 14

10000 -

6 8 10

Time (days)

12 14 16

Fig. 7. Changes in concentrations of glucose (A), fructose (B), pyruvate (c) andglutamine (D) in perfused ( ) and batch ( ) cultures of Vero cells. Perfusion of thecultures (100ml) began 24h after cell inoculation at 8 vol. day"1. Microcarrierconcentration in perfused and batch cultures was 20 and S g P 1 , respectively; exper-imental media were l-0x ( • ) , modified L15 (contained 4mM-glutamine + 10mM-fructose, instead of galactose) + 5 % foetal bovine serum; 0-5x (O), modified L15+5 %fnetnl bnvinp senim diluted 1:1 with PBS. Each circle reDresents a sinele value.


6 8 10Time (days)

12 14 16

TE"oEc

tion

a

entr

0

c8

EC3

Glu

i

3500

3000

2500

2000

C

1500

1000

500

1 1 1

{

1 \/ 1/ \

1 k \\ \

>\ \ \ K /

\\ \ * Y *

4 v ̂ ^\

0 0 •A/N S3

I I I I I

D-

_

-

A

\ >•

-

\ /6 8 10

Time (days)

Fig. 7C,D

12 14 16


4500

4000

3500

3000

2500

2000

1500

1000(

500<l

I 1 1

A

-'/ *

1 / ' ̂1 \

" / /J I-\l /TV1! >a_

Ii i i

i i i i i

A

^ \

\ \

\\\\\\

-̂o ^\

"-o •«-"\̂ 1*v / ^ ^

l 1 l l"" — -f

6 8 10 12 14 16

Time (days)

6 8 10

Time (days)

Fig. 8A,B

12 14 16


van Wezel, 1967; Horng & McLimans, 1975; Levineef al. 1977, 1979; Butler et al.1983). The data in Fig. 2 indicate that either deficiency of nutrients, conditioningfactor(s), or both,.were probably responsible for the poor growth and survival ofVero cells in cultures that were perfused with 12-5 % and 25 % L15 media. Doublingnutrient supply in both 0-125X and 0-25 X cultures by increasing the perfusion ratefrom 4 to 8 vol. day"1 resulted in no significant increase in cell growth (Fig. 2), sincethe rate of synthesis of conditioning factors apparently was less than the rate ofremoval. Similarly, 0-5X Vero cell cultures grew when perfused at 4vol. day"1

(Fig. 3), while no appreciable increase in cell number (Fig. 2) was observed incultures when nearly the same quantity of nutrients was supplied by decreasing themedium concentration (0-25X) and increasing the flow-rate (8 vol. day"1).

The cultures were evidently not limited in nutrients when they were supplied withthe 50% and 100% L15 media at either 400 or 800 ml day"1 rates of perfusion. Thefaster flow rate, however, provided a more favourable environmental condition forproliferation and survival of Vero cells in this study. The observed biphasic patternof cell growth (Fig. 3) under the optimum culture conditions was similar to thosereported for perfused cultures of rat Jensen sarcoma (Kruse et al. 1963) and MDCK

1EO

UIU

w

o

u

8

: co

rut

amat

i

O

900

800

700

600

500

400

300

200

100"

)

1

-

-

-

J1 /'-B /IFPLL/ t>-ci

1

i

f/ N/I/I

/

/>

i

/ \

A \J \ \

1 1

1 1

/

•

v̂ \ \

i i

i i

c-

s

-

/

sk -

-i i

10 12 14 16

Time (days)

Fig. 8. Changes in concentrations of lactate (A), ammonia (B) and glutamate (c) inperfused ( ) and batch ( ) cultures of Vero cells. Perfusion of the cultures (100 ml)began 24 h after cell inoculation at 8 vol. day"1. Microcarrier concentration in perfusedand batch cultures were 20 and SgI"1, respectively; experimental media were: 1-Ox ( • ) ,modified L15 (contained 4 mM-glutamine + lOmM-fructose, instead of galactose) + S %foetal bovine serum; 0-5X (O), modified L15+5% foetal bovine serum diluted 1:1 withPBS. Each circle represents a single value.


cells (Butler et al. 1983). However, the initial exponential cell proliferation (24 hdoubling time) lasted for 6 days, compared with the 2 days reported in the earlierinvestigations. Furthermore, 'record' Vero cell densities (3X107 cells ml"1) wereobtained in a perfused microcarrier culture. The next highest value in the literature(1X107 cells ml"1) was also reported from our laboratory, using perfused MDCKcells on microcarriers (Butler et al. 1983). Maximum cell density in batch cultures(3xl06cellsml~1) was 10 times lower than that in the perfused culture. A slightlylower value (2X106 cells ml"1) was reported for a batch MDCK cell culturemaintained under similar conditions (Butler et al. 1983).

Cell/bead ratio at the time of inoculation was 6-8, which ensured a very lowprobability of zero hits (0-04—0-008) between the two culture components, ascalculated by Butler & Thilly (1982). It was also reported that a culture with a5 mgrnl"1 microcarrier concentration could accommodate 2-8 (± 06) X 106 MDCKcells on the bead surface. Assuming an equal mean cell size for Vero and MDCKcells, it was deduced that the microcarriers in the cultures were completely coveredwith a monolayer of cells by day 6. Visual examination of the beads indicated that thiswas indeed the case. Moreover, in the perfused system with 20mgml~1 beadconcentration, the high cell concentrations (3xl07cellsml~ ) and visual examin-ation both indicated multilayering of the anchorage-dependent cells. At the higherbead concentration, confluent microcarrier monolayers were formed at 1x10cells ml"1 density. Abercrombie & Ambrose (1962) introduced the term 'contactinhibition' for a growth-cessation phenomenon observed in a monolayer confluentculture of anchorage-dependent cells. The data in the present study and those ofothers (Kruseef al. 1963, 1967, 1969, 1970; Kruse & Miedema, 1965; Knazeket al.1972, 1974) do not support this hypothesis, but they show a density-dependentretardation of cell growth. We suggest that Vero cells are not contact-inhibited, asthey proliferate and form multilayers on the carriers under the optimized conditionof the perfused culture. The observed retardation during the second phase of cellgrowth (between days 6 and 16) may be due to slower diffusion and exchange ofcellular microenvironmental components.

The most reasonable explanation for the difference in pH between the media in thereactor vessels and the waste reservoirs (Fig. 4) is production of volatile acidicmetabolites that were eliminated during passage of the spent medium through thesilicone tubing and storage in the waste reservoir. The volatile compound wasassumed to be carbon dioxide. If the assumption is true, then the pH data indirectlyindicate higher aerobic oxidative reactions in 1X than in 0-5X perfused cultures. Thediffusion and elimination rate of the respiratory waste product could not match therate of production, which led to its relatively excessive accumulation in the IXperfused culture. In addition, the production rates of the metabolic endproductssuch as lactate and ammonia (Fig. 8) seemed to neutralize each other, so that theyhad no apparent net effect on the medium pH. The rise in batch culture pH (Fig. 5)following day 6 was most probably due to utilization of lactate (Fig. 8), whileammonia accumulation continued.


1200

1000

800

600

400

+200

0

-200

400

600

800

1000

1200

1 1 1 1

-

-

• # #

8 a * ° o # ft0 0 * o O

o o

O

o o

~~ o

1 1 1 1

1 ' ' i—

• #

•-

—

Q - 9 o

—

—

—

o _

o —o

1 1 1 110 12 14 16

Time (days)

Fig. 9. Cumulative changes in concentrations of ammonia ( • ) , glutamine (O) andglutamate (©) in a batch spinner flask without cells. Microcarrier concentration was5gl~ ' ; experimental medium was modified LI5 (contained 4 mM-glutamine + 10 mM-fructose, instead of galactose) + 5 % foetal bovine serum. Each circle represents a singlevalue.

Data on dissolved oxygen (Fig. 6) suggest, although qualitatively, that the con-sumption of oxygen in 1X was higher than in 05 X perfused cultures, which supportsthe earlier suggestion for the higher rate of degradative reactions in 1X. The signifi-cance of 'available' oxygen in cell microenvironment has been discussed extensively(Graff & McCarty, 1957; Stevens, 1965; McLimansef al. 1968; Himmelfarb et al.1969; Richteref al. 1972; Knazekef al. \911; Jensen et al. 1974; Werrlein & Glinos,1974; ThiWyet al. 1982; Butler et al. 1983; Knop et al. 1984; Kashiwagura et al.1984). Cell growth and survival are known to be highly sensitive to changes in oxygentension in the culture environment, as discussed in the Introduction. According to a

A. T. Nahapetian, J. N. Thomas and W. G. Thilly

Time (days)

oo.

o

oo

-0-2

0-4

0-6

0-8

10

1-2

1-4

1-6

1-8

2-0

2-2

2-4

: i;12

r

i *

4T

6 8 10 12 14 16

I I ' II ' 1 1 ' LL ' |A ' $ I V

-

-J

-

-

-

-

I I I I I I I

2-0 -

•a

T +1-0

2 - 1 0

Time (days)

6 8 10 12 14 16

cCO

1 11 1

•! fi i

T

i

1<;?

c

1 1 1 1 1 1 1

> B

?iii

4\'\

i

1)I]

i

-2-0

30 -

Fig. 10. Utilization of glucose (A), fructose (B), glutamine (c) and pyruvate (D) inperfused ( ) and batch ( ) cultures of Vero cells. Perfusion of the cultures (100 ml)began 24 h after cell inoculation at 8 vol. day"1. Microcarrier concentration in perfusedand batch cultures was 20 and 5gl~ ' , respectively; experimental media were: 1-0X (# ) ,modified L15 (contained 4mM-glutamine + lOmM-fructose, instead of galactose) + 5 %foetal bovine serum; 0-Sx (O), modified L15+5 % foetal bovine serum diluted 1:1 withPBS. Each circle represents a single value.


Time (days)

2 4 6 8 10 12 14 16

- 0

0

1

1

2

2

2

3

3

4

4

4

•4

8

2

6

0 •

4 fi

8 -

2 -

6 -

0 -

4 -

8 -

]

—0-0

i

k I

T" I

I I I I I I I

Time (days)

oD.

Uooc

-0-1

0-2

0 - 3 •

0-4 •

0-5

0-6 •

0-7 •

0-8 •

0-9 -

10

1-1 j

i i

l

T 1

•0 ||

1

1

2

|

j

11

i

i

1

I

4I1

I

*6T

81

*

101

12l

141 1

h

16II

D

i i J

: ;

-

-

-

-

-

i i i i i

Fig. 10C.D

Net

cha

nge

in a

mm

onia

(pm

ol c

ell-

' da

y-')

+

--

--

+

6G

b&

do

Z

St

JZ

-

0-

i

I I

I I

Net

cha

nge

in l

acta

te (

pmol

cel

l-'

day-

')

I +

-+

Zo

Z+

G g

zg

xg

zg

zg


model proposed by McLimans et al. (1968), if respiration rate did not exceed thedelivery rate of oxygen, then the adaptive growth of cultured cells could take place.Our results suggest that the dissolved oxygen (Fig. 6) became limited at densities of3xl06cellsmr1 (Fig. 3) in IX and 0-5x perfused cultures on days 3 and 4,respectively. This cell density was close to the maximum observed in batch cultures.Similarly, oxygen has been reported as the most probable growth-limiting factor inbatch cultures of Vero cells (Jensen et al. 1974). The design of the oxygen supply inthe present study was such that the total quantity of gas supplied to 1X was abouttwice that in 0 5 X perfused cultures. Between days 3 and 6 (Fig. 6), even with thegreater rate of oxygen supply, DO in IX was much lower than 0-5 X perfusedcultures. In both perfused cultures, however, there was a proportional increase incell proliferation and respiration until day 6 when cell number reached- a value(IX 107 cells ml"1) required for formation of a confluent monolayer in a microcarrierculture with ZOmgP1 bead concentration. Following day 6, the respiration ratecontinued to increase, while the delivery rate of oxygen to the cells was graduallydiminished due to formation of multicell -layers, .and cell and bead aggregates.

i

eg•a

I3

5ana

u

0-8

0-7

0-6

0-5 '

0-4

0-3

0-2

+0-1

l

I

o

h

o 1 1 1 1 1

1

11

A1I

->o

ft

t

1

9

j

c -

-

-

-

-

-

-

10 12 14 16

Time (days)

Fig. 11. Production of lactate (A), ammonia (B), and glutamate (c) in perfused ( )and batch ( ) cultures of Vero cells. Perfusion of the cultures (100 ml) began 24 h aftercell inoculation at 8 vol. day"1. Microcarrier concentration in perfused and batch cultureswas 20 and 5gl"1 , respectively; experimental media were: 1-0X (# ) , modified L1S(contained 4 mM-glutamine and 10 mM-fructose, instead of galactose) + 5 % foetal bovineserum; 0-5X (O), modified L15+5% foetal bovine serum diluted 1:1 with PBS. Eachcircle represents a single value.


m

yruv

;

a.C3

3

45

40

35

30

25

20

15

10

5

(

1

-

-

-

—

-/

/

f— .

IP-

I'llfly*

i i

/ \

/ \/ \

•

1

11

•

j

//r AI / \

fO o—1

i i

i i

i\\\\\\\\\\*—-̂

0^5!

i i

i i

\

\ ^ JO

i i

-

-

—

—

—

-

~ — •

—

" • " • - •

i

6 8 10Time (days)

12 14 16

Fig. 12. Changes in lactate/pymvate ratios in perfused ( ) and batch ( ) culturesof Vero cells. Perfusion of the cultures (100 ml) began 24 h after cell inoculation at8 vol. day"1. Microcarrier concentration in perfused and batch cultures was 20 and 5gl" 1 , respectively; experimental media were: 1-0X ( • ) , modified L15 (contained 4mM-glutamine + lOmM-fructose, instead of galactose) + 5% foetal bovine serum; 0-5x (O),modified L15+S % foetal bovine serum diluted 1:1 with PBS. Each circle represents asingle value.

Moreover, presumably due to relatively lower consumption and thus higher avail-ability of oxygen, the cell proliferation rate in 0-5 X was higher than in IX perfusedcultures during the period following day 6.

Vero cells in batch spinner flasks (Fig. 3) proliferated until day 5, when they beganto slough off the microcarriers. Cessation of growth and cell detachment in the batchcultures seemed to be due both to deficiency of nutrients and accumulation of wasteproducts in the closed environment. An extremely high rate of glucose utilization,which resulted in almost complete utilization within the first 24 h (Fig. 7), confirmsthe results reported in other investigations (Graff & McCarty, 1957; Graff et al.1965; Himmelfarb et al. 1969; Zielke et al. 1978; Brand et al. 1984). Graff et al.(1965) suggested that the actual glucose requirement of cells in culture for eitherproliferation or maintenance was much less than the levels that cells could consumewhen glucose was supplied at relatively high levels. The excess was converted tolactic acid. However, if it was supplied at the required level, the cells becameconsumers rather than producers of lactic acid (Graff et al. 1965). Data presented inFig. 11 support this hypothesis. The fluctuations in concentration of glucose in 1X


perfused cultures (Fig. 7), which were small but detectable, were probably dueto changes in availability of dissolved oxygen in the culture microenvironment.Between days 1 and 6, glucose concentration in the 0-5 X perfused cultures (Fig. 7)were close to the requirement level reported by Graff et al. (1965). They observednegligible lactic acid production and uniformly good results when glucose concen-tration was maintained at lmglOOmP1 (S6nmolml~1) by regulating the rate ofglucose perfusion in their cytogenerators or chemostats. More recently, it wasreported that low levels of glucose (80 nmol ml"1) in media were essential for growthand survival of mammalian cells in culture (Zielke et al. 1978, 1984). The high rateof glucose utilization might be related in part to its high absorption. Gay & Amos(1983) found that glucose transport was enhanced when fructose was provided as themaintenance sugar, as was the case for the L15 medium used in the present study.

During the first week, there was no persistent high rate of fructose utilization(Figs 7, 10) similar to that discussed above for glucose. Fluctuations in fructoseconcentration/ especially in the IX perfused culture, can be explained by fluctu-ations of available oxygen in the culture microenvironment. The observed negative,

6 8 10Time (days)

12 14 16

Fig. 13. Changes in lactate/ammonia ratios in perfused ( ) and batch ( ) culturesof Vero cells. Perfusion of the cultures (100 ml) began 24 h after cell inoculation at8 vol. day"1. Microcarrier concentration in perfused and batch cultures was 20 and5 g l ~ \ respectively; experimental media were: 1-0X (#) , modified L15 (contained 4 mM-glutamine + lOmM-fructose, instead of galactose) + 5 % foetal bovine serum; 0-5X (O),modified L15+5 % foetal bovine serum diluted 1:1 with PBS. Each circle represents asingle value.


positive or zero net changes in fructose were probably related to the rates of glycolyticand gluconeogenic reactions in the perfused cultures. A decrease in oxygen concen-tration has been reported to lead to a decline in gluconeogenesis (Kashiwagura et al.1984). Net utilization of fructose was apparently low during the first week, especiallyin 05 X perfused cultures. Nevertheless, utilization of the sugar became apparenteven in the latter group during the second week. Similarly, fructose utilization in thebatch cultures was apparently higher at higher cell densities. In agreement with otherreports (Reitzer, Wice & Kennell, 1979; Imamura et al. 1982; Thilly et al. 1982),fructose was utilized more efficiently than glucose in the present study. The slow rateof fructose utilization might be related to limitation of a carrier-mediated transport ofthe monosaccharide (Van den Berghe, 1978).

There was a persistent utilization of glutamine in both batch and perfused culturesduring the entire period of the study (Fig. 10). The amino acid in the batch cultureswas almost completely consumed by the end of the first week (Fig. 7). Zielke et al.(1984) have suggested that 30-50 % of the energy requirement of mammalian cells ismet by aerobic oxidation of glutamine. It was also proposed that the amino acid couldbe the sole source of energy when the glucose concentration was low or when it wasreplaced by other carbohydrates. In contrast, Nagle & Brown (1971) have demon-strated that mouse L and HeLa cells could grow in a chemically defined mediumwithout glutamine. Moreover, in a study in our laboratory it was found that Verocells could grow and reach confluence on microcarriers and standard roller bottlescontaining the basal L15 medium supplemented with lOmM-fructose instead ofgalactose and 10 % foetal bovine serum, but not with glutamine (Thilly, Varunsatian& Nahapetian, unpublished data). The results were recently confirmed by others(Wolfrom, Polini, Decimo & Gautier, 1984). In both investigations, following aslow-growth period lasting from 1-2 weeks, there was a significant increase in cellproliferation in the glutamine-deficient group. There was no significant differencebetween the final cell numbers of test and control groups supplemented with theamino acid, by the end of the study. Since the test media were not supplementedwith glutamine or glucose, fructose was thus the only major source, apart from otheramino acids and pyruvate, that the cells could have utilized for production of energyand synthesis of triglycerides and nucleic acids required for cell proliferation. Thecontribution of fructose to the energy supply was probably as significant as that ofglutamine during the second week at the high cell densities produced under theconditions of the study.

There were significant fluctuations in concentration of glutamine (Fig. 7) thatwere very similar to changes in lactate/ammonia ratio (Fig. 13). The oscillations inconcentrations of ammonia and glutamate (Fig. 8) were inversely related to thechanges in concentration of glutamine. Glutamate accumulated in the batch cultureseven when the cultures were depleted of other sources of nutrients, including lacticacid (Fig. 8). Accumulation of glutamate has also been reported by others (Kruseet al. 1967; Butler et al. 1983). Ammonia concentration in the perfused cultures(Fig. 8) remained below the level (2000nmolml~1) known to be toxic for mam-malian cells in culture (Visek et al. 1972; Holley et al. 1978; Thilly et al. 1982).


However, the concentration of ammonia rose to toxic levels in a perfused culture ofMDCK cells (Butler et al. 1983), due apparently to the lower rate of perfusion usedin the earlier investigation. Concentration of ammonia in 1X and O'Sx batch culturesreached toxic levels by days 4 and 6, respectively.The final concentration of ammoniain IX batch culture (4500nmolml~') was more than twice the toxic level. The datafor batch cultures without cells (Fig. 9) suggest that spontaneous degradation ofglutamine (Tritsch & Moore, 1962) probably contributed to a small but significantportion of the rise in concentration of the metabolite under the environmentalconditions used in the present study.

Among the changes in metabolites, accumulation (Fig. 8) and production(Fig. 11) of lactate were the most significant differences observed between theoverfed (IX) and optimized (0-5x) perfused cultures. Relatively high accumulationand production of lactate in 1X perfused cultures was probably due to excessivesupply of nutrients. Owing to limitation of available oxygen and saturation ofrespiratory pathways, lactic acid accumulated. This phenomenon has also beenobserved when high concentrations of fructose (Van den Berghe, 1978; Holloway &Parsons, 1984; Vind & Grunnet, 1984) or glutamine (Zielke et al. 1980; Sumbillaet al. 1981) were supplied in the culture medium. The physiologically acceptablerange for the lactate/pyruvate ratio in mammalian cell culture has been suggested tobe between 6 and 15 (Imamura et al. 1982). In this study, the ratio was maintainedwithin or rose beyond the physiologically acceptable range in 0-5 X or IX cultures,respectively (Fig. 12).

Utilization (Fig. 10) and production (Fig. 11) of the metabolites, excluding that offructose, were inversely related to cell density (Fig. 3). Part of the decline was due tothe widening of the gap between cell population and nutrient supply. Particularly atthe high cell densities observed after the formation of cell multilayers in the perfusedcultures, the diffusion of nutrients into the cellular microenvironment, in addition tothe oxygen diffusion discussed earlier, probably played an important role in affectingutilization and production of the metabolites. However, fructose was more con-sistently utilized, especially in 0-5 X perfused cultures at higher cell densities.Metabolic modifications due to cell maturation and differentiation might have beenresponsible for the observed improvement.

Anaerobic oxidation of glucose and fructose would produce lactic acid, while theiraerobic oxidation would produce carbon dioxide. Aerobic oxidation of glutaminewould give rise to ammonia in addition to CO2. Comparison of data on production oflactate and ammonia shows the relative significance of the anaerobic and aerobicpathways. In our study, the production of lactate was always higher than ammonia inthe overfed perfusion system (Fig. 14). While the production pattern of ammonia inthe optimized perfused systems was similar to the one observed under the overfedcondition, that of lactate apparently was not (Fig. 15). During the initial 4 days,production of lactate was higher than that of ammonia in the optimized perfusedcultures. After this period, equal quantities of lactate and ammonia were producedfor the next 6 days. Finally, ammonia production was apparently higher than lactateproduction during the last 6 days of the experiment. The data suggest that there was

96 A. T. NahapetianJ. N. Thomas and W. G. Thilly

100I r

oD-

0-1

i i r

j _

8 10Time (days)

12 14 16

Fig. 14. Production of lactate (O O) and ammonia ( # • ) in overfed perfusedculture. Perfusion of the culture (100 ml) began 24 h after cell inoculation at 8 vol. day"1.Microcarrier concentration was 20gI"1; experimental medium was LI5 (containing4mM-glutamine and lOmM-fructose, instead of galactose) + 5 % foetal bovine serum.

a general decrease in both anaerobic and aerobic oxidative activities when expressedin pmol/celP'day"1. Moreover, at relatively low cell densities, anaerobic oxidationwas a more significant route for energy production under both optimized and overfedconditions. This trend was maintained throughout the study under overfed con-ditions. However, under optimized conditions, as cell density increased, aerobicoxidation assumed a more significant role for production of energy; by the end of thestudy almost all of the energy requirement of Vero cells was met by the aerobicpathway.

A possible mechanism for the observed changes is illustrated in Fig. 16. Thismechanism suggests that the capacity of Vero cells for conversion of excess nutrientsinto fat was limited under the conditions of our study and that excess intracellularnutrients were oxidized. Thus, the rate of oxidation under overfed conditions washigher than under optimized conditions due to excess supply of nutrients. Assuming


that the diffusion rate of oxygen was similar under the two conditions, the greateroxidation of nutrients in the overfed cultures led to greater consumption andlimitation of oxygen in this system. As oxygen diffused from one cell to the other, itslimitation became more critical for the innermost Vero cells. The disparity betweendissolved oxygen in the cellular microenvironments of overfed and optinmed con-ditions was increased by the phenomenon of multilayering. This could explain themuch greater lactate production under overfed conditions.

The observed fluctuations in concentration of glutamine in the perfused culturesseemed directly related to the available oxygen in the culture microenvironment.Werrlein & Glinos (1974) have reported the existence of oscillating concentrationgradients of oxygen above mammalian anchorage-dependent cells cultured in Petridishes. They showed that the oscillation amplitude increased at lower depths and

100

6 8 10Time (days)

12 14 16

Fig. IS. Production of lactate (O O) and ammonia ( • • ) in optimized perfusedcultures. Perfusion of the culture (100 ml) began 24 h after cell inoculation at 8 vol. day"1.Microcarrier concentration was 20gl~' ; experimental medium was L15 (containing4mM-glutamine and lOmM-fructose instead of galactose) + 5 % foetal bovine serumdiluted 1:1 with PBS.


O, O2

Fig. 16. Schematic presentation of the postulated mechanism for utilization andavailability of oxygen and production of carbon dioxide in the multicell-layeredmicroenvironment of overfed and optimized perfused cultures.

higher cell densities. They also demonstrated that the oscillations were the net resultof oxygen depletion or replacement due to cellular respiration or oxygen diffusion,respectively.

Similarly, our results indicate that oxidation of the nutrients oscillated (Fig. 7),perhaps due to the changes in available oxygen in the culture microenvironment.The oscillation amplitude was higher in 1X than in 0-5X perfused cultures, probablydue to a greater depletion of oxygen in the former culture. Since glucose andglutamine are known to be utilized by glycolysis and aerobic oxidation, respectively(Zielke et al. 1984), cell growth and survival would be favoured in an environmentsupplied with both, rather than either energy source alone, as concluded in a recentstudy of chick pigment epithelial cells in batch cultures (Barbehenn et al. 1984).

Patterns of changes in concentrations (Fig. 7) and utilization (Fig. 10) of glu-tamine and pyruvate in the cultures were quite similar. However, the apparentfluctuations in the amino acid concentration observed in 1X perfused cultures duringthe second week of the study were not detectable for pyruvate. Lack of accumulationof the latter metabolite during that period indicated its high utilization. Pyruvatecould be converted to either lactate or carbon dioxide plus water under anaerobic oraerobic conditions, respectively. Lactate production declined under the optimizedculture condition such that it reached an almost undetectable level by the 16th day(Fig. 11), while there was a significant utilization of pyruvate together with othernutrients (glucose, fructose, glutamine) in 0-5X perfused cultures until the last dayof the study. The results also suggest that, at least under optimized conditions,respiration progressively became the predominant route of energy production as cell


density increased during the course of the study. Data on changes in dissolvedoxygen (Fig. 6) and lactate/ammonia ratio (Fig. 13) support this hypothesis.

In addition to the explanations given above, the general decrease in utilizationunder optimized-conditions that was evident for all nutrients under investigationexcept fructose could be explained by the higher efficiency of energy production bythe aerobic pathway. The progressive dominance of aerobic over anaerobic oxidationobserved under optimized conditions in the present study might be due to a relativeincrease in intracellular fructose brought about by the development and synthesis ofproteins involved in fructose transport or metabolism during maturation and differ-entiation of Vero cells. Increased consumption of oxygen (Van den Berghe, 1978),stimulation of pyruvate kinase (Van den Berghe, 1978; Poole, Postle & Bloxham,1982), pyruvate dehydrogenase (Van den Berghe, 1978), and fatty acid syntheticenzymes (Van den Berghe, 1978; Spence & Pitot, 1982) by fructose have beenreported.

In summary, cell growth, pH, dissolved oxygen, changes in metabolites, lactate/pyruvate and lactate/ammonia ratios demonstrate that, under the conditions used inthe present study, perfusion of cultures with 50% L15 medium at 8 vol. day"1

provided the optimum microenvironment for Vero cell growth. The highest celldensity in the perfused cultures under the most favourable environmental conditionwas 3Xl07cellsml~1, which was 10 times higher than the maximum cell density(3X106 cells ml"1) obtained in a batch culture. Limitation of available oxygen andaccumulation of unknown metabolites in the cellular microenvironment were prob-ably responsible for the density-dependent inhibition of cell growth observed underthe optimized environmental condition. The inadequate supply of oxygen, togetherwith accumulation of bicarbonate, lactate, ammonia and unknown metabolites in thecellular microenvironment might have been responsible for the relatively higherdecline in growth rate observed in the 1X perfused culture at high cell densities. Wepostulate that glutamine was the major source of energy under optimized cultureconditions during the first week of the study. Significant utilization of fructosebecame evident during the second week when respiration became progressively thepredominant route of energy production. The relative significance of fructose eitheras a source of energy or as a precursor for nucleic acid and fatty acid synthesis in high-density cell cultures remains to be elucidated.

REFERENCES

ABERCROMBIE, M. & AMBROSE, E. J. (1962). The surface properties of cancer cells: A review.Cancer Res. 22, 525-548.

BARBEHENN, E. K., MASTERSON, E., KOH, S. W., PASSONNEAU, J. V. & CHADER, G. J. (1984). An

examination of the efficiency of glucose and glutamine as energy sources for cultured chickpigment epithelial cells. 7. cell. Physiol. 118, 262-266.

BARNGROVER, D., THOMAS, J. & THILLY, W. G. (1985). High density mammalian cell growth inLeibovitz bicarbonate-free media formula: Effects of fructose and galactose in culturebiochemistry. J. Cell Sd. 78, 173-189.

BERNT, E. & BERGMEYER, H. U. (1981a). D-Fructose. In Methods of Enzymatic Analysis (ed.H. U. Bergmeyer), pp. 1304-1307. Deerfield Beach, Florida: Verlag Chemie International.


BERNT, E. & BERGMEYER, H. U. (19816). L-Glutamate, UV-assay with glutamate dehydrogenaseand NAD. In Methods ofEnzymatic Analysis (ed. H. U. Bergmeyer), pp. 1704-1708. DeerfieldBeach, Florida: Verlag Chemie International.

BRAND, K., WILLIAMS, J. F. & WEIDEMANN, M. J. (1984). Glucose and glutamine metabolism inrat thymocytes. Biochem.J. 221, 471-475.

BURK, R. R. (1966). Growth inhibitor of hamster fibroblast cells. Nature, bond. 212, 1261-1262.BURROWS, M. T. (1924). Relation of oxygen to the growth of tissue cells. Am. J. Physiol. 68, 110

(abstract).BUTLER, M., IMAMURA, T., THOMAS, J. & THILLY, W. G. (1983). High yields from microcarrier

cultures by medium perfusion. J. Cell Set. 61, 351-363.BUTLER, M. & THILLY, W. G. (1982). MDCK microcarrier cultures: Seeding density effects and

amino acid utilization. In Vitro 18, 213-219.CHICK, W. L., LIKE, A. A. & LAURIS, V. (1975). Beta cell culture on synthetic capillaries: An

artificial endocrine pancreas. Science 187, 847-848.CZOK, R. & LAMPRECHT, W. (1981). Pyruvate, phosphoenolpyruvate and D-glycerate-2-phosphate.

In Methods of Enzymatic Analysis (ed. H. U. Bergmeyer), pp. 1446-1451. Deerfield Beach,Florida: Verlag Chemie International.

EAGLE, H. (1955). Nutrition needs of mammalian cells in tissue culture. Science 122, 501-504.EHRLICH, K. C , STEWART, E. & KLEIN, E. (1978). Artificial capillary perfusion cell culture:

metabolic studies. In Vitro 14, 443-450.FIXE, R. M., GLICK, J. L. & BURNS, A. A. (1977). Propagation of human lymphoid cell lines in

hollow fiber capillary units. In Vitro 13, 170 (abstract).GAY, R. J. & AMOS, H. (1983). Purines as "hyper-repressors" of glucose transport. Biochem. J.

214, 133-144.GEBHARDT, R. & MECKE, D. (1979). Perfused monolayer cultures of rat hepatocytes as an

improved in vitro system for studies on ureogenesis. Expl Cell Res. 124, 349-359.GRAFF, S. & MCCARTY, K. S. (1957). Sustained cell culture. Expl Cell Res. 13, 348-357.GRAFF, S., MOSER, H., KASTNER, 0 . , GRAFF, A. M. & TANNENBAUM, M. (1965). The signifi-

cance of glycolysis. J. natn. Cancer Inst. 34, 511-519.GRIFFITHS, J. B. (1972). The effect of cell population density on nutrient uptake and cell

metabolism: A comparative study of human diploid and heteroploid cell lines, j - Cell Sci. 10,515-524.

HIMMELFARB, P., THAYER, P. S. & MARTIN H. E. (1969). Spin filter culture: The propagation ofmammalian cells in suspension. Science 164, 555-557.

HOLLEY, R. W., ARMOUR, R. & BALDWIN, J. H. (1978). Density dependent regulation of growth ofBSC-1 cells in cell culture: Growth inhibitors formed by the cells. Proc. natn. Acad. Sci. U.SA.75, 1864-1866.

HOLLOWAY, P. A. H. & PARSONS, D. S. (1984). Absorption and metabolism of fructose by ratjejunum. Biochem. jf. 222, 57-64.

HORNG, C. B. & MCLIMANS, W. (1975). Primary suspension culture of calf anterior pituitary cellson a microcarrier surface. Biotechnol. Bioengng 17, 713-732.

HUE, L. & BARTRONS, R. (1984). Role of fructose 2,6-biphosphate in the control by glucagon ofgluconeogenesis from various percursors in isolated rat hepatocytes. Biochem. J. 218, 165-170.

IMAMURA, T., CRESPI, C. L., THILLY, W. G. & BRUNENGRABER, H. (1982). Fructose as a

carbohydrate source yields stable pH and redox parameters in microcarrier cell culture. Analyt.Biochem. 124, 353-358.

JENSEN, M. D., WALLACH, D. F. H. & LIN, P. S. (1974). Comparative growth characteristics ofVero cells on gas-permeable and conventional supports. Expl Cell Res. 84, 271—281.

JONES, M. & BONTING, S. L. (1956). Some relations between growth and carbohydratemetabolism in tissue cultures. Expl Cell Res. 10, 631-639.

KASHIWAGURA, T., WILSON, D. F. & ERECINSKA, M. (1984). Oxygen dependence of cellularmetabolism: The effect of Oj tension on gluconeogenesis. J. cell. Physiol. 120, 13-18.

KNAZEK, R. A. (1974). Solid tissue masses formed in vitro from cells cultured on artificialcapillaries. Fedn Proc. FednAm. Socs exp. Biol. 33, 1978-1981.

KNAZEK, R. A., GULLJNO, P. M., KOHLER, P. O. & DEDRICK, R. L. (1972). Cell culture onartificial capillaries: An approach to tissue growth in vitro. Science 178, 65-67.

Optimized environment for Veto cell culture 101

KNAZEK, R. A., KOHLER, P. O. & GULLINO, P. M. (1974). Hormone production by cells grownin vitro on artificial capillaries. Expl Cell Res. 84, 251-254.

KNAZEK, R. A., LIPPMAN, M. E. & CHOPRA, H. C. (1977). Brief communication: Formation ofsolid human mammary carcinoma in vitro. J. natn. Cancer Inst. 58, 419-422.

KNOP, R. H., CHEN, C. W., MITCHELL, J. B., RUSSO, A., MCPHERSON, S. & COHEN, J. S. (1984).

Metabolic studies of mammalian cells by 31P-NMR using a continuous perfusion technique.Biochim. biophys. Acta 804, 275-284.

KOVACEVIC, Z. & MORRIS, H. P. (1972). The role of glutamine in the oxidative metabolism ofmalignant cells. Cancer Res. 32, 326-333.

KREBS, H. A. (1950). Body size and tissue respiration. Biochim. biophys. Acta 4, 249-269.KRUSE, P. F. JR, KEEN, L. N. & WHITTLE, W. L. (1970). Some distinctive characteristics of high

density perfusion cultures of diverse cell types. In Vitro 6, 75-88.KRUSE, P. F. JR & MIEDEMA, E. (1965). Production and characterization of multiple-layered

populations of animal cells. J . Cell Biol. 27, 273-279.KRUSE, P. F. JR, MIEDEMA, E. & CARTER, H. C. (1967). Amino acid utilizations and protein

synthesis at various proliferation rates, population densities, and protein contents of perfusedanimal cell and tissue cultures. Biochemistry 6, 949-955.

KRUSE, P. F. JR, MYHR, B. C , JOHNSON, J. E. & WHITE, P. B. (1963). Perfusion system forreplicate mammalian cell cultures in T-60 flasks. J . natn. Cancer Inst. 31, 109-123.

KRUSE, P. F. JR, WHITTLE, W. & MIEDEMA, E. (1969). Mitotic and nonmitotic multiple layeredperfusion cultures. J . Cell Biol. 42, 113-121.

KUN, E. & KEARNEY, E. B. (1981). Ammonia. In Methods of Enzymatic Analysis (ed. H. U.Bergmeyer), pp. 1802-1806. Deerfield Beach, Florida: Verlag Chemie International.

LEIBOVITZ, A. (1963). The growth and maintenance of tissue-cell cultures in free gas exchangewith the atmosphere. Am.J. Hyg. 78, 173-180.

LEVTNE, D. W., WANG, D. I. C. & THTLLY, W. G. (1979). Optimization of growth surfaceparameters in microcarrier cell culture. Biotechnol. Bioengng 21, 821-845.

LEVTNE, D. W., WONG, J. S., WANG, D. I. C. & THTLLY, W. G. (1977). Microcarrier cell culture:New methods for research scale application. Som. Cell Genet. 3, 149-155.

LUND, P. (1981). L-Glutamine: Determination with glutaminase and glutamate dehydrogenase. InMethods of Enzymatic Analysis (ed. H. U. Bergmeyer), pp. 1719-1722. Deerfield Beach,Florida: Verlag Chemie International.

MCCARTY, K. (1962). Selective utilization of amino acids by mammalian cell cultures. Expl CellRes. 27, 230-240.

MCGOWAN, J. A., RUSSELL, W. E. & BUCHER, N. L. R. (1984). Hepatocyte DNA replication:Effect of nutrients and intermediary metabolites. Fedn Proc. Fedn Am. Socs exp. Biol. 43,131-133.

MCLIMANS, W. F., BLUMENSON, L. E. & TUNNAH, K. V. (1968). Kinetics of gas diffusion inmammalian cell culture systems. II. Theory. Biotechnol. Bioengng 10, 741-763.

MIEDEMA, E. & KRUSE, P. F. JR (1965). Enzyme activities and protein contents of animal cellscultured under perfusion conditions. Biochem. biophys. Res. Comrnun. 20, 528—534.

NAGLE, S. C. JR & BROWN, B. L. (1971). An improved heat-stable glutamine-free chemicallydefined medium for growth of mammalian cells. J. cell. Physiol. 77, 259-264.

PATTERSON, M. K. (1979). Measurement of growth and viability of cells in culture. Meth. Enzym.58, 141-152.

POOLE, G. P., POSTLE, A. D. & BLOXHAM, D. P. (1982). The induction of synthesis of L-typepyruvate kinase in cultured rat hepatocytes. Biochem. J. 204, 81-87.

QUARLES, J. M., MORRIS, N. G. & LEIBOVITZ, A. (1978). CEA production by human colorectaltumor cells in a matrix perfusion system. In Vitro 14, 335 (abstract).

QUARLES, J. M., MORRIS, N. G. & LEIBOVITZ, A. (1980). Carcinoembryonic antigen productionby human colorectal adenocarcinoma cells in matrix-perfusion culture. In Vitro 16, 113-118.

REITZER, L. J., WICE, B. M. & KENNELL, D. (1979). Evidence that glutamine, not sugar, is themajor energy source for cultured HeLa cells. J. biol. Chem. 254, 2669-2676.

RICHTER, A., SANFORD, K. K. & EVANS, V. J. (1972). Influence of oxygen and culture media onplating efficiency of some mammalian tissue cells, j . natn. Cancer Inst. 49, 1705-1712.


ROSE, G. G. (1967). The circumfusion system for multipurpose culture chambers. I. Introductionto the mechanics, techniques and basic results of a 12-chamber {in vitro) closed circulatorysystem. J . Cell Biol. 32, 89-112.

RUBIN, H. (1966). A substance in conditioned medium which enhances the growth of smallnumbers of chick embryo cells. Expl Cell Res. 41, 138-148.

RUSSELL, W. C , NEWMAN, C. & WILLIAMSON, D. H. (1975). A simple cytochemical technique fordemonstration of DNA in cells infected with mycoplasmas and viruses. Nature, Land. 253,461-462.

RUTZKY, L. P., TOMTTA, J. T., CALENOFF, M. A. & KAHAN, B. D. (1977). Matrix-perfusion

cultivation of human choriocarcinoma and colon adenocarcinoma cells. In Vitro 13, 191(abstract).

RUTZKY, L. P., TOMTTA, J. T., CALENOFF, M. A. & KAHAN, B. D. (1979). Human colonadenocarcinoma cells. III. In vitro organoid expression and carcinoembryonic antigen kinetics inhollow fiber culture. J. natn. Cancer Inst. 63, 893-899.

SANFORD, K. K., EARLE, W. R., EVANS, V. S., WALTZ, J. K. & SHANNON, J. E. (1951). The

measurement of proliferation in tissue cultures by enumeration of cell nuclei. J. natn. CancerInst. 11, 773-795.

SCHLEICHER, J. B. & WEISS, R. E. (1968). Application of a multiple surface tissue culturepropagator for the production of cell monolayers, virus and biochemicals. Biotechnol. Bioengng10, 617-624.

SMITH, M. A. & VALE, W. W. (1980). Superfusion of rat anterior pituitary cells attached to cytodexbeads: Validation of a technique. Endocrinology 107, 1425-1431.

SPENCE, J. T . & PTTOT, H. C. (1982). Induction of lipogenic enzymes in primary cultures of rathepatocytes. Relationship between lipogenesis and carbohydrate metabolism. Eur. J. Biochem.128, 15-20.

STEVENS, K. M. (1965). Oxygen requirements for liver cells in vitro. Nature, Land. 206, 199.STOKER, M. G. P. & RUBIN, H. (1967). Density dependent inhibition of cell growth in culture.

Nature, Lond. 215, 171-172.STRAND, J. M., QUARLES, J. M. & MCCONNELL, S. (1984a). A modified matrix perfusion-

microcarrier bead cell culture system. I. Adaptation of the matrix perfusion system for growth ofhuman foreskin fibroblasts. Biotechnol. Bioengng 26, 503-507.

STRAND, J. M., QUARLES, J. M. & MCCONNELL, S. (19846). A modified matrix perfusion-microcarrier bead cell culture system. II . Production of human (/3) interferon in a matrixperfusion system. Biotechnol. Bioengng 26, 508-512.

SUMBILLA, C. M., ZIELKE, C. L., REED, W. D., OZAND, P. T. & ZIELKE, H. R. (1981).

Comparison of the oxidation of glutamine, glucose, ketone bodies and fatty acids by humandiploid fibroblasts. Biochim. biophys. Acta 675,.301-304.

SUN, A. M. & MACMORINE, H. G. (1976). Cultivation of monkey and rat islets and the use ofartificial capillary units to diabetic animals. Diabetes 25, 339 (abstract).

THILLY, W. G., BARNGROVER, D. & THOMAS, J. N. (1982). Microcarriers and the problem of highdensity cell culture. In From Gene to Protein: Translation into Biotechnology, pp. 75-103. NewYork: Academic Press.

TILDON, J. T. & ROEDER, L. M. (1984). Glutamine oxidation by dissociated cells and homo-genates of rat brain: kinetics and inhibitor studies. J. Neurochem. 42, 1069-1076.

TRTTSCH, G. L. & MOORE, G. E. (1962). Spontaneous decomposition of glutamine in cell culturemedia. Expl Cell Res. 28, 260-364.

VAN DEN BERGHE, G. (1978). Metabolic effects of fructose in the liver. Curr. Topics cell. Res. 13,97-135.

VAN WEZEL, A. L. (1967). Growth of cell-strains and primary cells on micro-carriers inhomogeneous culture. Nature, Lond. 216, 64—65.

VIND, C. & GRUNNET, N. (1984). The reversibility of cytosolic dehydrogenase reactions inhepatocytes from starved and fed rats. Biochem. J. 222, 437-446.

VISEX, W. J., KOLODNY, G. M. & GROSS, P. R. (1972). Ammonia effects in cultures of normal andtransformed 3T3 cells. J . cell. Physiol. 80, 373-382.

WERRLEIN, R. J. & GLINOS, A. D. (1974). Oxygen microenvironment and respiratory oscillationsin cultured mammalian cells. Nature, Lond. 251, 317-319.


WESTFALL, B. B., EVANS, V. J., SHANNON, J. E. JR & EARLE, W. R. (1953). The glycogen contentof cell suspensions prepared from massive tissue culture: Comparison of cells derived frommouse connective tissue and mouse liver. J. tiatn. Cancer Inst. 14, 655-664.

WOLFROM, C , POUNI, G., DECMO, D. & GAUTIER, M. (1984). Comparative effects of glucosearid fructose on the growth of cultured human skin fibroblasts in the presence or absence ofglutamine. Biol. Cell 52, A35 (abstract).

ZlELKE, H. R., OZAND, P. T., TlLDON, J. T., SEVDALIAN, D. A. & CORNBLATH, M. (1978).Reciprocal regulation of glucose and glutamine utilization by cultured human diploidfibroblasts. J . cell. Physiol. 95, 41-48.

ZlELKE, H. R., SUMBILLA, C. M., SEVDALIAN, D. A., HAWKINS, R. L. & OZAND, P. T. (1980).Lactate: A major product of glutamine metabolism by human diploid fibroblasts. J. cell. Physiol.104,433-441.

ZlELKE, H. R., ZlELKE, C. L. & OZAND, P. T. (1984). Glutamine: A major energy source forcultured mammalian cells. Fedn Proc. FednAm. Socs exp. Biol. 43, 121-125.

(Received 5 August 1985 - Accepted 20 August 1985)

OPTIMIZATION OF ENVIRONMENT FOR HIGH DENSITY VERO CELL CULTURE

Documents